La verifica dosimetrica di trattamenti radianti: metodi, necessità ed efficacia. M. Iori, Reggio Emilia

Transcript

1 TORINO Novembre 2013 RADIOTERAPIA 1 RELAZIONI A INVITO La verifica dosimetrica di trattamenti radianti: metodi, necessità ed efficacia. M. Iori, Reggio Emilia Le nuove frontiere: l ottimizzazione radiobiologica è più facile di quel che sembra. M. Schwarz, Trento Le nuove frontiere: gestione del movimento respiratorio nell era dell imaging multimodale in radioterapia. C. Cavedon, Verona Radioterapia e HIFU: alternativa o sinergia? G. Borasi, Reggio Emilia Caratterizzazione dosimetrica e planning di/con fasci di piccola dimensione. P. Francescon, Vicenza Caratterizzazione di apparecchiature per IORT e tecniche di dosimetria in vivo. G. Taccini, Genova Il commissioning di fasci di protoni e ioni carbonio. M. Ciocca, Pavia Utilità del metodo Monte Carlo in radioterapia. E. Spezi, Cardiff (U.K.)

2 TORINO Novembre 2013 RADIOTERAPIA 1 RELAZIONI LIBERE Ripianificazione on Cone-Beam CT: metodologia di calibrazione e prime applicazioni cliniche. C. Carbonini, Milano Modellizzazione LINAC per radioterapia volumetrica ad intensità modulata: criticità, problematiche ed effetti delle griglie di dose. G. Guidi, Modena/Bologna Valutazione dell accuratezza di calcolo della dose per trattamenti IMRT nel distretto toracico: confronto tra pencil beam, convolution/superposition e analytical anisotropic algorithm. C. Sini, Cagliari/Milano Verifica dosimetrica pre-trattamento ed in-vivo dei piani di trattamento eseguiti in tomoterapia elicoidale (HT) mediante sistema di verifica dosimetrico accoppiato ai rivelatori HT. E. Mezzenga, Reggio Emilia Effetti di consumo del target in tomoterapia: uno studio Monte Carlo. A. Esposito, Roma Una rete dedicata alla radioterapia per applicazioni oncologiche. F. Di Rosa, Caltanissetta Implementazione di un modello di trasmissione del lettino di trattamento nel TPS in geometrie con elevato contributo di dose da fasci posteriori o obliqui. I. Solla, Cagliari Effetti dosimetrici degli errori di set up rotazionale sui trattamenti IMRT della prostata. C. Zucchetti, Perugia Validazione di un algoritmo di registrazione di immagini deformabile per la somma di dosi. M. Fusella, Torino Pianificazione IMRT sliding windows per mammella con boost simultaneo integrato (SIB): efficienza ed erogabilità in funzione dello smoothing sulle mappe di fluenza S. Naccarato, Negrar (VR) Boost adattivo simultaneamente integrato nella radiochemioterapia neoadiuvante per il carcinoma rettale: validazione prospettica dei margini al tumore comprendenti l impatto della deformazione. R. Raso, Milano Esperienza dell Azienda Ospedaliero-Universitaria Careggi nelle verifiche dosimetriche su singolo paziente: dalla metrica gamma al DVH. L. Marrazzo, Firenze Dose agli organi a rischio nel distretto dell addome superiore in pazienti trattati con campi estesi con tomoterapia elicoidale: analisi dell istogramma dose-volume e studio di tossicità. S. Bresciani, Candiolo (TO) Valutazione della dose agli organi in radioterapia per sistemi kv Cone Beam CT. F. Palleri, Bologna Effetti del riempimento vescicale sulla dose agli organi a rischio nella brachiterapia HDR endovaginale con pianificazione 3D. M. Piergentili, La Spezia Dosimetria in vivo mediante EPID per 3D-CRT, IMRT e VMAT: aggiornamento del progetto DISO. A. Fidanzio, Roma Metodi per la valutazione della robustezza di piani di trattamento con protoni e scanning attivo del fascio. L. Widesott, Trento/Zurich (CH)

3 TORINO Novembre 2013 RADIOTERAPIA 1 RELAZIONI LIBERE Criticità in radioterapia intraoperatoria con acceleratore mobile: aspetti geometrici e dosimetrici. S. Andreoli, Bergamo Applicazione del report AAPM TG 119 su un fantoccio dosimetrico 3-D per l implementazione di tecniche IMRT e VMAT. L. Trombetta, Milano Confronto fra tecniche per il trattamento adiuvante del carcinoma mammario. R. Cambria, Milano Confronto dosimetrico tra tecniche IMRT e tomoterapia elicoidale per il trattamento della prostata. A. Didona, Perugia Caratterizzazione dosimetrica di un diamante sintetico per la dosimetria di fasci di elettroni in radioterapia. M. D. Falco, Roma Utilizzo di film gafcromici per controlli di qualità per fasci clinici di fotoni e protoni. C. Stancampiano, Catania Integrazione di immagini DICOM nell applicazione Web-Geant4 iort_therapy: un ulteriore passo per supportare la tecnica di Radio-Terapia Intra-Operatoria (IORT). C. Casarino, Cefalù (PA) Dosimetria a risonanza paramagnetica elettronica con alanina per fasci clinici di ioni carbonio e protoni. A. Carlino, Palermo/Catania/Villigen (CH) Apparecchiatura di tipo pencil beam scanning per protonterapia: caratteristiche e prestazioni del fascio di Trento. S. Lorentini, Trento

4 r RADIOTERAPIA 1 POSTER r r r r r r r r r r r r O. Nibale, Rovigo rr r r r r r r M. L. Belli, Milano r r r r r r r r r G. R. Borzì, Catania/Viagrande (CT) r r r r R. Caivano, Rionero In Vulture (PZ) r r r r r r r r r r r r r M. Casale, Terni r r r r r r r r r r M. Bucciolini, Firenze r r r r r r r r M. Bucciolini, Firenze r r S. Clemente, Rionero In Vulture (PZ) r r r r r r r r r S. Cora, Vicenza r r r r r r M. Cozzolino, Rionero In Vulture (PZ) r r r r r r r D. Cusumano, Milano r r r r C. Cutaia, Candiolo (TO) r MapCheck r r r F. De Monte, Torino r r A. Esposito, Roma r r r L. Ferri, Genova r r r r r r M. Frigerio, San Fermo Della Battaglia (CO) IL TOPIC RADIOTERAPIA CONTINUA IN RADIOTERAPIA 2: TORNARE ALL'INDICE DEI TOPIC E CLICCARE SU RADIOTERAPIA 2.

5 La verifica dosimetrica dei trattamenti radianti: metodi, necessità ed efficacia Mauro Iori Servizio di Fisica Medica, ASMN-IRCCS di Reggio Emilia, Italia Le verifiche dosimetriche dei piani di trattamento radioterapici, spesso indicate come verifiche dosimetriche pre-cliniche, rappresentano ancor oggi, a più di dieci anni dall introduzione clinica dei trattamenti a fascio modulato, un argomento di grande interesse ed oggetto di approfondite discussioni [1, 2, 3]. Se la valutazione preclinica dei piani di trattamento di tipo convenzionale, meglio noti come terapie conformazionali (3DCRT), trova concorde la comunità scientifica nell abbandonare il controllo dosimetrico su fantoccio a fronte dell impiego di software indipendenti certificati per il calcolo delle unità monitor [4], lo stesso non avviene per le tecniche radianti ad intensità modulata (IMRT). Nella 3DCRT, dove i fasci radianti si contraddistinguono per il pattern uniforme delle fluenze, la verifica dosimetrica dei trattamenti si realizza principalmente verificando le unità monitor, cioè i valori di dose associati ad ogni campo di trattamento, utilizzando sistemi di calcolo certificati [5] grazie ai quali è agevole ricalcolare la dose di un dato piano di cui si conoscono il set-up dei fasci ed i relativi parametri dosimetrici. Questi sistemi, che si compongono di algoritmi di calcolo indipendenti e che utilizzano la stessa dosimetria delle unità radianti configurate nei sistemi di pianificazione dei trattamenti (TPS), non solo consentono il controllo di campi aperti e/o sagomati, ma grazie all'elevata qualità degli algoritmi che li compongono, si possono utilizzare anche per i fasci conformati di piccole dimensioni (CyberKnife) o per i fasci modulati di complessità crescente: filtri dinamici (dynamic wedge), IMRT a stativo fisso (slyding-window o step-and-shoot), IMRT di tipo rotazionale (IMAT), volumetrico (VMAT) o tomo terapico (Tomotherapy). Gli stessi sistemi di calcolo [6] si dimostrano inoltre estremamente efficaci anche nella valutazione dei trattamenti di brachiterapia dove, in questo caso, il controllo del piano di trattamento si realizza ricalcolando i tempi di permanenza delle sorgenti radioattive, per ciascuno degli applicatori che compongono il piano. Sebbene questi software consentano di individuare i principali errori associati ad un dato piano di trattamento, errori dovuti generalmente: ad una non corretta configurazione del TPS, ad una erronea modifica dei suoi parametri di configurazione, alla creazione di fluenze (di fascio) estremamente disomogenee e/o gradientate e ad una errata trascrizione/trasmissione dei dati del piano nel sistema di controllo e verifica (R&V), non consentono tuttavia di evidenziare possibili criticità correlate: alla meccanica dell'unità radiante, al corretto funzionamento dei suoi sistemi di erogazione/modulazione della dose o all errato set-up del paziente. Molto si è discusso e si discute in letteratura su quale sia il criterio e/o la modalità di verifica dosimetrica più corretta (sensibile ed efficace) per controllare e validare i piani di trattamento, impedendo che possano manifestarsi erogazioni scorrette di una singola terapia o eventi anche più catastrofici (errori sull intero ciclo di uno o più trattamenti). Ci si interroga sulla qualità degli strumenti dosimetrici di cui si dispone e della sensibilità dei test comunemente applicati nell'evidenziare incongruenze o criticità esistenti nell'esecuzione di un dato trattamento. Molti studi hanno evidenziato la scarsa efficacia delle metodiche di verifica attualmente in uso, sottolineando sia le limitazioni presenti nei più diffusi sistemi di rivelazione [4, 7, 8, 9]: forte diversità nel set-up costruttivo e nella tipologia di segnale fornito, dipendenza angolare nella risposta, limitata risoluzione spaziale, complessità nelle procedure di calibrazione, ecc.., sia nella limitata sensibilità [7, 10, 11, 12] dei metodi di analisi correntemente applicati: variabilità nella valutazione della risposta col cambio dei parametri di analisi selezionati, difficoltà nel correlare le differenze dosimetriche riscontrate al reale impatto clinico/dosimetrico sul trattamento del paziente, ecc... In merito a questo tipo di problemi, la letteratura scientifica propone diverse modalità di verifica [1, 2, 3], per ciascuna delle quali sono presenti punti di forza ed elementi di criticità. Parimenti i vari centri che attivano trattamenti con fasci modulati seguono modalità di verifica che rappresentano a volte il miglior compromesso tra quanto riportato in letteratura e quanto appartiene all esperienza dosimetrica e di QA di ogni centro.

6 Un primo metodo di verifica presente in letteratura si basa sulla conoscenza approfondita del proprio TPS, richiedendo che sia dedicata molta attenzione al commissioning dei suoi algoritmi di calcolo, fase che deve essere poi seguita da un percorso di assicurazione di qualità (QA) orientato alla verifica del TPS (ottimizzazione, sequencing e dosimetria su fantoccio) simulando condizioni complesse, critiche o 'estreme' nel processo di pianificazione. Questo metodo, che può essere realizzando seguendo quanto proposto al riguardo nella letteratura scientifica, accresce la conoscenza del comportamento del sistema IMRT nel suo complesso e consente di stimare, dopo aver verificato su fantoccio l'accordo dosimetrico tra il dato di pianificazione e quello di misura, le potenziali aree di criticità di ogni singolo TPS in relazione alle tipologie di pianificazione applicate. Tale processo dovrebbe consentire quindi la creazione di un proprio livello di attenzione, oltre che la conoscenza e la stima del livello di sicurezza dei propri sistemi dosimetrici e di pianificazione. Durante le fasi di pianificazione dei trattamenti su paziente, dovrebbe pertanto rivelarsi più semplice conoscere quali piani potrebbero rivelarsi potenzialmente più critici in fase di erogazione e necessitare, dopo una prima verifica con un sistema di calcolo certificato, un monitoraggio dosimetrico pre-clinico. A completamento di questo processo, diventa fondamentale seguire un stringente protocollo di QA sull unità radiante che tenga alta l attenzione sull'accuratezza e la qualità complessiva delle componenti meccaniche, dosimetriche, di conformazione/modulazione dell unità radiante. Alcuni lavori ritengono [1, 2] che un tale percorso di verifica sia sufficiente per garantire la qualità complessiva dei trattamenti e non accettano l'idea che solo una valutazione dosimetrica diretta del piano di trattamento rappresenti l'unico modo certo per garantire l efficacia e la sicurezza di un trattamento IMRT. Un secondo metodo di verifica, forse il più seguito, ritiene necessario verificare ogni singolo piano di trattamento a fasci modulati con misure dosimetriche specifiche realizzate prima dell erogazione della terapia (verifiche pre-cliniche). Seppure questo metodo non assicura che un buon accordo dosimetrico verificatosi in fase di inizio terapia si mantenga per l intero ciclo dei trattamenti, solo un sistema di monitoraggio dosimetrico posto direttamente sull unità radiante potrebbe confermarlo, è comunemente accettato che nella fase pre-clinica di QA del piano di trattamento possano essere identificati e corretti (nuova ottimizzazione del piano di trattamento con fluenze meno modulate, ecc..) gli errori dosimetrici più significativi. Benché l'esecuzione di verifiche QA pazientespecifiche non abbiano evitato che alcuni centri coinvolti nel test di accreditamento per l IMRT dell RTOG [13] non superassero il test, una maggiore cura nelle fasi di commissioning del TPS e di pianificazione dei trattamenti ha portato ad un miglioramento del livello di valutazione raggiunto nella prima fase di esecuzione dei test. Inoltre, la conoscenza dei risultati raggiunti dai migliori centri analizzati ha fornito un livello di riferimento a cui tendere e dei dati quantitativi con cui confrontarsi e poter valutare il proprio livello di accuratezza dosimetrica. L uso generale della verifica dosimetrica pre-clinica per tutti i trattamenti IMRT pianificati, verifica basata sul controllo della dose erogata su di un fantoccio omogeneo in cui è inserito un sistema di rivelazione planare, spesso utilizzando un set-up a stativo fisso (AP) dei campi di irraggiamento, lascia comunque alcuni dubbi sul grado di validità di un tale controllo soprattutto per quei piani che si contraddistinguono per le fluenze estremamente gradientate, spesso erogate su distretti anatomici molto disomogenei o con configurazioni complesse di target ed organi critici o dove i campi modulati sono inseriti all'interno di sequenze di gating o di tracking. Seppure la comunità scientifica sia concorde nel ritenere che le verifiche dosimetriche paziente-specifico siano affette da una serie di limitazioni e criticità, non per questo devono essere trascurate in quanto costituiscono ancor oggi lo strumento più efficace di cui si dispone per poter evitare gravi errori dosimetrici [14]. Un terzo metodo che viene applicato e che deriva dalla combinazione dei due precedenti, utilizza software dedicati per verificare lo scarto tra le fluenze e le dosi puntuali di un dato piano IMRT calcolate rispettivamente dal TPS e dal software indipendente certificato. Qualora le verifiche effettuate evidenzino una discrepanza superiore ai livelli di confidenza prefissati, spesso frutto dell esperienza dei singoli centri, viene attivata la verifica dosimetrica pre-clinica di alcuni fasci, spesso i più discordanti, o dell intero piano di trattamento. Alcuni lavori di letteratura sembrano confermare l esistenza di una corrispondenza tra quanto segnalato dalle verifiche dosimetriche virtuali effettuate con i software dedicati, spesso basati su metodi Monte Carlo, e quanto ottenuto con le verifiche dosimetriche. Anche con questo metodo di verifica dei piani di trattamento è fondamentale inserire specifici test di

7 QA per il controllo del corretto funzionamento dell unità radiante e dei suoi dispositivi di conformazione/modulazione della dose. Dopo aver effettuato un significativo periodo di verifiche pre-cliniche di trattamenti IMRT, per la cui verifica è opportuno utilizzare sistemi di rivelazione puntuali e bidimensionali accurati e con un elevata risoluzione spaziale [4], oltre ad aver pianificato un'ampia casistica di trattamenti, la scelta dei piani di cura da dosimetrare, la frequenza dei test di QA da come pure la frequenza scelta per la loro esecuzione sono il frutto dell'esperienza e della sensibilità del centro. Il controllo di processo statistico (CPS) [15, 16], utilizzato comunemente in campo industriale per mantenere sotto controllo la qualità di un processo, è stato applicato al QA dosimetrico dei piani di trattamento utilizzando le carte di controllo (confronto delle caratteristiche di un processo in relazione alle sue caratteristiche storiche) per verificare che il processo dosimetrico è sotto controllo (cosa che avviene quando le caratteristiche del processo sono prevedibili in senso statistico). Un tale metodo è stato utilizzato, applicandolo ai criteri di validità dosimetrica utilizzati come la % di punti con un indice γ 1 [4], per verificare il livello di prestazione di un dato controllo dosimetrico, cioè la sua appartenenza ad una data classe di accordo (cioè con un certo score in termini di γ 1). Grazie al controllo di processo statistico diventa possibile creare i propri livelli di confidenza dosimetrica (creazione di classi) ed individuare le criticità di un dato piano di trattamento qualora non rientri in una data classe di controllo. Certo è che l indicatore dosimetrico monitorato deve essere sufficientemente sensibile per individuare la presenza di una criticità. Il CPS [15, 17] dovrebbe fornire uno strumento per migliorare il monitoraggio dosimetrico dei vari centri e consentir loro di individuare, grazie al patrimonio dei loro dati storici, i limiti di accettabilità dosimetrica (livelli di azione [18]) più efficaci e statisticamente più sicuri. Un tale metodo può essere utilizzato anche per monitorare la qualità delle verifiche dosimetriche (confronto tra il calcolato) effettuate con i software certificati o per valutare le variazione dei parametri di trattamento registrati all interno dei log-file [19, 20] in fase di terapia. Recentemente sono apparsi sul mercato nuovi sistemi di acquisizione e di rielaborazione della dose misurata [7] che, servendosi del dato dosimetrico pre-clinico su fantoccio e di quello simulato in fase di pianificazione, utilizzano le differenze dosimetriche misurate per correggere la distribuzione di dose del piano di trattamento simulato sul paziente. Questa correzione della dose simula sul piano del paziente gli effetti dovuti alla reale erogazione del trattamento (misurati sul fantoccio) e consente di visualizzare/confrontare sulle immagini (e sui contorni in esse definiti) del paziente le distribuzioni di dose pianificata e misurata. Tale confronto può essere visualizzato in termini di differenze di dose o di istogrammi dose volume (DVH). Questi sistemi, spesso costituiti da veri e propri TPS indipendenti, non solo permettono un confronto pianificato/erogato in termini di % di indice γ, ma soprattutto consentono di stimare in 3D la dose erogata al paziente e di valutarne clinicamente la correttezza dosimetrica [21]. Seppure questi strumenti consentano di ottener un significativo miglioramento rispetto ai precedenti sistemi dosimetrici poiché forniscono una dosimetria 3D [7, 22] del piano di trattamento erogato sul paziente, permane la necessità della misurazione pre-terapia. A questo si aggiunge l importanza di far seguire ad una prima valutazione clinica del piano di trattamento effettuata sulla base della dose simulata, una seconda valutazione clinica incentrata sulla dose erogata. Una simile verifica dosimetrica, compiuta in questo caso durante l erogazione della terapia, può essere effettuata servendosi di sistemi di acquisizione e software dedicati [23] per la dosimetria in-vivo [22]. Tali sistemi che utilizzano il segnale di transito misurato dai rivelatori planari (EPID o MV detector) ed i dati di imaging acquisiti con la CB-CT o la MV-CT, consentono di ricostruire in 3D la dose effettivamente erogata al paziente. Se tali sistemi sono idonei per verificare la dosimetria dei trattamenti erogati su distretti anatomici pressoché omogenei, poco affetti da errori di set-up o da variazioni anatomiche intra-frazione, la loro applicazione su distretti anatomici disomogenei (polmoni, ecc..) non risulta ancora pienamente soddisfacente per la presenza di algoritmi di calcolo non sufficientemente evoluti. Tutti i dispositivi fin qui descritti permettono, in generale, una verifica dosimetrica off-line del trattamento. Altri sistemi invece [24], seppure non consentano una dosimetria in 3D del piano, grazie a dosimetrici planari installati sulla bocca di uscita del fascio ed al dettaglio delle informazioni sulle

8 varie fasi della modulazione, sono in grado di controllare in tempo reale e dosimetricamente la precisione e l accuratezza del trattamento in modo integrato con il sistema di R&V. Se il sistema di R&V e le verifiche dosimetriche citate in precedenza consentono di individuare ed evitare imprecisioni o errori durante l erogazione, esistono altri eventi che possono inficiare la qualità complessiva di una terapia: trattare un paziente sbagliato, centrare una sede anatomica errata, utilizzare un sistema di immobilizzazione improprio o non correttamente posizionato, ecc... Per questi tipi di errori sono disponibili sistemi identificativi basati su codici a barre e/o foto di riconoscimento (del paziente, dell area di set-up del trattamento, dell accessorio e/o del sistema di immobilizzazione usato) gestiti all interno dei sistemi di R&V. Di recente le ditte hanno introdotto nuovi sistemi di controllo che si servono del monitoraggio ottico, dell uso di transponder o di visori di sala forniti di codici colore o di altri dispositivi di simulazione/controllo finalizzati al supporto degli operatori nell eseguire correttamente le procedure di set-up e nel monitorare l unità radiante durante l effettuazione di terapie complesse (terapie con movimentazione dell unità, dei suoi dispositivi o del paziente). Bibliografia: [1] R.A. Siochi, A Molineu, CG Orton, Point/Counterpoint. Patient-specific QA for IMRT should be performed using software rather than hardware methods, Med. Phys. (2013), 40, [2] J.C. Smith, S Dieterich, CG Orton, Point/counterpoint. It is still necessary to validate each individual IMRT treatment plan with dosimetric measurements before delivery, Med. Phys. (2011), 38, [3] Markus Alber, ESTRO Booklet n. 9: Guidelines for the verification of IMRT (2008). [4] Low DA, Moran JM, Dempsey JF, Dong L, Oldham M, Dosimetry tools and techniques for IMRT, Med. Phys. (2011), 38, [5] E.E. Wilcox, G.M. Daskalov, G. Pavlonnis, R. Shumway, B. Kaplan, E. VanRooy, Dosimetric verification of intensity modulated radiation therapy of 172 patients treated for various disease sites: comparison of EBT film dosimetry, ion chamber measurements, and independent MU calculations, Med. Dosim. (2008), 33, [6] Y. Takahashi, M. Koizumi, I. Sumida, F. Isohashi, T. Ogata, Y. Akino, Y. Yoshioka, S. Maruoka, S. Inoue, K. Konishi, K. Ogawa, The usefulness of an independent patient-specific treatment planning verification method using a benchmark plan in high-dose-rate intracavitary brachytherapy for carcinoma of the uterine cervix, J. Radiat. Res. (2012), 53, [7] M. Stasi, S. Bresciani, A. Miranti, A. Maggio, V. Sapino, P. Gabriele, Pretreatment patient-specific IMRT quality assurance: a correlation study between gamma index and patient clinical dose volume histogram, Med. Phys. (2012) 39, [8] L. Masi, F. Casamassima, R. Doro, P. Francescon, Quality assurance of volumetric modulated arc therapy: evaluation and comparison of different dosimetric systems, Med. Phys. (2011), 38, [9] V. Feygelman, G. Zhang, C. Stevens, B.E. Nelms, Evaluation of a new VMAT QA device, or the "X" and "O" array geometries, J. Appl. Clin. Med. Phys. (2011), 31, [10] B.E. Nelms, H. Zhen, W.A. Tomé, Per-beam planar IMRT QA passing rates do not predict clinically relevant patient dose errors, Med Phys. (2011), 38, [11] J.J. Kruse, On the insensitivity of single field planar dosimetry to IMRT inaccuracies, Med, Phys. (2010), 37, [12] G. Budgell, Comment on "On the insensitivity of single field planar dosimetry to IMRT inaccuracies", Med Phys. (2010) 37, [13] G.A. Ezzell, J.W. Burmeister, N. Dogan, T.J. LoSasso, J.G. Mechalakos, D. Mihailidis, A. Molineu, J.R. Palta, C.R. Ramsey, B.J. Salter, J. Shi, P. Xia, N.J. Yue, Y. Xiao, IMRT commissioning: multiple institution planning and dosimetry comparisons, a report from AAPM Task Group 119, Med. Phys. (2009), 36, [14] P. Ortiz López, J.M. Cosset, P. Dunscombe, O Holmberg, J.C. Rosenwald, L. Pinillos Ashton, J.J. Vilaragut Llanes, S. Vatnitsky, ICRP publication 112. A report of preventing accidental exposures from new external beam radiation therapy technologies, Ann ICRP , 1-86.

9 [15] S.L. Breen, D.J. Moseley, B. Zhang, M.B. Sharpe, Statistical process control for IMRT dosimetric verification, Med. Phys, (2008), 35, [16] T. Pawlicki, S. Yoo, L.E. Court, S.K. McMillan, R.K. Rice, J.D. Russell, J.M. Pacyniak, M.K. Woo, P.S. Basran, J. Shoales, A.L. Boyer, Moving from IMRT QA measurements toward independent computer calculations using control charts, Radiother. Oncol. (2008), 89, [17] K. Gérard, J.P. Grandhaye, V. Marchesi, H. Kafrouni, F. Husson, P. Aletti, A comprehensive analysis of the IMRT dose delivery process using statistical process control (SPC), Med. Phys. (2009), 36, [18] G.M. Mancuso, J.D. Fontenot, J.P. Gibbons, B.C. Parker, Comparison of action levels for patient-specific quality assurance of intensity modulated radiation therapy and volumetric modulated arc therapy treatments, Med. Phys, 2012 Jul;39(7): [19] D.W. Litzenberg, J.M. Moran, B.A. Fraass, Verification of dynamic and segmental IMRT delivery by dynamic log file analysis, J. Appl. Clin. Med. Phys. (2002), 3, [20] C.E. Agnew, R.B. King, A.R. Hounsell, C.K. McGarry, Implementation of phantom-less IMRT delivery verification using Varian DynaLog files and R/V output, Phys. Med. Biol. (2012), 7, 57, [21] H. Zhen, B.E. Nelms, W.A. Tome, Moving from gamma passing rates to patient DVH-based QA metrics in pretreatment dose QA, Med. Phys. (2011), 38, [22] B. Mijnheer, S. Beddar, J. Izewska, C. Reft, In vivo dosimetry in external beam radiotherapy, Med Phys. (2013), 40, [23] W.D. Renner, 3D dose reconstruction to insure correct external beam treatment of patients, Med. Dosim. (2007), 32, [24] M.K. Islam, B.D. Norrlinger, J.R. Smale, R.K. Heaton, D. Galbraith, et al., An integral quality monitoring system for real-time verification of intensity modulated radiation therapy, Med. Phys. (2009), 36,

10 Biological plan optimization: it s not as difficult as it may seem. Marco Schwarz Agenzia Provinciale per la Protonterapia (ATreP) e Azienda Provinciale per i Servizi Sanitari, Trento. Introduction The idea of using radiobiological parameters in radiotherapy treatment planning is not new, as one can find already in the early '90s proposals for what we would now call biological optimisation (e.g. [18, 10]). The radiotherapy techniques available at that time, however, were not exible enough to significantly benfit from sophisticated metrics to judge the dose distributions or, even less, from automatic planning optimization techniques. Intensity Modulated Radiation Therapy (IMRT) was a significant change in this respect,making it possible to plan and deliver dose distributions using much more degrees of freedom than those available in 3D-CRT. In most cases the implementation of IMRT in clinical practice is associated with a number of issues that have little to do with radiobiological models, such as new commissioning work on the TPS, finer dosimetrical and mechanical checks of the MLC and the design of patient-specific dosimetric verfication. Only after most of the technical problems are solved, and IMRT is part of the clinical routine, people start asking themselves questions such as: "Now that I have IMRT, how am I going to fully benefit from it?". Then, a more worrying version of the dilemma might arise: "In IMRT treatment planning, we mostly get what we ask. Are we really sure about what to ask?". In other words, it takes some time to appreciate the importance of an appropriate and explicit definition of the treatment goals, which should be then translated into a cost (or objective) function. Before getting into the details about the use of parameters such as the Equivalent Uniform Dose (EUD), Tumor Control Probability (TCP) and Normal Tissue Complication Probability (NTCP) in treatment planning, we will shortly discuss three statements that are often brought up in the discussions on the use of radiobiological parameters in plan optimisation. In our opinion, these statements do more harm than good to the discussion, and can be summarized as follows: "We are not ready for 'biological' optimisation yet. Let's stick to 'physical' optimisation, which is safer." This statement implicitely assumes that it is possible to make a binary distinction between radiobiological and physical optimisation. Furthermore, such an argument often boils down to the idea that physical optimisation means using Dose Volume Histogram(DVH) points in the cost function, while if different parameters are used, such as the EUD, one steps into the world of radiobiological optimisation. I think that this issue should be looked at in a different way: as a matter of fact, any optimisation is "biological", i.e. it has an (either implicit or explicit) underlying dose{effect model driving the requirements on the dose distribution. For example, setting a DVH point objective to ensure that no more than a given rectal wall volume receives a given dose is a way to incorporate in the cost function data on dose-volume thresholds for serious late e_ect. This is per se neither more nor less "biological" than setting an EUD objective. It is true that in the past most analysis of clinical data were looking for relations between toxicity and DVH points, without providing fit for NTCP models. However, dose effect studies with NTCP fits are now available for several organs at risk, based on hundreds or even thousands of patients (see for instance [16] for data on pulmonary toxicity, [17] for the rectum and [5] for the parotid glands). The real issue, therefore, is not whether a "physical" or a "biological" parameter is used in the optimisation, but rather how good or bad a specific parameter is at a)describing the dose-effect relation for the volume of interest and b)being an efficient tool to steer the optimisation towards the desired results. "IMRT optimisation can not be based on radiobiological parameters obtained in the 3D-CRT era." It is unquestionable that biological model parameters should be used while being aware of the specific conditions in which they were obtained. However, this approach can lead to a circular argument: we can not use biological parameters to obtain IMRT dose distributions because they were derived with CRT, thus we keep using IMRT to

11 deliver CRT-like dose distributions and therefore we will never be able to collect biological model parameters specific of IMRT treatments. We should remember that models are built also to allow some extrapolation of present data, and that, by de_nition, the knowledge on dose response relations always predates the latest technique available. So, if we really want to try something new, there is no other way than a sensible use of the existing knowledge. "Radiobiological parameters are bad for plan optimisation but good for plan evaluation." Many centers find it difficult to start with EUD and NTCP-based optimisation, in particular when it comes to decide the parameter values. They may therefore decide to have a gradual transition, with an initial period where dose distributions are analyzed with biological models that are then going to be used in the optimisation. This approach seems very sensible but seems to forget two important facts: It's quite simple to calculate what EUD (or NTCP, or TCP) values are 'normal' for a specific radiotherapy clinic. You have been treating hundreds if not thousands of patients. For each of them you have the planning DVHs, which is the only thing needed (beside the number of fractions) to calculate most radiobiological quantities of interest. A few days of work are sufficient to re-evaluate past treatments and provide a baseline for radiobiological optimization. In general, separating what is good for the optimisation from what is good for plan evaluation is the consequence of a misunderstanding about the role of the cost function, which is often considered as a series of "tricks" to steer the optimisation towards an acceptable dose distribution, rather than as the mathematical expression of the doseresponse relation for the volumes of interest involved in the optimization problem. Using different parameters for these two steps of treatment planning makes it difficult to produce meaningful comparisons between competing plans. EUD and TCP-based cost functions for the target volume. Historically, the first quantity that turned out to be useful for dose optimization in the target volume was the Equivalent Uniform Dose (EUD) by A. Niemerko [19]. The concept of EUD is quite simple, i.e. it is the idea that for any heterogeneous dose distribution there is an homogenenous dose distribution giving the same biological e_ect. The problem, of course, is how we can define a 'conversion rule' from heterogeneous to homogeneous doses. Using EUD instead of TCP in treatment plan optimization has two practical advantages: 1. TCP optimisation assumes that target dose is a variable of a single plan. This is an interesting and sensible approach. However, in current clinical practice the prescribed dose is defined a priori for every patient, and variations are in principle allowed only in specific cases (e.g. whithin a clinical trial). 2. TCP optimization requires that the steepness of the dose-effect relation is well known, which is not trivial to derive from clinical data. So, even if we actually consider the target dose an optimisation variable, it may be more 'robust' to optimize based on EUD, which is still a dose as opposed to a probability of control. Last but not least, by optimizing EUD one is optimizing TCP, as there is a monotonic relation between the two quantities. Even then, while there has been some quite some interest in the recent years for EUD-based optimisation for the OARs, the clinical implementation of EUD/TCP optimisation for the target volumes progresses very slowly, if it does progresses at all, mostly in a few research-oriented departments (see e.g. [1, 27]. The reasons for such difference are at least two, strictly related one another:

12 1. The incentive to use EUD or TCP in the optimisation is not evident as long as IMRT is applied with the same dose prescription as CRT and with the same planning goals, i.e. getting homogeneous dose in the PTV and good dose conformity. As long as an homogeneous dose is prescribed in the PTV, two DVH points, constraining the minimum and the maximum dose, will usually lead to the desired dose distribution. The whole point of using EUD or TCP is to be able to judge/compare heterogenenous dose distributions in the PTV, which clinicians mostly tend to avoid. IMRT can achieve dose homogeneity in the target volume better and/or more easily than CRT, but it will not make a big difference whether DVH points or EUD objectives are used in the cost function. 2. The only EUD model implemented in a few commercial TPS is Niemierko's generalized EUD (see next section for more detail). In principle, this model can also be used to optimise the dose in the target volume by setting the volume parameter to a negative value, and a value of a between 10 and 20 has been suggested as appropriate for the target volume [29]. However, this is really nothing else than tweaking a parameter model until it produces dose distributions we are used to. At the moment, there aren't publications fitting the tumor control rate with the generalized EUD, and probably with a goodreason. In fact, although the radiobiological models are phenomenologicaland/or make some assumptions (i.e. simplifications) to treat the problem with a realistic approach and using available clinical data,still there are models that are more sound that other. When it comes to dose-effect relation for tumors, models based on Poisson statistics, i.e. assuming independent behaviour of tumor cells, are largely preferred to models using a volume parameter implemented as in the generalized EUD. As a consequence, EUD or TCP-based optimisationis more likely to enter the clinical practice when models such as the original EUD formulation[19] or the Webb-Nahum model[26] will be implemented in commercial TPS. Even then, a problem will remain, i.e. the current scarcity of analysis of clinical data producing fits for the model parameters. geud and NTCP-based cost functions for the organs at risk. Because of the limitations of DVH-based cost functions, in the last years there has been a growing interest in assessing the potential of new tools, with a particular interest in the so called 'generalized Equivalent Uniform Dose (geud)'[20], which is a way to "summarize" the whole dose distribution in a volume of interest into a single figure. This definition of EUD is identical to the effective uniform dose proposed by Mohan et al. [18] and can be derived from the DVH reduction scheme of the Lyman-Kutcher-Burman NTCP model 1. The geud has been effectively used as a method to control the irradiation of the organs at risk (OARs). (see e.g. [29, 28, 24]). What makes EUD particularly interesting is the presence of the volume dependence parameter (1/a in Niemierko's formulation, n in the LKB formalism)) that can be easily and effectively used to control the dose in a Volume of Interest (VOI), thus allowing to tune the balance between small volumes receiving high doses vs. large volume receiving low doses. Furthermore, the consistency between EUD de_nition and LKB formalism makes it possible to define EUD-based objectives where the volume parameter is not just a number that turns out to work, but rather the result of a dose-response relation available in the literature and calculated on a large number of patients. The EUD has useful features when used in optimising IMRT plans. First, it is an effcient optimisation tool. By expressing dose optimization objectives for OARs in terms of EUD, one can rapidly generate and evaluate a variety of dose distributions that would be much more cumbersome to produce with DVH-based optimisation Furthermore by combining the EUD as a parameter and a sigmoidal function a modifiers, it is possible to obtain a significant sparing of the OARs with respect to an optimisation based on DVH points and quadratic cost functions[29]. Of course, the EUD can also be used as a pure optimisation tool, using the parameter values that best 'steer' the optimisation in the desired direction. For instance, one can decide to optimize IMRT plans for prostate treatment in the clinical practice using two EUD objectives for the rectal wall, setting e.g n = 0:06 for the first objective and n = 1 for the second. The first

13 value of n has been proposed by Rancati et al.[22] as a good predictor of serious side effects (Grade _ 3). The EUD objective with n = 1, however, is not the result of a radiobiological study, but rather a useful way to control the rectum wall irradiation over the whole dose range. This approach, by the way, leads to treatment plans that usually comply with commonly used DVH thresholds even if these thresholds are not part of the cost function. In addition to that, it allows reducing the rectal wall irradiation for doses below 65 Gy, which could not be achieved using the aforementioned DVH points. This is just an example of how EUD can be used with or without an explicit 'biological' meaning, and how these two approaches can coexist even within the same cost function. Current and future developments Functional imaging & biological optimisation. When biological optimisation is discussed, the attention usually focuses on the use of biological models in the cost function. There is, however, another aspect to be considered, i.e. the inclusion of patient-specific biological (functional) data in the optimization problem. A thorough discussion of the issues involved in the integrations of functional imaging in the optimization loop is beyond the scope of this text. Here it is sufficient to say that the inclusion of functional information is the natural evolution of dose optimization based on biological measures, where patient-specific information are taken into account, not only from the anatomical point of view (as in the CT), but also from the functional aspects of the tumor and/or the healthy tissues. It is still very diffcult to reliably map the heterogeneous tumor activity. However, the field of functional imaging is in rapid development, in particular when it comes to Proton Emission Tomography (PET). The increasing availability of functional images thus spurred the interest for the so-called 'dose painting by number', an expression coined by S. Bentzen ([3] to indicate the possibility of 'painting' the dose in the target, voxel by voxel, following the functional patterns obtained via imaging. The path towards dose painting by number is far from easy, because obtaining reliable quantitative information via PET imaging is still largely an unsolved problem. Still, at least from the optimization problem, dose painting is the logical extension of TCP (or EUD) optimisation. Alber and colleagues ([2]), for instance, proposed to associate to each voxel a 'dose efficiency' value e. In theory, one could think of using a functional map to determine e, but the most realistic approach at the moment consist in using the intensity map only to determine an overdosage with respect to a prescribed dose. As a consequence, prescribed dose Dp and maximum dose Dmax are defined a priori, and efficiency will be calculated asv oxels with an intensity less than the average will have effciency equal to 1 and will not be overdosed. The remaing voxels, with higher signal and therefore higher tumor activity, will have an efficiency lower than 1, so the optimization will have an incentive in boosting the dose there in order to reach the desired TCP value. Setting Dmax a priori will ensure that no region will be irradiated at arbitrarily high doses. Biological optimisation in hadrontherapy Hadrontherapy, in particular when delivered with protons, is gaining significant interest these days. This technique has mostly been applied to quite rare tumors, where little data was available from photon therapy in terms of quantitative dose{response relations. It is likely that in the near future protontherapy will be applied, or at least tested, on the whole range of tumors treated in radiotherapy, thus raising questions about the applicability to protons of the dose{response data obtained in photon therapy, in particular those for organs showing a large volume e_ect. For example, if we consider lung irradiation, proton therapy can dramatically reduce the volume receiving low doses with respect to the conventional treatments, even when delivered with IMRT. The lung 'mean dose model', which is based on clinical data where a large lung volume receives a dose in the order of a few Gy, might not be that useful in protontherapy. It is likely that insights on the possible risk for lung radiation damage caused by proton therapy will come from the experience of stereotactic treatments (SRT) rather than from CRT and IMRT.

14 There is another issue with hadrons, due to the fact that their Relative Biological Effectiveness(RBE) is greater than one. Strictly speaking, the RBE is quite a different way to include radiobiological information in the planning procedure with respect to TCP and NTCP models: in principle, one could think of RBE as a 'dose modification factor' at a microscopic level, that is dependent on several parameters (radiation type and energy, dose, alfa/beta ratio of the irradiated tissue, etc.) and that in the end will affect the TCP or NTCP value when a radiation beam with an RBE different from one is applied. In this respect, RBE models the radiation effects at a much lower level than NTCP/TCP metrics, which summarize the expected result for the whole organ (or tumor) level. The issue of an accurate assessment of RBE affects ions much more than protons. In protontherapy is now common practice to assign a constant value of 1.1 to the RBE. There are in-vitro data suggesting that some variability does occur, in particular at the end of the spread-out Bragg Peak, but in-vivo data are much less conclusive, thus leading to the choice of a constante RBE. In protontherapy, which is the most popular form of hadrontherapy these days, one could think on introducing a variable RBE in the optimization process. A recent study analyzed the possibility of taking advantage of the increased linear energy transfer (LET) of protons at the end of range to increase the effectiveness of treatments [9]. Although the increase of LET for protons while slowing down is nowhere near what is seen in heavier ions, and although the LET can not be translated to RBE in a straightforward way, the authors could demonstrate that different IMPT delivery schemes (i.e. 3D-IMPT and Distal Edge Tracking [15]) producing basically the same physical dose distribution are associated to different LET distributions, both in the target and in the OARs. The potential clinical implications of these differences are still to beexplored, but, at least in principle, it is feasible to perform what the authors name 'biologically motivated optimization'. When it comes to ions, the issue is even more complex, partly because the RBE goes up to values of about 3 but, most importantly, because the RBE does change considerably along the Bragg curve, and it is not always higher in the target volume than in the healthy tissues. This makes it necessary that the plan optimisation for ion treatments includes a complete RBE model within the optimization loop[12], while, at least to some extent, an agreement has not been reached yet in the hadron therapy community as to which RBE model should be considered the gold standard. References [1] M. Alber, M. Birkner, W. Laub, and F. Nusslin. Hyperion - An integrated IMRT planning tool. In T. Bortfeld and W. Schlegel, editors, Proceedings of the 13th International Conference on the use of Computers in Radiation Therapy, pages 46{48, Heidelberg, Springer. [2] M. Alber, F. Paulsen, S. M. Eschmann, and H. J. Machulla. On biologically conformal boost dose optimization. Phys. Med. Biol., 48:N31 -N35, [3] S. Bentzen. Theragnostic imaging for radiation oncology: dose-painting by numbers. The Lancet Oncology, 6(2):112{117, Feb [4] G. Borst, M. Ishikawa, J. Nijkamp, M. Hauptmann, H. Shirato, R. Onimaru, M. van den Heuvel, J. Belderbos, J. Lebesque, and J. Sonke. Radiation pneumonitis in patients treated for malignant pulmonary lesions with hypofractionated radiation therapy. Radiotherapy and Oncology, 91(3):307{313, [5] J. O. Deasy, V. Moiseenko, L. Marks, K. S. C. Chao, J. Nam, and A. Eisbruch. Radiotherapy dose-volume effects on salivary gland function. International journal of radiation oncology, biology, physics, 76(3 Suppl):S58-63, [6] T. Dijkema, C. Raaijmakers, R. Ten Haken, J. Roesink, P. Braam, A. Houweling, M. Moerland, A. Eisbruch, and C. Terhaard. Parotid gland function after radiotherapy: the combined michigan and utrecht experience. International Journal of Radiation Oncology* Biology* Physics, 78(2):449{453, [7] A. Eisbruch, H. M. Kim, J. E. Terrell, L. H. Marsh, L. A. Dawson, and J. A. Ship. Xerostomia and its predictors following parotid-sparing irradiation of head-and-neck cancer. Int. J. Radiat. Oncol. Biol. Phys., 50:695{704, 2001.

15 [8] A. Eisbruch, R. K. Ten Haken, H. M. Kim, L. H. Marsh, and J. A. Ship. Dose, volume, and function relationships in parotid salivary glands following conformal and intensity-modulated irradiation of head and neck cancer. Int. J. Radiat. Oncol. Biol. Phys., 45:577{587, [9] C. Grassberger, A. Tro_mov, A. Lomax, and H. Paganetti. Variations in linear energy transfer within clinical proton therapy _elds and the potential for biological treatment planning. International Journal of Radiation Oncology* Biology* Physics, 80(5):1559{1566, [10] P. Kallman, B. K. Lind, and A. Brahme. An algorithm for maximizing the probability of complication-free tumour control in radiation therapy. Phys. Med. Biol., 37:871{890, [11] M. L. Kessler, D. L. McShan, M. A. Epelman, K. A. Vineberg, A. Eisbruch, T. S. Lawrence, and B. A. Fraass. Costlets: A generalized approach to cost functions for automated optimisation of IMRT plans. Opt. Eng., [12] G. Kraft and S. D. Kraft. Research needed for improving heavy-ion therapy. New Journal of Physics, 11(2):025001, [13] S. Levegruen, a. Jackson, M. J. Zelefsky, M. W. Skwarchuk, E. S. Venkatraman, W. Schlegel, Z. Fuks, S. a. Leibel, and C. C. Ling. Fitting tumor control probability models to biopsy outcome after three-dimensional conformal radiation therapy of prostate cancer: pitfalls in deducing radiobiologic parameters for tumors from clinical data. International journal of radiation oncology, biology, physics, 51(4):1064{80, Nov [14] S. Levegruen, a. Jackson, M. J. Zelefsky, E. S. Venkatraman, M. W. Skwarchuk, W. Schlegel, Z. Fuks, S. a. Leibel, and C. C. Ling. Analysis of biopsy outcome after three-dimensional conformal radiation therapy of prostate cancer using dose-distribution variables and tumor control probability models. International journal of radiation oncology, biology, physics, 47(5):1245{60, July [15] A. Lomax. Intensity modulation methods for proton radiotherapy. Phys. Med. Biol, 44:185{205, [16] L. B. Marks, S. M. Bentzen, J. O. Deasy, F.-M. S. Kong, J. D. Bradley, I. S. Vogelius, I. El Naqa, J. L. Hubbs, J. V. Lebesque, R. D. Timmerman, M. K. Martel, and A. Jackson. Radiation dose-volume effects in the lung. International journal of radiation oncology, biology, physics, 76(3 Suppl):S70{6, [17] J. M. Michalski, H. Gay, A. Jackson, S. L. Tucker, and J. O. Deasy. Radiation dose-volume e_ects in radiation-induced rectal injury. International journal of radiation oncology, biology, physics, 76(3 Suppl):S123{9, [18] R. Mohan, G. S. Mageras, B. Baldwin, L. J. Brewster, G. J. Kutcher, S. Leibel, C. M. Burman, C. C. Ling, and Z. Fuks. Clinically relevant optimization of 3-D conformal treatments. Med. Phys., 19:933{944, [19] A. Niemierko. Reporting and analyzing dose distributions: a concept of equivalent uniform dose. Med. Phys., 24:103{110, [20] A. Niemierko. A generalized concept of equivalent uniform dose (EUD). Med. Phys., 26:1100, [21] S. T. H. Peeters, M. S. Hoogeman, W. D. Heemsbergen, A. A. M. Hart, P. C. M. Koper, and J. V. Lebesque. Rectal bleeding, fecal incontinence, and high stoll frequency after conformal radiotherapy for prostate cancer: normal tissue complication probability modeling. International Journal of Radiation Oncology, Biology, Physics, 66(1):11{19, [22] T. Rancati, C. Fiorino, G. Gagliardi, G. M. Cattaneo, G. Sanguineti, V. C. Borca, C. Cozzarini, G. Fellin, F. Foppiano, G. Girelli, L. Menegotti, A. Piazzolla, V. Vavassori, and R. Valdagni. Fitting late rectal bleeding data using different NTCP models: results from an Italian multi-centric study (AIROPROS0101). Radiother. Oncol., 73:21{32, [23] J. M. Roesink, M. A. Moerland, J. J. Battermann, G. J. Hordijk, and C. H. Terhaard. Quantitative dose-volume response analysis of changes in parotid gland function after radiotherapy in the head-and-neck region. Int. J. Radiat. Oncol. Biol. Phys., 51:938{946, [24] M. Schwarz, J. V. Lebesque, B. J. Mijnheer, andand E. M. F. Damen. Sensitivity of treatment plan optimisation for prostate cancer using the equivalent uniform dose (EUD) with respect to the rectal wall volume parameter. Radiother. Oncol., 73:209{218, 2004.

16 [25] Y. Seppenwoolde, J. V. Lebesque, K. De Jaeger, J. S. Belderbos, L. J. Boersma, C. Schilstra, G. T. Henning, J. A. Hayman, M. K. Martel, and R. K. TenHaken. Comparing di_erent NTCP models that predict the incidence of radiation pneumonitis. Normal tissue complication probability. Int. J. Radiat. Oncol. Biol. Phys., 55:724{735, 2003 [26] S. Webb and A. E. Nahum. A model for calculating tumour control probability in radiotherapy including the effects of inhomogeneous distributions of dose and clonogenic cell density. Phys. Med. Biol., 38:653{666, [27] M. G. Witte, J. van der Geer, C. Schneider, J. V. Lebesque, M. Alber, and M. van Herk. IMRT optimization including random and systematic geometric errors based on the expectation of TCP and NTCP. Medical Physics, 34(9):3544, [28] Q. Wu, D. Djajaputra, Y. Wu, J. Zhou, H. H. Liu, and R. Mohan. Intensitymodulated radiotherapy optimization with geud-guided dose- volume objectives. Phys. Med. Biol., 48:279{291, [29] Q. Wu, R. Mohan, A. Niemierko, and R. Schmidt-Ullrich. Optimization of intensity-modulated radiotherapy plans based on the equivalent uniform dose. Int. J. Radiat. Oncol. Biol. Phys., 52:224{235, 2002.

17 Le nuove frontiere: gestione del movimento respiratorio nell era dell imaging multimodale in radioterapia New frontiers: management of respiratory motion in the era of multimodality imaging in radiation oncology C. Cavedon Azienda Ospedaliera Universitaria Integrata - Verona 1. Introduction Control of respiratory motion has been a topic of special interest in radiation therapy (RT) in recent years. Such interest was primarily driven by the translation of stereotactic approaches from intracranial applications to anatomic regions affected by respiratory motion, where control over time with a resolution of tens or hundreds of milliseconds is a key factor for treatment success or at least thought to be such, while more clinical evidence is desired [1]. Undoubtedly, regimes of high dose administered in few fractions proved to be effective to enhance local control rates and though less dramatically overall survival, for example, in selected lung tumors [2]. Based on the reasonable assumption that smaller irradiated volumes would allow higher dose to be delivered while keeping the toxicity to healthy tissues acceptable, major manufacturers of medical technology in radiation oncology have proposed devices for the explicit management of respiratory motion, ranging from gating techniques [3,4] to sophisticated motion tracking systems some of which already in use [5,6], others announced but not available yet [7,8]. The rapidly evolving field of particle therapy is also likely to contribute to the development of respiratory motion management techniques, as spatial inaccuracy translates into dosimetric uncertainty in particle therapy more than with any other radiotherapy modality [9]. 2. Background Roughly, control of respiratory motion can be divided in two categories: passive and active methods. Passive methods do not make use of explicit actions following the measurement of a respiration state. A typical passive method in radiation oncology consists of estimating the most probable trajectory of a target object and planning so that the envelope of all positions is likely to be encompassed by the treatment within an acceptable probability level. A typical active control of motion is respiratory gating [3]. Gating techniques consist in determining whether a parameter that describes the respiratory state is within a pre-defined window of acceptability, and in taking actions (typically, a binary decision between beam on or beam off ) as a consequence of the measurement. Gating techniques suffer from several drawbacks such as the poor correlation between a descriptive parameter (the surrogate signal) and the actual respiratory state [4], and the discretization of the respiratory cycle in finite bins, inevitably causing a residual geographical miss to persist reduced only at the expense of longer treatment times and poorer duty cycles [10]. Tracking methods were introduced to obviate to some of the problems mentioned above. A tracking strategy consists of actively modifying the treatment geometry in real time to conform to the actual breathing phase without losing beam-on time. The most clinically-applied tracking strategy in radiation oncology consists of redirecting the treatment beam by means of a robot [5], however other strategies now seem to be ready for routine clinical application such as gimbals-mounted linacs [6], couch-tracking [11-12], and MLC-tracking [7-8,13]. While eliminating or reducing problems related to the poor duty cycle, however, tracking methods suffer from difficult real-time adaptation penalized by relatively long latency times (currently in the range ms) that might cause the treatment to fall frequently off-target, if not adequately accounted for. This drawback (less critical for the

18 simpler gating techniques) led to the development of refined prediction algorithms [14-16] capable of estimating the position of the target after the latency time and of redirecting the treatment accordingly. Synthetically, the most advanced methods of explicit control of respiratory motion in radiation therapy are composed by four steps: 1. detection of a surrogate (descriptive) signal 2. estimation of the actual target position through a correlation model 3. prediction of position after latency interval 4. treatment adaptation (steps 2 and 3 might be simultaneous or exchanged in order). 3. Respiratory motion control in imaging for treatment planning and verification. Given the advanced methods available today and foreseeable in the near future for treatment delivery, it becomes more and more important to adapt imaging modalities and procedures to motion compensation, in order to provide a consistent and reliable framework for treatment planning and verification. CT is still the main modality for treatment planning, also because of its 3D spatial resolution and freedom from geometric distortion [17]. Active control of respiratory motion in CT is available since at least 10 years ago. Many examples exist of techniques aimed at accounting for respiratory motion in CT, including amplitude- or phasebased 4DCT, whose development has been mostly driven by the ever-increasing demand for accuracy in RT imaging techniques, where description of anatomical modifications over several respiratory cycles is increasingly recognized as a factor for possible treatment success. 4DCT datasets are used today in treatment planning, for motion encompassing techniques as well as gating and tracking approaches. However, it is recognized that other treatment modalities add significant information not only to the diagnostic process but also to planning of therapeutic procedures, especially in high-precision radiation therapy where a differentiation between regions of a planned target might be used for individualized prescription (e.g., simultaneous integrated boost (SIB) irradiation, dose painting by numbers strategies etc.) In particular, the need for multimodality imaging with control of respiratory motion is mostly driven by the ever-increasing need for quantitative information [17]: motion is a problem in morphological imaging itself and becomes a dangerous source of error if quantitative information is used to define a target volume. Nuclear medicine is the typical field where imaging combined to quantification is the main purpose and source of information. The clinical introduction of PET-CT scanners at the beginning of the 21st century [18] and their rapid development in the following years has allowed physicians to extract ever-increasing information from positron emission tomography in the last decade. PET-CT imaging in radiation oncology has been used since the appearance of PET-CT [19] and has been subject to development of motion control techniques in parallel to radiation therapy devices [20]. Available motion control techniques today span from gating [21], to advanced motion compensation methods based on the analysis and processing of raw data [22-23], with a variety of hybrid methods that aim at taking advantage from both approaches [24]. While gating suffers from poor signal-to-noise ratio due to the low statistics of binned counts, clinical implementations of methods capable of recovering the full information are still sparse and, frequently, not validated yet. However, methods that allow the whole count statistics to be preserved and used for image formation would enable acquisitions brief enough to be routinely adopted in the clinical setting [25]. These might include methods based on deformable image registration, image reconstruction based on priorknowledge, and the combination of both [26-27], though much work has yet to be done before these techniques become clinically available. The matching between 4DCT and 4DPET data both for spatial localization and attenuation correction purposes - will also be subject to further study. This point is critical especially in the development of PET-MR systems [28], where motion control for the associated MR dataset is far from trivial (see also below).

19 Parallel to the use of PET-CT in radiation oncology, motion control has already started to provide more accurate quantitative information in diagnostic PET-CT [29]. This is of special importance in the quantification of the standardized uptake value (SUV) or other quantitative or semi-quantitative metrics. In fact, motion blurring might act both enlarging the perceived uptake volume and reducing it. Enlargement occurs for example when visual evaluation is adopted, while reduction happens if quantitative criteria such as a SUV threshold are used (other oversimplified quantification algorithms might actually result in the opposite) [10,12]. Hence, volumes estimated based un SUV maps might be under- or over-estimated as a function of the method used for analysis, while SUV values suffer from underestimation due to motion in general (fig. 1). Fig. 1: SUVmax as a function of the number of phase bins in gated 4D-PET-CT. Lung nodule with total excursion 18 mm in sup-inf direction ml volume as seen in CT. The dotted line shows a cubic spline interpolation. As in CT and PET, motion control in MRI is recognized as a promising technique for accurate treatment planning and dose calculation in radiation oncology [30]. However, 4D-MRI is not widely used in treatment planning yet, but it is conceivable that developments in MRI motion control made for diagnostic purposes will be translated to RT as well. Multimodality deformable image registration not an easy task, and still research-demanding before being widely introduced into the clinical practice today seems the most probable candidate as a tool for the full integration of motion-controlled MRI in treatment planning [31]. Motion control in MRI is commonly used in cardiac imaging to compensate for cardiac motion; compensation of breathing movements in anatomical sites that move with respiration is less frequently used in the clinical practice [32-33]. However, major manufacturers of MRI technology have been equipping their systems with motion control devices or methods since the introduction of clinical scanners, from strain-gauge belts to pencil-beam excitation used as an internal surrogate signal [34-35]. Third-party manufacturers offer devices to interface to MRI systems and provide respiratory motion control. The use of multi-element coils and parallel imaging is increasingly facilitating the development of motion-correlated MRI [36]. Researchers have proposed methods based on image registration and image processing to extract time-series from MRI data [37-38]. Some of the above methods, however, are still cumbersome to use in clinical practice or not commercially available yet. On the other hand, the demand for motion control is implicitly increasing due to the ever-wider use of quantitative methods in MRI, including estimation of perfusion, diffusion and oxygenation [39], methods that were originally developed for the brain and are now increasingly used in full-body applications. Similarly to CT, extraction of accurate quantitative

20 information in MRI cannot be performed without motion control: at least, robust and reliable methods for subtraction and/or comparison of two or more spatially-correlated images are required. As quantification gains importance in routinely-performed examinations, respiratory movements must be accounted for in an ever simpler, more practical way: this will require further research and additional effort by the industry. Given the importance of multi-parametric imaging including several quantification techniques, this will probably happen in the next few years. 4. Conclusion Control of respiratory motion in radiation therapy has seen a rapid development in recent years and new methods are increasingly being introduced in the clinical practice. At the same time, quantitative imaging is growing and providing new information to be used in the treatment planning process. It is essential that imaging techniques for treatment planning and verification be adapted to the available delivery options: the temporal dimension is becoming essential not only in CT (e.g. in 4DCT), but also in PET-CT and MRI. Temporal accuracy achieved by means of multiple strategies and mediated between different modalities by deformable registration in particular - will probably allow users to exploit the full potential of modern treatment systems. References: [1] P. Keall, G.S. Mageras, J.M. Balter, R.S. Emery, K.M. Forster, S.B. Jiang, J.M. Kapatoes, D.A. Low, M.J. Murphy, B.R Murray, C.R. Ramsey, M.B. Van Herk, S.S. Vedam, J.W. Wong, E. Yorke, The management of respiratory motion in radiation oncology report of AAPM TG 76. Med. Phys. 33, (2006). [2] R. Timmerman, R. Paulus, J. Galvin, J. Michalski, W. Straube, J. Bradley, A. Fakiris, A. Bezjak, G. Videtic, D. Johnstone, J. Fowler, E. Gore, H. Choy, Stereotactic Body Radiation Therapy for Inoperable Early Stage Lung Cancer, JAMA 303, (2010). [3] H.D. Kubo, B.C. Hill, Respiration gated radiotherapy treatment: a technical study, Phys. Med. Biol. 41, (1996). [4] R.I. Berbeco, S. Nishioka, H. Shirato, G.T.Y. Chen, S.B. Jiang, Residual motion of lung tumours in gated radiotherapy with external respiratory surrogates, Phys. Med. Biol. 50, (2005). [5] A. Schweikard, H. Shiomi, J. Adler, Respiration tracking in radiosurgery, Med. Phys. 31, (2004). [6] T. Depuydt, D. Verellen, O. Haas, T. Gevaert, N. Linthout, M. Duchateau, K. Tournel, T. Reynders, K. Leysen, M. Hoogeman, G. Storme, M. De Ridder, Geometric accuracy of a novel gimbals based radiation therapy tumor tracking system, Radiother. Oncol. 98, (2011). [7] A. Sawant, R. Venkat, V. Srivastava, D. Carlson, S. Povzner, H. Cattell, P. Keall, Management of threedimensional intrafraction motion through real-time DMLC tracking, Med. Phys. 35, (2008). [8] M.B. Tacke, S. Nill, A. Krauss, U. Oelfke, Real-time tumor tracking: automatic compensation of target motion using the Siemens 160 MLC, Med. Phys (2010). [9] C. Grassberger, S. Dowdell, A. Lomax, G. Sharp, J. Shackleford, N. Choi, H. Willers, H. Paganetti, Motion interplay as a function of patient parameters and spot size in spot scanning proton therapy for lung cancer, Int J Radiat Oncol Biol Phys. 86, (2013). [10] Q. Ren, S Nishioka, H Shirato, R Berbeco, Adaptive external gating based on the updating method of internal/external correlation and gating window before each beam delivery, Phys. Med. Biol. 57, N145 N157, [11] W. D. D Souza, S. A. Naqvi, and C. X. Yu, Real-time intra-fractionmotion tracking using the treatment couch: A feasibility study, Phys. Med. Biol. 50, (2005).

21 [12] J. Wilbert, K. Baier, A. Richter, C. Herrmann, L. Ma, M. Flentje, and M. Guckenberger, Influence of continuous table motion on patient breathing patterns, Int. J. Radiat. Oncol., Biol., Phys. 77, (2010). [13] P.R. Poulsen, W. Fledelius, B. Cho, P. Keall, Image-based dynamic multileaf collimator tracking of moving targets during intensity-modulated arc therapy, Int J Radiat Oncol Biol Phys. 83, e265-e271 (2012). [14] F. Ernst, R. Dürichen, A. Schlaefer, A. Schweikard, Evaluating and comparing algorithms for respiratory motion prediction, Phys Med Biol. 58, (2013). [15] S.S. Vedam, P.J. Keall, A. Docef, D.A. Todor, V.R. Kini, R. Mohan, Predicting respiratory motion for four-dimensional radiotherapy, Med Phys. 31, (2004). [16] M.J. Murphy, S. Dieterich, Comparative performance of linear and nonlinear neural networks to predict irregular breathing, Phys Med Biol. 51, (2006). [17] J.M. Boone, Radiological interpretation 2020: Toward quantitative image assessment, Med. Phys. 34, (2007). [18] T. Beyer, D.W. Townsend, T. Brun, P.E. Kinahan, M. Charron, R.Roddy, J. Jerin, J. Young, L. Byars, and R. Nutt, A combined PET/CT scanner for clinical oncology, J. Nucl. Med. 41, (2000). [19] J. Rosenman, Incorporating functional imaging information into radiation treatment, Semin Radiat Oncol. 11, (2001). [20] L.A. Dawson, M.B. Sharpe, Image-guided radiotherapy: rationale, benefits, and limitations, Lancet Oncol. 7, (2006). [21] G. Chang, T. Chang, J.W. Clark, O.R. Mawlawi, Design and performance of a respiratory amplitude gating device for PET/CT imaging, Med Phys. 37, (2010). [22] X. Jin, C. Chan, T. Mulnix, V. Panin, M.E. Casey, C. Liu, R.E. Carson, List-mode reconstruction for the Biograph mct with physics modeling and event-by-event motion correction, Phys Med Biol. 58, (2013). [23] D. Wallach, F. Lamare, G. Kontaxakis, D. Visvikis, Super-resolution in respiratory synchronized positron emission tomography, IEEE Trans Med Imaging. 31, (2012). [24] C. Liu, A.M. Alessio, P.E. Kinahan, Respiratory motion correction for quantitative PET/CT using all detected events with internal-external motion correlation, Med Phys. 38, (2011). [25] T. Pan and O. Mawlawi, PET/CT in radiation oncology, Med. Phys. 35, (2008). [26] F. Qiao, J. Clark, T. Pan, O. Mawlawi, Expectation Maximization Reconstruction of PET Image with Non-rigid Motion Compensation, Conf Proc IEEE Eng Med Biol Soc. 4, (2005). [26] C. Cavedon, E. Berna, P. Francescon, S. Cora, P. Scalchi, P. Polloniato, N. Satariano, P. Zanco, Building a PET Time Series Using Information From 4DCT Data, Med. Phys. 33, 2027 (2006). [28] D.A. Torigian, H. Zaidi, T.C. Kwee, B. Saboury, J.K. Udupa, Z.H. Cho, A. Alavi, PET/MR imaging: technical aspects and potential clinical applications, Radiology 267, (2013). [29] B.K. Teo, B. Saboury, R. Munbodh, J. Scheuermann, D.A. Torigian, H. Zaidi, A. Alavi, The effect of breathing irregularities on quantitative accuracy of respiratory gated PET CT, Med Phys. 39, (2012). [30] D. Boye, T. Lomax, A. Knopf, Mapping motion from 4D-MRI to 3D-CT for use in 4D dose calculations: A technical feasibility study, Med. Phys. 40, (2013). [31] K.K. Brock, L.A. Dawson, M.B. Sharpe, D.J. Moseley, D.A. Jaffray, Feasibility of a novel deformable image registration technique to facilitate classification, targeting, and monitoring of tumor and normal tissue, Int J Radiat Oncol Biol Phys. 64, (2006). [32] C. Plathow, S. Ley, C. Fink, M. Puderbach, W. Hosch, A. Schmähl, J. Debus, H.U. Kauczor, Analysis of intrathoracic tumor mobility during whole breathing cycle by dynamic MRI, Int. J. Radiat. Oncol. Biol. Phys. 59, (2004).

22 [33] M.C. Liszewski, F.W. Hersman, T.A. Altes, Y. Ohno, P. Ciet, S.K. Warfield, E.Y. Lee, Magnetic resonance imaging of pediatric lung parenchyma, airways, vasculature, ventilation, and perfusion: state of the art, Radiol Clin North Am. 51, (2013). [34] R.L. Ehman, M.T. McNamara, M. Pallack, H. Hricak, C.B. Higgins, Magnetic resonance imaging with respiratory gating: techniques and advantages, American Journal of Roentgenology 143, (1984). [35] M. Ishida, A. Schuster, S. Takase, G. Morton, A. Chiribiri, B. Bigalke, T. Schaeffter, H. Sakuma, E. Nagel, Impact of an abdominal belt on breathing patterns and scan efficiency in whole-heart coronary magnetic resonance angiography: comparison between the UK and Japan, J Cardiovasc Magn Reson. 13, 71 (2011). [36] M. von Siebenthal, G. Székely, U. Gamper, P. Boesiger, A. Lomax, Ph. Cattin, 4D MR imaging of respiratory organ motion and its variability, Phys. Med. Biol. 52, (2007). [37] G. Vaillant, C. Prieto, C. Kolbitsch, G. Penney, T. Schaeffter, Retrospective rigid motion correction in k- space for segmented radial MRI, IEEE Trans Med Imaging Jun 14 (Epub ahead of print) [38] A.P. King, C. Buerger, C. Tsoumpas, P.K. Marsden, T. Schaeffter, Thoracic respiratory motion estimation from MRI using a statistical model and a 2-D image navigator, Medical Image Analysis 16, (2012). [39] T. Henzler, G. Schmid-Bindert, S.O. Schoenberg, C. Fink, Diffusion and perfusion MRI of the lung and mediastinum, Eur J Radiol. 76, (2010).

23 Radiotherapy and HIFU: alternative or complementary techniques? Giovanni Borasi 1, Giorgio Russo 2, Filippo Alongi 3, Alan Nahum 4, Giuliana Candiano 5, Alessandro Stefano 6, Maria Carla Gilardi 7, Cristina Messa 8 1) Università di Milano Bicocca, Milano, It ; 2) LATO, Palermo, It ; 3) Istituto Clinico Humanitas, Milano, It ; 4) Clatterbridge Cancer Centre, Bebington, UK; 5) LATO, Palermo, It; 6) LATO, Palermo, It; 7) IBFM-CNR, Milano, It; 8) CNR, Milano, It Corresponding Author: Giovanni Borasi (giovanni.borasi@unimib.it) Radiotherapy and High Intensity Focused Ultrasound To date Radiotherapy (RT) and High-Intensity Focused Ultrasound (HIFU or FUS) have gone in totally different directions in the broad field of Oncology. External-Beam Radiotherapy (EBRT), relying on the ionizing effect of high-energy radiation, has been developed continuously over the last 115 years, from low-energy X-rays, through Cobalt-60 gamma rays to megavoltage x-rays and electrons from linear accelerators (LINACS). High energy photon and electron beams have progressively increased the therapeutic ratio (loosely defined as the local tumor control probability for a given, low complication rate). Nowadays, external photon beams can be delivered to precise, irregular targets via many modalities such as three-dimensional conformal therapy (3DCRT), intensity-modulated radiotherapy (IMRT) and intensity modulated Arc therapy (IMAT). All these modalities further concentrate energy deposition in the target, sometimes at the expense of an increased volume of tissue exposed to low dose (with as yet no evidence of negative effects). Interstitial and endocavitary RT have evolved from Radium needles and tubes to modern computerized Brachytherapy devices[1]. Parallel to improved sources and delivery methods, progress in radiotherapy has also benefited from the progressive availability of CT, MR and CT/PET 3D digital maps. The different anatomical and metabolic content of these images enables tumor characterization down to the biological level (the biological target ). In addition to computer-controlled multileaf collimators, which have made intensity modulation practicable, modern LINACS include a cone beam CT imaging system [2;3] or their geometry is directly borrowed from a CT-Scanner [4]. Also an industrial robot, incorporating a short in-line LINAC, is in use for stereotactic therapy[5]. On the other hand, ionization at the subatomic level (with direct damage to cell DNA and an indirect damage to water molecules, creating aggressive chemical species) is spatially sparse (i.e. low LET), making the radiation effect highly dependent on chemical damage, i.e. on the local Oxygen content. As a consequence, hypoxic regions of tumours, which are almost always present, are less sensitive to ionizing radiation and represent a primary source of local failure and adverse patient outcome [6]. To partially overcome this fundamental limitation, high-let radiation modalities are under development (principally heavy charged particles). With few exception (for example Protons for eye tumors), the high cost of these facilities makes it improbable that these will ever be available solution for more than a small fraction of cancer patients [7]. It is believed by some that metabolic administration of short range alpha emitters in targeted vectors could play a greater role in cancer treatment, even in the metastatic phase, with a personalized approach[8]. A comprehensive treatment can be found in Modern Practices in Radiation Therapy [9]. Just as radiotherapy can be thought of as the application of x-rays to therapy, HIFU can be considered as the therapeutic version of ultrasound (US) echography. In the latter case, the energy is not carried by a high-energy photon or electron beam but by an elastic longitudinal wave, of millimeter wavelength, which is focused on the target. The source is a piezoelectric, spherically curved, vibrating shell, known as a transducer. This source in contained in a degassed water sink (or pillow) and should be tightly coupled to the patient skin. As for

24 echography, air represents a totally reflecting and bone a highly absorbing medium. Like photons, the HIFU penetration of organic matter depends on frequency (in an inverse way!) but is much more dependent on the nature of the traversed medium (absorption, reflection and refraction) and the effects are of a highly diverse nature (mechanical stress, heat production, cavitation, micro streaming, etc.). HIFU is much younger than RT, beginning in the 1940s. The renaissance of interest in HIFU dates, however, from the 90s, when imaging techniques, like MR and US, made it possible to visualize the beam focus and even measure its temperature. It s worth mentioning that in RT it s impossible to view the beam directly and its path can only be obtained by calculation. Another fundamental difference is that even the best hadron therapy depth-dose curve cannot compete with the focused energy concentration obtainable with HIFU. These fundamental characteristics stimulated the development of HIFU systems along three main directions: US guided, Prostate dedicated[10;11], US[12;13] or MR[14;15] guided total body systems. HIFU is employed to treat several oncological pathologies cured curable also with RT, like bone metastasis, liver, pancreas, breast, prostate, rectum, etc..one of the most promising fields of application is drug delivery ; this means the localized activation, thanks to heat or mechanical effects, of drugs included in specialized vectors (liposomes, nanoparticles, micro bubbles, etc.). The relevant experiments, restricted, until now, to animal models, frequently involve chemotherapeutic agents and one of the main target is the reversible opening of the blood brain barrier (BBB) [16;17]. In summary, there is an exponentially growing interest in HIFU. Excellent reviews of the principles can be found in [18] and [19]. FUS/RT: a new weapon in oncology As they are based on very different physical principles, the two techniques have complementary strengths and weaknesses. We call the combination of the two modalities FUS/RT. There are, at least, four good reasons to implement FUS/RT in the clinical practice. 1) The different, sometimes opposite, behavior in penetrating human tissues, in particular bone and air. This was evident in our experience of treating bone metastases with MRgFUS for palliative purposes. A comprehensive discussion on this point can be found in our previous paper[20]. 2) HIFU is insensitive to hypoxia, which is, as we stated above, a primary source of local failure and adverse patient outcome for EBRT. In fact, when the tumour has become quite large, the central part, somewhat compressed by angiogenesis, has a poorer blood supply and tends to become hypoxic. In these conditions, photon radiation loses a great deal of its destructive capability, which relies on well oxygenated clonogens. This seems exactly the perfect target for HIFU, whose effect is not significantly dependent on the oxygen content. On the other hand, the peripheral regions of the tumour, where the oxygen supply is good, but, also, cell proliferation may be high, is certainly a better target for radiation, and a good sterilization of quite large volumes may be possible in a reasonable time. The precise localization of the hypoxic region inside the tumour, obtained with contrast MR and/or CT PET, opens the door to truly personalized treatments. 3) The hyperthermia field, generated by the HIFU treatment, could represent a potent enhancer of the radiation effect. But what would be the optimal timing between HIFU and EBRT? The application of HIFU generates a lot of heatwhich diffuses away from the focus toward external tissue regions, while the temperature progressively decreases. It is well known that mild hyperthermia (HT) increases the blood supply which can make ionizing radiation much more effective. A good example comes from the treatment of primary, locally advanced, Cervix Cancer, where the beneficial effect of adding HT to EBRT (Thermoradiotherapy) has been confirmed on a large patient population (378 cases) [21]. Of course the positive effect of HT doesn't stop immediately when the heated tissues return to the normal body temperature. The effect may last sometime, depending on several conditions, and this could explain the good results obtained from

25 combining HIFU and EBRT, even at intervals of several hours[22]. However, the optimal timing recipe would be: "EBRT immediately after HIFU, or, even better, at the same time"[23]. In this way the heat produced by HIFU, in a time of just a few minutes, would became a powerful enhancer for the concomitant EBRT. Calculations are in progress to quantify the Hyperthermia field generated in different tissues by different ablated volumes. On the other hand, HIFU systems seem to be evolving towards the capability of generating and monitoring specific low-energy pulse sequences, in the hyperthermia range. 4) The Hyperthermia field, generated by the HIFU, or by specific pulse sequences, may make possible heat-mediated drug delivery[24] implementation. A recent experiment[25] involved: (1) pulsed Ultrasound (pfus), (2) EBRT, (3) an antitumor drug (Docetaxel), and combinations (1)+(3), (2)+(3), (1)+(2)+(3) on mice bearing prostate tumours. Quite reasonably the combination of all the three weapons gave the best tumour control. Quite inexplicably, (1)+(2) was not tested. But considering here mainly HIFU and EBRT, this strategy would allow also a great sparing of time and of radiation dose. A lower radiation level means, in turn, a potentially significant reduction of sequels. Animal experiments with FUS/RT are planned. Conclusion While some form of FUS/RT can be quite easily implemented in clinical practice (for example in Prostate treatments using HIFU-dedicated systems), developing a new total body FUS/RT device, or even retrofitting an existing LINAC with HIFU capabilities, would require considerable technical development. However, a FUS/RT device could represent a potent new weapon against cancer and, thanks to drug delivery, also against many degenerative diseases. It would offer the possibility to integrate, in real time and in the same session, target and temperature imaging, radiation sterilization, hyperthermia, ablation and drug delivery. Reference List 1 Nucletron-Elekta AB: SE Stockholm, Sweden, 2 Varian Medical System: 3100 Hansen Way, Palo Alto, CA , USA, 3 Elekta AB: SE Stockholm, Sweden, 4 TomoTherapy: 1240 Deming Way, Madison, WI , USA, 5 Cyberknife: Accuray,1310 Chesapeake Terrace, Sunnyvale, CA 94089, // 6 Vaupel P, Mayer A: Hypoxia in cancer: significance and impact on clinical outcome. Cancer and Metastasis Reviews 2007;26: De Ruysscher D, Lodge MM, Jones B, Brada M, Munro A, Jefferson T, Pijls-Johannesma M: Charged particles in radiotherapy: A 5-year update of a systematic review. Radiotherapy and Oncology 2012;103: Allen BJ, Raja C, Rizvi S, Li Y, Tsui W, Zhang D, Song E, Qu CF, Kearsley J, Graham P, Thompson J: Targeted alpha therapy for cancer. Physics in Medicine and Biology 2004;49: Various Authors, Edited by: Gopishankar Natanasabapathi: Modern Practices in Radiation Therapy. ed Free online edition, Sonablate 500: Sonacare Medical,801 E. Morehead St.Suite 20,Charlotte, NC 28202, 11 Ablatherm HIFU: EDAP TMS S.A., Parc d'activités la Poudrette - Lamartine, 4, rue du Dauphine Vaulx-en-Velin - France,

26 12 JC Model: Chongqing Haifu Medical Technology Co., Ltd, Chongqing, China, FEP-BY: Beijing Yuande Bio-Medical Engineering Co.,Ltd, China Medical Technologies, InSightec Ltd: Tirat Carmel, Israel, 15 Sonalleve MR HIFU: Philips Healthcare, P.O. Box , 5680 DA Best,TheNetherlands, Hynynen K, McDannold N, Vykhodtseva N, Jolesz FA: Noninvasive MR imaging-guided focal opening of the blood-brain barrier in rabbits. Radiology 2001;220: Alkins R, Burgess A, Ganguly M, Francia G, Kerbel R, Wels WS, Hynynen K: Focused Ultrasound Delivers Targeted Immune Cells to Metastatic Brain Tumors. Cancer Research 2013;73: Malietzis G, Monzon L, Hand J, Wasan H, Leen E, Abel M, Muhammad A, Price P, Abel P: Highintensity focused ultrasound: advances in technology and experimental trials support enhanced utility of focused ultrasound surgery in oncology. Br J Radiol 2013;86: Physical Principles of Medical Ultrasonics: ed Second Edition, John Wiley & Sons, Ltd, Borasi G, Russo G, Alongi F, Nahum A, Candiano G.C, Stefano A, Gilardi M.C, Messa C: Highintensity focused ultrasound plus concomitant radiotherapy: a new weapon in oncology? Journal of Focused Ultrasound 2013;1: Franckena M, Lutgens LC, Koper PC, Kleynen CE, van der Steen-Banasik EM, Jobsen JJ, Leer JW, Creutzberg CL, Dielwart MF, van NY, Canters RA, Van Rhoon GC, van der Zee J: Radiotherapy and hyperthermia for treatment of primary locally advanced cervix cancer: results in 378 patients. Int J Radiat Oncol Biol Phys 2009;73: Liu CX, Gao XS, Xiong LL, Ge HY, He XY, Li T, Zhang HJ, Bai HZ, Lin Q, Zhang M, Zhao J, Xiong W, Bai Y, Asaumi J: A preclinical in vivo investigation of high-intensity focused ultrasound combined with radiotherapy. Ultrasound Med Biol 2011;37: Sapareto SA, Raaphorst GP, Dewey WC: Cell killing and the sequencing of hyperthermia and radiation. Int J Radiat Oncol Biol Phys 1979;5: Partanen A, Yarmolenko PS, Viitala A, Appanaboyina S, Haemmerich D, Ranjan A, Jacobs G, Woods D, Enholm J, Wood BJ, Dreher MR: Mild hyperthermia with magnetic resonance-guided highintensity focused ultrasound for applications in drug delivery. Int J Hyperthermia 2012;28: Mu Z, Ma CM, Chen X, Cvetkovic D, Pollack A, Chen L: MR-guided pulsed high intensity focused ultrasound enhancement of docetaxel combined with radiotherapy for prostate cancer treatment. Phys Med Biol 2012;57:

27 Dosimetric characterization and planning of/with fields of small sizes P. Francescon San Bortolo Hospital, Medical Physics Department, Vicenza, Italy INTRODUCTION In modern radiation therapy small megavoltage (MV) photon fields of dimensions less than 3 3cm 2 are being increasingly used. The dosimetry of small photon field sizes down to 1x1.cm 2, for the special case of stereotactic radiotherapy/radiosurgery (SRT/SRS) has been discussed in AAPM Report There are several challenges in the dosimetry of small MV photon fields. These primarily arise because of the occlusion of the direct photon beam source at small collimator settings and the lack of lateral charged particle equilibrium or transient equilibrium (CPE/TCPE). Moreover in small fields where the dose experiences a steep fall-off in dose profiles with no flat part at their centre most widely available detectors are too large to resolve the beam profile and the penumbra. To a lesser extent, variations in radiological parameters due to changes in the particle spectrum with decreasing field size also pose considerations. Although the physics of small field dosimetry has been discussed with reference to the above challenges 2-4 there is an on-going need for guidance on the selection of suitable detectors and methodologies for the determination of small field dosimetric parameters. Notwithstanding the fact that the detectors produced in recent times are getting smaller and smaller, they still cause significant fluence perturbations due to the comparable dimensions of the sensitive volume with the field size, the presence of non-water equivalent materials surrounding the sensitive volume and the increased signal-to-noise ratio due to the smaller sensitive volume. Numerous experimental studies and Monte Carlo (MC) simulations have been performed to investigate the suitability of the various types of detectors in the measurement of percentage depth doses (PDD), tissue maximum ratios (TMR), tissue phantom ratios (TPR) and field size factors (output factors) in small photon fields However it lacks a detailed analysis of the behavior of available dosimeters enabling the choice of an appropriate detector in terms of size, composition and construction. Moreover the additional correction factor k fclin, fmsr Qclin,Q,recently introduced by the IAEA and AAPM task group to account for possible changes in detector msr response when the dimensions of the field are small, up to now has been calculated only to determine S c,p or the f clin, f Ω msr, in the notation of Alfonso et al MC has proven to be an effective tool to calculate correction Q clin,qmsr f factors clin, fmsr k In particular, with MC it is possible to simulate very accurately the detector with its design, Qclin,Qmsr materials and composition, and calculate with high accuracy both the dose deposited within the cavity or sensitive volume and the perturbation correction factors of dosimeters in non-reference conditions Monte Carlo (MC) simulation of dose to water and dose to detector has been used to calculate the correction factors needed for dose calibration and output factor measurements on the CyberKnife system and standard linac 23,24. Moreover Monte Carlo simulation has been used to compare the TMRs, PDDs and OARs calculated in different type of dosimeters and in water. MATERIALS AND METHODS MonteCarlo simulation The simulations were made using the data for the head of the linac and for the dosimeters reported in 24. As reported in 23 we assumed that the reading of the detector corrected for influence quantities is proportional to the dose absorbed in the sensitive volume, then the PDD det, TMR det, OAR det can be computed by MC simulation using the following relationship:

28 where D(x,y,z) represents the total dose per initial history scored within the sensitive volume of the modeled detector. The egs_chamber 35 user code was used to simulate, OAR det, TPR det and PDD det. The cross section enhancement factor was set to 2048 in a volume that extends about 1.5 cm around the sensitive volume of the detector. The egsnrc cross section options and transport parameters were set to default values. The particle production and transport threshold energies were 521 kev (e+/e-) and 1 kev (gamma). Since each OAR det TMR det PDD det value is defined respectively as the ratio of D det (x,y,z 0 )/ D det (0,0,z 0 ), D det (0,0,z)/ D det (0,0,z max ), we calculated these ratios within a single simulation by using the correlated sampling technique 35. This technique, takes advantage of the Intermediate Phase Space Storage volume (IPSS), a user-defined volume which includes the positions of the detector along the profile, or along the depth or around the point of dose calculation. The simulation starts in the geometry outside the IPSS and stops in its surface where the particle phase space is stored. Then, this phase space is used for each position of the detector inside the IPSS. As the state of the number generator is also stored at the IPSS, there is the maximum correlation between the calculated doses. The number of histories was chosen to obtain an statistical uncertainty lower than 0.1 % (1σ) for each point of the profile, of the PDD and of TMR. Therefore the overall statistical uncertainty of every point was about 0.25% (1σ). Since the dosimetric quantities PDD, TMR, OAR are ratios of absorbed dose calculated at different positions, the overall statistical uncertainty of each calculated point is 0.14%. About the PDD, TMR, OAR calculated in the water it seems appropriate to point out that the concept of dose to a point, like it is usually required in international dosimetry protocols, is ill-defined since the dose is the deposition of energy in a mass of tissue. Moreover, in MC simulation the dose deposition depends on the size of the scoring volume, and the associated uncertainty increases with decreasing the size of this parameter. Therefore, in the following the dose in a small volume of water is intended as a volume of 0.5x0.5x0.5 mm 3 with an associated uncertainty of 0.1% (1 standard deviation). This choice was based on the consideration that the dose distribution in a patient cannot be calculated with a resolution better than the underlying CT information, which is usually not lower than the above voxel size 36. RESULTS MonteCarlo simulations show: 1. the need to apply a correction factor to the responses of the dosimeters except that to the scintillating detector, to obtain the correct values of the omega factor for fields sizes smaller than 2.5 cm. 2. the PDD and TMR in water can be obtained directly from the experimental values without applying any correction factor when using the stereotactic diodes, the microlion and the scintillating detector; instead the micro-chambers can be used without any correction factor only for fields sizes greater than 1.5 cm. 3. The dose profiles in water can be obtained using only the diodes stereotactic diodes and the scintillating detector because the micro-chambers and the microlion significantly underestimate the dose gradient in the penumbra region due to the size of the sensitive volume. REFERENCES 1. AAPM Report 54, Stereotactic Radiosurgery Report of Task Group 42 Radiation Therapy Committee. (American Insitute of Physics, New York, NY, 1995). 2. I. J. Das, M. B. Downes, A. Kassaee and Z. Tochner, "Choice of radiation detector in dosimetry of stereotactic radiosurgery-radiotherapy," J Radiosurg 3, (2000). 3. T. C. Zhu, B. E. Bjarngard and H. Shackford, "X-ray source and the output factor," Med Phys 22, (1995).

29 4. A. Wu, R. D. Zwicker, A. M. Kalend and Z. Zheng, "Comments on dose measurements for a narrow beam in radiosurgery," Med Phys 20, (1993). 5. I. J. Das, M. B. Downes, A. Kassaee and Z. Tochner, "Choice of radiation detector in dosimetry of stereotactic radiosurgery-radiotherapy," J Radiosurg 3, (2000). 6. K. De Vlamynck, H. Palmans, F. Verhaegen, C. De Wagter, W. De Neve and H. Thierens, "Dose measurements compared with Monte Carlo simulations of narrow 6 MV multileaf collimator shaped photon beams," Med Phys 26, (1999). 7. F. Araki, "Monte Carlo study of a Cyberknife stereotactic radiosurgery system," Med Phys 33, (2006). 8. A. Wu, R. D. Zwicker, A. M. Kalend and Z. Zheng, "Comments on dose measurements for a narrow beam in radiosurgery," Med Phys 20, (1993). 9. R. K. Rice, J. L. Hansen, G. K. Svensson and R. L. Siddon, "Measurements of dose distributions in small beams of 6 MV x-rays," Phys Med Biol 32, (1987). 10. P. Francescon, S. Cora, C. Cavedon, P. Scalchi, S. Reccanello and F. Colombo, "Use of a new type of radiochromic film, a new parallel-plate micro-chamber, MOSFETs, and TLD 800 microcubes in the dosimetry of small beams," Med Phys 25, (1998). 11. P. Alfonso, P. Andreo, R. Capote, M. S. Huq, W. Kilby, P. Kjall, T. R. Mackie, H. Palmans, K. Rosser, J. Seuntjens, W. Ullrich and S. Vatnitsky, "A new formalism for refeence dosimetry of small and nonstandard fields," Med Phys 35, (2008). 12. B. E. Bjärngard, J.-S. Tsai and R. K. Rice, "Doses on the central axes of narrow 6-MV x-ray beams," Med Phys 17, (1990). 13. R. Capote, F. Sanchez-Doblado, A. Leal, J. I. Lagares, R. Arrans and G. H. Hartmann, "An EGSnrc Monte Carlo study of the microionization chamber for reference dosimetry of narrow irregular IMRT beamlets," Med Phys 31, (2004). 14. H.-R. Lee, M. Pankuch and J. Chu, "Evaluation and characterization of parallel plate microchamber's functionalities in small beam dosimetry," Med Phys 29, (2002). 15. X. A. Li, M. Soubra, J. Szanto and L. H. Gerig, "Lateral electron equilibrium and electron contamination ni measurements of head-scatter factors using miniphantoms and brass caps," Med Phys 22, (1995). 16. C. McKerracher and D. I. Thwaites, "Assessment of new small-field detectors against standardfield detectors for practical stereotactic beam data acquisition," Phys Med Biol 44, (1999). 17. P. Alfonso, P. Andreo, R. Capote, M. S. Huq, W. Kilby, P. Kjall, T. R. Mackie, H. Palmans, K. Rosser, J. Seuntjens, W. Ullrich and S. Vatnitsky, "A new formalism for refeence dosimetry of small and nonstandard fields," Med Phys 35, (2008). 18. E. Pantelis, C. Antypas, L. Petrokokkinos, P. Karaiskos, P. Papagiannis, M. Kozicki, E. Georgiou, L. Sakelliou and I. Seimenis, "Dosimetric characterization of CyberKnife radiosurgical photon beams using polymer gels," Med Phys 35, (2008). 19. E. Chung, H. Bouchard and J. Seuntjens, "Investigation of three radiation detectors for accurate measurement of absrobed dose in nonstandard fields," Med Phys 37, (2010). 20. K. E. Rosser and J. L. Bedford, "Application of a new dosimetry formalism to volumetric modulated arc therapy (VMAT)," Phys Med Biol 54, (2009).

30 21. Cranmer-Sargison, S. Weston, J. A. Evans, N. P. Sidhu and D. I. Thwaites, "Implementing a newly proposed Monte Carlo based small field dosimetry formalism for a comprehensive set of diode detectors," Med Phys 38, (2011). 22. G. Cranmer-Sargison, S. Weston, N. P. Sidhu and D. I. Thwaites, "Experimental small field 6MV output ratio analysis for various diode detector and accelerator combinations," Radiother Oncol 100, (2011). 23. P. Francescon, S. Cora and N. Satariano, "Calculation of k(q(clin),q(msr) ) (f(clin),f(msr) ) for several small detectors and for two linear accelerators using Monte Carlo simulations," Med Phys 38, (2011). 24. P. Francescon, W. Kilby, N. Satariano and S. Cora, "Monte Carlo simulated correction factors for machines specific reference field dose calibration and output factor measurement using fixed and iris collimators on the CyberKnife system," Phys Med Biol 57, (2012). 25. H. Bouchard, J. P. Seuntjens, J. Carrier, I. Kawrakow, A Monte Carlo method to evaluate the impact of positioning errors on detector response and quality correction factors in nonstandard beams Phys. Med. Biol , L.A. Buckley and D.W.O. Rogers, Wall correction factors, P wall, for thimble ionization chambers, Med. Phys. 33 (2), (2006). 27. J. Wulff, J.T. Heverhagen, and K Zink, Monte Carlo based perturbation and beam quality correction factors for thimble ionization chambers in high energy photon beams, Phys. Med. Biol. 53(11), (2008). 28. F. Crop, N. Reynaert, G. Pittomvils, L. Paelinck, C. De Wagter, L. Vakaet, and H. Thierens, The.influence of small field sizes, penumbra, spot size and measurement depth on perturbation factors for microionization chambers, Phys. Med. Biol. 54(9), (2009). 29. I. J. Das, C. W. Cheng, R. J. Watts, A. Ahnesjö, J. Gibbons, X. A. Li, J. Lowenstein, R. K. Mitra, W. E. Simon and T. C. Zhu, "Accelerator beam data commissioning equipment and procedures: Report of the TG-106 of the therapy physics committee of the AAPM," Med Phys 35, (2008). 30. J. Shi, W. E. Simon and T. C. Zhu, "Modeling the instantaneous dose rate dependence of radiation diode detectors," Med Phys 30, (2003). 31. S. M. Sze, Physics of Semiconductor Devices. (Wiley, New York, 1969). 32. A. S. Saini and T. C. Zhu, "Dose rate and SDD dependence of commercially available diode detectors," Med Phys 31, (2004). 33. E. E. Wilcox and G. M. Daskalov, "Evaluation of GAFCHROMIC EBT film for Cyberknife dosimetry," Med Phys 34, (2007). 34. A. Micke, D. F. Lewis, X. Yu "Multichannel film dosimetry with nonuniformity correction" Med Phys 38, (2011). 35. D.W.O. Rogers, B.R.B. Walters, and I. Kawrakow. BEAMnrc Users Manual. NRC Report PIRS 509(a) rev. I, H. Bouchard, J. P. Seuntjens, J. Carrier, I. Kawrakow, A Monte Carlo method to evaluate the impact of positioning errors on detector response and quality correction factors in nonstandard beams Phys. Med. Biol , L. A. Buckley, I. Kawrakow, D. W. O. Rogers An EGSnrc investigation of cavity theory for ion chambers measuring air kerma Med Phys 30, (2003).

31 38. L.A. Buckley and D.W.O. Rogers, Wall correction factors, P wall, for thimble ionization chambers, Med. Phys. 33 (2), (2006). 39. J. Wulff, J.T. Heverhagen, and K Zink, Monte Carlo based perturbation and beam quality correction factors for thimble ionization chambers in high energy photon beams, Phys. Med. Biol. 53(11), (2008). 40. F. Crop, N. Reynaert, G. Pittomvils, L. Paelinck, C. De Wagter, L. Vakaet, and H. Thierens, The.influence of small field sizes, penumbra, spot size and measurement depth on perturbation factors for microionization chambers, Phys. Med. Biol. 54(9), (2009).

32 Caratterizzazione di apparecchiature per IORT e tecniche di dosimetria in vivo Gianni Taccini (1) IRCCS A.O.U. San Martino IST U.O. Fisica Medica e Sanitaria - Genova Electron beams dosimetry of IORT devices are not completely described using the IAEA TSR 398 protocol. An overview of several dosimetric instruments, starting from absolute Fricke dosimeter and going on to more useful daily-use instrumentation, will be presented mainly keeping in account the determination of several correction factors such as kqq and ksat for different chambers. Between September 2009 and September 2012 we have treated with IOERT 459 patients affected by early-stage breast unifocal cancer. During IOERT treatment, patients can receive: 21 Gy intraoperatively or, if they aren t eligible to full dose irradiation (i.e. tumor diameter >2,5 cm, lymph node ) 10 Gy as a boost, or 16 Gy if a nipple sparing mastectomy [12] was performed. IOERT was delivered by a dedicated mobile accelerator LIAC Sordina Italy, electron beams: 4, 6, 8 and 10 MeV nominal energy. LIAC was calibrated using Fricke dosimeter [13] in inter-comparison with two plane-parallel ion chamber [14,15] (Roos and Adv. Markus, PTW Freiburg). Radiation is collimated by cylindrical applicator at different diameter between 4-10 cm (R90, namely therapeutic ranges for 5 cm diameter: 12, 15, 20, 26 mm); a steel-ptfe shielding disk (3mm + 3mm) is place between the deep face of the residual breast and the pectoralis muscle to minimize the thoracic wall irradiation. To avoid superficial underdosage due to entrance dose that is lower at lower energy (-5.4%: 4MeV; -5.6% : 6 MeV; -3.1%: 8MeV; -0.7%: 10MeV) we preferred higher electron energy (8 or 10 MeV); obviously before irradiation the surgeon carefully check that the shielding disk is correctly centred below the applicator end. MicroMOSFET 502-RDM (Best Medical Canada Ltd) detectors are employed to monitor exit dose. The detectors are placed inside a thin and sterile catheter (6Fr closed-end brachytherapy cather) and fixed at the centre of the PTFE side of the shielding disk before the insertion in the breast, in this way the dose collected by the deeper part of the target is measured. To perform treatment, optimized Monitor Units (MU) are calculated according the method described by Agostinelli S. et al.[8], irradiation is split in two parts in order to allow dose correction with a real time proper action level. Of all the patients treated with IOERT, the 52% of has been monitored using micromosfet in vivo dosimetry. The detectors were calibrated with the 4 Linac electron beams in the same setup of daily dosimetry check [8] using PMMA calibration Jig (TN-RD Best Medical) inserted in a slab solid phantom (RW3, PTW, Germany). Reproducibility of the response, linearity, sensitivity of detectors has been previously tested by Ciocca M. et al. [6], which estimated overall uncertainty of in vivo dosimetry of 3,6%, our further checks, using LIAC set-up, confirm these data. MD-V2-55 radiochromic films has been employed in a sample of 19 patients together with micromosfet in order to have information about spatial dose distribution along the target. Before IOERT procedure a radiochromic film (3 x 3 cm2) was wrapped in a sterile film and positioned between MOSFET and PTFE side of the shielding disk, in the centre of the plate. The field perturbation caused by the film envelope was less than 1% even at low electron energy. Films were scanned with an Epson flat-bed color scanner in transmission mode (EPSON Expression 1000XL), Image Acquisition software was used to scan all the films in the 48 bit RGB modality (16 bit per colors). All the images were scan with an image resolution of 75 dpi (pixel dimension 0.34x0.34 mm) and saved as tagged image file format (TIFF) image files. For each film we acquired 2 images. TIFF images were then imported into Picodose 8.0 software (Tecnologie Avanzate s.r.l.) to extract the red component of the RGB scanned images, elaborate, set calibration and analyze films, average and median filter (in 3x3 pixel regions) were applied. To obtain absolute dose [16, 17] a calibration curve is performed: film were exposed at LIAC electron beam collimated by PMMA flat-ended 10 mm diameter applicator in a slab RW3 solid phantom at the depth of dose maximum (12mm: 8MeV and 15,5mm: 10MeV) in a dose range: 0-33 Gy. Two different calibration curves are performed to keep into account the film post-irradiation stability with time: one after 24h and one after 72h. The dependence of film response on the energy, tested using electron beams produced by both CLINAC 2100 (6MV)

33 and LIAC (all available energies) was well within +3% as found by Piermattei et al. [17], moreover dose per pulse independence of film response was performed. Since films were kept in contact with the patient for a period of minutes, verify a correction factor equal to 0.95 was applied to take into account temperature dependence of film response during irradiation. Correction factor was evaluated by comparing the response of MD-V2-55 wrapped in sterile film above PTFE side of the shielding disk (to simulate surgery), irradiated in water at calibration temperature (21 C) and irradiated in water at human body temperature (about 37 ); the same correction factor has been previously reported [9 e AAPM]. The estimated uncertainty in the determination of absorbed dose to water at the depth of dose maximum was about 3% and include film calibration and temperature correction factor uncertainty. Each exposed film (double scanned) was imported into Picodose 8.0,visualized in dose modality and analyzed in term of mean, median, modal dose along the film, standard deviation, maximum (M) and minimum (m) dose values. A step forward will be concern in vivo dosimetry problems with the use both of Mosfets and Gafchromic films. Finally quality assurance protocol and time-schedule will be presented. Radioprotection problems will be approached regarding patients and operators and showing procedures used in our institution to satisfy radioprotection criteria. References [1] Veronesi U, Orecchia R, Luini A, et al. Intraoperative radiotherapy during breast conservering surgery: a study on 1822 cases treated with electrons. Breast Cancer Res Treat 2010; 124: [2] Orecchia R, Leonardo MC, Intraoperative radiation therapy: is a standard now?. Breast 2011 Oct; 20 Suppl3: S111-5 [3] Ruano Ravina A, Cantero Munoz P, Eraso Urien A, Efficacy and safety of intraoperative radiotherapy in breast cancer: a systematic review. Cancer Lett. 2011; Dec 26; 313 (1): [4] Njeh CF, Saunders MW, lanton CM, Accelerated partial breast irradiation (APBI): a review of available tecniques. Radiat Oncol 2010; 5:90. [5] Veronesi U, Orecchia R, Luini A, et al. Intraoperative radiotherapy during breast conservering surgery: a study on 1822 cases treated with electrons. Breast cancer Res Treat 2010; 124: [6] Ciocca M, Piazzi V, Lazzari R, et al. Real-time in vivo dosimetry using micromosfet detectors during intraoperative electron beam radiation therapy in early-stage breast cancer. Radither Oncol 2006; 78:213-6 [7] Soriani A, Landoni V, Marzi S, et al. Setup verification and in vivo dosimetry during intraoperative radiation therapy (IORT) for prostate cancer. Med Phys 2007; 34: [8] Agostinelli S, Gusinu M, Cavagnetto F, et al. On-line optimization of intraoperative electron beam radiotherapy of the breast. Radiother Oncol 2012; 103: [9] Ciocca M, Orecchia R, Garibaldi C, et al. In vivo dosimetry using radiochromic films during intraoperative electron beam radiation therapy in early-stage breast cancer. Radiother Oncol 2003; 69: [10] Beddar AS, Salehpour M, Briere TM, Hamidian H, Gillin MT. Preliminary evaluation of implantable MOSFET radiation dosimeters. Phys Med Biol 2005; 50: [11] Cavagnetto F, Agostinelli S, Guenzi M, Gusinu M, Zeverino M, Taccini G. In vivo dosimetry in IORT cancer treatment. 2011; 99: S [12] Petit JY, Veronesi U, Orecchia R, LuiniA, et al. Nipple-sparing mastectomy in association with intra operative radiotherapy (ELIOT): a new type of mastectomy for breast cancer treatment. Breast Cancer Res Treat 2006; 96: [13] Rosi A, Viti V. Guidelines for quality assurance in intra-operative radiation therapy, ISTISAN 2003Report No. 03/1 EN.

34 [14] Di Martino F, Giannelli M, Traino A.C, Lazzeri M. Ion recombination correction for very high dose-per-pulse high energy electrons beams. Med Phys 2005; 32: [15] Laitano R.F, Guerra A.S, Pimpinella M, et al. Charge collection efficiency in ionization chambers exposed to electron beams with high dose per pulse. Phys Med Biol 2006; 51: [16] Niroomand-Rad A, BlackweelCR, Coursey BM, et al. Radiochromic film dosimetry: recommendations of AAPM Radiation Yherapy Committee Task Group 55. Med Phys 1998; 25: [17] Piermattei A, Delle Canne S, Azario L, et al. Linac Novac7 electron beam calibration using GAF-chromic film. Phys Med 1999; XV:

35 Il commissioning di fasci a scansione di protoni e ioni carbonio M. Ciocca 1. (1) Fondazione CNAO (Centro Nazionale di Adroterapia Oncologica), Pavia Introduzione La Fondazione CNAO, istituita nel 2001 dal Prof. Veronesi, rappresenta il primo esempio in Italia di centro ospedaliero di adroterapia. La Fondazione ha lo scopo di trattare pazienti oncologici con fasci di particelle cariche pesanti di alta energia e di condurre programmi di ricerca e sviluppo in campo clinico, radiobiologico e fisico. Presso il CNAO è installato un sincrotrone di circa 25 metri di diametro, in grado di accelerare protoni e ioni carbonio in un intervallo di energia pari rispettivamente a e MeV/u, corrispondenti a profondità del picco di Bragg in acqua comprese tra 3 e oltre 30 cm (27 cm per gli ioni carbonio), con uno step di 2 mm [1]. La modalità di dose delivery è di tipo attivo (pencil beam scanning, con variazione attiva anche dell energia). L irraggiamento del volume tumorale avviene mediante una successione di irraggiamenti di slice iso-energetiche. Il centro è dotato di tre sale di trattamento con linee di fascio fisse (una linea orizzontale in ogni sala, oltre ad una linea verticale nella sala centrale). Ogni sala di trattamento dispone di un sistema per il posizionamento del paziente, a 6 gradi di libertà, e di uno a raggi x per la verifica del set-up [2], oltre che di una sala esterna di preposizionamento del paziente. Nella pratica clinica viene utilizzato un sistema elettronico di gestione della cartella clinica del paziente e di record&verify dei trattamenti. Riguardo al progetto dell Alta Tecnologia del CNAO (sincrotrone nel suo complesso, linee di estrazione dei fasci, sistemi di controllo, di sicurezza, di distribuzione e monitoraggio dosimetrico della dose, allestimento delle sale di trattamento) si tratta di una realizzazione assolutamente custom, eseguita da personale interno ad alta specializzazione, Università ed Istituzioni scientifiche italiane ed estere (quali INFN, CERN e GSI), industria [3]. Attualmente il CNAO sta trattando pazienti in regime di sperimentazione clinica, autorizzata dal Ministero della Salute. Il CNAO ha avviato l attività clinica con fasci rispettivamente di protoni e ioni carbonio il 22/9/2011 e il 13/11/2012. Sono stati sinora trattati oltre cento pazienti (settembre 2013) [4, 5]. Caratterizzazione dosimetrica e Quality Assurance dei fasci a scansione di particelle cariche pesanti La caratterizzazione fisico-dosimetrica dei fasci a scansione del CNAO e il commissioning del TPS impiegato per la pianificazione dei trattamenti sono stati effettuati mediante acquisizioni sperimentali e simulazioni Monte Carlo (codice FLUKA) [6, 7]. Sono state dapprima misurate le distribuzioni di dose in profondità integrate lateralmente, per ciascun pencil beam mono-energetico, mediante un sistema dedicato ad alta risoluzione spaziale, costituito da una doppia colonna sigillata e motorizzata d acqua e da una coppia di camere a ionizzazione piatte ad ampia superficie (Bragg Peak chambers) [8]. Successivamente, sono stati acquisiti i profili trasversali dei fasci al variare della loro energia, a diverse distanze e profondità in acqua, mediante pellicole radiocromiche tarate in termini di dose assorbita in acqua. Sono state poi determinate le curve di conversione delle Unità Hounsfeld in stopping power relativo all acqua per ogni protocollo di imaging utilizzato (testa-collo e pelvi), così come la curva di taratura in dose del sistema di Dose Delivery installato su ciascuna linea di trattamento. Per la determinazione della dose in condizioni di riferimento è stato utilizzato il protocollo IAEA TSR-398, modificato secondo il formalismo impiegato presso il GSI per fasci di ioni carbonio a scansione. Infine, sono stati determinati le procedure e i valori di riferimento per i controlli di qualità periodici dei fasci, relativi per esempio alla stabilità dell energia dei fasci, alla riproducibilità e proporzionalità del sistema di Dose Delivery, all accuratezza di deflessione dei fasci da parte dei magneti di scansione e alla costanza delle dimensioni del pencil beam. Per questi controlli, vengono impiegate camere a ionizzazione di tipo Farmer, fantocci d acqua oppure solidi, pellicole radiocromiche. A livello di procedure specifiche per ogni paziente, ogni piano di trattamento viene sempre verificato in fantoccio prima di poter essere erogato al paziente. La verifica dosimetrica consiste nella misura della distribuzione di dose

36 fisica in punti pre-stabiliti, per ogni campo di trattamento, mediante un sistema di camere a ionizzazione multiple di tipo Pin-point fissate su un apposito adattatore secondo una disposizione tridimensionale, in fantoccio d acqua motorizzato: ciascuna delle dosi misurate viene poi confrontata col proprio valore atteso, calcolato mediante un modulo dedicato del TPS[8, 9]. Prima dell avvio dell attività clinica, sono stati anche effettuati esperimenti di validazione radiobiologica dei fasci, mediante irraggiamenti sia in vitro (colture cellulari) che, per ioni carbonio, in vivo (topi), in collaborazione con i gruppi di radiobiologia dell INFN e col NIRS di Chiba. Bibliografia 1. Rossi S. The status of CNAO. Eur. Phys. J. Plus 2011, 126: Desplanques M, Tagaste B, Fontana G et al. A comparative study between the imaging system and the optical tracking system in proton therapy at CNAO. J. Radiat. Res. 2013, 54 (Suppl. 1):i129-i Giordanengo S, Donetti M, Garella MA et al. Design and characterization of the beam monitor detectors of the Italian National Center of Oncological Hadron-therapy (CNAO). Nucl. Instrum. Meth. A 2013, 698: Orecchia R, Srivastava A, Fiore MR et al. Proton beam radiotherapy: report of the first patient treated at the Centro Nazionale di Adroterapia Oncologica (CNAO). Tumori 2013, 99:e34-e Tuan J, Vischioni B, Fossati P et al. Initial clinical experience with scanned proton beams at the Italian National Center for Hadrontherapy (CNAO). J. Radiat. Res. 2013, 54 (Suppl. 1):i31-i Parodi K, Mairani A, Brons S et al. Monte Carlo simulations to support start-up and treatment planning of scanned proton and carbon ion therapy at a synchrotron-based facility. Phys. Med. Biol. 2012, 57: Tessonnier T, Mairani A, Cappucci F et al. Development and application of tools for Monte Carlo based simulations in a particle beam radiotherapy facility. Appl. Radiat. Isot Jan 4. doi:pii: S (12) /j.apradiso Karger CP, Jaekel O, Palmans H et al. Dosimetry for ion beam radiotherapy. Phys. Med. Biol. 2010, 55:R193-R Molinelli S, Mairani A, Mirandola A et al. Dosimetric accuracy assessment of a treatment plan verification system for scanned proton beam radiotherapy: one-year experimental results and Monte Carlo analysis of the involved uncertainties. Phys. Med. Biol. 2013, 58:

37 L Utilità del Metodo Monte Carlo in Radioterapia E. Spezi Velindre Cancer Centre, Cardiff, UK La radioterapia é una delle tecniche di trattamento dei tumori piú utilizzate ed efficaci e con un rapporto costo beneficio relativamente basso [1]. In radioterapia l accuratezza delle pianificazione e del delivery del trattamento é di straordinaria importanza sia per la salvaguardia del paziente che per l efficacia del trattamento. Gli algoritmi di calcolo basati sul metodo Monte Carlo sono ampiamente considerati come gli strumenti piu accurati disponibili in radioterapia [2]. Un numero considerevole di codici Monte Carlo cosiddetti general purpose come Fluka [3], MCNP [4], EGSnrc [5], Penelope [6] e Geant [7] sono disponibili pubblicamente e sono stati molto usati per ricerca e sviluppo in applicazioni medicali. Altri codici come XVMC [8], VMC++ [9], e DPM [10] sono stati sviluppati specificatamente per sostituire gli algoritmi convenzionali nella pratica clinica grazie ad una riduzione sostanziale dei tempi di calcolo. Recentemente Rogers [11] ha pubblicato una review delle tecniche di simulazione per il trasporto di elettroni e fotoni con una particolare attenzione al codice EGS4/EGSnrc. Ma e Bijang [12] hanno rivisto le metodologie di simulazione per fasci clinici di elettroni, mentre Verhaegen e Seuntjens [13] si sono concentrati sulla modellizzazione di fasci esterni di fotoni. Un lavoro piú recente di Reynaert et al [14] ha rivisto e discusso le procedure di commissioning del metodo Monte Carlo in radioterapia mentre Spezi e Lewis [15] hanno rivolto la loro attenzione a codici Monte Carlo veloci per la pianificazione dei trattamenti. Anche Chetty et al [5] ha compilato una lista dei codici Monte Carlo più frequentemente utilizzati per applicazioni cliniche e ha fornito utili raccomandazioni per l implementazione clinica di questi programmi. Zaidi e Sgouros [16] hanno esaminato le applicazioni di simulazioni Monte Carlo in medicina nucleare and terapia radiometabolica, mentre El Naqa et al [17] hanno rivisto l utilità di questo metodo in radiobiologia. Questo lavoro rivisiterà i progressi più e meno recenti nel settore delle tecniche Monte Carlo in fisica medicale e discuterà l utilità del metodo Monte Carlo in radioterapia dal punto di vista clinico, industriale e di ricerca e sviluppo. References [1] Department of Health (2007) Cancer Reform Strategy (DoH, London) [2] Chetty I et al. (2007) Report of the AAPM Task Group No. 105: Issues associated with clinical implementation of Monte Carlo-based photon and electron external beam treatment planning Med. Phys [3] A. Fasso et al. (2005), FLUKA: a multi-particle transport code, CERN , INFN/TC_05/11, SLAC-R-773 [4] Brown FB (2003) MCNP A general Monte Carlo-particle transport code, version 5 Report LA-UR , Los Alamos National Laboratory, Los Alamos, NM [5] Kawrakow I and Rogers DWO (2000) The EGSnrc code system: Monte Carlo simulation of electron and photon transport Technical Report PIRS-701, National Research Council of Canada, Ottawa, Ontario [6] Baro J et al. (1995) PENELOPE - An algorithm for Monte-Carlo simulation of the penetration and energy-loss of electrons and positrons in matter Nucl. Instrum. Methods Phys. Res. A [7] Agostinelli S (2003) GEANT4 A simulation toolkit Nucl. Instrum. Methods Phys. Res. A [8] Fippel M (1999) Fast Monte Carlo dose calculation for photon beams based on the VMC electron algorithm Med. Phys [9] Kawrakow I (2001) VMC++, electron and photon Monte Carlo calculations optimized for radiation treatment planning Advanced Monte Carlo for Radiation Physics, Particle Transport Simulation and Applications: Proceedings of the Monte Carlo 2000 Meeting (Lisbon) ed. A. Kling et al. (Springer, Berlin) [10] Sempau J et al. (2000) DPM, a fast, accurate Monte Carlo code optimized for photon and electron radiotherapy treatment planning dose calculations Phys. Med. Biol [11] Rogers DWO (2006) Fifty years of Monte Carlo simulations for medical physics Phys. Med. Biol. 51, R287 R301

38 [12] Ma CM and Jiang SB (1999) Monte Carlo modelling of electron beams from medical accelerators Phys. Med. Biol. 44 R157 R189 [13] Verhaegen F and Seuntjens J (2003) Monte Carlo modelling of external radiotherapy photon beams Phys. Med. Biol. 48 R107 R164 [14] Reynaert N et al. (2007) Monte Carlo treatment planning for photon and electron beams Radiat. Phys. Chem [15] Spezi E and Lewis DG (2008) An overview of Monte Carlo treatment planning for radiotherapy, Radiat. Prot. Dosimetry [16] Zaidi H and Sgouros G Eds (2002) Therapeutic Applications of Monte Carlo Calculations in Nuclear Medicine, Taylor & Francis. [17] El Naqa I, Pater P, Seuntjens J. (2012) Monte Carlo role in radiobiological modelling of radiotherapy outcomes, Phys. Med. Biol. 57, R75-97

39 Ripianificazione con Cone-Beam CT: metodologia di calibrazione e prime applicazioni cliniche Replanning on Cone-Beam CT: method of calibration and first application on patients C. Carbonini 1, M.G. Brambilla 1, M. Minella 1, A. F. Monti 1, A. Torresin 1, V. Valsecchi 2 (1)Department of Medical Physics, Ospedale Niguarda Ca Granda Milan, Italy (2)Department of Physics, University of Milano Bicocca Milan, Italy Purpose: Image Guided Radiotherapy (IGRT) with on-board cone-beam CT (CBCT) for radiation therapy delivery devices offers the potential for adaptive therapy and dose replanning. In order to use the CBCT images to calculate a dose distribution, it is necessary to convert the images from arbitrary gray scale to Hounsfield Unit (HU). In this work, we report the method for image conversion directly on the CBCT, its the validation on phantom and its final application on patients. Methods and materials: The treatment unit was an Elekta Synergy equipped with XVI on board CBCT imager. The software used for analysis of the images was ImageJ The dose distribution was calculated on a RTPS Elekta-Nucletron Oncentra External Beam (OEB) ver The first step was to find the conversion function for a specific acquisition protocol, by analyzing the Catphan 600 phantom images. We compared the HU obtained after the application of the conversion function with nominal HU. After a conversion from gray level value to HU, we validated the dose re-calculated with OEB on phantom with three insert of different material, in terms of absolute dose. We also investigated more complex 3D CRT and IMRT H&N plans. As suggested by ICRU 50 and ICRU 83, the Conformity Index (CI) was used for PTV, while the dose near maximum (D 2% ) and the mean dose (D mean ) were considered for OARs. Results: CBCT images can be converted into HUs directly on the XVI system. The dose distribution should be carefully evaluated before replanning in simple and complex geometries. Re-planning on a phantom showed a variation on D mean (Gy) in all the inserts <3%. Re-planning on the patients CBCT and DVH evaluation showed a substantial agreement with the reference CT. The analysis of the parameters chosen for the PTV and ORAs have differences that seem to be related more to the difficulty to reproduce the same contours on the CB images than a real calculation defect: Conformity Index (CI (Δ%)): 1.06 ± 1.46 D 2% (Δ%): ± 3.4 D mean (Δ%): 0.21 ± 3.05 Conclusion: The method used for the direct conversion from gray level value to HU on the XVI system is simple, but strongly dependent on the scatter, it is therefore necessary to find a conversion function dedicated to the different anatomical district.

40 References: [1] Depuydt, T., J. Hrbacek, P. Slagmolen, and et al. "Cone-beam CT Hounsfield unit correction methos and application on images of the pelvic region." Radiother. Oncol 81 (2006): (Suppl 1): 29. [2]Guan, H., and H. Dong. "Dose calculation accuracy using cone-beam CT (CBCT) for pelvic adaptive radiotherapy." Phys. Med. Biol. 54 (2009): [3]ICRU-50. "Prescribing, Recording and Reporting Photon Beam Therapy." International Commission on Radiation Units and Measurements, Baltimore, Maryland, USA, [4]ICRU-83. "Prescribing, Recording, and Reporting Photon - Beam Intensity - Modulated Radiation Therapy (IMRT)." International Commission on Radiation Units and measurements, Baltimore, Maryland, USA, 2010, [5]Letourenau, D., R. Wong, D. Moseley, and et al. "Online planning and delivery technique for radiotherapy of spinal metastases using cone.beam CT: image quality and system performance." Int. J. Radiat. Oncol. Biol. Phys. 67 (2007): [6]Richter, A., et al. "Investigation of the usability of conebeam CT data for dose calculation." Rad. Oncol. 16 (2008): [7]Siewerdsen, JH, and DA Jaffray. "Cone-beam computed tomography with a flat-panel imager: magnitude and effects of x-ray scatter." Med Phys 28 (2001): [8]Tanyi, J.A., and M. H. Fuss. "Volumetric image - guidance: Does routine usage prompt adaptive replanning? An institutional review." Acta Oncol 47 (2008): [9]Yoo, S., and FF Yin. "Dosimetric feasibility of cone-beam CT-based treatment panning compared to CTbased treatment planning." Int J Radiat Oncol Biol Phys 66 (2006):

41 Modellizzazione LINAC per VMAT da parte dei fornitori: criticità, problematiche ed effetti delle griglie di dose VMAT LINAC modeling from factory service: criticality, issues and dose grid effects. G.Guidi 1,2, G.Gottardi 1, A.Bernabei 1, T.Costi 1 (1) Az. Ospedaliero-Universitaria di Modena - Policlinico, Modena (2) Università degli Studi di Bologna, Bologna Purpose: VMAT modelling physics tools are not always accessible to physicists and LINAC characterization is provided by the manufactures. Most of parameters and measures follow different calibration protocols; anyway, it is requested to commission complex procedures indirectly, using reports or tech manuals (e.g. IAEA, AAPM). We highlight criticality and dose grid effects for VMAT commissioning. Methods and materials: VMAT SW could not include 3DCRT, IMRT or other biological functions. Many tests become impracticable. Rotational techniques could be more specificity than checks detailed by TG119. MU, segments Homogeneity (HI) and Conformity Index (CI) could be significantly influenced by dose grids and model tweaked. Models approvals should be obtained by using hard constraints (e.g. Gamma Index (GI) <2%@2mm) instead of standard clinical routine (e.g. local dose comparison without dose cut-off). Results: From first model provided (v.2.20) up to last commissioned (v.3.30) we got 5 tweaked models, 3 VMAT MLC calibration and 1 magnetron replacement. By repeating tests and local patient QA, we could prove models inefficiency by comparison with other TPS, increasing GI by >40%. Dose grid (2,3,4 and 5mm) analysis showed different dose optimization (>5%), increasing MU# calculation (>10%) and number of segments (>6%). HI and CI are reasonable for targets, but display differences for OARs due to grids size (>1-10%). Unexpected hot and cold spots could emerge in patients and MU increments should be investigated whenever suspected to relate to machine models. Therefore access to physics tools must be available. Conclusion: external services is not an optimal solution for complex software and commissioning. Bugs and unexpected errors became corrected with progressing versions but model should be re-optimized and validated to avoid drifted models from the original one. Physicist must have full access to the modelling tools, to ensure LINAC commissioning and safe treatments. Indirect measures, SW and LINAC upgrade and maintenance (e.g. MLC, energy and output calibration) could have high impact on patient plans and produce unexpected discrepancies. 3mm dose grid is reasonable in terms of time consuming but must probed in depth for SBRT or RS plans. A secondary TPS or Monte Carlo characterization should be available, when modelling physics tools are absent. References: [1]C.G. Orton, Controversies in Medical Physics: a Compendium of Point/Counterpoint; 2012, Debates (Vol. 2)

42 Valutazione dell accuratezza di calcolo della dose per trattamenti IMRT nel distretto toracico: confronto tra pencil beam, convolution/superposition e analytical anisotropic algorithm. Dose calculation accuracy for IMRT treatments in the thorax: comparing pencil beam, convolution/superposition and analytical anisotropic algorithms. C. Sini 1,2, S. Broggi 2, C. Fiorino 2, G. M. Cattaneo 2, R. Calandrino 2 (1) School of Specialization of Medical Physics, University of Cagliari, Cagliari, (2) Medical Physics Department, San Raffaele Scientific Institute Milano Purpose: to investigate the accuracy of three dose algorithms in critical geometries. Pencil beam (PB), Anisotropic Analytical Algorithm (AAA) and Convolution Superposition (CS) were considered in heterogeneous media. Methods and materials: the diode array ArcCHECK (SunNuclear Corp.) and the cylindrical solid water Cheese phantom were used as homogeneous phantoms. Three heterogeneous phantoms were then considered: Multiplug ArcCHECK equipped with inhomogeneous inserts; the same ArcCHECK slightly modified with a low density lung shape insert and a slab heterogeneous phantom simulating the thorax region. Linac IMRT (IMRT), arc volumetric RapidArc (RA) and Helical Tomotherapy (HT) lung treatment plans were calculated with PB/AAA, AAA, CS, respectively. Absolute dose measurements with ion chambers and planar dose maps, with ArcCHECK array and EBT3 films, were performed. Percentage point dose differences and quantitative analysis of gamma function distribution (acceptance criteria (AC): 3%/3mm) were analyzed. Results: average absolute deviations between measured and calculated dose were <2% and <3% in homogeneous and heterogeneous phantoms, respectively. Worse deviations (around -4%) were found in inhomogeneous conditions for IMRT-PB plans. Good results were found for planar dose distributions acquired with diodes array: at least 95% of points satisfy the gamma AC in homogeneous and in all inhomogeneous arrangements, except for IMRT-PB plans, where slightly worse results were found (91% of points satisfy the AC). A different behavior was found between homogeneous (Cheese phantom) and inhomogeneous conditions (slab thorax phantom), when EBT3 were used. An excellent agreement was found in Cheese phantom for all dose algorithms: on average 97% of points satisfy the AC. A worse agreement was obtained in the slab thorax phantom where the AC was satisfied for 63%, 69%, 80% and 75% of points for IMRT-PB, IMRT- AAA, RA and HT, respectively. It was probably due to the difference between the calculated dose in water and measured dose in media. The disagreement is more evident for points placed in lung interfaces or in small lesions in low density media, where also the advanced algorithms fail. Conclusion: a dose overestimation (5-10%) was confirmed for PB plans, above all in complex inhomogeneous arrangements (small lung lesions, media interface). Acceptable accuracy was found with more advanced algorithms, with better results of CS compared with AAA in critical geometries.

43 Verifica dosimetrica pre-trattamento ed in-vivo dei piani di trattamento eseguiti in Helical Tomotherapy (HT) mediante il sistema di verifica Dosimetry Check accoppiato ai rivelatori HT Pre-treatment and in-vivo dosimetry of Helical Tomotherapy (HT) treatment plans by using the Dosimetry Check system coupled to HT detectors 1 E. Mezzenga, 1 E. Cagni, 1 A. Botti, 1 M. Orlandi, 2 W.D. Renner, 1 M. Iori 1 Medical Physics Unit, ASMN-IRCCS of Reggio Emilia, Italy; 2 MathResolution LLC, Columbia, MD, USA Purpose: The Dosimetry Check software (DC, Math Resolution, LLC, Maryland) is one of the medical devices used to ensure that the dose delivered to a patient during a radiotherapy treatment is correct and strongly agrees with that simulated by his treatment plan. DC is capable of evaluating either pre-treatment or in-vivo patient plan dosimetry by using only the on-board imaging detectors of the treatment unit. The purpose of the study is to assess the proper functioning of the DC software on an Helical Tomotherapy (HT) unit either in pre-treatment and in-vivo dosimetry. Materials & methods: 10 different treatment plans were selected for brain, head & neck, thorax and prostate tumour. For the pre-treatment dosimetry, each plan was delivered without the presence of the treatment couch, while for in-vivo dosimetry the couch and the patient were present. In both situations, the delivered fluence fields were measured by the HT on-board MV detectors. The CT scans, the structures used in planning and the recorded MV detector signal were imported into the DC software to calculate the absolute doses and the dose-volume histograms, that were compared with the planned ones by means of the gamma-index analysis method into the DC software, using a 3%/3mm pass criteria and a dose threshold of 10% on the calculated dose. Results: the gamma-index values between the planned dose and that calculated by the DC tool ranged from 88% to 100% for the pre-treatment dose verification method, while for the in-vivo dosimetry the same agreement ranged between 88% and 99.61%. The lowest values have been observed for the thorax treatment, where the in-homogeneities were more present than in all the other treated anatomical sites. This effect was much less strong in prostate and brain tumours where the results were the best in terms of dose agreement reaching values around 97% 99%. Conclusions: The Dosimetry Check software coupled with the MV Tomotherapy detectors have proved to be an invaluable tool for the volumetric pre-treatment and in-vivo dosimetry verification of the HT treatments. The dose agreement reached for brain and prostate treatments is very high also for the in-vivo verification methods. However, because the DC tool is based on a pencil beam algorithm, which is fast but over-estimate the dose values where in-homogeneities are present, cautions should be used for thorax and head & neck treatments where the pre-clinical method is still more safe and reliable.

44 1.Target degradation effects in Hi-Art Tomotherapy: a Monte Carlo study A.Esposito 1,2,3, B. Caccia 1,2, C.Andenna 4, A. Sarnelli 2,5, E. Menghi 5, M.Benassi 5, L.Strigari 6 (1) Istituto Superiore di Sanità, Roma, (2) INFN, Joint Group of Istituto Superiore di Sanità, Roma (3) Scuola di Specializzazione Fisica Medica, Università La Sapienza, Roma (4) INAIL, Roma (5) IRST, Meldola (Forlì), (6) IFO-IRE, Roma Introduction Helical Tomotherapy (HT) combines a 360 fan-beam delivery of intensity-modulated radiation with megavoltage computed tomography (MVCT) imaging for the treatment of complex tumours: the MVCT facility reproduces a 3D image of the patient s anatomy for image-guided treatment positioning, and is paired to a system of modulated radiation delivery along a helical trajectory around the patient [1,2]. Through the HT unit, complex dose distributions may be planned with an optimal degree of normal tissue sparing, but the evaluation of dose distribution for such a complex technique is an outstanding problem that requires sophisticated computing technologies to optimize clinical results and Quality Assurance (QA) requirements. HT presents new challenges in Monte Carlo (MC) simulation because the simultaneous movement of the couch, the multi-leaf, and the gantry must be accurately simulated. Besides, several parameters related to the open static fields and the HT beam should be accounted for to optimize the MC simulation of the HT unit [3, 4]. In HT, the beam-on time needed to perform a treatment can be up to 15 times longer than in conventional techniques [5]. In a HT unit, the tungsten x-ray target rotates continuously in a water bath for dissipation of heat: the aqueous environment and long beam-on times can cause the target to deteriorate more rapidly than in a conventional linear accelerator. As a consequence, the tungsten target degrades and needs a periodical replacement [6]. As the target degrades, modifications of fluencies are expected with consequent distortion of dose profiles in water phantom. It comes out that the target consumption is a parameter that has to be taken into account in HT commissioning to assess the effect on patients during treatments. The specific aim of this work was to fully simulate an HT through Geant4 MC by taking into account the possible effect of the target degradation in dose distribution. Methods and materials Geant4 [7] is a MC toolkit implemented in C++ that allows modelling and simulation of the interaction of particles and matter. The code is freely distributed under an open software license. The Geant4 code has been chosen because of its flexible combinatorial geometry capability, which is required for detailed modelling of the system. A full MC Geant4 simulation of a HT treatment unit has been carried out. The Geant4-based model has been commissioned by comparing simulated dose distributions with the measured ones, adopting a standard iterative procedure. Four different degraded targets have been simulated in two different geometrical models. For this study, the Geant4 (release 9.6 patch 2 distributed at code was used to simulate the HT unit, the phantom, and the dose distribution therein. Geant4 physics is encoded into Physics Lists, and the user is allowed to select the relevant physics processes for a specific application. Two sets of electromagnetic models are available: the "Standard" and the "Low-energy". All the data presented in this work were simulated using the Low-energy Penelope package [8]. Results Starting from a target of 1.4 mm thickness (considered as a fresh target), Geant4 MC simulations were carried out for the following thickness values, corresponding to different levels of tungsten target thinning: 1.0 mm, 0.7 mm, 0.5 mm and 0.3 mm. The target degradation has been simulated with two different geometrical models. Starting from a 1.4 mm thick tungsten block as fresh target, the consumption was simulated both by a depth-parameterised half-toroid hole on one hand and by a simple tungsten cylinder with different thickness on the other. To highlight

45 the degradation effect, the different depth dose curves are normalized to the value obtained simulating a HT unit with a fresh target (1.4 mm in our study). As shown in Figure 1 for the cylinder model, there is a steepening of the depth dose curve with time, explained by a change in the photon spectrum to a lower mean energy. MC simulated data show a significant reduction of dose in the water phantom, up to about 12% at depth 10 cm for a 0.3 mm cylinder-modelled target thickness. All the dose reduction effects encountered with a cylinder-modeled target are lightly less evident using half-toroid hole model (see Figure 2). Figure 1: Different depth dose curves (obtained with five different target thickness values) normalized to the value obtained simulating a HT unit with a target with a thickness of tungsten of 1.4 mm. A simple tungsten cylinder with different thickness was adopted as model. Figure 2: Different depth dose curves (obtained with five different target thickness values) normalized to the value obtained simulating a HT unit with a target with a thickness of tungsten of 1.4 mm. In this case the consumption is simulated by a depth-parameterised half-toroid hole.

46 Dose reduction with target usage was also verified by Kampfer et al [6] (although they did not measure different target thickness values). By comparing simulated data with the experimental data [6], the cylinder-model of the target seems to be more adequate to the real behaviour of the dose as a function of the target reduction. Conclusion Simulations show that the tungsten target consumption has a non-negligible influence on the beam characteristics and the dose in water phantom, and therefore it should be considered and properly modelled in the commissioning of any MC code for the HT unit. MC simulations could help to appropriately interpret pre-treatment measurements for patient plan verification, during a target lifetime. The disagreement between dose distribution calculated with a fresh target and dose distribution measured with a degraded target over time could reflect a possible inaccuracy in treatment delivery, which needs to be investigated. References: [1] T.R. Mackie, Tomotherapy: a new concept for the delivery of dynamic conformal radiotherapy, Med. Phys. 20, (1993). [2] R. Jerai, T.R. Mackie, J. Balog, G. Olivera, D. Pearson, J.Kapatoes, K. Ruchala and P. Reckwerdt, Radiation characteristics of helical tomotherapy, Med. Phys. 31(2), (2004). [3] N. Reynaert, S.C. van der Marck, D.R. Schaart, W. Van der Zee, C. Van Vliet-Vroegindeweij, M. Tomsej, J. Jansen, B Heijmen, M. Coghe, and C. De Wagter, Monte Carlo treatment planning for photon and electron beams Radiat. Phys. Chem. 76, (2007). [4] Y. Zhao, M. Mackenzie, Monte Carlo calculation of helical tomotherapy dose delivery, Med. Phys. 35(8), (2008). [5] R.J. Staton et al, S.L. Meeks, Dosimetric effects of rotational output variation and x-ray target degradation on helical tomotherapy plans, Med. Phys. 36(7), (2009). [6] S. Kampfer et al, Measurements to predict the time of target replacement of a helical tomotherapy, Journal of Appl. Clin. Med. Phys. 12(4), (2011). [7] GEANT4 collaboration, GEANT4 a simulation toolkit, Nucl. Instrum. Meth. A 506, (2003). [8] E. Sterpin, F. Salvat, R. Cravens, K. Ruchala, G.H. Olivera and S. Vynckier, Monte Carlo simulation of helical tomotherapy with PENELOPE, Phys. Med. Biol. 53, (2008).

47 A Radiotherapy network for oncology solutions F. Di Rosa, R. Costa, V. A. La Monaca, G. Politi Azienda Sanitaria Provinciale di Caltanissetta Fisica Sanitaria U.O.C. Radioterapia Introduction Radiotherapy treatment consists of a complex sequence of data all referred to a defined case that the therapist needs in structured patient-related design. So, a solution to improve patient care, to reduce errors and share files would be to extend the information on a central communication system (i.e. radiotherapy network). The use of this system becomes essential for a management of a complex Radiotherapy Department where multi-modality imaging, clinical and physical information are involved in the processes of planning and delivery of radiation therapy requiring a dedicated DICOM standard. DICOM RT is an extension to the current DICOM standard specifically created to accommodate the fast evolving radiotherapy procedures. It consists of five extended Information Object Definitions (IOD), namely RTIMAGE, RTSTRUCT, RTPLAN, RTDOSE and RTRECORD to report clinical activities of radiotherapy. A dedicated network for radiotherapy purposes could be conceived as an extended DICOM server implementing requirements that are faced in a specialized RT Department and not addressed by standard radiology PACS systems. It should be able to manage and store all types of DICOM and non-dicom data under the same patient entry making them available for any other RT system located in the network. Our particular organization of two facilities, located in distant geographical areas, belonging to a single RT Department, needs a connection to secure a massive storage of multi-modal DICOM RT images and a technology support assuring dataset coherence. We have designed a server infrastructure to perform a real time connection for the complete management of Department. The server architecture is based on a PACS system dedicated to radiotherapy requirements employing advanced DICOM forwarding rules. With this system, all DICOM RT data and all clinical and physical information related to the same patient can be requested and retrieved by the physicians and the physicists in any of the two facilities. Methods and Materials Our RT network is implemented as a multi-node server infrastructure to carry out a redundant connection based on Multi Protocol Label Switching to create a Virtual Private Network (MPLS-VPN). Its full compatibility with the TCP/IP technologies gets results a flexible system able to transport and route several types of network traffic. Moreover, additional backup line connection is expected as safeguard. An IP address is assigned to each Client (i.e. TPS, R&V, Linac, CT Simulator and RT Dicom Viewer) in order to detect and identify any device univocally and to permit the full management of all equipments installed in our Radiotherapy Department. Network architecture is managed by a central PACS server, DICOM RT PACS, embedded with ConQuest DICOM server (v ) [1] and dedicated to radiotherapy purposes. ConQuest includes a collection of libraries and applications implementing large parts of the DICOM standard. The software is designed to be field/run time programmable. This user-programmable feature allows the software to be custom tailored to a particular Clinical or Research area. The server supports a wide range of databases with ODBC or without, including a programmable SQL database to satisfy all traditional PACS features. The most important of these provide a complete DICOM interface, an easy installation procedure, the possibility to create many servers on a single PC and a solution to satisfy the requirements for a robust and long term storage. A database browser, integrated in our PACS system, is available for the full management of the patient images: delete, change and edit data patient, anonymize studies and series, print images, image format conversion, send and receive images over the entire network connection. It is possible to handle directories using a simple query/move user interface to grab missing data from another server. A DICOM web access for viewing and data administration is implementable. Finally, PACS organization becomes a powerful vehicle that manages, compares and records the quality control results performed by images with an integrated tool to convert non-dicom objects to supported files and viceversa.

48 The peculiarity of DICOM RT objects requires a dedicated Viewer installed in the network. As main Viewer we have chosen Fratoria DICOM RT Studio. We have used other commercial and non-commercial Viewers in order to offer an useful and complete comparison among these systems. The key properties of the tested Viewers are a fast transferring rate both in send and in receiving mode, a prompt loading and browsing of a considerable number of images. They also process all kind of DICOM RT files using a complete support of protocols, allow images analysis with a list of 2D graphic function and display the datasets in volume rendering with 3D post processing function. All software are able to view and/or calculate DVHs (dose-volume histograms) by IOD. The selected hardware for the PACS server is a multi-core workstation with 8 GB of RAM and 4 TB of storage to store up to 25 millions of images. We have installed a RAID system with four identical hard disks formatting in order to perform a RAID 5 system. It guarantees fast performances and a secure system (no data are lost if a drive fails). A remote Network Attached Storage (NAS RAID) assures the coherence and the safety of RT data. Discussion and Results A server infrastructure was designed to perform a non-stop connection and communication of the two facilities for the full management of our Department. Each step of radiation treatment process produces DICOM and non- DICOM objects that must be recorded, shared and stored. This PACS makes more feasible the supply of radiotherapy procedures eliminating the incompatibility among different multi-vendor systems. It also allows to streamline data storage and clinical work-flow. Archived data on RT PACS can be reviewed by any Client workstation of RT Department, thus the staff can easily access to all patient data and to a single patient waiting list. A treatment can be planned, valued, approved and delivered in both facilities independently. At the same time, the patient could carry out first consultation, multi-modal images acquisition, daily treatment and follow-up checks in the facility that he prefers, minimizing logistical difficulties. This increases quality and provides a continuous service avoiding treatment interruptions caused by machine maintenance or downtime, too. DVH COMPARISON AMONG TPS AND THREE DIFFERENT VIEWERS

49 We have carried on an investigation on reliability of DVH calculation by two releases of commercial software Fratoria DICOM RT Studio (v.2.16 and v.3 beta), a Viewer undercostruction ( Viewer 3) and a non-commercial software released by University of Arkansas for Medical Sciences, DICOMan. DVH calculated by TPS were compared to DVH calculated by Viewers: differences in the maximum dose, in the mean dose and in the V 95% (target volume value receiving 95% of prescribed dose) were evaluated. We have performed the study considering Planned Target Volume and Organ At Risk: ipsilateral lung for breast treatment and rectum for prostate treatment. We have also considered two cases of Glioblastoma to investigate the operation of DICOM viewers with small structures evaluating calculation on both eye lenses and optic chiasm. We have evaluated the differences in OAR histograms in V 20Gy for lung, in V 50Gy for rectum, and in maximum and mean dose for small structures. The table above shows that the accordance among TPS and Viewers is always within 2,3% for PTV, up to 10% and 23% for lung and rectum respectively. About small structures, differences in D max are up to 15%. The observed deviations are probably due to the different calculation around the edges of the ROIs and small volumes. Conclusions A RT network implementation, in our territorial area, joining two far facilities of the same RT Department, answers to the needs and cares of the oncologic patient. It represents an innovative, easy and affordable solution providing a clinical service that we cannot find in our area. The development of our network architecture is still subject to many iterative stages of testing, refinement and evaluation before its complete implementation in the clinical environment. Another enhancement of this system, an App for mobile device, is under construction. This application will permit overview of the patient clinical course. It also will facilitate data migration to a cloud service. Eventually, the server support a modality worklist that can be loaded with HL7 data and queried by the modalities to load patient demographic information that can be reconsolidated with the pertinent images. All these applications can be entirely and functionally integrated in a user interface. A system integration infrastructure based on standards is crucial for streamlining clinical work-flow and for the establishment of medical research related to outcomes for future radiation therapy patients. The open based architecture of the RT-Server facilitates consultation or clinical research across other Radiotherapy Department and Institutions and permits a better collaboration between radiation oncologist and medical physicist of the same Department. A unique RT data management is a right way to go for a better organization of human resources, including the institution of Medical Physics Services that could constitute an independent Department in the territorial area. The benefits of the implementation of such network are multiple: it can be used as a standard to perform an effective and an efficient clinical service and as a common platform for radiotherapy data exchange and expert consultation. References [1] ConQuest freely available for download at

50 Implementazione di un modello di trasmissione del lettino di trattamento nel TPS in geometrie con elevato contributo di dose da fasci posteriori o obliqui. Incorporation of the couch transmission model in the treatment planning system: impact on beam arrangement with few or high weighted posterior oblique fields. I.Solla 1, S.Zucca 1, M.Rinaldi 1, E.Deiana 1 and G.Meleddu 1. (1) Fisica Medica, Ospedale Oncologico Businco Azienda ASL n 8 Cagliari Purpose: Carbon fiber (CF) top is widely used as patient support component for external beam radiotherapy. However, in beam arrangement with few or high weighted posterior oblique beams the couch interference could be significative in dependence on couch model, vendor and beam quality [1,2,3]. The aim of this study was to evaluate attenuation of the IMRT/IGRT carbon fiber (CF) couch installed on the Radiotherapy department at Cagliari. Materials and methods: The image of the treatment couch was acquired on a CT (Sensation Open, Siemens) (Figure 1-Left), and a virtual couch model was created on Pinnacle (Philips) using autosegmentation tools and added to organ model library (Model Based Segmentation). A single button procedure was created using TPS scripting language in order to automatically load couch model on CT images. The couch modelling was validated by absolute dose measurements with a ion chamber (PTW cm 3 ) in a cylindrical phantom (D=32 cm h=15 cm) (Figure 1-Right). Isocenter was set in the center of phantom and dose calculation was performed for open beams of 6 MV and 18 MV with different field gantries and field sizes. The attenuation of treatment couch was then evaluated for different beams arrangement: a) single posterior beam at G180, b) single oblique beam at G230, c) two equal weighted opposed beams at G0 and G180, d) three equal weighted beams at G0, G230 and G130. Figure 1 (Left) Axial, Sagital and Coronal view of carbon fiber treatment couch. (Right) 3D view of cylindrical phantom (D=32 cm h =15cm) with couch model and beams configuration d). The two posterior oblique fields crossed the reinforced side of the couch (shaded areas). Results: A maximum couch attenuation of 4.7% for 6 MV and 3.2% for 18 MV was measured for oblique beam with gantry angle of 230 (crossing the couch through the reinforced side). Using a density of 0.31 g/cm 3 for CF model and 1.5 g/cm 3 for reinforced areas, the agreement between the transmission measured with ion chamber and calculated on TPS was within 1.5% for both 6MV and 18 MV. The Table 1 shows measured (D iso,measured ) and calculated (D iso,calculated ) isocenter dose in cylindrical phantom for 6 MV and 18MV.

51 Percentage dose differences (Diso,calculated /Diso,measured-1) are reported in brackets for beam configuration a), b), c), d). If the dose calculation is performed without the couch attenuation model, percentage dose difference reach a value of 6% and 3% in oblique posterior beam at G230 with 6 MV and 18 MV respectively (case a) and b)). The agreement between calculated and measured dose is within 1.5 % if the couch model is properly included in dose calculation. The couch effects is higher in plans with high weighted posterior beams, in particular for those beams which crossed the reinforced side of the couch. Couch attenuation also depend on photon energy and beam size. Table 1 Measured and calculated isocenter dose in cylindrical phantom (D=32 cm h =15cm) for different field size, energy and beams configuration: a) single posterior beam 180 ; b) single oblique beam 230 ; c) two equal weighted opposed beams G0 and 180 ; d) three equal weighted beams G0, G230 and G130. Percentage differences from measured dose are reported in bracket. Conclusion: Treatment couch attenuation was evaluated and couch model was created on treatment planning system using contours with appropriate density that simulate the CF couch attenuation. A single button automatic procedure was then created to apply couch model to our TPS (Pinnacle v9.2) improving the accuracy of dose calculation in beam arrangement with few or high weighted posterior oblique beams. References 1. McCormack S, Diffey J, Morgan A. The effect of gantry angle on megavoltage photon beam attenuation by a carbon fiber couch insert. Med Phys. 2005; 2. Mihaylov IB, Corry P, Yan Y, Ratanatharathorn V, Moros EG. Modeling of carbon fiber couch attenuation properties with a commercial treatment planning system. Med Phys. 2008; 3. Gerig LH, Niedbala M, Nyiri BJ. Dose perturbations by two carbon fiber treatment couches and the ability of a commercial treatment planning system to predict these effects. Med Phys

52 Dosimetric effects of rotational setup errors on prostate IMRT treatments. C. Zucchetti 1, A.C. Dipilato 2, A. Didona 1, M. Iacco 1, M. Marcantonini 1, C. Aristei 2, G. Gobbi 1 (1) Medical Physics Department, Santa Maria della Misericordia Hospital, Perugia (2) Radiotherapy Department, Santa Maria della Misericordia Hospital, Perugia Introduction: The intensity-modulated radiotherapy (IMRT) makes it possible to devise highly conformal treatment plans to the target volume and spare adjacent critical structures. However, the efficacy of this tecnique can be compromised by errors in the treatment setup of the patient which affects the delivered radiation. Many tools are available to verify the right setup of patients but while translation errors can be corrected with a simple couch shift, rotational errors cannot be easily accounted for and therefore exist throughout the course of treatment. Such rotational errors represent a systematic difference between patient simulation and treatment, and the dosimetric consequences need to be carefully evaluated. The aim of this work is to determinate dose delivery errors that could result from systematic rotational set up errors for prostate cancer patients treated with IMRT. Methods and materials: Five prostate cancer patients who had undergone IMRT in our institution were selected for the study. The IMRT plans are elaborated with Pinnacle3 Treatment Planning System (TPS) using 5 beams of energy 15 MV. For all patient dose distribution of each beam is calculated by TPS on CT of a phantom composed by sandwich of water solid slabs 5 cm thick with MatriXX detector (IBA Dosimetry) inside; the detector consists of a two dimensional 1020 ion chamber array with a sensitive volume of 0.08 cm 3 and detector spacing 7.6 mm (fig.1). beam MatriXX solid water solid water solid water fig.1 Phantom composed by solid water slab and MatriXX detector We have calculated for all beams the planar dose distribution with gantry at 0 so that the central axis of the beam result perpendicular to the phantom; and we have taken this distributions as reference. The rotational setup errors are simulated by calculating, for all beams and each patient, the dose distribution with the gantry rotated by 0.5, 2.5 and 5 respectively. The reference dose distribution is compared with those calculated at 0.5, 2.5 and 5 by OmniProI mrt (IBA Dosimetry) software and, for each simulated rotational error, the percentage of points with a value > 1 (Dose Difference, DD, 1% - Distance To Agreement, DTA, 1mm) is measured. We have repeated the same procedure taking as reference the dose distribution measured with MatriXX and compare this with the dose distributions calculated at 0.5, 2 and 5 ; in this case too, the percentage of points with a value > 1 is measured for each simulated rotational errors.

53 Results: The mean percentage of points with a value > 1 (P ), derived by comparison between calculated dose distributions, is (0.08 ± 0.05)%, (0.28 ± 0.05)%, and (0.68 ± 0.18)% for a simulated rotational error of 0.5, 2 and 5 respectively (Tab. 1). The same parameter derived by comparison between measured (0 ) and calculated at 0.5, 2 and 5 dose distribution is (15.6 ± 1.5)%, (15.5 ± 1.5)%, and (15.3 ± 1.4)% respectively (Tab. 2). Conclusion: Our results indicate that rotational setup errors introduce modifications into calculated dose distributions that increase with the angle value even if the mean percentage of points that fail the tolerance normally accepted. The mean percentage of points that fail distribution at 0 and calculated ones, shows no dependence by the simulated rotational errors, due to the detector spatial resolution; high mean percentage values originate from the low limits of DD and DTA used to evaluate function. As points that fail test are distributed on regions of high gradient, the greater the complexity of the dose distribution the greater the error introduced by rotation. Patient P 0.5 vs 0 P 2 vs 0 P 5 vs mean Tab. 1 percentage of points with a value > 1 derived by comparison between calculated reference dose distribution (gantry angle 0 ) and calculated dose distributions at gantry angle 0.5, 2 and 5 Patient P 0.5 vs 0 P 2 vs 0 P 5 vs ,82 13,71 13, ,37 17, ,71 14,34 14, ,82 16,58 16, ,41 15,6 15,4 mean 15,63 15,49 15, Tab. 2 percentage of points with a value > 1 derived by comparison between measured reference dose distribution (gantry angle 0 ) and calculated dose distributions at gantry angle 0.5, 2 and 5

54 Validazione di un algoritmo di registrazione di immagini deformabile per la somma di dosi Validation of a deformable image registration algorithm for dose accumulation M. Fusella, C. Fiandra, F.R. Giglioli, R. Ragona (1) School of Medical Physics, University of Torino, Torino, Italy (2) University of Torino, Department of Oncology, Radiation Oncology Unit, Turin, Italy (3) Medical Physics Unit, Azienda Ospedaliera Città della Salute e della Scienza, Turin, Italy Obiettivo del lavoro è validare un algoritmo di deformazione della dose attraverso confronti di matrici di dose su fantocci virtuali tra calcolo Monte Carlo e risultati di deformazione, e attraverso misure su fantoccio antropomorfico con pellicole radiochromiche. Sono numerose le situazioni nell ambito del percorso radioterapico, in cui si rivela utile e necessario ricorrere a fusioni di immagini, attraverso l utilizzo di algoritmi di registrazione deformabile: i movimenti e le deformazioni anatomiche degli organi, che intercorrono fra e durante le sedute di radioterapia, portano ad avere incertezze geometriche e dosimetriche sia in fase di planning sia nella esecuzione del trattamento stesso. Ciò comporta una variabilità nella erogazione della corretta dose al tumore ed in quella ricevuta dagli organi sani circostanti. Tra le varie tecniche utilizzate per correggere e/o valutare queste variabilità, c è la Image Guided Adapative Radiation Therapy (IGART), il cui obiettivo principale è di estrarre le informazioni dinamico-temporali del paziente in corso di terapia per poi utilizzarle nelle eventuali variazioni del piano di trattamento, ottenuto rispetto all esame TC basale. Ruolo centrale in questo processo lo riveste la tecnica di dose accumulation / adaptive monitoring. Questa strategia permette di studiare la dose cumulativa totale del paziente, opportunamente corretta nelle diverse sedute di trattamento, grazie alle informazioni morfologiche e di posizionamento fornite dalle immagini CBCT. Gli algoritmi di registrazione delle immagini sia di tipo rigido sia di tipo elastico forniscono un valido aiuto per ottenere dei risultati affidabili. Gli algoritmi necessari per la deformazione delle immagini e le relative matrici di trasformazioni elastiche, possono essere utilizzate anche per la conseguente deformazione della distribuzione della dose fra i due set di immagini. Si tratta di applicare alla matrice 3-D della dose, la stessa matrice di deformazione ricavata dalle modifiche anatomiche. Per cui non si tiene conto di alcun fattore fisico riguardo l interazione della radiazione con la materia, trattandosi solo di un rimodellamento della matrice di dose. Proprio sullo studio dei limiti e dell appropriatezza di questo assunto si basa il lavoro di questa lavoro. Attualmente in letteratura esistono lavori per la valutazione della affidabilità di questi algoritmi, e fanno soprattutto uso di gel 3-D deformabili, i quali però presentano le ovvie limitazioni legate al limitato numero di eterogeneità tissutali presenti nelle immagini e all incertezze dosimetriche del gel stesso. Obiettivo del lavoro è di validare uno di questi algoritmi di deformazione elastica della dose. Si utilizzeranno il fantoccio antropomorfico Alderson-Rando, su cui si effettuerà un esame TC per ogni deformazione anatomica introdotta, e un set di fantocci virtuali con variazioni geometriche note. In particolare vengono testate due macro presenti nel software commerciale VelocityAI (Velocity Medical Solutions, 1350 Spring Street Atlanta, GA 30309): somma di dosi tra CT diverse, e adaptive monitoring. MATERIALI E METODI E stato usato l algoritmo di registrazione che si basa su combinazioni di funzioni radiali, come le B-Spline, implementato nel software commerciale VelocityAI. Queste funzioni hanno la forma di: ( x) pi i ( x) Dove pi è un fattore di scala, mentre βi una funzione radiale base, generalmente di tipo polinomiale. La formulazione con B-Spline cubiche è la più diffusa. L algoritmo lavora su una griglia di nodi, ad ognuno dei quali è

55 associato un vettore, che viene deformato a seconda della metrica scelta per la registrazione delle immagine. Il fattore peso p i è il parametro che viene variato durante il processo di registrazione. Per lo scopo del lavoro sono stati creati due fantocci virtuali CT, e, per ognuno di loro, quattro versioni modificate. Il primo set di fantocci si basa su una semplice figura cubica con inserite all interno tre diverse strutture: un organo a rischio sferico (OAR), un volume bersaglio sferico (PTV) e midollo, di forma cilindrica. Il secondo set di fantocci si basa su un semplice modello del distretto toracico, con i due polmoni (costituiti da due sfere e un cilindro), un midollo e vertebre (costituiti da due cilindri concentrici), e il volume bersaglio (sferico). Entrambi sono stati generati secondo lo standard DICOM. Ad ogni fantoccio è stato aggiunto del rumore bianco di tipo gaussiano, che simulasse una CT standard. Per quanto riguardo il fantoccio Alderson Rando, tre versioni diverse sono state utilizzate per la validazione della deformazione della dose. E stato usato il distretto capocollo, e le variazioni tra le CT è stata ottenuta simulando un edema omolaterale a livello della laringe. Il calcolo della dose è stato eseguito con TPS Monaco v3.2 (Elekta). I set di fantocci, con variazioni note tra loro, sono stati registrati tra loro. Attraverso analisi di indici di conformità (indice di Dice), è stata valutata la qualità delle co-registrazioni. Poiché gli algoritmi di registrazione di immagini, non eseguono un reale voxel-tracking, e quindi non lavorano su strutture, non risulta appropriato valutare i risultati dosimetrici utilizzando i DVH. Per questo motivo sono stati usati dei marker (circa 40 per ogni fantoccio). Quindi il confronto della dose calcolata, deformata e misurata, è stato effettuato basandosi su dei marker di posizione nota nei fantocci. Soltanto dopo aver ottenuto delle buone registrazioni di immagini, si è passati ad analizzare la deformazione della dose, ottenuta da calcolo Monte Carlo implementato sul TPS. Lo studio si è focalizzato su due aspetti con rilevanza clinica: somma di dosi, e valutazione della dose giornaliera. Con un test statistico sulle popolazioni calcolato e deformato Wilcoxon Signed-Rank, si sono analizzati i risultati. RISULTATI Manipolando opportunamente i parametri liberi dell algoritmo, sono stati ottenuti buoni risultati: l indice di Dice medio è risultato di 0.95±0.04. La somma di dosi su differenti set di immagini CT, forniscono dei risultati accettabili. Un test di Wilcoxon è stato eseguito fra dosi deformate e dosi calcolate, fornendo come risultato e 0.25, rispettivamente per il fantoccio cubico e il fantoccio toracico. Per quanto riguarda i risultati del test sulla macro adaptive monitoring, i risultati sono: e 0.43, rispettivamente per il fantoccio cubico e il fantoccio toracico. CONCLUSIONI Dal confronto dei valori di dose calcolata e deformata nei vari marker presenti in ogni fantoccio, si sono potute trarre le conclusioni di questo lavoro riguardo l affidabilità della deformazione della dose. L analisi statistica su circa cento punti di misura ha mostrato che, nei limiti dei fantocci utilizzati e delle situazioni analizzate, la somma di dosi si dimostra accurata. Per la valutazione di dose giornaliera il numero di punti di misura analizzati è stato di 92, per i fantocci virtuali, e 25 per Rando. I risultati indicano che la per la valutazione di dose giornaliera, così come proposta dal software testato, non risulta sufficientemente accurata. Basandosi su questi risultati, si può dedurre che per la valutazione giornaliera della dose ( Adaptive Monitoring ), risulta più indicato effettuare un ricalcolo del piano di trattamento sul set di immagini deformate. Queste nuove dosi possono essere sommate tra di loro in maniera affidabile utilizzando il software Velocity AI.

56 Dalle indicazioni fornite da questo lavoro, mediante l utilizzo è infine possibile trarre spunto per la stesura di protocolli clinici di Image Guided Adaptive Radiation Therapy, per la personalizzazione dei trattamenti radioterapici.

57 Sliding window IMRT plans for breast with simultaneous integrated boost (SIB): efficiency and deliverability as a function of the smoothness of the beam fluence maps. S. Naccarato 1, G. Sicignano 1, R. Ruggieri 1. (1) U.O. Radioterapia Ospedale Sacro Cuore Don Calabria Negrar (Verona) Purpose: For dynamic sliding window IMRT, treatment planning systems let the user to simplify the beam fluences, computed from the optimizer, by smoothing before leaves motion be calculated. In this study we investigated whether some action level in the use of such smoothing can de defined to improve efficiency and the deliverability of breast IMRT plans with SIB. Methods and materials: For each of 6 breast patients to be treated with SIB of increasing complexity, we prepared four sliding windows IMRT plans with 7 fields, by Eclipse TPS (v. 10.0, Varian) with AAA dose calculation algorithm (2 mm grid), with increasing X-Y smoothing priority value (25%, 50%, 75%, and 100% of the PTV priority value), and equivalent level of PTV coverage and OAR sparing. The resulting 24 IMRT plans comprised 252 fields (84 fields splitted fields), for which MU, and average leaf pair opening (ALPO) were computed. 2D dose distributions, both for each field and for all fields together with gantry equal 0, were in-phantom calculated and measured with an i.c. 2D-Array Seven 29 ( PTW). Measured and calculated dose comparison (pass%= %pixels with <1) was performed with different -criteria (3%/3mm, 3%/1mm). Results: For increasing X-Y smoothing level from 25% to 100% we observed linear variations of MU and ALPO. Wilcoxon paired-sample test was used to test the smoothing differences. Statistical significance was considered at p<0.05. The extremes for smoothing=[25%, 100%] were MU FIELD =[133±59, 102±38] (p<0,0001) MU PLAN =[1355±246, 970±127] (p<0.0001), ALPO(cm)=[1.3±0.61, 1.64±0.74] (p<0.0001), %pass (3%/3mm)=[98.7±1.94, 99.1±1.4] (p=0.04), and %pass (3%/1mm)=[85.4±7.4, 89.6±6.5] (p<0.0001). Given the strong linear correlation (R=0.6) between ALPO(cm) and %pass (3%/1mm), we identified ALPO as indicator of the complexity of the beams. For a threshold ALPO=1.5cm sliding window breast IMRT beams accurately delivered, at 97%pass (3%/3mm) and 85%pass (3%/1mm), with 100% specificity in both cases 15%(3%/3mm) or 32% (3%/1mm) sensitivity. Conclusion: Increasing X-Y smoothing level decreases the total number of MU, and increases ALPO, %pass (3%/3mm) and %pass (3%/1mm). Smoothing priority value at 75% of the PTV priority value produce effective increment in efficiency and deliverability of the plan without affecting PTV coverage or OAR sparing. ALPO, which is accessible in the field s properties, shows the ability to distinguish the complexity of the beam. By incrementing the smoothing value to obtain ALPO > 1,5 cm can improve deliverability of the beam and sometimes of the whole plan. References: [1] S. V. Spirou et al., Smoothing intensity-modulated beam profiles to improve the efficiency of delivery, Med. Phys (2001) 28(10), [2] G. Nicolini et al., What is an acceptability smoothed fluence? Dosimetric and delivery considerations for dynamic sliding windows IMRT, Radiation Oncology (2007) 2:42 [3] A.L. McNiven et al., A new metric for assessing IMRT modulation complexity and plan deliverability, Med. Phys (2010) 37(2),

58 Boost adattivo simultaneamente integrato nella radiochemioterapia neoadiuvante per il carcinoma rettale: validazione prospettica dei margini al tumore comprendenti l impatto della deformazione Daily image-guided Adaptive simultaneous integrated boost in neo-adjuvant Radiochemotherapy for rectal cancer: prospective validation of tumor margins including the impact of deformation R. Raso 1, C. Fiorino 1, P. Passoni 2, G. Rizzo 3, N. Di Muzio 2, R. Calandrino 1 (1) Medical Physics, San Raffaele Scientific Institute, Milano (2) Radiotherapy, San Raffaele Scientific Institute, Milano (3) Istituto di Bioimmagini e Fisiologia Molecolare, CNR, Segrate (MI) Purpose: An adaptive concomitant boost (ACB) technique for neo-adjuvant rectal cancer patients (RCP) treated with Helical Tomotherapy (HT) was activated at our Institute in the last two years. Before the clinical activation, an analysis focused on motion data from daily MVCT of 10 previously treated consecutive RCP was attained: combining coverage probability maps (CPM) of the rectum (after rigid set-up correction) with 3D local distance measurement permitted to quantify margins for different treatment phases. Results referred to the treatment second half suggested 7mm for the anterior direction and 5mm elsewhere in order to cover the rectum in at least 90% of the fractions (frs) for 90% of the patients. Current study was focused on the prospective validation of these margins in the ACB phase. Methods and materials: Twenty RCP (10 males, 10 females) treated with HT (18x2.3Gy), delivering an ACB (3.0 Gy/fr) on the residual GTV in the last 6 frs were chosen for the ACB margin validation. The ACB was planned based on CT/MRI imaging taken at half-therapy (hct). 120 MVCTs of the last 6 ACB frs were bone-matched with the corresponding planning CT (pct) and hct. The rectum was contoured by a single observer on each MVCTs only in the slices where the residual GTV was present and then expanded by 7mm in the anterior part and 5mm elsewhere (PTV(5,7 ANT )). All contours were transferred to the pct: the union of them (UN) and the 90% CPM (volume covered by 5/6 frs) were created. The percentage volume (%) of the UN/90%CPM missed by the PTV(5,7 ANT ) was measured and used as an estimate of the appropriateness of the applied margin. Male and female patients were analysed separately. Results: The fraction of 90% CPM outside the PTV(5,7 ANT ) was <2% for all male and <5% for 9/10 female patients. It reduced to <2% for 9/10 females when a 7mm isotropic margin was applied. For males, the UN fraction outside PTV(5,7 ANT ) was on average 1.8% (maximum: 6.9%): when considering 5/6 frs, the mean and max % missing were only 0.4% and 1.5%. Females had larger variations: average % of UN missing was 5.3% with a max of 18.9%; these values decreased to 1.8% and 9.7% respectively when considering 5/6 frs. In order to reduce the % of UN of female patient n 5 below 2%, a symmetric 10mm margin was necessary. Conclusion: A 7mm anterior and 5mm in all other directions margin has been prospectively found as adequate in our ACB context; female patients show a larger residual error.

59 Esperienza della Azienda Ospedaliero-Universitaria Careggi nelle verifiche dosimetriche su singolo paziente: dalla metrica gamma al DVH. AOUC experience on patient-based dose verification: from gamma passing-rate to DVH. L. Marrazzo 1, M. Casati 1, C. Arilli 1, A. Compagnucci 1, S. Pallotta 1,2, C. Talamonti 1,2, E. Vanzi 1, M. Bucciolini 1,2. (1) A.O.U. Careggi, Firenze (2) University of Florence Purpose: the aim of this work is to evaluate the performances of a commercial software, which perturbs the calculated dose distribution according to the dose discrepancies detected with planar or volumetric measurements in order to predict the actually delivered 3D dose distribution in the patient. The software was then used to evaluate the perturbed DVH for a sample of patients and to investigate the correlation between the gamma passing rate and the DVH modifications. Methods and materials: several data sets were created by inducing different types of errors in clinical prostate step&shoot IMRT plans, by using an approach already proposed in the literature [1]. The error-free plans were used as simulated measurements for generating the IMRT QA dose planes to be compared to the corresponding data calculated by the error-induced plans. In addition, 3 prostate plans, 2 accelerated partial breast irradiations, 2 head and neck cases planned with step&shoot IMRT were delivered on both a planar and a cylindrical commercial diode matrix detectors. The measurements were used as input of the software together with patient RTDose, RTStructure, RTPlan and CT images and the corresponding perturbed 3D dose distribution in the patient was calculated. Results: the validation tests show that the software is able to reproduce very precisely the induced DVH modifications for all the proposed errors. Concerning patient analysis, in most of the cases there were only weak to moderate correlations between the gamma metrics and the clinical metrics. Moreover, some of the largest clinically significant dose differences occurred in the cases of high gamma passing rates, while low gamma passing rates do not necessarily imply large differences in the DVH. Planar and volumetric measurements led to similar perturbed dose distributions. Conclusion: our first results suggest that this software is a reliable tool for predicting DVH modifications on the base of the patient based pre-treatment QA. Patient analysis show a lack of correlation between gamma passing rates and clinically relevant dose errors. The most common acceptance criteria should then be supported by further analysis allowing a stronger predictive power. References: [1] B. E. Nelms, Per-beam, planar IMRT QA passing rates do not predict clinically relevant patient dose errors, Med. Phys. (2011) 38(2),

60 Dose to organs at risk in the upper abdomen in patients treated with extended fields by helical TomoTherapy: a dose-volume histogram analysis and acute toxicity study. Sara Bresciani 1, Elisabetta Garibaldi 2, Gabriella Cattari 2, Angelo Maggio 1, Amalia Di Dia 1, Pietro Gabriele 2 and Michele Stasi 1 (1)Department of Medical Physics, Institute for Research and Treatment of Cancer (IRC@C) at Candiolo, Turin, Italy. (2)Department of Radiation Oncology, Institute for Research and Treatment of Cancer (IRC@C) at Candiolo, Turin, Italy. Purpose: the aim of this work was to determine the technical feasibility and safety of extended-field radiotherapy (EF), performed by Helical TomoTherapy, in patients with positive pelvic and/or para-aortic nodes, through the analysis of dose-volume histogram (DVH) parameters and their correlation to toxicities, including assessments of organ function by complete blood counts and the laboratory tests. Dosimetric data were collected and acute/ sub-acute toxicities of the upper abdominal organs at risk (OAR) were evaluated. Methods: 29 patients suitable for EF irradiation for local and nodal disease in the pelvic or para-aortic area were treated by HT units. The series included 17 patients with nodal recurrences of prostate cancer, 4 patients with very high-risk prostate cancer, 5 patients with stage II-III of cervical cancer and 3 patients with postoperative stage II-III endometrial carcinoma. Median age was 65.8 years (range years). The prescription dose was 50.4/54 Gy ( Gy/fraction) for prophylactic lymph nodes (N-) and Gy ( Gy/fraction) for clinically evident gross disease (N+). Modulation factor, pitch and field width were chosen to optimize dose distribution and treatment duration. Dose-volume histograms (DVHs) of the PTVs and the critical normal structures were analyzed. For PTVs, we evaluated the minimum doses delivered to 95% (D95%) and the average dose of the PTV T, PTV N+, PTV N-. For OAR, the mean and maximum dose of small bowel, pancreas, spleen, stomach, kidneys and liver were examined. V45 of small bowel was registered according to QUANTEC recommendations. The length of the treatment field, the N+ and N- volumes, and treatment duration were reported. Hematological, hepatic, renal and pancreatic functions were evaluated by complete blood count and laboratory tests before, during and after treatment. Toxicities were graded according to the National Institute Common Toxicity Criteria for Adverse Events (CTCAE), version 3.0 scale. Results: In most cases, plan parameter values used for these treatments were FW=2.5 cm, pitch= cm and MF= 2.2. In two cases we used a FW of 5 cm to reduce treatment time. The value of these parameters were chosen in order to produce a treatment that could be delivered in a reasonable length of time (<15 min). In one re-treatment case we used a FW of 1.05 cm because the new target was just superior to a previously treated PTV and the smaller field has a smaller penumbra in longitudinal direction. The median actual MF was 1.8 (range: ). The average length of treatment was 32.5 cm. Median treatment time was 660 sec (range: sec). Excellent PTV coverage was obtained: in general, the mean value of D95% for PTVs of primary targets was 96.5%, ranging between 94% and 98%. Mean absolute dose and D95% were 65.3±3.5 Gy and 63.7±3.4 Gy to PTVN+, 54.5±2.1 Gy and 52.2±2.0 Gy to PTVN-, respectively. We irradiated 1-6 positive pelvic and/or lumbar-aortic nodes simultaneously, with a mean volume of 76.6±48.3 cc. Mean volumes of irradiated prophylactic nodes was 770.6±307.1 cc. The mean volume of small bowel that received more than 45 Gy (V45) was 157±83 cc (4.3±3.7%). Maximum dose to small bowel was less than

61 60 Gy and the mean V55 was 1.8±2.2cc (0.4±0.5%). No patients failed to meet QUANTEC published dose objectives for the rectum, bladder, femoral heads, liver, kidney and spinal cord. Only two patients treated for postoperative endometrial cancer exceeded the QUANTEC constraint for the intestinal cavity (V 45 >230 cc). The median follow-up time was 8.4 months (range: 5-22 months). For two of the 29 patients it was not possible to evaluate sub-acute toxicity because of the short follow up (<6 months). Overall the treatment was well tolerated: all patients but one completed treatment without interruption. One patient who received concurrent chemotherapy for cervical carcinoma required a break of 6 days due to severe urinary infection. Of these 29 patients, 10 (34.5%) experienced G1 and 3 (10.3%) G2 acute gastrointestinal toxicity. Acute hematological toxicity was the following: G1 in 7 patients (24.1%), G2 in 4 patients (13.8%; 2 of these patients received chemotherapy, one concurrently, and 2 received only hormonal therapy), G3 in 4 patients (13.8%; all of whom received chemotherapy, 3 concurrently). In 3 (10.3%) patients we observed an early slight increase of pancreatic enzyme (G1 acute toxicity) and in 5 (17.2%) patients an early slight increase of hepatic enzymes (G1 acute toxicity). No acute renal toxicity was observed. Table 2 summarizes acute toxicity as a function of the patient, disease, and treatment characteristics (i.e. history of prior RT or surgery). About sub-acute toxicity, on the 27 evaluable patients, no gastrointestinal or renal toxicity was observed. G1 hematological toxicity occurred in 1 (3.7%) patient and G2 in 2 (7.4%) patients. Only one (3.7%) patient had a persistent slight increase of pancreatic enzyme and 2 (7.4%) patients a slight increase of hepatic enzymes six months after radiotherapy (G1 toxicity). All patients with hematological, pancreatic and hepatic toxicity received chemotherapy during radiation treatment. The only one patient that experienced G1 acute and late pancreatic toxicity showed a pancreatic dose greater than 60 Gy (D1cc=67.8 Gy). The values of DVH for pancreas were evaluated by comparing the mean relative volumes at selected doses values between the patient with toxicity and the others 28 without toxicities. Independent samples 2-sided t-tests were performed at each 10 Gy dose level; no p values reached statistical significance (p>0.05), until 60 Gy. The only patient whose pancreas dose exceeded 60 Gy was the patient with toxicity. Conclusion: with our treatment design and dose regimen we found that EF IMRT by TomoTherapy could be delivered with minimal acute and sub-acute toxicities in the upper abdomen area. Dosimetric analysis among our patient set indicated that EF-IMRT by HT seems to be a safe treatment approach. Indeed, therapy was well tolerated and there were no sub-acute Grade 3 and 4 toxicities. In our study, no patient to date has exhibited any renal toxicity and no patient has reported hepatic and/or pancreatic toxicities greater than G1. Longer follow-up is required to validate these favorable long-term toxicity findings. References 1 Du XL, Sheng XG, Jiang T, et al: Intensity-modulated Radiation Therapy Versus Para-aortic Field Radiotherapy to Treat Para-aortic Lymph Node Metastasis in Cervical Cancer: Prospective Study. Croat Med J 2010, 51(3): Salama JK, Mundt AJ, Roeske J, et al: Preliminary outcome and toxicity report of extended-field, intensitymodulated radiation therapy for gynecologic malignancies. Int J Radiat Oncol Biol Phys 2006, 65: Portelance L, Chao KS, Grigsby PW, et al: Intensity-modulated radiation therapy (IMRT) reduces small bowel, rectum, and bladder doses in patients with cervical cancer receiving pelvic and para-aortic irradiation. Int J Radiat Oncol Biol Phys 2001, 51:

62 4 Poorvu PD, Sadow CA, Townamchai K, et al: Duodenal and Other Gastrointestinal Toxicity in Cervical and Endometrial Cancer Treated With Extended-Field Intensity Modulated Radiation Therapy to Paraaortic Lymph Nodes. Int J Radiat Oncol Biol Phys. 2012, article in press. 5 Marnitz S, Köhler C, Burova E, et al: Helical tomotherapy with simultaneous integrated boost after laparoscopic staging in patients with cervical cancer: analysis of feasibility and early toxicity. Int J Radiat Oncol Biol Phys. 2012, 82(2):e Jackson A, Marks LB, Bentzen SM, et al: The lessons of QUANTEC: recommendations for reporting and gathering data on dose-volume dependencies of treatment outcome. Int J Radiat Oncol Biol Phys 2010, 76(3 Suppl):S

63 Valutazione della dose agli organi in Radioterapia per sistemi kv Cone Beam CT Evaluation of organ doses in Radiation Therapy from kv-cone Beam CT F. Palleri 1, C. Ghetti 2, R. Rossi 2, E. Calabri 2, S. Magi 1, M. Palombarini 1, W. Gaiba 1,G. Frezza 3 (1) UOS Fisica Sanitaria, Azienda USL di Bologna (2) Servizio di Fisica Sanitaria, Azienda Ospedaliero-Universitaria di Parma (3) UOC Radioterapia Ospedale Bellaria, Azienda USL Bologna Purpose: Estimation of organ and effective doses of patients undergoing radiation therapy from two kv Cone Beam Computed Tomography(CBCT) systems installed on linear accelerators for patient position verification. Matherials and Methods: The kv-cbct systems examined were On Board Imager(OBI, Varian) and X- ray Volumetric Imager(XVI, Elekta). OBI is mounted on Varian Clinac DHX 2100 at Azienda Ospedaliero- Universitaria of Parma and XVI on Elekta Synergy at AUSL of Bologna(Bellaria Hospital). For both CBCT systems clinical acquisition protocols for head and neck(h and N), thoracic and pelvic regions were examined. Since Computed Tomography Dose Index(CTDI 100 ) is not appropriate for wide cone beam dosimetry, the method suggested in IAEA Human Health Reports No 5 was applied and in-house made solutions for experimental set-up have been specifically realized. For the three types of acquisition protocols CTDI 100 was measured in air and in PMMA(head or body phantom depending on the protocol used) with a 100 mm pencil ion chamber and CTDI weighted(ctdiw) was calculated. Dose measurements were used to estimate organ and effective doses for CBCT acquisitions with Imaging Performance Assessment of CT scanner(impact) calculator. Results: CTDIw for H and N, thoracic and pelvic acquisition protocols were 5, 9, 21 mgy for OBI and 1, 21, 27 mgy for XVI. For H and N scan, eye lenses, oral mucosa and salivary glands doses result respectively 6, 2.5, 2.4 mgy(obi) and 1, 0.6, 0.6 mgy(xvi). Regarding to thoracic acquisition, lung, thymus, heart and breast doses were 11,17,12,9 mgy(obi) and 17,28,15,22 mgy(xvi). As regards pelvic region, bladder, prostate and gonads doses were 34, 34, 27 mgy(obi) and 33, 33, 29 mgy(xvi). Effective doses in H and N, thoracic and pelvic acquisition were 0.5, 4, 7 msv(obi) and 0.04, 7, 6 msv(xvi). Conclusion: Although Impact calculation present several limitations, a useful estimation of organ and effective doses for clinical CBCT protocols used in two different centers, has been done. Patients undergoing radiation therapy at AOU of Parma are exposed to about 8 CBCTs in overall treatment. At Bellaria Hospital instead patients are exposed to a daily CBCT. In both situations estimated kv imaging doses have not a relevant influence on the total dose received by organs during the treatment. However, for young patients, a more accurate evaluation of benefit respect to the risk deriving from a daily additional radiation dose, should be done.

64 Effect of bladder filling on dosimetry for organs at risk (OAR) in high dose rate (HDR) vaginal cuff in brachytherapy M. Piergentili 1, O. Ferrando 1, F. Foppiano 1, M. Vanoli 2, T. Scolaro 2 (1) S.C. Fisica Sanitaria ASL5, La Spezia (2) S.C. Radioterapia ASL5, La Spezia Purpose: To investigate the effect of bladder filling on dosimetry and to determine the best bladder dosimetric parameter for HDR vaginal cuff brachytherapy Methods and Materials: Up to 2010, in our centre, we made adjuvant HDR vaginal brachytherapy (BT) for patients with endometrial cancer using two orthogonal radiography and the Treatment Planning System (TPS) Plato Nucletron. Since 2011 every brachytherapic treatment was planned on CT images with TPS Oncentra Brachy. Classic 2D dosimetry calculated from orthogonal radiographs offers limited OAR information, but the wide availability of 3D planning resources allows an accurate record of such data. We made 14 comparisons on patients with endometrial carcinoma who underwent CT scans with vaginal cylinder applicator inside. Each patient underwent EBRT to the pelvis and received three HDR treatments. The cylinder diameter was 2.5 cm. Treatment was administered using a 192 Ir HDR afterloader (Micro-Selectron, Nucletron, Veenendaal,The Netherlands). We made CT scans every treatment day in 2 conditions: empty bladder and bladder filled with 100cc physiological solution. All CT scans were performed with 1 mm slices from the lumbosacral junction to the ischial tuberosity. Medical doctor contoured Gross Tumour Volume (GTV), bladder and rectum in both bladder filling conditions on each slice of the 2 CT scans coregistered in the 3D treatment planning system. As GTV we considered the cylinder vaginal applicator extending in the upper half of the residual vagina. We planned, for every patient and in both bladder filling conditions, a standardized treatment planning in order to estimate, with the same conditions, dose differences for OAR. This treatment planning consist of 21 source position activated, dose normalization and optimization at 0,5 cm from cylinder surface and 500 cgy per fraction as dose prescription. Dose volume histograms (DVHs) were generated. Results: A set of rectal doses (D 0.5 cc, D 1 cc, D 2 cc ) and bladder dose (D 2 cc ) were assessed. Contouring of GTV on CT scans brought to a reduction of treated volumes in respect of the technique based on the two radiological orthogonal projections: mean irradiated volume length passed from 5 cm to 4 cm. GTV volumes in case of filled bladder resulted smaller than corresponding GTV volumes in case of empty bladder in 86% of the comparisons. Mean intersection of GTV volumes in case of empty and filled bladder was 91% of GTV volume. Standardized treatment planning showed a dose reduction to 0.5 cc of the rectum in case of filled bladder in 86% of the comparisons and a dose increasing to 2 cc of the bladder in 79% of the comparisons.

65 % different 0.5cc rectum and 2cc bladder 15% % diff dose (full-empty) 0.5 cc rectum % diff dose (full-empty) 2 cc bladder 10% 5% 0% -5% % -15% -20% n patient As a mean result: if the treatment was performed in filled bladder condition 0.5 cc of the rectum receive 4% of dose less and the bladder 3% of dose more in respect of empty bladder condition. mean rectal doses mean bladder doses full bladder empty bladder D0,5 cc D1 cc D2 cc D2 cc full bladder empty bladder Conclusion: Variations in bladder filling condition effect in a variable behaviour on rectum and bladder dose. It seems to be worthwhile to evaluate, on dosimetric basis, for every individual patient if it s better to threat vaginal cuff with empty or filled bladder.

66 Tab 1: Dose-volume histogram D 0,5 cc D 1 cc D 2 cc Bladder empty bladder ±13 full bladder ±13 Rectum empty bladder 107±9 99±9 90±9 full bladder 103±8 98±8 86±8 D 0,5 cc, D 1 cc, D 1 cc dose percentage related to the prescribed dose (5Gy) References: [1] S. Sabater, M. Sevillano, I. Andres, R. Berenguer, S. Machin-Hamalainen, K. Müller, M. Arenas, Reduction of rectal doses by removal of gas in the rectum during vaginal cuff brachytherapy, Strahlentherapie und Onkologie (2013) 4, Epub ahead of print [2] A. Stewart, R. Cormack,H. Lee, L. Xiong, J.L. Hansen, D. A. O Farrell, A.N. Viswanathan, Prospective clinical trial of bladder filling and threedimensional dosimetry in high-dose-r-ate vaginal cuff brachytherapy, Int. J. Radiation Oncology Biol. Phys.(2013) Vol. 72, No. 3, pp , [3] H. Yamashita, K. Nakagawa, K. Okuma, A. Sakumi, A. Haga, R. Kobayashi, K. Ohtomo, Correlation between bladder volume and irradiated dose of small bowel in CT-based planning of intracavitary brachytherapy for cervical cancer, Jpn J Clin Oncol (2012) 2

67 In vivo dosimetry by EPID for 3D-CRT, IMRT and VMAT: an update of Diso-INFN project A. Fidanzio 1,2,6, S. Cilla 3,6, F. Greco 1,6, L. Azario 1,2,6, M. Russo 4,6,S. Zucca 5,6, A. Piermattei 1,2,6 (1) Istituto di Fisica, Università Cattolica del S. Cuore, Rome, Italy. (2) U.O. di Fisica Sanitaria, Università Cattolica del S. Cuore, Rome, Italy. (3) U.O. di Fisica Sanitaria, Fondazione di Ricerca e Cura Giovanni Paolo II, Università Cattolica del Sacro Cuore, Campobasso, Italy. (4) Unità Operativa di Radioterapia, Ospedale Belcolle, Viterbo, Italy. (5) U. O. di Fisica Sanitaria, Presidio Oncologico Businco, Cagliari, Italy. (6) Istituto Nazionale di Fisica Nucleare, Sezione Roma 3, Rome, Italy. Purpose: The IVD plays an important role in the chain of dosimetric verification in a Radiotherapy Department. In this work a EPID-based in vivo dosimetry (IVD) for 3D-CRT, IMRT and VMAT, developed in the DISO project supported by the Istituto Nazioneale di Fisica Nuclare (INFN), is described. The EPID-based IVD, that supplies the isocenter dose reconstruction, in combination with 2D gamma analysis of the portal images, is a fast and accurate tool to guarantee the accuracy of the delivered dose. Material/methods: In the last three years, the IVD procedures developed in DISO have been applied in 8 radiotherapy centers for pelvic, thorax, breast and head 3D-CRT treatments. Recently three centers performed IVD for IMRT treatments and one center performed IVD for VMAT treatments. Tolerance levels of ±5% for all the treatments were estimated in the comparison, between the reconstructed isocenter dose, D iso, and the D iso,tps computed by the TPS. The 2D image gamma analysis was performed with 5%, 3 mm criteria to agreement. Dedicated software, interfaced with the Record & Verify system in use in each center, were developed to automate the IVD analysis. Results: About tests for about 1500 patients have been checked for 3D-CRT treatments, using Varian, Elekta and Siemens linacs. 50 patients were checked for step and shoot and sliding windows IMRT treatments with Elekta and Varian linacs and 20 patients were checked for VMAT treatments with a Varian linac. The dosimetric discrepancies were essentially due to patient s set-up errors, patient s morphological changes, attenuating media accidentally interposed between the source and the patient, beam output fluctuations, linac laser misalignments and TPS implementation errors. The results of the multicenter application will be used to develop the action-guidelines for the IVD checks out of tolerance. Conclusion: EPID-based IVD is until now performed in a limited number of centers and the first reason that IVD verification is not yet applied on a large scale is that dedicated software only recently became commercially available and second reason is that software should be easy to implement and rapid in supplying the response in clinical practice.

68 Methods for robustness evaluation of proton scanning beams treatment plans Metodi per la valutazione della robustezza di piani di trattamento con protoni e scanning attivo del fascio L Widesott 1,2,4, G Gargano 1, A J Lomax 3,4, M Schwarz 1,2 1 Agenzia Provinciale per la Protonterapia, Trento, Italy 2 Azienda Provinciale per i Servizi Sanitari, Trento, Italy 3 Center for Proton Radiation Therapy, Paul Scherrer Institute, 5232 Villigen PSI, Switzerland 4 Department of Physics, Swiss Institute of Technology, 8092 Zurich, Switzerland Abstract Several variables can influence the robustness, reliability, of a radiotherapy plan, such as accuracy of the dose calculation algorithm, uncertainties in beam monitoring and calibration, anatomical changes (inter and intra fraction), set-up error, misalignments between imaging and delivery systems, etc. For any given patient, these errors can be either systematic (i.e. with the same module and direction throughout the treatment) or random (i.e. subject to statistical fluctuations in either direction). In the practice of radiotherapy, the estimation and reporting of uncertainty has historically been at best implicit. The theoretical high precision of the intensity modulated proton plans necessarily requires an explicit evaluation of their accuracy/reliability. Hence, the need for tools that can aid clinicians in the evaluation of the treatment plan, hoping to find the best compromise between completeness of information and ease of use. The aim of this study is: 1) compare different tools to analyze plans robustness in active scanning proton therapy, pointing out their strengths and weaknesses; 2) assess the advantages of an accurate sampling of the uncertainties space. We simulated range uncertainty as systematic error and setup uncertainty as both random and systematic error. Both range and setup errors are modelled via a discrete set of 273 possible error scenarios. Overshoot and undershoot is modelled by a ±3% scaling of the CT Hounsfield numbers. The following methods were evaluated: a) agreement within a dose interval (ΔD), b) Probability-Volume Histograms, c) worst case scenario, d) worst case scenario within a given probability threshold, e) dose-volume (DV) bands, and f) probability distribution of a dosimetric index. Worst case analysis is expected to be too conservative (shows information that may occur with very low probability) that is why we believe it is better to use the worst case analysis at 95% probability or ΔD dose agreement, to avoid being influenced by events very unlikely. The introduction in clinical software of robustness evaluation tools, like DVH with percentile bands and probability distribution of dosimetric indexes as D1%, D99% and EUD, will allow the clinician to have a quick and more realistic estimate of the indices chosen for evaluating the radiotherapy plan. These tools are also necessary to compare the CTV coverage obtained with different techniques (i.e. photons vs. protons) and/or compare different PTV definitions.

69 Critical issues in IORT with mobile linac: geometric and dosimetric aspects. S. Andreoli, M. Fortunato, P. Colleoni, R. Moretti. A.O. Papa Giovanni XXIII, Bergamo Purpose: To explain the solutions adopted in our center to overcome the geometric and dosimetric issue in IORT with mobile linac. Methods and materials: A mobile LINear ACcelerator (Novac7, NRT) has been installed at Papa Giovanni XXIII Hospital, Bergamo; it produces electron beams with mean energy up to 7,2 MeV and it is employed for IntraOperative Radiation Therapy (IORT) of early breast cancer. The beam collimation is performed by perspex flat cylindrical applicators (in clinical practice: diameter 4, 5 and 6 cm with Sorce Skin Distance 80 cm), mounted on the exit window head. From 02/2006 to 09/2013, 737 female patients have been irradiated with the maximum available energy (E 0 = 7.2 MeV with R 90 = 21 mm) (dose prescription: 12/21 Gy to 90 % isodose; target thickness: cm). Internal shields (Al, Pb) have been used to protect the organs at risk under the target (the aluminium shield in contact to the target, to limit the backscattered dose from lead). There are two critical aspects in this practice: geometrical issue, for the non-matching between the treatment and the beam characterization set-ups (due to the irregular morphology of the target surface and the herniation of tissue inside the applicator) and dosimetric issue related to the output reproducibility (depending of the linac management: daily warm-up, shut-down, restrike for treatment), the evaluation of backscattered dose from internal shields and the absolute dose measurement. To solve the geometric aspect, a simple approach has been adopted: the use of a plex disk between the applicator and the target surface. The disk diameter is 2 cm larger than the correspondent applicator; so, when the applicator is positioned, the disk lightly compresses the target surface and guarantees a uniform dose distribution in the target and a useful build-up effect (fig.1). Considering the light compression determined from the plex disk, the correct measure of the target thickness is carried out by a needle inserted in the holes of a perforated disk temporary positioned on the target surface, until it intercepts the protection shields under the treatment volume. So, fixed the treatment energy, the thickness of the target volume determines the choice of the disk thickness (available: 2 mm and 5 mm). Moreover, to verify the accuracy in dose delivering in the clinical practice, a systematic in vivo dosimetry has been implemented and a μmosfet detector (TN-502-RDM, Best Medical) can be firmly sandwiched between the disk and the target surface [Radioth and Oncol 81(suppl.1): S504; Radioth and Oncol 92(suppl.1): S231]. Altogether, about one hundred µmosfets have been employed in clinical practice. The life span of a μmosfet is limited and cumulated dose-dependent; setting the standard sensitivity (about 1 mv/cgy), the cumulative total dose is about 200 Gy and the sensitivity changes about +1% every 20 Gy of absorbed dose up to 140 Gy and remains costant at higher doses. Before the clinical use, each detector was calibrated in a slab phantom (RW3), at the energy in use and in presence of the plex disk (2 mm and 5 mm), with a dose of 5 Gy. As reference dosimeter, a Roos chamber was placed to buildup. For the dosimetric characterization, a set of ten μmosfet was irradiated (with the energy in use) in a slab phantom, at depht of 5 mm: five to determine the sensitivity to cumulated dose relationship (exposure of 5 Gy up to their breakdown) and the remaining to verify linearity and reproducibility. To solve the dosimetric aspect related to reproducibility, a conditioning modality of linac has been implemented and systematic simulations of treatment are performed in slab phantom to check the output with a similar timing presenting in the clinical practice (daily warm-up, shut-down, restrike after two hours and a single exposure). To evaluate the contribution of shields backscattering, MonteCarlo simulations (Fluka code) and experimental validations with gaf-chromic have been performed (fig.2) [Journal of Physics, Conference Series, 74: The optimisation process of plex thickness, considering also the backscatter dose contribution of shields, brought to the following treatment set-up: for target thickness up to 15 mm, a 5 mm plex disk; for target thickness between 16 and 23 mm, a 2 mm plex disk. For target volume thickness between 24 and 27, again a 2 mm plex disk but with the lead shield in contact with the target).

70 Result: In clinical practice, the measured vs expected dose difference has been -0.4%±2.6% (1 DS) (range: %; 75 th percentile: 2.9%); the output reproducibility in simulations has been within ±3.0% (0.1%±0.9% (1 DS)) (fig.3). Conclusion: The adopted set-up for treatment and in-vivo dosimetry ensures a high level of accuracy of the overall treatment. The use of μmosfet is feasible and does not affect the surgical operativeness and the dose distribution inside the target. usual setup applicator target adopted setup plex disk Internal shields fig.1 PDD (%) PDD (no shields) 100/0 0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 target thickness (cm) PDD (%) fig PDD (no shields) 100/0 0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 target thickness (cm) misures (%) N = 737 (treatments) -7% - 5% - 3% - 1% 1% 3% 5% 7% misures (%) N = 737 (treataments) 0 0% 1% 2% 3% 4% 5% 6% 7% 8% Δ % measured VS prescribed) (clinical practice) Δ % measured VS prescribed) (clinical practice) misures (%) N = 340 (simulations) -7% - 5% - 3% - 1% 1% 3% 5% 7% Δ % (measured VS prescribed) (simulations with chamber) misures (%) fig N = 340 (simulazions) 0 0% 1% 2% 3% 4% 5% 6% 7% 8% Δ % (measured VS prescribed) (simulations with chamber)

71 Applicazione del report AAPM TG 119 su un fantoccio dosimetrico 3-D per l implementazione di tecniche IMRT e VMAT Application of AAPM TG 119 report on a 3-D dosimetric phantom for IMRT and VMAT commissioning Luca Trombetta 1, Angelo F. Monti 1, Maria Grazia Brambilla 1, Claudia Carbonini 1, Maria Bernadetta Ferrari 1, Alberto Torresin 1 (1) Department of Medical Physics, A.O. Niguarda Ca Granda Milan, Italy Purpose: The capabilities and dose calculation accuracy of TPS and treatment delivery should be carefully verified for an overall clinical commissioning. The American Association of Physicists in Medicine- Task Group 119 (TG119) proposed a water equivalent square slab phantom (30x30x15 cm 3 ) to be used with four particular IMRT test plans. TG119 defines structures, beams arrangements, IMRT goals and methods to analyze the dosimetric results on its phantom. All test suites, including DICOM- RT images and structures, can be downloaded from the AAPM web- site. The TG119 phantom is easily reproducible in every department, but it allows only single point or planar measurements. In this work, we proposed to use a new 3D dosimetric phantom in order to apply the TG119 report in a more sophisticated situation and ride over this limitation. All tests were thus re- optimized to satisfy goals defined into the TG119 with IMRT (static and dynamic) and VMAT techniques. Methods and materials: TG119 was initially used as is in order to test the capability of our clinical arrangement. After this, TG119 structures were superimposed on the CT images of a cylindrical PMMA phantom surrounding two orthogonal matrixes with 1069 total diodes (Delta4; Scandidos, SWE). TG119 tests were thus calculated and optimized using a Monte Carlo TPS (Monaco 3.20; Elekta, SWE) for 6 and 10 MV photon beams, with IMRT and VMAT techniques. Delta4 phantom was used in order to carry out comparison between measured and planned 3D absolute dose distributions. A global 3%- 3mm gamma test with a 10% of maximum dose threshold was performed for plans analysis. Results: Goals proposed in TG119 were satisfied for each plan and technique. For all static and dynamic IMRT plans of all energies, the percentages of gamma passing points were within the range of 99.4%- 100%, 0.1 with a mean value of %. Similar results were obtained for VMAT plans for both 6 and 10 MV, with a 0.2 range of gamma passing points of 97.4%- 100% and a mean value of 99.1 ± 0.9%. Conclusions: TG119 structures and plans were found to be easily adaptable to Delta4 phantom CT enabling a rigorous procedure for IMRT and VMAT plans verification. TG119 applied with the Delta4 phantom could thus be a practical commissioning procedure for modulated arc therapy too even if this solution is unfortunately not available in every department for intercomparison purposes.

72 TomoDirect e 3DCRT: un confronto fra tecniche per il trattamento adiuvante del carcinoma mammario TomoDirect and 3DCRT, a techniques comparison for the adjuvant treatment of breast carcinoma Cambria R 1, Cattani F 1, Luraschi R 1, Pansini F 1, Bazani A 1, Morra A 2, Leonardi MC 2, Pedroli G 1, Orecchia R 2,3. (1) Medical Phys Dep, (2) Radiation Oncol Dep, Istituto Europeo di Oncologia, Milano; (3) Università degli Studi di Milano, Milano Purpose: The TomoDirect was set up in our Institute in January 2012 and then implemented as a technique for the adjuvant breast radiotherapy. The purpose of the study is to compare the TomoDirect treatment modality to the standard 3DCRT. Methods and materials: We compared the treatment plans of 30 patients consecutively treated with the TomoDirect (TD) modality. Clinical target volumes (CTV) and organs at risk (OARs) were contoured for all the patients by the same physician to avoid interobserver variability; a PTV (planning target volume) was generated by adding a 5 mm uniform margin to the CTV. Patients underwent the whole breast irradiation and a simultaneous integrated boost to the tumor bed region. The prescribed doses were Gy/fr up to a total dose of 45-50Gy (20 fr) to the whole breast and to the postsurgical area respectively. Plans for TD and 3DCRT were both optimized, according to our Institutional protocol, in terms of dose coverage to target and constraints to the OAR: PTV(breast), V 95% 95%, D 50% 108%, V 100%(boost dose) 30%, D max 115%; PTV(boost), V 95% 95%, D max <115%; Heart (right breast), D max <16-20Gy, V 8Gy <10-15%; Heart (left breast), D 5% <16-20Gy, V 8Gy <30-35%; Heart (right/left breast) D mean <3.2-4Gy; Ipsilateral lung V 16Gy <15-20%, V 8Gy <35-40%, V 4Gy <50%, Controlateral lung V 4Gy <10-15%; Controlateral breast D max < Gy. Results: The dosimetric results which are statistically significant (P<0.05) are reported as median (%) plus/minus the standard deviation (%) for 3DCRT and TD respectively: PTV(breast), V95%: 96±3%; 99±1%; D max : 116±2 %; 113±32 %; V100 (boost dose): 7±5%; 0±1%; Ipsilateral Lung, V8Gy: 13±6%; 17±4%; V4Gy: 31±9%; 24±8%; Heart (left breast): D mean 3±1%; 1±1%; Controlateral breast: D max 3±3Gy; 3±2Gy. Concerning the treatment time and the planned monitor units, the firsts where 95±15s and 378±55s while the monitor units were 278±13 and 5199±822 in the in the 3DCRT andtd cases respectively. Conclusion: TD was investigated as an alternative technique to the 3D conformal one for the postoperative breast radiotherapy. DVHs show an improvements in PTV coverage and heart sparing. On the contrary, the treatment time and the monitor units of the TD technique are about 4 and 18 times those of the 3DCRT. The real advantage of the TomoDirect is the possibility of performing the image guided radiotherapy and in some cases results suggest that the use of the TD technique could be favourable.

73 Dosimetric comparison of prostate IMRT and Helical Tomotherapy treatment planning. A. Didona 1,M. Marcantonini 1, M. Iacco 1, C. Zucchetti 1, A.C. Dipilato 2, R. Bellavita 2, G. Gobbi 1 (1) Medical Physics Department, Santa Maria della Misericordia Hospital, Perugia (2) Radiotherapy Department, Santa Maria della Misericordia Hospital, Perugia Purpose: There are indications from published studies that the use of high doses (even above 80 Gy) radiation therapy appears to be effective for patients with prostate cancer leading to better local tumour control. However, radiation induced complications can occur during the course of treatment (such as urinary incontinence and rectal bleeding) or some months or years after it. The newer forms of treatment such as Intensity-modulated radiotherapy (IMRT) and Helical Tomotherapy (HT) lead as advantage the possibility of a dose escalation to the tumour while minimizing toxicity to normal tissue. HT is an innovative intensity-modulated radiation therapy technology which combines the characteristics of a 6 MV linear accelerator with the facilities of a computed tomography unit; in particular, the helical beam delivery in HT is expected to offer a better conformation of the dose to the target when compared to conventional IMRT treatments. The purpose of this study is to compare Prostate (P) and Prostate with Seminal Vesicles (P&SV) IMRT Step and Shoot and HT treatment planning techniques. Methods and materials: We selected a sample of 10 prostate cancer patients (5P, 5P&SV) with a wide variety of prostate, bladder and rectal volumes. All patients were CT scanned and prostate and seminal vesicles CTVs and Organs At Risk (OARs), including rectum, femoral heads and bladder, were countered on CT images. To obtain prostate PTV (PTV1) the relative CTV was expanded of 1cm in all directions except 0.5cm posteriorly. The seminal vesicles PTV (PTV2) was obtained from the relative CTV in the same way avoiding PTV1. The dose prescription for this study was 74.25Gy for PTV1 and 62.04Gy for PTV2 in 33 fractions and it was required that PTVs D 95% be 95% of prescribed dose. HT planning utilized an incident 6MV beam, a field width of 2.5 cm, a pitch of and a modulation factor of 2. Treatment planning was achieved by a proprietary inverse planning system specific of HT based on minimization of an objective function which made use of precedence, importance and penalty factor parameters. In addition, OAR overlaps with PTVs were considered as target volumes with a prescription of D 90% of PTV prescribed dose and a maximum dose of 95% of PTV prescribed dose. IMRT plans with 5 coplanar 15MV beams (180, 255, 325, 35 and 105 ) were generated using Pinnacle3 TPS with maximum number of segments 25 (P) and 35 (P&SV) and 8cm 2 minimum segment area. As a starting point for optimization dose constraints used for OARs for both techniques were V 40Gy <60%, V 60Gy <40% and V 70Gy <25% for rectum and bladder and D 2% <47.5Gy for femoral heads. However, after reviewing the dose distributions, the objectives could be modified to obtain more satisfactory dose distributions and more OARs sparing. Results: In Table 1 a summary of the planning data is reported. Homogeneity Index (HI) is calculated as: D2% D98% HI D50% A HI of zero indicates that the absorbed dose distribution is completely homogeneous. Paired t-test was used to determine if the dosimetric results from a comparison was significant based on significance being defined as p<0.05. The HT and IMRT plans were compared in terms of V 40Gy, V 60Gy and V 70Gy for bladder and rectum, D 2% for femoral heads and HI for PTVs. Comparison results are shown in Table 2.

74 Prostate Prostate & Seminal Vesicles IMRT TomoTherapy IMRT TomoTherapy PTV1 D 2% (Gy) 75.7± ± ± ±0.2 D 50% (Gy) 74.1± ± ± ±0.4 D 98% (Gy) 69.7± ± ± ±1.5 Mean Dose(Gy) 73.8± ± ± ±0.4 HI 0.07± ± ± ±0.02 PTV2 D 2% (Gy) 70.6± ±3.8 D 50% (Gy) 63.0± ±0.4 D 98% (Gy) 56.9± ±1.4 Mean Dose(Gy) 63.4± ±1.0 HI 0.217± ±0.07 Femoral Heads (R/L) D 2% (Gy) 34.2±8.2/32.5± ±0.9/20.1± ±8.8/33.6± ±1.2/25.6±2.0 Rectum V 40Gy (%) 34.7± ± ± ±17.5 V 60Gy (%) 11.1± ± ± ±9.8 V 70Gy (%) 2.6± ± ± ±3.3 Bladder V 40Gy (%) 17.0± ± ± ±14.5 V 60Gy (%) 10.8± ± ± ±7.6 V 70Gy (%) 5.9± ± ± ±3.9 Table 1 Summary of planning data Table 2 Comparison Results Parameter P value HI PTV1 p = HI PTV2 p = Right Femoral Head D 2% p = Left Femoral Head D 2% p = Rectum V 40Gy p = Rectum V 60Gy p = Rectum V 70Gy p = Bladder V 40Gy p = Bladder V 60Gy p = Bladder V 70Gy p = Conclusion: HT appears to provide plans that have at least equivalent quality to IMRT in terms of targets coverage. IMRT PTV1 shows better HI, probably due to the different planning protocols; in any case, differences

75 are not statistically significant. On the contrary, PTV2 dose distribution is more homogeneous in HT plans (statistically significant) due to the better modulation achieved by an helical technique. Statistically significant reduction of the maximum dose to the femurs is achieved by HT. No significant differences are shown in V 40, V 60, V 70 for bladder and rectum between IMRT and HT plans, but for the latter high doses are better controlled and V values are lower. In conclusion, this study demonstrates that HT provides some dosimetric improvements in terms of femoral heads sparing and seminal vesicles dose distribution homogeneity in patients with prostate cancer. However, IMRT and HT plans are equivalent regard rectum and bladder sparing, probably due to the limited sample size and its heterogeneity. References: [1] Rodrigues G, Yartsev S, Chen J, Wong E, D'Souza D, Lock M, Bauman G, Grigorov G, Kron T, A comparison of prostate IMRT and helical tomotherapy class solutions, Radiother Oncol Sep;80(3): [2] ICRU Report 83, Prescribing, Recording, and Reporting Intensity-Modulated Photon-Beam Therapy (IMRT) [3]Zelefsky MJ, Pei X, Chou JF, Schechter M, Kollmeier M, Cox B, Yamada Y, Fidaleo A, Sperling D, Happersett L, Zhang Z, Dose escalation for prostate cancer radiotherapy: predictors of long-term biochemical tumor control and distant metastases-free survival outcomes. Eur Urol Dec;60(6):

76 Dosimetric characterization of a synthetic single crystal diamond diode for radiotherapy electron beam dosimetry M.D. Falco 1,2, P. Bagalà 1, C. Di Venanzio 2, A.S. Guerra 3, Marco Marinelli 2, E. Milani 2, M. Pimpinella 3, G. Prestopino 3, R. Santoni 1, A. Tonnetti 2 and G. Verona-Rinati 2 1 Department of Diagnostic Imaging, Molecular Imaging, Interventional Radiology and Radiotherapy, Tor Vergata University General Hospital, Viale Oxford 81, Rome, Italy 2 INFN Department of Industrial Engineering, University of Rome "Tor Vergata", Via del Politecnico 1, Rome, Italy 3 Istituto Nazionale di Metrologia delle Radiazioni Ionizzanti, ENEA-INMRI C R Casaccia,Via Anguillarese 301, Roma, Italy The aim of this work is to investigate the dosimetric properties of a synthetic single crystal diamond diode prototype developed at Rome Tor Vergata University laboratories, in megavoltage clinical electron beams from a linear accelerator. Particular care was devoted to the device response in electron small field sizes. Diamond dosimeters based on such a device are now commercialized by PTW-Freiburg (microdiamond n ) and Tor Vergata University General Hospital is one of the four European PTW-Freiburg beta test facility for such diamond dosimeters and the only one in Italy. Synthetic single crystal diamond (SCDD) dosimeter was fabricated in a Schottky photodiode configuration. In all of the tests reported it was operated in photovoltaic-mode, i.e. with no external bias voltage applied. The detector sensitive volume, of about mm 3, is defined by the depletion region extending through the whole thickness of the thin intrinsic diamond layer (1.0±0.1 µm thick), below the circular metallic contact. A picture of the investigated detector is showed in Fig. 1. The SCDD was tested for regular electron beam field sizes from 6 6 to cm 2 obtained using standard applicators and small electron beams shaped by using commercial tubular applicators 2, 3, 4 and 5 cm in diameter, with energy in the range 6 to 15 MeV, generated by an Elekta Precise linear accelerator (Elekta Crawley, UK). Linearity, dose rate dependence, output factors, percentage depth dose and lateral field profiles were investigated. Measurement results from the diamond detector were compared to those from reference detectors routinely used for therapeutic electron beam dosimetry, i.e. two plane parallel ionization chambers (PTW Advanced Markus type and Scanditronix Wellhöfer type PPC05), a cylindrical ionization chamber (PTW Semiflex type 31010) and a p-type silicon detector (PTW Dosimetry Diode E type 60017). A good linear behaviour of the SCDD response as a function of the delivered dose is observed with deviations below ± 0.3% in the dose range from 0.02 Gy to 10 Gy (Fig. 2). The R 2 parameter of the linear best fit was found to be 1 with a precision of From the slope of the linear fit, a sensitivity of nc/gy for the SCDD was obtained. The normalized percentage ratio, R, of the SCDD signal to the SF-IC signal is shown in Fig. 2(b) as a function of delivered dose, where deviations no larger than ± 0.3% can be observed down to the minimum delivered dose of 0.02 Gy. In addition, the detector response is dose rate independent, with deviations below 0.3% in the investigated dose rate range from 0.17 to 5.45 Gy/min. Percentage depth dose curves obtained from the diamond detector are in good agreement with the ones from the reference dosimeters both for standard and tubular applicators. As an example, in Fig. 3 the percentage depth dose curves measured by Si-D, SCDD and PPC05 for 6, 8, 10, 12 and 15 MeV for a cm 2 field size are displayed. Difference plots in % between the diodes and PPC05 are also shown. Lateral beam profile measurements evidence a better spatial resolution of solid state detectors with respect to that of ionization chambers, being the one from SCDD comparable to that of the silicon diode. Fig. 4 shows the in-plane electron beams normalized profiles measured at the R 100 depths for 6 MeV (a) and 15 MeV (b) for all the utilized tubular applicators.the R 100 values were derived for both Si-D and SCDD detectors from the PDDs acquired in the same irradiation conditions. A good

77 agreement within experimental uncertainties was also found in terms of output factor measurements between SCDD and reference dosimeters. As concerns the tubular applicators, two different effects were evidenced by both the solid state detectors: (i) they measured, in the case of the highest beam energy and 5 cm diameter tubular applicator, OFs of about 1.3, that is 30% greater than the values obtained at the 10x10 cm 2 reference field and (ii) a noticeable increase of the OF values with increasing beam energy, up to about 100 % in the case of the smaller applicator. The observed dosimetric properties indicate that the tested diamond detector is suitable for relative dosimetry in clinical electron radiotherapy. Small electron beams shaped by using commercial tubular applicators have OFs which depend considerably on the field size; we recommend to measure them for each field separately rather than attempt to interpolate. References: [1] Ciancaglioni I, Marinelli M, Milani E, Prestopino G, Verona C, Verona-Rinati G, Consorti R, Petrucci A and De Notaristefani F. Dosimetric characterization of a synthetic single crystal diamond detector in clinical radiation therapy small photon beams. Med. Phys. (2012) 39(7) [2]Di Venanzio C, Marinelli M, Milani E, Prestopino G, Verona C, Verona-Rinati G, Falco M D, Bagalà P, Santoni R and Pimpinella M. Characterization of a synthetic single crystal diamond Schottky diode for radiotherapy electron beam dosimetry. Med. Phys. (2013) 40 (2) [3] Zani M, Bucciolini M, Casati M, Talamonti C, Marinelli M, Prestopino G, Tonnetti A, Verona-Rinati G. A synthetic diamond diode in volumetric modulated arc therapy dosimetry. Med. Phys. (2013) 40(9) [4] Pimpinella M, Ciancaglioni I, Consorti R, Di Venanzio C, Guerra AS, Petrucci A, Stravato A, Verona-Rinati G. A synthetic diamond detector as transfer dosimeter for Dw measurements in photon beams with small field sizes. Metrologia (2012) 49(5) S207-S M SCDD (nc) a) Fig.1 Synthetic single crystal diamond diode. (R/ R 1 Gy -1) 100 b) dose (Gy) Fig.2: Measured SCDD charge as a function of the dose for a 10 MeV electron beam. (b) Percentage deviation of the ratio, R, between SCDD and ionization chamber signal normalized to the value obtained with a delivered dose of 1 Gy.

78 SSCD Si-D PPC05 PDD Difference MeV Si-D - IC; SCDD - IC MeV MeV MeV MeV Depth (mm) Fig. 3: Percentage depth dose curves measured by Si-D, SCDD and PPC05 for 6, 8, 10, 12 and 15 MeV for a 10x10 cm 2 field size. Difference plots in % between the diodes and PPC05 are also shown. Normalized Charge (a.u.) (a) 6 MeV SCDD 5 cm SCDD 4 cm SCDD 3 cm SCDD 2 cm Si-D all cones Normalized Charge (a.u.) SCDD 5 cm (b) 15 MeV SCDD 4 cm SCDD 3 cm SCDD 2 cm Si-D all cones Inplane Position (mm) Inplane Position (mm) Fig. 4: In-plane electron beams normalized profiles measured by Si-D and SCDD for 6 MeV (a) and 15 MeV (b) for all the field size diameters.

79 GAFCHROMIC EBT3 FILMS FOR ROUTINE QUALITY ASSURANCE OF CLINICAL PHOTON AND PROTON BEAMS C. Stancampiano 2,3, L. Raffaele 1,2,3, G. Politi 1,2, F. Romano, 2,3, A. Gueli 3, S. Spampinato 1,3 [1] A.O.U Policlinico Vittorio Emanuele, Catania [2] INFN-LNS, Catania [3] Scuola di Specializzazione in Fisica Medica, Facoltà di Medicina, Catania PURPOSE The usefulness of new generation GAFchromic EBT3film (Z eff =6.73) has been investigated for radiotherapy quality assurance (QA) of megavoltage photon beams (6 MV) and narrow eye proton beams (62 MeV) of the Catana facility [1]. Batches with sheet dimensions of cm 2 were used for this study. METHODS AND MATERIALS A commercial flatbed scanner EPSON Expression 10000XL and the software package FilmScan (PTW, Freiburg) were used for the digitization of irradiated EBT3 films. Images were acquired in transmission mode and landscape orientation; RGB-positive images were collected at a depth of 16 bits per color channel at 72 dpi resolution, corresponding to a pixel size of mm 2, data were saved in a TIF format. Each scan was taken over an area corresponding to the sheet dimension minus 0.5 cm on all sides ( cm 2 ) even if only a small piece of film was scanned; all film samples were placed in the middle of the scan window with the aid of a positioning frame and scanned in a consistent orientation. The red channel of the scan images was extracted and raw signal processed to dosimetric evaluations by using PTW FilmAnalyze software. Because of the post irradiation growth, as suggested by the manufacturer film scans were performed 2 h after irradiation to stabilize the film response. The standard geometry used for film calibration was an isocentric perpendicular setup: film samples sized 3 3 cm 2 were irradiated with a 6 MV photon beam from a Siemens Oncor Linac, at a depth of 5 cm in a solid water phantom (RW3); a cm 2 field size at isocenter was used. Proton irradiation was accomplished on the horizontal beam line of the Catana proton therapy facility, using 62 MeV mono-energetic beams; strips 3 3 cm 2 were irradiated in a RW3 solid phantom at 1 mm depth in the plateau region of the Bragg curve, corresponding to a residual range (R res ) of about 30 mm. Films were irradiated with the reference 25 mm diameter circular collimator. For both radiation types films were irradiated to different doses in the range of 0.15 to 4 Gy; for each sheet used for calibration measurements some film pieces were left unexposed for background determination (zero dose point). Dose rates at calibration depth were 3 and 15 Gy/min., respectively for photon and protons The scanned calibration films were calibrated in terms of pixel value (PV) directly to applied dose using PTW FilmCal software); direct calibration minimizes uncertainties in calibration chain, by removing the conversion to and from optical density. Four fiducial markers were added to EBT3 films using a fine marker pen. Markers were drawn to indicate the position of the light-projected crosswires for reference in dosimetry and QA measurements. Scanner non-uniformity, especially in the direction perpendicular to the scan direction, was measured, with a maximum variation in PV relative to scan center of about 4%. A flattening correction for the scanner nonuniformity has been carried out from the images of a homogeneous exposed film (2 Gy, 6 MV x-rays, focus to film distance=235 cm, cm 2 at isocenter), smoothed by means of a median filter (6 6 pixels). The flattening correction, evaluated over the scan area ( cm 2 ), reduces for the scanner non-uniformity to ± 2.0 % in dose values.

80 RESULTS To evaluate the intrasheet uniformity 48 small pieces (3 3 cm 2 ) cut from a full-size film sheet were irradiated to 2 Gy one after the other. A region of interest (ROI) of 1 1 cm 2 at sample center was selected to obtain the mean pixel value with its standard deviation in the ROI. The intrasheet uniformity in the red channel [( PV max PV min) /(PV max+ PV min) 100 ] resulted to be within 1%, assuring uniformity response on the total surface of EBT3 film. Also the sheet to sheet homogeneity (intersheet umiformity), obtained by analyzing 96 small pieces from two EBT3 film sheets of the same batch resulted to be within 1%. Red channel calibration curves of EBT3 films in the investigated range fit to a third order polynomial (R 2 =0.9999) for both type of radiations (Fig.1); the differences in PV response with respect to radiation types are within 2 %, indicating only a weak dependence on radiation type (quasi water-equivalent detector)[2]. Moreover experimental results show that EBT3 film response does not vary by more than 2% for residual ranges proper to ocular proton therapy beams (6 25 mm); as a consequence only one calibration file is needed to evaluate EBT3 films exposed at different depths in SOBPs. For red channel, a 1% uncertainty in PV results in dose uncertainties of around 2% for doses up to 2 Gy, for larger doses, where calibration curve begins to flatten, dose uncertainty can increase considerably. Consequently a dose of 2 Gy was delivered to EBT3 films for QA and dosimetry applications in photon and proton beams; time difference between irradiation and scanning was always the same as for calibration. Fig.1 Comparison between photon and proton beam calibration in the red channel Experimental results demonstrate that EBT3 films are suitable for routine for routine quality assurance of high-energy photon beams, providing accurate quality controls as indicated in fallowing table. Light and Radiation Field Coincidence Radiation Field Centre to Crosshair Position Numerical and Radiation Field Size Asymmetrical Fields Congruence of Opposed Radiation Fields Leaf Position Accuracy Strip Test (pre IMRT Test) Radiation Isocentre (Star shot)

81 EBT3 films are also a valuable tool for quality control of Optifocus multileaf collimator (82 leafs) valid for conformal and modulated radiation treatments (pre-imrt test), especially regarding leaf position accuracy (radiological calibration), relative leaf positions (strip test) and transmission (intraleaf, interleaf, side-end). The leaf position accuracy was evaluated by irradiating EBT3 films using a shaped MLC field in which one of the edges is at 45 to the edge of the unblocked field, according to the IPEM Report n.81 (Fig.2)[3]. The films were than scanned and for each profile thus obtained (Fig.2) the position of the 50% of the dose relative to the beam axis dose was related to the actual leaf position; the value for the nominal (planned) leaf coordinate was taken from the MLC control unit. The use of calibrated EBT3 films provides an accurate calibration method of the MLC allowing detection of leaf-positioning errors of 0.5 mm compared to a tolerance of 1 mm. Fig.2 MLC leaf position test The test was extended to µmlc collimators (3D line, leaf size at isocenter = 4.7 mm), by irradiating EBT3 films with an irregularly-formed field approximated to a circle of 8 cm in diameter (Fig.3). The error in leaf position resulted to be not greater than 0.5 mm at isocenter, as prescribed by the manufacturer. EBT3 films were found to be suitable for small field 6 MV photon beam dosimetry.

82 Fig.3 µmlc collimators (3D line) test The value of S cp for a 1 1 cm 2 square field (1.096) is consistent with the corresponding value (1.102) measured with Edge Diode Detector (Sun Nuclear) assumed to be the reference detector for small beam dosimetry ( mm 2 ). Characterization of narrow proton beams for ocular tumor radiotherapy, with collimator areas up to 40 mm 2, requires special efforts, since the use of finite size detectors can lead to distortion of the measured dose. EBT3 films can be used as reference detector for accurate determination of lateral off-axis profiles and field-size dependent-factors of narrow eye proton beams [4]. Advantages of EBT3 include a very high spatial resolution, short beam-time, and the ability to provide proton dose distributions in a single exposure. EBT3 film is used before each treatment session at Catana facility to check the full energy (62 MeV) proton beam in terms of field size (W 50%), lateral penumbra, lateral flatness and lateral symmetry (Fig.4), with reference to the standard circular collimator (ϕ=25 mm). Fig.4 Inplane profile of reference standard circular collimator

83 Moreover EBT3 films are used to test each patient clinical setup, in particular for measuring the transverse proton beam parameters at mid-sobp, as W 95% treatment width, lateral penumbra and field flatness on principal and diagonal axes [5]; all the beam parameters are strongly dependent on modulator, range shifter and field shape adopted in the clinical practice. Also 2D isodose distributions of irregularly shaped collimators can be obtained at mid-sobp (Fig.5). Fig.5 2D isodose distribution of irregularly shaped collimators The dose per monitor unit (O.F.) for modulated clinical proton beams depends on the irradiated area, with a more significant change for smaller fields; furthermore the drop in dose per monitor unit with decreasing beam area is energy dependent; the O.F. has to be measured at the middle of SOBP, which represents the dose prescription point (ICRU point). The O.F.s for the narrowest shaped ocular proton beams (cgy/u.m) are routinely measured with EBT3 films, to avoid an underestimation which may occur by using other types of small detectors (stereotactic PFD diode detector, CVD diamond), as shown in (Fig.6). It should be stressed that alignment in small fields is not critical for EBT3 film techniques. Fig. 6

84 CONCLUSION Based on experimental results, we use EBT3 films for QA and dosimetry of high energy photon beams [6] and modulated clinical proton beams. BIBLIOGRAPHY [1] G. Cuttone, at all. CATANA protontherapy facility: the state of art of clinical and dosimetric experience, Eur. Phys. J. Plus (2011) 126, 65; [2] S. Reinhardt, at. all Comparison of Gafchromic EBT2 and EBT3 films for clinical photon and proton beam Med. Phys.(2012) 39, 8; [3] IPEM Report 91; [4] S M Vatnitsky, at all Dosimetry techniques for narrow proton beam radiosurgery, Phys. Med. Biol. (1999) 44, ; [5] Prescribing, Recording, and Reporting Proton-Beam Therapy (ICRU Report ); [6] L Richley, at all., Evaluation and optimization of the new EBT2 radiochromic film dosimetry system for patient dose verification in radiotherapy, Phys. Med. Biol. (2010) 55,

85 Integration of the DICOM images in the Geant4 Web-based iort_therapy application : a further step to support Intra-Operative Electron Radio-Therapy (IOERT). C. Casarino 1, S. Guatelli 2, G. Russo 1, G. C. Candiano 1, F. Romano 3, G.A.P. Cirrone 3, G. La Rocca 4, R. Barbera 5,4, G. Borasi 1,6, C. Messa 1,6,7, M.C. Gilardi 1,6,8. (1) IBFM CNR LATO, Cefalù, Italy. (2) CMRP University of Wollongong, Australia. (3) INFN LNS, Catania, Italy. (4) INFN Catania, Italy. (5) DPA University of Catania, Italy. (6) University of Milano Bicocca, Italy. (7) San Gerardo Hospital, Monza, Italy. (8) San Raffaele Scientific Institute, Milano, Italy. Introduction: IOERT permits to deliver a single high dose of radiation directly to the tumour bed, or to the exposed tumour, during surgery. The IOERT treatment is the result of a sequence of manually handled actions performed by an heterogeneous staff [1]. Contrary to what happens for other more complex radiotherapy techniques, a treatment planning system (TPS) is not yet routinely employed, and the only one system available is based on adapted version of the pencil beam algorithm. Otherwise, the urgent demand by radiotherapy-medical community for more precise TPS software opens significant challenges to reconcile the accuracy of Monte Carlo (MC) methods with the long computing times [2]. In this perspective, a Geant4 Web-based iort_therapy application was developed to support IOERT research and clinical activities at G. Giglio Hospital in Cefalù (Italy) [3,4]. It simulates the collimation system of the NOVAC7 mobile linear accelerator (linac), and a detector placed inside a standard water phantom to calculate the delivered dose. Actually, the application facilities and the web-work-flow permits us to address some technical IOERT requirements as the optimization of the collimation system, specific research activities on patient radio-protection, and dose distribution evaluation in some critical clinical case. Nowadays, the application is fully operative and available on the GARR Science Gateway [5] and running on the Italian Grid Infrastructure (IGI). This work reports about the way the DICOM images were integrated in iort-therapy. This represents an essential further step to adapt the application for a practical-clinical use. In fact, this will allow us to evaluate the dose distribution for treatment planning or for special post-surgery cases as to verify the correct placement of protection disc. Methods and materials: Iort_therapy application was merged with the principal facilities of the DICOM medical Geant4 example. This example converts a DICOM file to a simple ASCII file, transforming Hounsfield numbers in the corresponding densities and materials. This information is then used to construct a physical voxellized dicom volume so that it can be exploited in the Geant4 environment. Both codes, iort_therapy and DICOM, have been modified and adapted to be compatible to each other. In addition, the rotation and translation of the DICOM container and beamcollimator system were implemented, to emulate the clinical setup. As output files, the dicom.out and the.gdd file were produced. The first records the integral dose at each voxel, the second, opened with the gmocren toolkit (a volume viewer for Geant4 medical simulations) [6], visualizes the dose on dicom images as in a conventional TPS. The calibration curve to convert Hounsfield numbers to physical densities was obtained analyzing scans of a CIRS (model 062) Electron Density Phantom from our PET-CT machine (Discovery STE 8, GE Medical System). Three tests have been performed off-line to evaluate the dose deposited resolution obtainable by forcing to limit the cpu performance (Intel Q8400 Quad Core 2.7Ghz, 4Gb Ram) and gmocren display management capabilities (NVIDIA 220 GT 1GB video card). The first two tests by loading a CT water phantom 24 DICOM slices (25x25x6 cm 3 and 512x512x24 pixels). The third by loading a female trunk 50 DICOM slices (50x50x13 cm 3 and 512x512x50 pixels) and positioning the simulated NOVAC7 collimation system as in a typical breast-ioert treatment.

86 Results: Figure 1c shows the Hounsfield numbers vs. densities calibration curve obtained using a CIRS (model 062) Electron Density Phantom (figure 1a) from our PET-CT machine (Discovery STE 8, GE Medical System). This curve, given as input to our code, allows it to be able to transform the dicom Hounsfield numbers into density values. Figure 1. a) CIRS (model 062) Electron Density Phantom; b) DICOM; c) Calibration Curve CT vs. density. Figure 2 is relative to tests 1-2. After the CT water phantom (figure 2a) 24 DICOM slices (figure 2b) have been uploaded, the software constructs a physical voxellized volume (figure 2c), where to each voxel is assigned a material (water in this case). At this point, the particle paths are simulated, reproducing the interactions with matter traversed and then calculating the energy released voxel-by-voxel (figures 2d-2e-2f-2g). The last two pictures on the right reproduce the dose deposit when a metal disc is inserted inside the water phantom. Working configuration was obtained rescaling to 256x256x24 the original CT images, thus reducing the voxel volume to 1.0x1.0x2.5 mm 3. Figure 2. Test 1-2. a) Water phantom; b) DICOM; c) Voxellized Geant4 Geometry; d-e-f-g) 3D, 2D dose distributions (without and with the metal protection disc). Figures 3-4 are relative to test 3. In this case a female trunk 50 DICOM slices (figure 3a) were loaded to reproduce a typical clinical setup in breast IOERT. Working configuration was obtained rescaling to 256x256x50 the original CT images, thus reducing the voxel volume to 1.9x1.9x2.5 mm 3. It is evident from figures 3c, 3d, 4a and 4b, as without the protection metal disc, a considerable dose is absorbed by lung, heart and rib cage.

87 Figure 3.Test 3. a) Female trunk 50 DICOM; b) Voxellized Geant4 Geometry; c-d) 3D, 2D dose distributions. Figure 4. Test 3. a-b) 3D dose distributions, colour map: lung and bone. Conclusion: Many other tests must still be performed. Particular attention must be dedicated to implement the updated application for the grid environment, solving the heavy demand for fresh working ram, and obtaining the necessary computing power to correctly validate the application. References: [1] U. Veronesi et al., Eur J Canc 37 (2001) [2] P. Downes, et al., Phil Trans R Soc 367 (2009) [3] G. Russo et al., J App Clin Med Phys, 13 (2012) [4] C. Casarino et al., Proc 5th Int Workshop on Science Gateways (IWSG23013), ISSN , Vol-993, urn:nbn:de: [5] (last visited September 2013) [6]

88 Dosimetria EPR con alanina per fasci clinici di ioni carbonio e protoni Suitability of alanine EPR technique in clinical carbon ion and proton beams. A.Carlino1,2,4, M. Marrale1,2, A. Bolsi4, M. Durante3, M. Kramer3, T. Lomax4, A. Longo1,2, S.Panzeca1, E. Scifoni3, M. Brai1,2 (1) Department of Physics and Chemistry, University of Palermo, Italy (2) INFN Section of Catania, Italy (3) GSI Helmholtzzentrum für Schwerionenforschung GmbH (Darmstadt, Germany) (4) Paul Scherrer Institute PSI (Villigen, Switzerland) Purpose: Although particle beam radiotherapy ultimately requires dose prescription in terms of biological dose, absorbed dose is still the quantity mostly used in clinical Quality Assurance and to dosimetrically characterize the beam [1]. Among solid state detectors the alanine EPR detectors present several advantages such as: tissue equivalence, linearity of its dose-response over a wide range, high stability of radiation induced free radicals, no destructive read-out procedure. The main goal of the present work is to investigate the response behaviour of alanine EPR pellets in clinical proton and carbon ion beams delivered by active scanning and passive scattering techniques. Moreover, a model based on the Local Effect Model (LEM) together with the dedicated Treatment Planning System TPS for particles TRiP98 has been applied to calculate alanine pellets response. A comparison among alanine measurements, TPS dose and model predictions has been performed. Methods and Materials: 12C ions and proton irradiations have been carried out at GSI (Germany) and PSI (Switzerland), respectively. All measurements have been carried out with the alanine pellets 4.5 mm in diameter and 3 mm in thickness. The dose distribution has been optimized and computed by two different dedicated TPS. Results: Regarding the passive scattering modality for proton beams, Output Factor measurements have been carried out and the results are in agreement with Hi-p semiconductor diode up to 10 mm collimator diameter. Moreover, regarding the active scanning technique (raster scanning for 12C and spot-scanning for protons) the dose linearity alanine response in the clinical dose range has been successfully verified and the alanine response at selected locations in depth has been measured and compared with TPS planned dose in different quasi-clinical scenarios. A dosimeter quenching more evident for 12C ions than for protons has been measured. The LEM-like model predicts quite well the experimental response. Furthermore the signal fading for each radiation beam has been analyzed. Conclusion: This work proved that alanine is a robust and promising dosimeter in clinical particle therapy. The improvement of alanine detector response model will make alanine a good candidate for QA and commissioning of a particle beam. References: [1] R. Herrmann et al., Dose response of alanine detectors irradiated with carbon ion beams, Med. Phys., (2011), 38,

89 Apparecchiatura di tipo pencil beam scanning per protonterapia: caratteristiche e prestazioni del fascio di Trento. Pencil beam scanning system for proton therapy: characteristics and performances of Trento s facility. S. Lorentini 1, F. Fracchiolla 1,3, F. Fellin 1,2, D. Ravanelli 1, M. Schwarz 1,2, C. Algranati 1,2. (1) Agenzia Provinciale per la Protonterapia (ATreP), Trento (2) Azienda Provinciale per i Servizi Sanitari (APSS), Trento (3) Post Graduate School of Medical Physics Sapienza University of Rome. Purpose: To describe the main features and performances of the pencil beam scanning (PBS) equipment installed at the proton therapy center of Trento. Methods and materials: Our cyclotron-based facility has clinically available energies ranging from 226MeV to 100MeV (shallower ranges are achieved by using a range shifter). Three settings of the spot size (named Tune ID ) are available, two of which are obtained by inserting a thin scattering foil across the beam path. The PBS performances were characterized in terms of beam energy, shape, size, divergence, and positioning accuracy. Range measurements in water, using a large plane parallel Bragg peak chamber, and acquisitions of spots images along the transversal plane with a scintillation screen coupled to a CCD camera were performed. Further measurements were carried out to assess the uniformity within large field and velocity in energy changes. Results: Cross sectional beam shape, measured at different depths in air and positions over the transversal plane is circular-shaped with Gaussian-like profiles. The beam size (measured as 1 sigma (σ) in air at isocentre plane) ranges from 2.6 (at 226MeV) to 5.1mm (at 100 MeV) for spot tune ID 1, from 4 to 8.3mm for spot tune ID 2, while for spot tune ID 3 σ goes from 6.5 to 13.3mm. The spot symmetry is better than ±10% regardless of energy and beam size. The relative spot positioning accuracy (i.e. difference between expected and measured distance between spots centroid) is on average about 0.5mm for the smallest spot size. Absolute spatial accuracy is about 1mm on average. For spot sizes larger than 5mm, the relative accuracy is better than 10% of the spot size for all energies and gantry angles. When multiple pencil beams are combined to get a modulated field (e.g. 10x10x10 cm 3 ), uniformity is better than ±2.5% in the transversal beam direction and better than 2% in the longitudinal direction within the spread out Bragg peak area. Range accuracy is always less than 1mm over the entire energy range: this value is highly reproducible (SD=0.1mm in average) over time, regardless the beam size and energy. The time to switch between neighbouring energy levels (range step ~0.6cm in water) is 2s. Conclusions: Our PBS system is able to deliver small proton pencil beams with satisfactory accuracy and reproducibility. Further investigations will aim at testing the limits of the system in view of complex delivery schemes such as volumetric repainting.

90 Protocollo di controllo del set-up del paziente in uso presso la SOC di Radioterapia Oncologica dell Ospedale di Rovigo su acceleratore lineare Elekta Synergy Agility per trattamenti con frazionamento standard O. Nibale 1, E. Bellan 1, M. Gava 1, G. Pavanato 2, G. Virdis 2, C. Capirci 2, G. Mandoliti 2 (1): SOC Fisica Sanitaria Ospedale S. Maria Misericordia Rovigo (2): SOC Radioterapia Oncologica Ospedale S. Maria Misericordia Rovigo Introduzione: All inizio dell anno 2013 presso la Struttura Complessa di Radioterapia Oncologica dell Ospedale di Rovigo è stato installato un nuovo acceleratore lineare Elekta Synergy Agility: le immagini di verifica del setup del paziente in questo sistema possono essere acquisite sia in modalità bidimensionale con pannello EPID (Elekta IViewGT ver. 3.4) sia in modalità volumetrica attraverso un acquisizione kv cone-beam CT (Elekta XVI ver. 4.5). Tutte le componenti della nuova apparecchiatura di trattamento sono collegate attraverso la rete record&verify Elekta Mosaiq ver Scopo di questo lavoro è descrivere quali sono i passaggi coinvolti nella correzione del setup del paziente, mediante acquisizione delle immagini di verifica, per i trattamenti con frazionamento standard (almeno 20 sedute). Materiali e Metodi: Una volta deciso di candidare un paziente ad un trattamento radiante, viene effettuata la procedura di simulazione virtuale dal medico radioterapista insieme al tecnico TSRM di radioterapia: viene eseguita una scansione CT del paziente ed identificato un isocentro nel volume acquisito. La procedura di simulazione virtuale è implementata nell apparecchiatura Siemens Somatom Sensation Open con software di gestione CMS Focal ver. 4.7, mediante l impiego del sistema a laser mobili Diacor Centralite. Il paziente giunge, quindi, alla prima seduta di terapia con un isocentro di trattamento identificato da tre punti in china sulla cute o su maschera di immobilizzazione (uno in direzione antero-posteriore e due laterali a destra e a sinistra). Il piano di cura viene quindi costruito dal fisico medico sull isocentro identificato in simulazione virtuale. Alla prima seduta di terapia, presso la consolle del linac, viene eseguita una scansione cone-beam CT centrata sull isocentro definitivo del trattamento. Per l acquisizione mediante sistema XVI sono disponibili vari protocolli (o preset) impostati sulla workstation dedicata, al variare della zona anatomica interessata al trattamento. Il volume acquisito mediante scansione cone-beam CT viene poi co-registrato on-line con il volume acquisito in fase di centramento con la CT Siemens Somatom Sensation Open e utilizzato poi per il calcolo di dose del piano di cura. La fase di registrazione permette di selezionare differenti protocolli di fusione di immagini identificabili in due categorie: algoritmi basati su reperi anatomici e algoritmi basati sui livelli di grigio (fusione per correlazione). Nella valutazione della registrazione di immagini, viene visualizzato con colore viola il volume di pianificazione del TPS, mentre viene visualizzato con colore verde il volume acquisito mediante sistema XVI. Inoltre, attraverso lo strumento clipboard, il medico ha la possibilità di decidere sulle tre ricostruzioni assiale, coronale e sagittale un volume di riferimento, più piccolo del totale scansionato, su cui valutare il risultato del matching. Attraverso il matching dei due volumi, si ottiene uno spostamento (offset) da applicare ai tre assi laterale, longitudinale e verticale attraverso il movimento automatico del lettino portapaziente: è possibile effettuare lo spostamento automatico in remoto dalla consolle del linac solo se la correzione rientra in un offset massimo di 2.5 cm dall isocentro. La verifica dell isocentro di trattamento alla prima seduta, mediante acquisizione cone-beam CT, rappresenta un modo per eliminare errori grossolani di posizionamento del paziente.

91 Alla seconda, terza e quarta seduta viene applicato un protocollo di correzione off-line di tipo NAL (No-Action- Level). Posizionando il paziente all isocentro di trattamento identificato in simulazione virtuale (eventualmente corretto per errori grossolani alla prima seduta), vengono acquisite le immagini di due campi aperti di dimensioni 20x20 cm con dispositivo EPID IviewGT per angoli del gantry a 0 e 90. Per ognuno dei tre giorni di verifica, le immagini portali vengono inviate alla rete R&V Mosaiq dove viene effettuata una registrazione di immagini stereoscopica con le immagini di riferimento DRR. Per registrazione stereoscopica si intende una registrazione di immagini con la possibilità di osservare e sovrapporre due viste diverse con le rispettive DRR (proiezione a 0 e proiezione a 90 ) contemporaneamente: in questo modo si riesce a gestire simultaneamente lo spostamento lungo le tre coordinate spaziali. Alla fine delle tre acquisizioni portali, nell applicativo Mosaiq viene calcolato uno spostamento medio lungo i tre assi laterale, longitudinale e verticale: se tale offset calcolato supera una soglia prestabilita (nel nostro caso pari a 0.2 mm per ogni asse), viene inviata automaticamente la correzione della posizione dell isocentro di trattamento al site-setup del paziente (modulo di Mosaiq che contiene informazioni sul posizionamento del paziente). In questo modo, dalla quinta seduta di trattamento, il tecnico alla macchina può applicare in automatico la correzione al posizionamento del paziente attraverso il modulo RATM (remote automatic table movement), centrando inizialmente il paziente sui punti in china derivanti dalla simulazione virtuale. L offset medio calcolato su tre sedute successive, applicato per correggere la posizione dell isocentro per il resto del trattamento, minimizza la componente di errore sistematico nel posizionamento del paziente su tutte le sedute di terapia dopo la quarta. Per il resto del trattamento (dalla quinta seduta), viene effettuato un controllo settimanale off-line di posizionamento del paziente attraverso l acquisizione delle due immagini portali 20x20 cm alle proiezioni di gantry 0 e 90. Tale controllo serve a confermare o correggere l offset applicato dalla quinta seduta in poi. Dall applicativo Mosaiq, attraverso il modulo D&I (Diagnoses and Interventions) e relativamente al trattamento in corso, è possibile osservare lo storico degli spostamenti calcolati attraverso il comando localization trend review : in questo modo è possibile effettuare un analisi statistica degli offset calcolati e applicati. Risultati: Per studiare l accordo tra lo spostamento fornito dalla registrazione di immagini volumetriche nella consolle XVI e lo spostamento fornito dalla media di valutazione su immagini portali in tre sedute successive, si sono suddivisi i pazienti per area anatomica irradiata creando tre categorie: pazienti del distretto testa-collo con maschera termoplastica di immobilizzazione, pazienti del distretto torace-addome superiore soggetti a notevole spostamento dovuto al movimento respiratorio, pazienti del distretto pelvico. L analisi è stata effettuata su un campione di 49 pazienti così suddivisi: 11 pazienti del distretto testa-collo, 24 pazienti del distretto torace-addome superiore, 14 pazienti del distretto pelvico. La correzione maggiore di spostamento si ha per i pazienti del distretto toracico e, complessivamente, in valore assoluto l offset per singolo asse è sempre contenuto entro 8 mm. La correzione minore di spostamento si ha per i pazienti del distretto testa-collo e, complessivamente, in valore assoluto l offset è sempre contenuto entro 3 mm. In media, su tutte le categorie di distretto anatomico, l offset applicato dalla quinta seduta ricade nel range 3-5 mm per singolo asse spaziale, mentre il vettore di spostamento (calcolato su tutti e tre gli assi) in media è intorno a 5 mm. La direzione in cui si ha il maggior numero di spostamenti è quella longitudinale. Sicuramente la procedura di simulazione virtuale applicata durante la fase di centramento aiuta già di per sé a minimizzare gli spostamenti durante le sedute di terapia, utilizzando la stessa tipologia di lettino porta-paziente e gli stessi accessori per agganciare eventuali sistemi di immobilizzazione. Le due tecniche di registrazione di immagini (una basata su CT volumetrica e l altra basata su immagini portali bidimensionali) forniscono spostamenti in accordo tra loro lungo i tre assi (stessa direzione). Come ci si poteva aspettare, lo spostamento, valutato nel suo valore assoluto, è maggiore nella prima seduta con applicazione XVI (componente sistematica più componente casuale di errore di posizionamento) rispetto alla media calcolata con

92 registrazione stereoscopica su immagini portali bidimensionali in Mosaiq dove il risultato è modulato su tre sedute e ha l obiettivo di minimizzare la sola componente sistematica di errore. Conclusioni: La verifica del setup del paziente implementata nel nostro centro di Radioterapia si può dividere in due parti: verifica on-line alla prima seduta con acquisizione cone-beam CT e la presenza del medico radioterapista alla consolle della macchina di trattamento, verifica off-line nelle tre sedute successive per il calcolo della componente d errore sistematico nel posizionamento del paziente. Il controllo on-line richiede al medico radioterapista un impegno minimo (solo alla prima seduta e quando lo ritenga opportuno), mentre, per il controllo off-line, l organizzazione del tempo da dedicare alla registrazione stereoscopica delle immagini è personalizzabile a seconda delle esigenze dei singoli. La media degli spostamenti calcolata in automatico in Mosaiq aggiorna in tempo reale l offset totale da applicare al paziente nel modulo site setup della rete R&V ed è impostabile in automatico direttamente da consolle della macchina se contenuto entro la tolleranza di 2.5 cm. In questo modo si possono evitare eventuali errori di calcolo di correzione manuale da parte dell operatore. La verifica portale settimanale conferma nella quasi totalità dei casi l offset medio calcolato alla quarta seduta. References: [1] The Royal College of Radiologists Institute of Physics and Engineering in Medicine Society and College of Radiographers, On target: ensuring geometric accuracy in radiotherapy, pdf pubblication (2008).

93 Impact of residual error on lymphnodes location during image-guided Tomotherapy for N + head-and-neck cancer (HNC) patients Impatto dell errore residuo sulla locazione di linfonodi positivi in trattamenti image-guided con Tomoterapia per tumori N + in pazienti testa collo (HNC). Belli Maria Luisa 1,2, Dell Oca Italo 3, Raso Roberta 1, Zerbetto Flavia 3,4, Chiara Anna 3, Cattaneo Giovanni Mauro 1, Gianolli Luigi 5, Calandrino Riccardo 1, Di Muzio Nadia 3, Fiorino Claudio 1 1 Medical Physics Department, San Raffaele Scientific Institute, Milano, Italy 2 School of Specialization of Medical Physics, Università degli Studi di Milano, Italy 3 Radiotherapy Department, San Raffaele Scientific Institute, Milano, Italy 4 School of Specialization of Radiotherapy, Università Milano-Bicocca, italy 5 Nuclear Medicine Department, San Raffaele Scientific Institute, Milano, Italy Purpose Positive lymphnodes (PLs) of patients treated with image-guided RT (IGRT) for HNC are subject to changes in dimension, shape and position with a potential risk of geographical miss, especially in the case of boost on PET PLs. Aim of this study was to assess if the margin used in our simultaneous-integrated-boost (SIB) approach with Helical Tomotherapy (HT) is appropriate. Material and methods 37 consecutive HNC N2/N3 patients (pts) were considered. All pts were treated with HT with SIB delivering 54,66 and 69 Gy in 30fr on PTV(N),PTV(T+N + ) and PTV of the PET-positive T+N +. Regarding PLs position: 33, 7 and 2 PLs were respectively in levels II,III and V. For each pt, weekly(7) MVCTs were matched with the planning kvct (pl_kvct) on bone anatomy. 7pts were excluded because PLs were not visible on MVCT; 42PLs of 30pts were contoured by 3 experts on the pl_kvct and all MVCTs. Intraobserver variability was assessed by blind re-delineation of 16PLs. Time-trends of volumes normalized to fr1 MVCT, was performed (Spearman s test). The shift of PLs was assessed by the center of mass (CM) shift with respect to fr1 MVCT. For each PL, the % fraction of the union (UN) of all PL positions over the whole treatment that was missed by the clinical PTV (pl_kvct: PL contour+5mm) was assessed. For pts with some missing, larger margins were tested to find the margin covering>99% of UN. The same procedure was applied for the 1 st and the 2 nd half of the treatment. Results PLs were sufficiently well visible on MVCT(median DICE 0.805±0.134). 27/42 PLs showed a significant volume shrinkage (average reduction:70%, range:27-94%, median Spearman r=-0.93;range ;p<0.05). 3D average CM shifts on the whole population were larger in the 2 nd part of the treatment (medial shift in the 1 st and 2 nd part:0.1 vs 1.6 mm,p<0.001). 99%covering of UN was obtained with 5mm margin for 77%pts,6mm for 83%,7mm for 90%,8mm for 93%,9mm for 97%. Conclusions Results show a significant, although small, residual error in the localization of PLs during IGRT for HNC. 64%PLs show a significant shrinkage. A margin of 5mm covers all possible positions of PLs in 77% of pts, extending to 93-97% with a margin of 8-9 mm. Interestingly, the shrinkage seems to counterbalance the shift of PLs due to deformations, more pronounced in the last fractions. Seeing these results, adaptive re-planning aiming to avoid to miss PLs should not be recommended (excepting very selected pts) also in the case of PET PLs boost.

94 Stereobody for lung cancer: X-ray Voxel Monte Carlo vs Pencil Beam based dose calculation G. R. Borzì 1, V. Landoni 2, S. Strolin 2,V. Bruzzaniti 2, A. Soriani 2, D. D Alessio 2, L. Strigari 2 (1) REM Radioterapia, Catania Istituto Oncologico del Mediterraneo (IOM), Viagrande (CT) (2) Laboratorio di Fisica Medica e Sistemi Esperti, Istituto Nazionale Tumori Regina Elena, Roma INTRODUCTION The widespread of technologies allowing the safe delivery of high doses per fraction and the fact that Sterotactic Hypofractionated Radiotherapy for lung cancer has revealed to be a promising technique in terms of clinical results have again posed attention to the importance of the accuracy of heterogeneity correction algorithms for the calculation of dose. It is well known that calculation performed with the Pencil Beam (PB) algorithm is reasonably accurate for tumours located in homogeneous regions but doses tend to be overestimated where large inhomogeneities exist. Recent codes based on Monte Carlo (MC) algorithm such as X-Ray Voxel Monte Carlo (XVMC) have allowed the use of fast and accurate algorithms for dose calculation in the clinical practice. The purpose of this study is to evaluate the differences between dose distributions calculated with PB and XVMC algorithm implemented in a commercial treatment planning system on a cohort of patients treated for lung cancer with stereotactic body radiotherapy (SBRT) with IMRT and HybridArc (HA) techniques in terms of dose volume histograms (DVHs) on tumour and organs at risk (OARs), in normal and deep breathing condition and in terms of tumour control probability (TCP) and normal tissue complication probability (NTCP). Also the dependence on tumour volumes and locations of the observed differences in calculations has been investigated. Furthermore dosimetric validation using gafchromic EBT3 has been performed in the Rando anthropomorphic phantom. MATERIALS AND METHODS CT data of 45 lung cancer patients were used for this study. Four out of 45 patients had also been scanned under deep-breath condition to plan a gated radiotherapy. Patients were immobilized using a custom-made cast; CT scans were acquired from the larynx to the level of the abdomen, with a thickness of 2 mm. In table 1 patients characteristics are reported. PTV (cc) Spinal cord (cc) Lung (cc) Heart (cc) Mean Median Std dev Range Table 1. Volumes of PTV and OARs for the 45 patients investigated. For these CT datasets an IMRT or HA plan was developed to give 40 Gy at 8 Gy/fraction with 5 no coplanar 6 MV IMRT beams or 3-4 dynamic conformal arcs with 3-5 IMRT beams distributed per arc (for HA plans). The spatial resolution that defines the size of the dose calculation grid was set to 2 mm. The mean variance that gives the statistical uncertainty of the MC dose calculation was set to 2%. Plans were optimized to give at least 95% and not more than 107% of the prescribed dose to the tumour normalizing to the tumour mean dose. The dose volume constraints on lung, considering both lungs as one organ, heart and spinal cord were set on the basis of QUANTEC, re-evaluated for the different fractionation according to the isoeffect equation and reported in table 2. LUNG V 10Gy (%) < 30-35; MLD (Gy) < HEART V 11Gy (%) < 10 SPINAL CORD D max (Gy) < 16 Table 2. Dose volume constraints on lung, heart and spinal cord.

95 Treatment plans were calculated with PB algorithm (plan PB). Then each optimized plan was recalculated with the XVMC algorithm maintaining the same monitor units (MUs) (plan MC). Finally XVMC plans were renormalized to give the prescribed dose to the tumour (plan MC-normalized). To estimate the impact of the observed differences on treatment outcome, TCP and NTCP were calculated. Differences between MU to be delivered according to the PB plan and the renormalized MC plan were evaluated by means of the t-test for independent samples; differences were considered statistically significant if p value was less than Finally CT data of the Rando phantom (thorax region) were acquired and a typical IMRT treatment plan was performed on contours delineated by the physician; doses were calculated with PB and MC algorithms. The two plans were irradiated with gafchromic EBT3 positioned in significant planes in the phantom; measured and calculated doses were compared to by means of gamma index. RESULTS Figure 1 shows the comparison of DVHs for the three plans (PB, MC and MC-normalized) calculated for patients with PTV volume of cc (a), 9.39 cc (b) and in deep-breathing condition (c) (a ) (b ) (c ) Volume (%) PB MC MC-normalized Volume (%) PB MC MC-normalized Volume (%) PB MC MC-normalized Dose (Gy) Dose (Gy) Figure 1. Comparison of DVHs for PB (lightly shaded line), MC (dark solid line) and MC-normalized (dashed line) plans for patients with tumour volume of cc (a), 9.39 cc (b) and in deep-breathing condition (c). Figure 2(a) shows the data comparison graph relative to the mean tumour dose for the 45 plans calculated with PB and MC algorithms and figure 2(b) for patients in deep-breathing condition Dose (Gy) (a) (b) Figure 2. Mean tumour doses for the 45 treatment plans calculated with PB and MC algorithms (a) and for the 4 patients in deep-breathing condition (b). Table 3 shows the statistical results of PB vs MC and PB vs MC-normalized analysis.

96 PTV PB MC MC-normalized PB vs MC PB vs MC-normalized Dmax (Gy) 41.3 ± ± ± 1.2 p < p < Dmean (Gy) 39.9 ± ± ± 0.5 p < p = D98 (Gy) 38.5 ± ± ± 1.1 p < p < D95 (Gy) 39.1 ± ± ± 1.0 p < p < Table 3. Summary of the study parameters relative to PTV among the 3 different plans. Differences between mean tumour dose calculated with PB and XVMC were about (10 ± 4)%, even larger in deepbreathing condition, while differences between doses to significant volumes for OARs were generally lower. Differences between mean tumour dose calculated with PB and XVMC correlated with tumour volume were found statistically significant (r = 0.37, p = 0.01) and are shown in figure 3. Figure 3. Correlation between mean PTV dose calculated with PB and XVMC and tumour volume. We also analyzed these dose differences in function of the X-ray path of the beams using the tissue ratio: Figure 4 shows the correlation between deviations of mean PTV dose with PB and XVMC and X-ray path of the beams (tissue ratio). Figure 4. Correlation between deviations of mean PTV dose calculated with PB and XVMC and tissue ratio.

97 After normalization, MUs for XVCM were always higher than for PB though not significantly (p = ). TCP ranged from (99.9 ± 0.1)% to (95.3 ± 4.9)% for PB calculated plans respect to XVMC while NTCP on OARs did not vary significantly. In the deep-breathing condition TCP ranged from (99.9 ± 0.3)% to (66.9 ± 17.2)% for PB and XVMC, respectively. In figure 5 the irradiated EBT3 for a typical SBRT plan is shown (a) together with the results of the comparison between calculated and measured dose in terms of gamma index (b). Figure 5. Irradiated EBT3 (a) and gamma index analysis on the calculated and measured dose (b). (a) (b) The dosimetric evaluation confirmed the better accuracy in calculation of XVMC. The agreement in terms of absolute gamma function (γ < 1, 3%, 3 mm) was about 94% for XVMC and lowered down to 67% for PB. DISCUSSION In this study, lung cancer patient planning calculations were performed to demonstrate the significant improvement in dose calculation accuracy using XVMC algorithm in comparison with PB for SBRT treatments. The difference between mean tumour dose computed with the PB and MC algorithms was statistically significant (table 3), while for OARs it was not significant. In figure 1 the reduction in dose and the considerable dose heterogeneity within the PTV is noted in the MC DVHs. It should be noted that for the plan with PTV volume of cc, the agreement between MC and PB calculations was found to be better than that where the PTV volume was 9.39 cc. This inverse correlation is illustrated in figure 3 and suggests that field size strongly influences the impact of electron scattering on the dose to the target. This is not a surprising result given the physical processes leading to electronic disequilibrium. In figure 2 we note that mean doses calculated with the MC algorithm are much more spread out than those calculated with PB. In fact, since MC calculation diversifies radiological paths of the beams within different tissues, it takes into account the volume of the PTV, the ratio between PTV and lung volume, the position of PTV within the lung. The correlation showed in figure 4 is direct, ie the increase of the X-ray path increases the difference of the mean doses calculated with the two algorithms. The reduction of TCP for plans calculated with MC algorithm respect to PB is probably due to the fact that PB does not take account of the lateral scattering of secondary electrons set in motion by primary radiation. Finally, the better agreement obtained for the MC plan created for the dosimetric verification in the Rando phantom compared to the PB plan (figure 5) is a further demonstration of the fact that MC dose calculation is more accurate and close to reality. CONCLUSIONS Results showed that PB calculation leads to overestimate the dose with respect to XVMC especially for the points inside the tumour. In fact, for each case the major discrepancies were observed along the boundary between tissue and air. The increase in MU due to the renormalization of the plans to have comparable mean doses to the tumour was not significant and the calculated NTCP values for OARs were far below the allowed tolerable values. MC algorithms are becoming often available in commercial TPS and their employment should be encouraged to accurately calculate dose in presence of high inhomogeneities especially when delivering high doses per fraction. However the introduction of MC calculation in the process of IMRT optimization is still to come and the PB based optimized plan can finally result suboptimal. For this reason efforts should be done to find methods to control the optimization process according to the dose calculated with MC so to obtain a result that is as close as possible to the best solution.

98 Per-patient quality assurance for head and neck IMRT plans by the novel system COMPASS Controlli di qualità per-paziente di piani IMRT per il trattamento testa-collo utilizzando il nuovo sistema COMPASS R.Caivano 1 M. Cozzolino 1, G. Califano 1, C. Oliviero 1, P. Pedicini 1, S. Clemente 1, A. Fiorentino 1, C. Chiumento 1 and V. Fusco 1. 1 Department of Radiation Oncology, IRCCS CROB 1 Padre Pio Street, Rionero in Vulture, PZ, Italy Purpose: IMRT requires a pretreatment verification procedure aiming to assess the correct delivery of the treatment plan of each patient by comparing the dose distribution calculated by TPS and an independent dosimeter. The purpose of the present work was to explore and to evaluate a novel patient-dose DVH-based method for pretreatment dose QA: the COMPASS system (IBA Dosimetry, Germany). Methods and materials: The COMPASS system consists of a dedicated control and analysis software interfacing a detector mounted at the head of the linac: the 2D array ionization chamber MatriXX (IBA Dosimetry, Germany). The COMPASS software controls and acquires the detector measurements but it also includes a beam model of the linac. The DICOM file for patient s CT and plan structures are fed into the software that computes the dose distribution in the patient geometry (computed dose). With the beam model optimized on the actual delivery, the software reconstructs the actually delivered dose distribution in the patient geometry (reconstructed dose). We used MatriXX mounted at the head of a Trilogy (Varian Medical Systems) and a set of 20- IMRT plans for head and neck cancer patients with 7 beams calculated by TPS Eclipse (Varian Medical Systems). A comparison was performed between the planned dose distributions, the computed and the reconstructed ones using the gamma index (GI) method. The GI analysis was performed using both the 3%/3mm and the 2%/2mm acceptance criteria on whole grid patient and on single structure. A DVH-based analysis was accomplished for target and for most relevant organs-at-risk (OAR). Results: No significant DVH-deviation was observed for target and OAR. Considering the 3%/3mm criteria the mean GI%<1 for the body was 97.48±0.21. The mean GI%<1 for the considered structures were: 91.81±3.68, 93.03±3.84, 94.04±2.42, 98.15±3.69, 98.24±3.26 and 96.65±4.35 for PTV1, PTV2, PTV3, spinal cord, enlarged spinal cord and parotids respectively. While considering the 2%/2mm criteria for the analysis, GI values reduced significantly. Reconstructed doses have been further compared with the computed ones showing smaller differences with the planned dose distributions. Conclusions: These results show a good agreement between the COMPASS dose reconstructions and the reference measurements. Further analysis are in progress to clinically validate the COMPASS System as QA-method based on DVH metrics, especially to define new benchmarks for clinical plans verification.

99 Interconfronto multicentrico sulla caratterizzazione dosimetrica di un collimatore microlamellare per applicazioni in radioterapia stereotassica. A multicentre intercomparison of dosimetric characterization of a dynamic multileaf collimator for stereotactic radiotherapy applications. M. Casale 1, M. Italiani 1, V. Magaddino 2, F. Vittorini 4, M. Muti 1,E.Buono 1, N. Franza 3 (1) A.O. Santa Maria, Terni (2) CHUV, Lausanne (3) Dosimetrica, Nocera Inferiore (4) Ospedale San Salvatore, L Aquila Purpose: to compare experimental results from different sites in small fields dosimetry for linac based stereotactic radiotherapy by means of an Apex Dynamic Multileaf Collimator (DMLC). Methods and materials: the dynamic mini-mlc Apex from Elekta (Stockholm,Sweden) is add-on MLC consisting of 112 leaves (56 pairs) with a physical leaf width of 1.5 mm giving a nominal width of 2.5 mm at the isocentre. The minimum nominal field at the isocentre is 0.5 cm 0.5 cm. A dosimetric intercomparison has been carried out for 3 radiotherapy centers equipped with Apex DMLC. The dosimetric features investigated in this study are percentage depth dose (PDD) and total scatter factors (Scp) acquired with the add-on Apex DMLC on Elekta linacs. All measurements were performed with beam energy of 6MVand similar linear accelerators.different kinds of detectors were used to measure the radiation characteristics: Scanditronix-Wellh ofer CC13, CC01, IC 04 and stereotactic diode (IBA Dosimetry, Schwarzenbruck, Germany), Gafchromic EBT2 film (ISP, NJ, USA), PTW PinPoint and diode detectors. Results: results from different sites with several dosimetric equipment will be shown. For all fields sizes greater than 2.0 cm the difference between the measured Scps of the 3 centers were within 0.5%. Same results were observed for PDD curves. For fields sizes less than 2.0 cm the differences in Scp values were within 3%. Analysis of PDD for such a small field sizes has been more cumbersome due probably to the critical alignment of the detectors inside the water phantom. Conclusion: in the dosimetry of small beams is important the choice of the right detectors and the experimental set up due to the dimension of the detectors compared to the small field. Data comparison between different centers has been considered of a great importance in order to minimize possible systematic errors and improve algorithm parameters optimization.

100 Definizione di nuovi criteri di accettabilità per verifiche pre-trattamento di piani IMRT con rivelatori 3D: studio preliminare Pretreatment dosimetric verification of IMRT plans with 3D diode matrices: preliminary study for the establishment of new tolerance levels. M. Casati 1, C. Arilli 1, M.Bucciolini 1, A. Compagnucci 1, L.Marrazzo 1, S.Pallotta 1, C.Talamonti 1, E.Vanzi 1 (1) A.O.U. Careggi, Firenze Purpose: Due to the introduction of the rotational radiotherapy tecniques, the geometry of the radiation detectors has been evolved from 2D arrays to 3D arrays. The aim of this work is to compare 2D and 3D diode arrays for pretreatment verification of step&shoot IMRT plans to determine if the acceptance criteria used for the evaluation of the agreement between dose distributions measured with 2D diode matrixes and calculated by TPS are also valid for 3D diode arrays. Materials and methods: Two commercial dosimetry systems have been compared, both based on diodes. The planar array detector was irradiated in the plane containing the isocenter at 5cm water equivalent depth, with the gantry at 0. The three dimensional detector contains diodes spiralling around a cylinder surface and embedded in a PMMA phantom with an internal hole, which can be filled with an insert for ionization chamber. In this way, by measuring without the insert, the capability of the TPS of calculating in low density regions is tested, while the insertion of the chamber allows absolute dose verification at the isocenter. The cylindrical geometry allows for verification of both entrance and exit dose. The plan can be verified in clinical conditions, with variable gantry angles, thus testing also gravity effects and couch attenuation. Ten step&shoot IMRT plans, already verified with the fixed gantry tecnique and the planar diode array, have been tested again with the 3D diode array and a ionization chamber inside. The agreement between measured and calculated dose distributions was evaluated in terms of gamma index passing rate, with 3% and 3mm tolerance levels. Results: The three dimensional diode array, which is the only choice for rotational plans verification, has been proved to be suitable also for the verification of IMRT plans, with an improved capability of revealing criticalities in the three dimensional distribution of dose. Conclusion: The three dimensional diode array tested in this study against a planar diode array, can be conveniently used for non-fixed gantry pre-treatment dose verifications of IMRT plans, provided that appropriate acceptance levels are defined for the evaluation of the agreement between measured and calculated dose distributions. References: [1] AAPM TG-119 report, Med. Phys. 36, 11, (2009) [2] IMRT Quality assurance Evolution from 2D to3d Med. Phys. 38, 3528 (2011)

101 Validazione del software commerciale EPIDose per verifiche dosimetriche pretrattamento di piani IMRT step&shoot con LINAC Synergy BM e EPID iviewgt. Pre-treatment verification of IMRT fields produced by an Elekta Synergy BM LINAC with an iview GT EPID and a commercial dosimetric software: validation of the SunNuclear EPIDose system. M. Casati 1, C. Talamonti 1,2, E. Vanzi 1, M. Bucciolini 1,2. (1) Azienda Ospedaliera Universitaria Careggi, Firenze. (2) Università degli Studi di Firenze, Firenze. Purpose: Electronic Portal Imaging Devices (EPID) potentially offer an attractive alternative to other detectors such as chamber or diode arrays for IMRT pre-treatment verifications, because of their high resolution and because there is no detector set-up time, so the measurement is less LINAC-time-consuming. The purpose of this work is to commission and test a commercial software for the conversion of EPID images to dose distributions at the isocenter at 5cm water equivalent depth. Methods and materials: The SunNuclear EPIDose system was commissioned for an Elekta Synergy Beam Modulator LINAC and an iview GT EPID. The first step was the acquisition of measurements with different field sizes and varying MUs, both with the MapCHECK diode array and with the EPID, without any phantom. The LINAC output was tuned to the reference value used by our TPS. Basing the EPIDose physics modelling on these measurements, the images acquired in future sessions with the EPID will be converted to absolute dose. In every patient plan pre-treatment verification session with EPIDose, a reference image has to be acquired. The images are acquired in IMRT mode, Single exposure, Frame averaging maximum. A DICOM image is saved for each IMRT segment. The EPIDose software corrects for the pixel frame factor and the automatic normalization. This work aims to test the overall dosimetric accuracy of the dose distributions resulting from the EPIDose system, which depends both on the intrinsic limits of the EPID and on the goodness of the commissioning and of the EPIDose physics modelling. Time lag effects and PRF dependence are not explicitly addressed by EPIDose model, and the possibility to neglect these effects has been tested in terms of overall contribution. In this study the model was optimized and tested on ten clinical patient plans. Results: The gamma passing rates in the comparison of the measured dose distributions with the calculated one are similar to those obtained with MapCHECK. Conclusion: EPIDose has proven to be a possible alternative to MapCHECK for pre-treatment verification of IMRT plans, provided that the physics modelling has been commissioned for the specific LINAC and EPID. The LINAC time necessary for data acquisition is reduced, but the data transmission and elaboration process has to be improved to become competitive respect to MapCHECK. Further optimization of the EPIDose procedure and investigation of PRF dependence and time lag effects could be appropriate.

102 Tecniche IGRT per il posizionamento dei pazienti testa e collo: ExacTrac vs CBCT Is ExacTrac X-ray system an alternative to CBCT for positioning patients with head and neck cancers? S. Clemente 1,V. Simeon 2, C. Oliviero 1, M.Cozzolino 1, G. Califano 1, R. Caivano 1, C. Chiumento 1, A. Fiorentino 1, V. Fusco 1. (1) Department of Radiation Oncology, IRCCS CROB, Rionero in Vulture (PZ) (2) Laboratory of Preclinical and Translational Research, IRCCS CROB, Rionero in Vulture (PZ) Proper daily patient alignment is one fundamental pre-requisite for patients with head and neck (HN) cancer undergoing intensity modulated radiotherapy (IMRT) due to the high conformality of the dose distribution. Set-up uncertainties and anatomic variations represent critical points for HN cancer because of the complexity of the HN anatomy, proximity of cancer to several normal structures, different relative motion among HN structures (i.e. mandible, upper neck region, lower neck region). Several studies have shown a different displacement among different bony structures in the HN region [1]. Patient position accuracy has been assessed with megavolt (MV) X-rays, a two-dimensional (2D) radiographs projection technique after traditional immobilization with standard thermoplastic face masks, bite blocks or vacuum bags. Determination of set up errors have been performed off-line using anatomic bony landmarks due to poor visualization of soft tissues in the planar projection X-ray images. Therefore, patient s position was adjusted at a subsequent treatment fraction. However, while off-line correction ameliorates the systematic component of set-up errors, it is less effective than on-line correction to minimize both systematic and random setup errors [2]. Nowadays several systems are available in clinical practice for daily Image Guided RT (IGRT). Daily pretreatment acquisition of a cone-beam computed tomography (CBCT) represents a widely adopted option for set up verification of patients undergoing IMRT for HN cancers. CBCT offers a three-dimensional (3D) view with a better visualization of anatomical structures and soft tissues than 2D imaging options. However, its application is limited by a relatively long image acquisition time and a relatively high radiation dose to the patient. A potential alternative to CBCT is offered by the ExacTrac (ET) Robotics IGRT system. The 6D ET is composed of an Infrared (IR) tracking system, an X-ray system (consisting of two diagnostic kv X-ray tubes and asi detectors) and a robotic couch capable of 6D correction positioning including pitch, roll and yaw. It offers potential clinical benefits over CBCT including faster patient positioning, an alignment using 6D degree of freedom, the ability to monitoring patient motion during the treatment and a reduction in image-based radiation delivered to the patient. However, to our knowledge, ExacTrac has not been directly compared to CBCT for set up of patients with head and neck cancers. The purpose of the present paper is to clarify this issue. Twelve patients with HN cancer treated by IMRT at IRCCS CROB from January 2012 to July 2012 were selected for the present study. A 3 dose level IMRT (66, 60, 54 Gy) by means of a simultaneously integrated boost (SIB) approach was used for all patients. In order to verify the accuracy of patient positioning, our Trilogy Linac is equipped, with both a Varian KV on-board-imager and a BrainLAB 6D ExacTrac system. All patients were positioned on the treatment table using a thermoplastic face mask extended to the shoulders, in combination with bite blocks and foot brace devices. Patient position was imaged daily upon two different protocols: CBCT and ET protocol as shown in figure 1. A total of 149 fractions in 6 patients and 130 fractions in 6 patients were analyzed for the CBCT and ET protocols, respectively. The present study aims at comparing the performance of both methods, in terms of 6D residual setup errors. All registrations were performed by one physician (C.C.) to eliminate inter-observer variability of the process. Images were aligned automatically using a bony anatomy-based registration, accomplished by aligning cervical vertebrae at level of the planning isocenter (C3 vertebrae) as shown in figure 2. Data imaged with either approach in two non-overlapping periods were compared in terms of both residual errors after correction and punctual displacement of selected regions of interest. On average, both protocols showed small residual errors and their magnitude were comparable. Compared to ET protocol, CBCT correction generally showed reduced residual errors in translations, but not in rotations; a significant shift in the Pitch direction was

103 observed for the CBCT protocol (p<0.001), on the overall treatment duration and specifically in the first half of the treatment. However, no significant positional changes occured during the first and the second half of the treatment period, therefore both set-up tecniques showed to have an high interfractional reproducibility. The magnitudes of displacements (residual errors) from each techniques are not expected to produce relavant dosimetric differences in surrounding critical structures and target coverage. The observed differences in shift vectors between the two protocols (describing the deviation of the treatment position from the reference position at simulation time), showed that the anisotropic relative movement (or the flexibility) between different ROIs of HN bony structures introduces extra setup errors that seems to be corrected by either strategy in a different way. CBCT tends to weight more C2 and C6 at the expense of the mandible, while ET tends to average more differences among the different ROIs. Zhang et al. [3] using CBCT, showed local setup variations of HN using three bony regions, C2, C6 and the palatine process of the maxilla (PPM) in 14 patients. Similarly to our findings, C2 showed the smallest setup variations indicating that C2 is the most stable region in AP and CC directions (an anatomic pivot for head motion). Of note, our experience shows that the vector displacement of C2 is significantly lower with CBCT over ET. Conversely, PPM was found to have the largest CC motion, which in turn impacted Pitch rotation. Our data are consistent with these findings. With CBCT, the (whole) mandible showed the largest vector displacements from the simulation position and, consistently, the residual error for Pitch rotation was extremely variable. With ET, the average displacement of the mandible was smaller and the residual Pitch rotation was less variable than with CBCT. We hypothesize that ET, compared to CBCT, provides a more averaged interpretation of shifts, that may derive from both a more comprehensive view of the ROI and the possibility to correct also for rotations over CBCT. We believe that the more consistent C2 alignment on CBCT comes from being reconstructed as a 3D structure as opposed to ET that uses, within each projected radiograph, a region with the best visible bony landmark, which may not be consistently the same throughout the two radiographs and may be different from the ROI reconstructed by CBCT. CBCT system allowed only corrections in translations and thus we were unable to perform them in rotations. On the other hand, the ET system registers the acquired 2D planar X-ray images with DRR and achieves best match by tuning the image registration with both translations and rotations, but acquired images may not always be optimal for image registration due to substantial overlapping bony structures. Again CBCT and ET alignment use different ROIs, 3D versus 6D couch movement correction and different images quality. On the other hand, the volumetric reconstruction (3D) of structures along with a better visualization of anatomical structures and soft tissues by CBCT outweighed the disadvantage of not being able to perform 6D corrections. Overall, CBCT seems more reliable than ET on both C2 and C6, but somewhat less performing on the mandible. Therefore, CBCT may be better indicated when, within a comprehensive treatment of the whole neck, the high dose target volume is in close proximity to the vertebral bodies (pharyngeal and laryngeal cancers) as opposed to the oral cavity. Obviously, these considerations do not take into account the fact that CBCT is able to see soft tissues and their deformation, providing additional advantages in situations where this is clinically relevant. In conclusion, while both protocols achieved reasonably low residual errors after initial correction, CBCT, even without 6D correction capabilities, seems preferable to ET for better C2 and C6 alignment and the capability to see soft tissues. Therefore, in our experience, CBCT represents a benchmark for positioning head and neck cancer patients. Further clinical investigations are needed to validate the adequacy of these findings and to determine whether ET is a legitimate alternative or a complement to CBCT. References: [1] S. van Kranen, Setup uncertainties of anatomical sub-regions in head-and-neck cancer patients after offline CBCT guidance, Int. J. Radiat. Oncol. Biol. Phys. (2009) 73(5), [2] D. Yan, The use of adaptive radiation therapy to reduce setup error: A prospective clinical study, Int. J. Radiat. Oncol. Biol. Phys. (1998) 41(3),

104 [3] L. Zhang, Multiple regions-of-interest analysis of setup uncertainties for head-and-neck cancer radiotherapy. Int. J. Radiat. Oncol. Biol. Phys.(2006) 64(5), Figure 1: Correction set-up protocols adopted in the study. Figure 2: Reconstruction using BEV (beam's eye view), of ROIs used for the automatic registration: a) CBCT ROI; b) ET ROI.

105 Evaluation and comparison of different commercial systems for IMRT and VMAT Quality Assurance, by using EPID (Electronic Portal Imaging Device). S. Cora 1, G. Murtaza 2,P. Francescon 1, P. Scalchi 1, V. Santangelo 1, N. Satariano 1, M. Bignotto 1, 1 A.O. San Bortolo, Vicenza, Italy; 2 University of Islamabad, Pakistan. Purpose: Since the introduction of IMRT (Intensity Modulated Radiation Therapy) and VMAT (Volumetric Modulated Arc Therapy) techniques, several commercial systems have become available with the purpose of verify the delivery of patient plans in terms of absolute dose, as part of a pre-treatment patient-specific Qualiy Assurance (QA) program. The use of electronic portal imaging devices, EPIDs, for dosimetric verification of treatment plans has been suggested by several groups [ref].the purpose of this study is to analyse some of these commercial systems, which are based on the images acquisition of the delivered beams with the EPID and their conversion to a dose distribution. These systems can provide either a 2D planar dose or a 3D reconstructed dose (within the CT scan of the patient). In this analysis, the accuracy of each system will be evaluated together with the sensitivity to measurement errors. Methods and materials: VMAT and IMRT plans for different anatomic localizations (prostate, Head&Neck, and intracranial lesions) have been calculated with the Pinnacle Treatment Planning System (TPS) ver. 9.2, with the module SmartArc TM for VMAT optimization. The plans have been delivered with an ELEKTA SynergyS TM linear accelerator. The portal imaging system associated is an iviewgt TM. The commercial systems that are analysed in this work are : a) EPIDose TM, which allows a conversion from Epid image to planar dose and it can be used as a pre-treatment verification system; b) EPIgray TM ; c) Dosimetry Check TM. These last two systems can be used for in-vivo dosimetry since they reconstruct a 3D dose in the CT of the patient from EPID images. For the purpose of comparison, a complete different system, which doesn t use the EPID as a dosimetric device, has been included in the study. It is the Compass TM system, which uses a 2D array of ionization chambers (I mrt MatriXX). This system allows a 3D reconstruction of the dose in the CT of the patient. Results: The comparison is still on going and results will be provided, based on the gamma analysis of the dose matrices, obtained with the planar dose from EPIDose and the corresponding planar dose from the TPS. For the systems which provide a 3D reconstruction, a comparison between the DVHs obtained from the dose reconstruction and the original DVHs obtained with the TPS, will be performed. Conclusion: The availability of many commercial systems for pre-treatment verifications of IMRT and VMAT plans and for in-vivo dosimetry, is recently increased. This requires a clear understanding of their functioning and of their possible pitfalls or limitations. The goal of this work is to provide some information from the point of view of the user. References: Steciw, S, Warkentin, B, Rathee, S and Fallone BG. Three-dimensional IMRT verification with a flat- panel EPID. Med. Phys. 32: , Van Esch A, Depuydt, T and Huyskens DP. The use of an asi-based EPID for routine absolute dosimetric pretreatment verification of dynamic IMRT fields. Radiother. Oncol. 71: , Van Zijtveld M, Dirkx MLP, de Boer HCJ and Heijmen BJM. 3D reconstruction for clinical evaluation of IMRT pretreatment verification with an EPID. Radiother. Oncol , 2007.

106 J Godart, EW Korevaar, R Visser, D J L Wauben and A A van t Veld Reconstruction of high- resolution 3D dose from matrix measurements: error detection capability of the COMPASS correction kernel method, Phys. Med. Biol. 56 (2011), extendedabstract.doc

107 Controlli di qualità ad impatto clinico per piani RapidArc della prostata: analisi gamma e valutazione dei parametri del DVH. Clinically relevant quality assurance (QA) for prostate Rapid Arc plans: gamma maps and DVH-based evaluation. M. Cozzolino 1, R. Caivano 1 C. Oliviero 1, G. Califano 1, S. Clemente 1, P. Pedicini 1,, C. Chiumento 1, A. Fiorentino 1, and V. Fusco 1. (1) Department of Radiation Oncology, IRCCS CROB, 1 Padre Pio Street, Rionero in Vulture, PZ, Italy. At our institute RapidArc (RA) is the technique of choice for treatment of prostate cancer: the radiation is delivered in one or two rotations of the gantry around patient and the treatment volume is created by a continuous change of the MLC positions during the rotation, varying the dose rate and the speed of gantry rotation [1]. According to many comparison studies, the target coverage and sparing of normal tissue was better than or as good as intensity modulated radiotherapy (IMRT) [2,3,4] with lower MU and shorter treatment time. This kind of treatment delivery represents a new level of complexity, and a through dosimetric verification is therefore highly desirable. Few publications addressed the employ of 2D detectors arrays, such as MapCHECK (Sun Nuclear) and MatriXX (IBA Dosimetry), to validate planar dose distributions of elaborate plans thanks to simple use and immediate readout [5,6] although their response was gantry angle dependent. Specifically developed products are currently available for arc techniques. A 3D diode array (Delta4, ScandiDos AB) that consists of 1069-p-type Silicon diodes in a crossed array inside a cylindrical PMMA phantom has lately been described [7,8] to override angle-dependence. A new 3D diode array (ArcCHECK, Sun Nuclear) has been developed for routine QA of IMRT and VMAT [9] and tested in a recent study by Li et al. [10]. While phantom-calculated based gamma supplies a basic understanding of whether the linac is operating as planned, it provides no-information regarding the end impact of treatment errors in the patient. A recent study showed that per-beam planar gamma passing rates doesn t predict the clinical impact on the patient in terms of changes in DVH values for target and organs-at-risk (OAR), which has thrown into question the feasibility of QA procedures based on gamma passing rate (GP) [11]. The QA process can be improved in terms of dosimetric accuracy and quantification of anatomical impact of delivery errors, using software that estimate patient dose errors. A novel QA system called COMPASS (IBA Dosimetry, Germany) allows to recalculate and to reconstruct dose distribution directly on the patient anatomy based on measured fluence at the gantry. At our department, the standard QA for RA plans was performed by the detector MatriXX but recently the COMPASS system was implemented to provide, from detector s measurements, a reconstruction of 3D dose distribution in the patient CT model. This QA system shows no-angular dependence, because the detector is kept normal to the beam at all times during beam delivery. Moreover it renders the results as dose distribution in patient anatomy, giving the possibility to evaluate clinically relevant dose discrepancies for each region. Previous studies [12, 13] evaluated the performance of COMPASS system for the verification of IMRT prostate plans, comparing to QA reference-procedures: absolute dose measurements with ion chambers and radiographic film measurements in various homogeneous and anthropomorphic phantoms. A maximum deviation of 0.4% with ion chambers and nearly 100% passed gamma criteria 3%/3 mm with films were found. Korevaar et al. [14] compared film QA of 24 head and neck IMRT patients to COMPASS QA, confirming a good agreement and sustaining clinical introduction. The aim of this paper is to describe our experience introducing COMPASS for the verification of RA prostate plans. Twenty plans were verified by COMPASS system that provides an independent angle response and a reconstruction of dose distribution in patient CT model. Plan data were imported from treatment planning system via DICOM. The fluencies were measured by the MatriXX ionization chambers array, mounted on gantry accelerator, and then were used by the COMPASS to forward calculate dose in CT patient and reconstruct the dose volume histogram (DVH). A 3D gamma comparison between the measured and TPS dose maps was performed. As the acceptance criteria, the common tolerance level was chosen: dose difference of 3% and distance to agreement

108 of 3 mm. The gamma passing rate was evaluated for whole grid patient (GP W ) and for each structure (GP S ) in terms of percentage of volume passing for selected gamma criteria; average gamma of PTV and OAR. The correlation between GP W,S and DVH discrepancies was performed using Pearson s test. Sensitivity, specificity and accuracy of GP W and GP S method were also calculated. Target coverage and OAR sparing were compared between TPS and actual delivery, evaluating the following DVH based metrics: dose received by 95% and 5% of PTV (D95, D5); the volume of rectum receiving 50 Gy, 60 Gy, 70 Gy (V50, V60,V70); volume of bladder receiving 60 Gy (V60), mean dose to femoral heads and to penile bulb. In addition, mean doses to PTV and normal structures (D mean ) were also evaluated. Statistical correlation of DVH deviations with GP was studied with Pearson s coefficient (r), the relative difference of DVH index was assumed to be correlated with GP when the p-value, obtained from r, was <0.05. To compare whole body and specific structure gamma methods, the number of false-negative (FN), true positive (TP), falsepositive (FP) and true-negative (TN) were calculated, assuming that a DVH deviation<5% corresponding to GP<95% is a FP and a DVH deviation<5% corresponding to GP>95% is a TN. So sensitivity, specificity and accuracy were obtained. The sensitivity and the specificity indicate the probabilities to identify dosimetric alterations >5% or <5%, respectively. Instead, the accuracy refers to number of cases correctly identified over total cases. The mean passing rate GP W for whole body was 99.17% while the average whole gamma is The mean GP S was for: PTV 94.01%; rectum 95.95%, bladder 98.05%, femoral head 97.43% and penile bulb 88.00%. The specific average gamma were for: PTV 0.44, rectum 0.43, bladder 0.34, femoral heads 0.33 and penile bulb The significant r-values were negative both for whole body and per-structure gamma. The most correlated DVH index resulted D mean to PTV, bladder and femoral heads (r = -0.77;-0.67;-0.71, p<0.01). Weak correlation between GP W,S and dosimetric deviations was observed, all significant r-values were negative. From the analysis of the sensitivity, specificity and accuracy, GP w resulted more efficient to identify the dosimetric alterations of different DVH indices. Actually delivered DVH metrics respected the objectives set in planning phase, but PTV coverage resulted slightly worse and bladder was on average less saved, while better sparing of rectum, penile bulb and femoral heads was observed. This underlines the importance of an accurate per-patient analysis based on DVH goals, with support of clinicians to approve the plan quality, but we claim the need for new benchmarks for clinical plans validation. New action levels must be defined for each clinic and each QA system. References: [1] K. Otto, Volumetric modulated arc therapy: IMRT in a single gantry arc. Med Phys (2008); 35: [2] F. Kjaer-Kristoffersen, RapidArc volumetric modulated therapy planning for prostate cancer. Vol. Acta Oncologica (2009); 48: [3] D.Palma, Volumetric modulated arc therapy for delivery of prostate radiotherapy: comparison with intensitymodulated radiotherapy and three-dimensional conformal radiotherapy. Int J Radiat Oncol Biol Phys (2008); 72: [4] Alexander AS, Wells D, Berrang T, Parsons C, Mydin A, Shaffer R, Wong F, Sayers D, Otto K. Volumetric arc therapy (VMAT) reduces treatment time compared to conventional IMRT (cimrt) while maintaining similar plan quality in whole pelvic gynecologic radiotherapy. Int J Radiat Oncol Biol Phys (2008); 72:S366. [5] Jursinic PA, Nelms BE. A 2-D diode array and analysis software for verification of intensity modulated radiation therapy delivery. Med Phys (2003); 30(5):870e9. [6] Wagner D, Vorwerk H. Two years experience with quality assurance protocol for patient related Rapid Arc treatment plan verification using a two dimensional ionization chamber array. Radiation Oncology (2011); 6:21. [7] Feygelman V, Forster K, Opp D, Nilsson G. Evaluation of a biplanar diode array dosimeter for quality assurance of step-and-shoot IMRT. J Appl Clin Med Phys (2009); 10(4):3080 [8] Sadagopan R, Bencomo JA, Martin RL, Nilsson G, Matzen T, Balter PA. Characterization and clinical evaluation of a novel IMRT quality assurance system. J Appl Clin Med Phys (2009); 10(2):2928.

109 [9] Yan G, Lu B, Kozelka J, Liu C, Li JG. Calibration of a novel four-dimensional diode array. Med Phys (2010); 37(1): [10] Li G, Zhang Y, Jiang X, Bai S, Peng G, Wu K, Jiang Q. Evaluation of the ArcCHECK QA system for IMRT and VMAT verification. Phys Med (2013); 29(3): [11] Nelms BE, Zhen H, Tomé WA. Per-beam, planar IMRT QA passing rates do not predict clinically relevant patient dose errors. Med Phys (2011); 38(2): [12] Boggula R, Lorenz F, Mueller L, Birkner M, Wertz H, Stieler F, Steil V, Lohr F, Wenz F. Experimental validation of a commercial 3D dose verification system for intensity-modulated arc therapies. Phys Med Biol (2010); 55(19): [13] Boggula R, Jahnke L, Wertz H, Lohr F, Wenz F. Patient-specific 3D pretreatment and potential 3D online dose verification of Monte Carlo-calculated IMRT prostate treatment plans. Int J Radiat Oncol Biol Phys (2011); 81(4): [14] Korevaar EW, Wauben DJ, van der Hulst PC, Langendijk JA, Van't Veld AA. Clinical introduction of a linac head-mounted 2D detector array based quality assurance system in head and neck IMRT. Radiother Oncol (2011); 100(3):

110 I nuovi film radiocromici EBT3 in radioterapia stereotassica: valutazione dell accuratezza dosimetrica Evaluation of dosimetric accuracy of stereotactic radiotherapy with new radiochromic EBT3 film D.Cusumano 1, M.Fumagalli 1, F.Ghielmetti 1, V.Pinzi 1, L.Fariselli 1, E.De Martin 1. (1) IRCCS C.Besta, Milano, Italia Purpose: Aim of this study is to examine the feasibility and dosimetric accuracy of using the new Gafchromic EBT3 model radiochromic film for stereotactic radiotherapy (SRT) quality assurance (QA). Materials and Methods: Measurements were performed using Gafchromic EBT3 films in conjunction with an Epson Expression 10000XL scanner. Uniformity and reproducibility were investigated analysing five different whole (8'x10') unexposed films (five times each). All scans were performed based on information by the manufacturer as well as published studies: transmission mode, 150 dpi resolution, RGB tagged image file format and acquiring the whole plate area [1]. The images were then processed using the red channel and dividing each whole film in 1 cm wide profiles in both landscape and portrait direction. The pixel values variations across these profiles were analysed to identify the region of the scanner plate with the higher uniformity (to appropriately position patient QA dose distribution films). Uniformity and sensibility dose dependence were also investigated irradiating 4x4 cm 2 films with doses ranging from 1 to 8 Gy (e.g. 0.8 Gy, 1.0 Gy, 1.2 Gy, 2.0 Gy, 2.2 Gy etc.). For each film, ADC values were extracted and evaluated from 2x2 cm 2 areas. Results: For unexposed films, uniformity tested for 1 cm wide profiles across the landscape direction had an average value of 0.58% (2σ values ranging from 0.50% to 0.83%, with the higher value further from the scanner center). Higher nonuniformity in the portrait direction is reported in literature and our results showed an average value of 0.9% [2]. These results did not appreciably vary for the irradiated films. Sensibility of the irradiated films was found to correspond to a few cgy for doses up to ~4 Gy and decreasing for doses from 5 to 8 Gy. Conclusions: This study evaluates the response of the new Gafchromic EBT3 film for typical SRT resolution and doses. The results showed an adequate level of accuracy for all the analysed dose levels, thus confirming the feasibility of using EBT3 films for SRT QA. Sensibility was found to be higher in low dose regions. References: [1] Huet C. et al., Characterization and optimization of EBT2 radiochromic films dosimetry system for precise measurements of output factors in small fields used in radiotherapy, Rad. Meas (2012), Vol.47, [2] Alnawaf H. et al., Comparison of Epson scanner quality for radiochromic film evaluation, Jour of App (2012), vol.5,

111 Experimental characterization of a x-rays IORT device C. Cutaia, S. Bresciani, M. Stasi Physics Department, Institute for cancer Research and Treatment (IRCC) at Candiolo (TO) Purpose: The aim of this work is to perform a dosimetric characterization of a x-ray IORT device (Intrabeam, Carl Zeiss Surgical, Oberkochen, Germany), made up by a 50 kv x-ray source which could be used for intra-operative radiotherapy delivering dose from inside a tumour cavity. Its principal application are partial breast irradiation and brain treatment.the Intrabeam device accelerates electrons that collide on a small golden probe producing x-rays in approximately isotropic distribution and is capable of currents up to 40 μa. Both the physical and the dosimetrical properties of the device are not similar to any other kilovoltage units used for therapy (Roentgen or plesiotherapy y). As a matter of fact the device delivers dose as it was an external beam device, the way of dose administration is similar to a conventional brachitherapy (the higher dose is near the bare probe or in contact with the applicators surface) and the typical energies used are in a radiological range. Methods and materials: All measurements were performed in a water phantom in whichh the source is inserted. The tank is also equipped with two waterproof detector holders, one placed horizontally with respect to the support plane used to measure the dose at different distances from the bare probe and one placed vertically useful for the isotropy measurement. A micrometer screw controls the motion of the bare probe source with an accuracy of 0.1 mm. We used a parallel plate ionizationn chamber PTW TN to perform basic measurements. This cm 3 soft X-ray chamber has an extremely small sensitive volume and its usage is indicated to measure therapeutic X-ray beams with very small field sizes or with steep fluence gradients. The calibration is typically done at 15 to 50 kv. The energy response within this range is ± 2 %. The membrane material is polyethylene of 0.03 mm thickness. The source was moved at distances between 5 and 47 mm with respect to the detector and depth dose curves were obtained. The step used between different measurements was 3 mm, reducing it to 1 mmm approaching the bare probe.to check for planar isotropy a 360 rotation of the source was performed and measurements were sampled every 45 at a fixed distance from the source. Moreover depth dose and isotropy curves weree measured in order to characterize all the applicators with different diameters (5, 4.5, 4, 3.5, 3, 2.5, 2, 1.5 cm). Results: The reproducibility of the measurements was around 2%. Considering the depth dose curve the dose rate droppedd from 27.5 Gy/min at 5 mm from the source up to Gy/min at 45mm from the x-ray generator (the dose rate at 10 mm was 4 Gy/min) using the bare probe as shown in figure 1. Figure 1. Dose rate Vs distance from bare probe

112 Thhe difference between thee dose rates measured m at a fixed distan nce from thee eight appliccators surfacee had a deppendence onn their size. At A a distance of 3.4 mm from fr the appllicator surfacce, for exampple, the dose rate was Gyy/min for thee 4.5 cm diam meter applicaator and Gy/min for the 1.5 cm one o as shownn in figure 2. Figuree 2. Dose rate V Vs distance from m applicators surface Thhe planar isottropy of the bare b probe w was constant w within a 2.2% %, while the isotropy meeasured for ev very app plicators shiffted in a rangge between 00.8% and 3.77%. The moree stable resuults were obtaained when the t smaller app plicators werre inserted (ssee figure 3).. Figure 3. IIsotropy Vs Source angle Co onclusion: nar isotropy measuremen m nts confirmedd the ones deeclared by the company Bo oth the dose ddepth curvess and the plan out dosimetrric characteriistic. abo Reeferences: [1]] D.J.Eaton, Q Quality assu urance and inndependent ddosimetry forr an intraoperratve x-ray ddevice, Mediccal Physics (20012), vol.39, N 11, TORNA ALL'INDICE DEI TOPIC TORNA ALL'INDICE DI RADIOTERAPIA 1

113 Implementazione clinica di un metodo di compensazione della dipendenza angolare del dispositivo MapCheck 2 per verifiche di piani di cura IMRT. F. De Monte *, O. Rampado *, A. Di Dio *, U. Nastasi * ( * ) A.O. Città della Salute e della Scienza di Torino, Torino Scopo: questo studio presenta l implementazione clinica di un metodo di compensazione della dipendenza angolare del dispositivo MapCheck 2 (MCK) (Sun Nuclear Corporation, Melbourne, Florida), composto da una matrice 2D di diodi, al fine di utilizzarlo per effettuare verifiche rotazionali di piani di cura IMRT (e non solo verifiche standard con gantry dei fasci a 0 ) e ricavare quindi informazioni dosimetriche volumetriche superiori rispetto alla misura di dose puntuale con camera a ionizzazione. Materiali e metodi: nella prima parte dello studio, è stata studiata la dipendenza angolare del MCK: sono state create 2 configurazioni per le verifiche rotazionali, con il MCK inserito tra strati di solid water (SW) (SP34 QA Phantom, IBA Dosimetry) fino a raggiungere uno spessore complessivo di 10 cm per verificare piani IMRT del testa/collo e di 20 cm per i piani IMRT della pelvi, con il livello dei diodi posizionato a metà spessore. Dalle specifiche del costruttore, sappiamo che il MCK ha un buildup inerente acqua-equivalente di 2.0 cm ed un backscatter inerente di 2.8 cm. Quindi, per ottenere 10 cm di spessore sono stati aggiunti 3.0 e 2.2 cm di SW rispettivamente sopra e sotto il MCK, mentre per ottenere 20 cm, oltre a questi, sono stati aggiunti 5.0 cm sia sopra sia sotto. Sono state eseguite scansioni TC di entrambe le configurazioni (spessore slice 0.1 cm), importate le immagini nel TPS Oncentra MasterPlan v.4.0, contornate le varie regioni e infine assegnate le densità: agli strati di SW è stata assegnata la densità dichiarata ( SP34 =1.045 g/cm 3 ), mentre al MCK è stata assegnata una densità omogenea; inoltre il lettino di trattamento è stato modellizzato come un bolus di densità lettino 0.5 g/cm 3 e spessore 1 cm. L isocentro è stato posizionato nel diodo centrale. Il MCK è stato calibrato per il fascio di fotoni da 6 MV generato dal Linac Elekta Precise su cui si eseguono trattamenti IMRT. La dipendenza angolare del MCK è stata studiata misurando le mappe di dose nel piano passante per l isocentro per i campi quadrati 5 5, 10 10, cm 2 al variare degli angoli di gantry ( ) ed erogando 100 UM/campo. I confronti tra le mappe di dose misurate e quelle calcolate dal TPS sono stati eseguiti utilizzando il software del MCK (analisi 3%-3 mm, threshold 10%, measurement uncertainty attiva). I calcoli, eseguiti con l algoritmo Collapsed Cone, sono stati ripetuti più volte assegnando densità diverse al MCK, da 1.0 a 1.5 g/cm 3 per cercare di stabilire quale valore fornisse il miglior accordo con le mappe di dose misurate. Nella seconda parte dello studio, dopo aver definito un appropriata procedura di calibrazione, abbiamo verificato 69 piani di cura IMRT (41 pelvi e 28 testa/collo da ottobre 2012 ad agosto 2013) usando sia il metodo standard (gantry 0 ) sia il metodo rotazionale. Risultati: considerando inizialmente la configurazione MCK di 10 cm, le tabelle seguenti mostrano rispettivamente: tab.1) per gantry 0 e campo cm 2, il confronto tra la dose misurata e quella calcolata all isocentro per diverse densità del MCK; tab.2) per gantry ruotati e campo cm 2, i risultati dei confronti tra le mappe di dose misurate e quelle calcolate all isocentro in termini di percentuali di punti che passano il confronto (passing points PP). Dai risultati risulta un range critico di angoli di gantry, nello specifico da 85 a 95, e in tutti i confronti misurato-calcolato si è notato che la dose misurata è maggiore di quella calcolata. La criticità è dovuta all assenza di materiale di buildup laterale nel MCK in corrispondenza dell ingresso dei fasci a tali angoli ed alle disomogeneità conseguenti alla presenza dei diodi e delle componenti circuitali sul piano di rivelazione. Dallo studio dei profili di dose assoluta ottenuti con gantry a 90 sull asse dei campi cm 2 (fig.1), si deduce che bisognerebbe applicare un fattore di correzione alla dose misurata per poter ottenere il miglior accordo con il calcolato. Visti nel complesso i risultati ottenuti, si decide di assegnare una densità omogenea al MCK pari a 1.2 g/cm 3, in accordo con quanto già riportato altrove (1). Questa scelta implica che, per compensarne la dipendenza angolare con gantry prossimi a 90, bisogna creare un apposito file di calibrazione in dose assoluta per il MCK per tali angoli: quindi si divide il valore della PDD del fascio per il fattore correttivo ricavato dalle misure (dopo aver

114 esaminato tutti i risultati per i campi in esame e per entrambe le configurazioni del MCK di 10 e 20 cm, abbiamo scelto un fattore correttivo medio tra tutte le misure: f=1.15). Riportiamo in tab.3 i risultati finali dei confronti ottenuti per entrambe le configurazioni del MCK con SW (l ultima riga della tabella indica il confronto tra la mappa di dose misurata ottenuta sommando i contributi delle mappe dei singoli fasci, tramite la funzione combine del software del MCK, e la mappa di dose calcolata ottenuta dal TPS sommando i contributi dei singoli fasci). Tab. 1: confronto tra dose misurata e dose calcolata all isocentro per la configurazione MCK di 10 cm, gantry 0 e campo cm 2. Dose misurata all isocentro - gantry 0 (cgy): 95.9 Densità assegnata al MapCheck 2 (g/cm 3 ): Dose calcolata (cgy): Differenza dal misurato (%): Densità assegnata al MapCheck 2 (g/cm 3 ): Angolo di gantry ( ): 1,0 1,1 1,2 1,3 1, Media PP (%) Media PP senza (%) Tab. 2: percentuali di PP ottenute dai confronti tra mappe di dose misurate e calcolate all isocentro per la configurazione MCK di 10 cm e campo cm 2. Fig. 1: correlazioni tra profili di dose misurati e calcolati ottenuti con gantry a 90 sull asse dei campi cm 2 per varie densità del MCK (configurazione MCK di 10 cm).

115 Densità assegnata al MCK 1.2 g/cm 3 : Configurazione MCK di 10 cm Configurazione MCK di 20 cm Angolo di gantry ( ) 25x25 cm 2 10x10 cm 2 5x5 cm 2 25x25 cm 2 10x10 cm Media PP (%) Media PP senza (%) PP dalla composizione (%) Tab. 3: percentuali di PP ottenute dai confronti tra mappe di dose misurate e calcolate all isocentro per le due configurazioni del MCK di 10 e 20 cm (MCK di densità assegnata pari a 1.2 g/cm 3 ). Si noti la criticità delle misure a 90, dove il miglior confronto con il calcolato è stato del 85%. Abbiamo quindi applicato il metodo di verifica rotazionale e quello standard a 69 piani di cura IMRT, di cui solo 3 prevedevano fasci con gantry nel range critico (90 /270 ). Per ciascun piano di cura, stabilita la configurazione del MCK con SW da utilizzare (10 cm per testa/collo e 20 cm per la pelvi): abbiamo prima acquisito le mappe di dose dei fasci con gantry a 0 e poi le mappe di dose agli effettivi angoli di trattamento; quindi, abbiamo confrontato le mappe misurate con quelle calcolate e ricavato la media dei confronti (in termini di PP) ottenuti con i due metodi; inoltre, solo per il metodo rotazionale, abbiamo confrontato la mappa di dose totale (ottenuta dalla composizione delle singole mappe misurate) con la mappa di dose totale calcolata dal TPS. Riportiamo nel grafico sottostante i risultati ottenuti per i 69 piani di cura ordinati in termini di media crescente dei PP ricavati con metodo standard. Considerando tutti i piani, il numero medio di PP è stato: - ( )% con il metodo standard (gantry 0 ); - ( )% con il metodo rotazionale (gantry effettivi di trattamento); - ( )% con la composizione delle mappe di dose ottenute con il metodo rotazionale.