ESTRO 2020 Abstract book

S329 ESTRO 2020

PD-0548 Quantitative Analysis of SCOPE 2 Trial 4DCT pre-accrual benchmark cases J. Kahan 1 , R. Carrington 2 , S. Cox 1 , G.J. Lewis 2 , S. Gwynne 3 1 Velindre Cancer Centre NHS trust, Clinical Oncology, Cardiff, United Kingdom ; 2 Velindre Cancer Centre NHS trust, Medical Physics, Cardiff, United Kingdom ; 3 South West Wales Cancer Centre, Clinical Oncology, Swansea, United Kingdom Purpose or Objective We have previously reported results of a qualitative analysis of the RTQA 4DCT benchmark cases within the SCOPE 2 trial (UK phase 2/3 trial looking at PET response and dose Escalation). Here we present a complementary quantitative analysis to explore whether automated assessment using conformity indices can expedite the time-to-approval of centres for use of 4DCT in the trial. Material and Methods Centres underwent pre-submission training and had access to a clear protocol. Submissions were classified as having acceptable or unacceptable variation based on pre- defined criteria. Cases were compared to a reference volume created from outlines from five experienced clinicians using the STAPLE algorithm. Future use of 4DCT was only permitted once a benchmark case had been satisfactorily completed. To date 21 cases were available for analysis, including the 14 original submissions. Of these, 3 (21%) were acceptable. Of the resubmissions, 5 (83%) were acceptable. One centre required a third attempt which was then considered acceptable. Qualitative analysis has previously shown that unacceptable variation usually occurs in the creation of ITV (the composite of CTVB volumes, edited to account for motion seen in 4DCT phases). Therefore, this analysis was limited to ITV. Investigator ITV (iITV) was compared to reference ITV (rITV) using Jaccard Conformity Index (JCI) and Geographical Miss Index (GMI). The GMI indicates the volume of rITV that is missed within the iITV, 0 indicates no volume has been missed. JCI indicates the concordance of rITV and iITV. 1 represents perfect concordance while 0 represents none. JCI of 0.7 is considered to represent excellent concordance in oesophageal cancer. Results For all cases submitted the median iITV JCI was 0.79 indicating very good conformity. Within the acceptable submissions; median JCI was 0.8, 100% had JCI >0.7 and 44% had JCI >0.8 (see table 1). Median GMI in these cases was 0.07 suggesting very little of the rITV was missed from the iITV. However, in those with unacceptable variation (12 cases) they also displayed high JCI values. Median iITV JCI was 0.76, 83% had a JCI >0.7 and 17% >0.8. These results were gained despite failure of volumes to account for appropriate motion in ITV which was identified on analysis of the volumes by the trials team. GMI in this case was 0.12 which indicates a larger volume missed from the iITV.

Conclusion Assessment of whole volume concordance with JCI does not reliably distinguish between acceptable and unacceptable variations in the iITV. This was not improved by increasing of JCI to 0.8. The addition of GMI may aid analysis but there is no agreed significant value. Further work is needed to explore whether slice-by-slice analysis would lead to better discrimination and provide automated decision making. However, the qualitative review with clinician feedback remains an essential component of a RTQA programme. It can lead to improved clinician skills, delineation and identification of ambiguities in the protocol. PD-0549 Improving radiotherapy accuracy with EPID in- vivo dosimetry: results from a multicentric study. M. Esposito 1 , A. Piermattei 2 , S. Bresciani 3 , L.C. Orlandini 4 , M.D. Falco 5 , S. Giancaterino 5 , C. Savino 6 , A. Ianiro 6 , R. Nigro 7 , L. Botez 3 , S. Riccardi 7 , A. Fidanzio 2 , F. Greco 2 , E. Villaggi 8 , S. Russo 9 , M. Stasi 3 1 USL CENTRO TOSCANA, S.C. Fisica Sanitaria, Bagno a Ripoli, Italy ; 2 Radioterapia Oncologica ed Ematologia- Fondazione Policlinico Universitario “A. Gemelli” IRCCS, UOC di Fisica Sanitaria-, Roma, Italy ; 3 Candiolo Cancer Institute – FPO IRCCS, Medical Physics, Candiolo, Italy ; 4 Sichuan Cancer Hospital, Department of Radiation Oncology, Chengdu, China ; 5 Università di Chieti, Dipartimento di Radioterapia, Chieti, Italy ; 6 Fondazione di ricerca e cura “Giovanni Paolo II”, Medical Physics Unit, Campobasso, Italy ; 7 OGP S. Camillo de Lellis, Medical Physics Unit, Rieti, Italy ; 8 AUSL Piacenza, Medical Physics Unit, Piacenza, Italy ; 9 USL Toscana Centro, S.C. Fisica Sanitaria, Bagno a Ripoli, Italy Purpose or Objective The Working Group of Associazione Italiana di Fisica Medica on in-vivo dosimetry (IVD) carried out a program to investigate critical aspects and effectiveness of IVD obtained by electronic portal imaging device (EPID). Material and Methods 8 Centers, equipped with three commercial systems (SOFTDISO (SD), Dosimetry Check (DC), and PerFRACTION (PF)), collected IVD results for a total of 2010 patients and 32516 tests. In the table, data are summarized for IVD software, radiotherapy technique and treated anatomical site. Every Center reported the number of patients and tests carried out and the percentage of tests out of the tolerance level (T%). T% was categorized as due to: incorrect setup, incorrect