ESTRO 2020 Abstract Book

S266 ESTRO 2020

Physics, Florence, Italy ; 12 SS. Annunziata Hospital, Medical Physics, Chieti, Italy ; 13 University of Chieti- SS. Annunziata Hospital, Department of Radiation Oncology “G. D'Annunzio”, Chieti, Italy ; 14 Ospedale San Pietro FBF, Medical Physics, Rome, Italy ; 15 HSR San Raffaele Milano, Medical Physics, Milan, Italy ; 16 Policlinico S. Orsola-Malpighi, Medical Physics, Bologna, Italy ; 17 Fondazione di Ricerca e Cura “Giovanni Paolo II” Università Cattolica del Sacro Cuore, Medical Physics, Campobasso, Italy ; 18 Regina Elena National Cancer Institute, Medical Physics, Rome, Italy ; 19 Azienda Ospedaliera Ordine Mauriziano, Medical Physics, Turin, Italy ; 20 Humanitas Clinical and Research Hospital, Medical Physics Unit of Radiation Oncology, Milan, Italy Purpose or Objective The rising interest to implement stereotactic body radiation therapy (SBRT) for prostate carcinomas (PCa) with simultaneous boost (SIB) to the dominant intraprostatic lesion (DIL) highlights several challenges concerning planning methodology and dosimetric variability. The Italian Association of Medical Physics (AIFM) and the German Society for Medical Physics (DGMP) working groups for SBRT joined conducting a study to evaluate the feasibility and variability of SIB planning for prostate SBRT with DIL in an international environment and with different planning platforms. Material and Methods An in-house web-platform ( https://sbrtvirtualaudit.it ) was developed to share and analyze two pre-contoured PCa cases with DIL boost volumes. For each case, 2 plans with different prescriptions were optimized: (I) 35 Gy (7Gy x 5fr) homogeneous to the planning target volume (PTV) which includes the PTV-DIL; (II) 50 Gy (10Gy x 5fr) as SIB to the PTV-DIL and 35 Gy to the remaining PTV (plan objectives are listed in Table 1). Participants uploaded their solutions to the web-platform using DICOM-RT format. Statistical methods to characterize crowd data were developed using R. Dose-Volume-Histogram (DVH) curves were recomputed from the original DICOM-RT data using the RadOnc R-package and grouped to build crowd- DVHs. Crowd-DVHs were presented as a shaded area indicating the range of the collected DVHs and a solid line indicating the median DVH (Fig.1). A two-sample t-test was used to quantify significant differences (p < 0.01) among crowd-DVH curves of the two different prescriptions. Furthermore, the same test was also performed between crowd-DVHs and individual DVHs to quantify significant differences between participants.

Forty-eight treatment plans with various technologies (2 robotic-based, 1 tomotherapy and 9 c-arm-based Linacs) from Italy, Germany, Spain and Australia were collected. Despite the significant increase of dose between the classical and the DIL prescriptions, most planners reached crowd-derived pareto-optimal dose distributions to the PTVs and OARs (median, 1 st and 3 rd quartile shown in Table 1). Figure 1 shows the crowd-DVHs of PTV-DIL and rectum with statistical significance of differences in the upper panel (yellow areas indicate dose levels where differences are significant, p < 0.01). No difference between centers from different nations was observed.

Conclusion Treatment planning for a prostate SBRT with SIB to DIL was feasible using various technologies with no major difference compared to classical prescription. Representation of crowd-DVHs as described in this study can be useful in clinical trial settings to facilitate analysis of submitted cases that are automatically benchmarked against prior submissions as the cohort of participants grows. This can help to reduce variability and harmonize

Results