Late toxicity-associated patient-specific subregions in plan optimization for prostate cancer
MO-0475
Abstract
Late toxicity-associated patient-specific subregions in plan optimization for prostate cancer
Authors: Lisa Alborghetti1, Roberta Castriconi1, Carlos Sosa Marrero2, Alessia Tudda1, Maria Giulia Ubeira-Gabellini1, Sara Broggi1, Javier Pascau3, Lucia Cubero3, Cesare Cozzarini4, Renaud De Crevoisier2, Tiziana Rancati5, Oscar Acosta2, Claudio Fiorino1
1IRCCS San Raffaele Scientific Institute, Medical Physics, Milano, Italy; 2CLCC Eugène Marquis, INSERM, LTSI—UMR1099, F-35000, Univ Rennes, Rennes, France; 3Universidad Carlos III de Madrid, Bioengineering Department, Madrid, Spain; 4IRCCS San Raffaele Scientific Institute, Radiotherapy, Milano, Italy; 5Fondazione IRCCS Istituto Nazionale dei Tumori (INT), Progetto Prostata, Milano, Italy
Show Affiliations
Hide Affiliations
Purpose or Objective
An association between the dose received by local subregions of bladder/rectum and toxicity in prostate cancer radiotherapy has been recently evidenced by voxel-wise analyses. An exploratory study was performed to assess if a knowledge based (KB) plan prediction model can be further optimized using multi-criteria optimization (MCO), reducing the dose to symptom-related subregions (SRS) without any detriment of the dose distribution in targets and OARs.
Material and Methods
Forty-five high-risk prostate cancer patients previously treated in Institute A to 74.2Gy/28fr were selected. Using deformable registration, according to previous studies, contours of SRS were generated in institute B and returned to institute A for planning. Five patients were selected to reasonably represent the different anatomy (i.e. the location of SRS in relation to PTVs). A KB RapidPlan (Varian, Inch.) model was used to obtain clinically suitable plans using VMAT with four complete arcs (KB plan). Then, keeping the same beam ballistic, plans were further optimized using MCO (Trade-Off Exploration, Varian Eclipse v. 16.1) to reduce the dose to the four SRS shown in Fig.1. Three are vesical regions linked to specific late symptoms, namely dysuria (DYS_LAT), hematuria (HEM_LAT) and retention (RET_LAT). The last in the rectum relates to rectal bleeding (SRR). First, a set of clinical goals was established: for low (pelvic nodes), intermediate (seminal vesicles) and high dose (prostate) PTV, the minimum value of V95% was set to 95%. Moreover, objectives for MCO were chosen. Compared to the KB plan: a) dose homogeneity for PTVs should be unchanged; b) DVHs of the whole rectum/bladder and other OARs cannot be worsened; c) mean dose of SRS should decrease as much as possible while respecting the previous criteria. Finally, a database of optimal plans was generated and navigated to find the best-compromised plan, giving the same importance to all SRS.
Results
DVH variations from KB to KB+MCO plan and differences in dose parameters were evaluated. Mean dose difference ranges between -2,4 and -1,5 Gy for DYS_LAT, -2,1 and -1,2 Gy for HEM_LAT, -4,0 and -1.9 Gy for RET_LAT, -2,3 and -0,5 Gy for SRR. Dose to 1% of the volume (D1%) also decreased in most cases: the ranges were between -1,3 and -0,1 Gy for DYS_LAT, -6,4 and -1,9 Gy for HEM_LAT, -5,4 and -1,7 Gy for RET_LAT, -2,3 and 0,2 Gy for SRR. The better sparing of SRS was clearly visible on DVH; at the same time, PTVs and OARs DVH not only was not worsened, but in some cases even improved (Fig.2).
Conclusion
A selective sparing of vesical and rectal substructures which does not compromise the appropriate coverage of the targets seems feasible. An extension of the procedure to a larger population of patients is therefore justified and could be the basis for the development of a KB plan prediction model incorporating the sparing of SRS.
This study is supported by ERAPERMED-2020-110-JTC.