ESTRO 2023 - Abstract Book

S126

Saturday 13 May

ESTRO 2023

Conclusion Our results show that PSA dynamics within the first 3-years of RT are prognostic of future BCR. Further, we demonstrate that including postRT PSA dynamics in predictive models improves model performance, with all postRT PSA variables more prognostic than baseline PSA. These results demonstrate the value of short-term postRT PSA information for prediction of future BCR in men with PCa and provide a mechanism for better targeted follow-up. PD-0165 Combining patient and population rectal shapes for re-irradiation of recurring prostate cancer S. Pilskog 1,2 , A. Størksen Stordal 3 , L. Bolstad Hysing 1,4 , C. Ekanger 1 , Ø. Lunde Rørtveit 1,4 1 Haukeland University Hospital, Department of oncology and medical physics, Bergen, Norway; 2 University of Bergen, Faculty of Mathematics and Natural Sciences, Bergen, Norway; 3 NORCE Norwegian research centre, Department of Energy, Bergen, Norway; 4 University of Bergen, Department of oncology and medical physics, Bergen, Norway Purpose or Objective Balancing risk of toxicity against effect is challenging in re-irradiation of local recurrent prostate cancer, as dose-response data for these patients are non-existing. A pragmatic solution in order to avoid toxicity in our clinic is to crop the target volume to the rectal shape of the planCT, and thereby constraining target dose. Knowing that rectal motion varies among patients and that rectal shape changes can affect our decision, the objective was to develop a model that combines patient specific rectal shapes with population motion patterns to assess the consequence of changes. Materials and Methods A Bayesian deformation model was used to combine the two patient-specific rectal shapes with the shape changes obtained from a population (Figure 1). The patient-specific rectal shapes in this case were shapes from the primary and the re irradiation planCT, respectively (Figure 1, right). These patients belong to a cohort 39 patients treated for recurring prostate cancer. The rectum shapes from the two planCTs were deformably co-registered. Thereafter, the rectum shape of the second planCT was deformably registered to a population rectal shape (PRS). The PRS was obtained from >300 CTs in another cohort of 37 prostate cancer patients. The resulting deformation fields allow for analysis of probable motion patterns (Figure 1, left) and generation of new probable rectum shapes. The estimated shape of the mean rectum in the re-irradiation patients was found by shrinkage estimation, weighting the two patient-specific rectum shapes with the PRS. To estimate variations around this estimated mean shape, the Bayesian deformation model was used to interpolate probable deformation patterns that move the rectum into and out of the high dose volume. The mean rectum shape was deformed in this way by +/-1SD . To evaluate the impact of the changes we compared DVHs for the modelled mean rectum shape +/- 1SD to the rectum shape in the planCT used for re-irradiation.