ESTRO 2022 - Abstract Book

S368

Abstract book

ESTRO 2022

In terms of DICE overlap the ANN output was 0.91±0.03, 0.94±0.04 and 0.82±0.09 while for DIR 0.92±0.03, 0.92±0.08, 0.86±0.05 for the CTV, bladder and rectum respectively. Similarly, APL results where 3479±1835, 7356±4391 and 6832±2362 for ANN and 2853±1687, 8571±4654 and 6395±2663 voxels for DIR.

Conclusion Patient specific artificial neural network models trained on a single dataset are feasible with comparable accuracy to DIR.

OC-0424 Target volume time trends during daily adaptive 5-fraction MR-guided radiotherapy

E. Koper 1 , M. Kamer 1 , O. Bohoudi 1 , M.A. Palacios 1 , S. Senan 1 , F.L. Schneiders 1 , B. Slotman 1 , F. Lagerwaard 1 , A. Bruynzeel 1

1 AmsterdamUMC, Radiotherapy, Amsterdam, The Netherlands

Purpose or Objective Stereotactic MR-guided adaptive radiation therapy (or SMART) allows for recontouring of both target volumes and organs at risk prior to each fraction. Since SMART allows for optimizing treatment delivery in response to inter-fractional anatomical and volumetrical changes, subsequent fractions are reviewed separately. This approach will incorporate radiation-induced increases in the GTV into planning. However, editing daily contours can potentially reduce target volumes, and result in the unintended use of ‘shrinking field’ radiation. In this study we evaluated time trends in GTV’s (and CTV for prostate cancer), during the course of SMART with daily plan re-optimization. Materials and Methods All patients treated with five fractions of SMART (total dose 35-50 Gy) within an overall treatment time of 2.5 weeks at a single institute using the MRIDIAN system (ViewRay Inc) were selected for this study. Target volume data for subsequent fractions were obtained for prostate cancer (N=463 pts), pancreatic cancer (N=129 pts), lung tumors (N=40 pts), renal cell cancer (N=79 pts) and adrenal metastases (N=58 pts). An analysis of time trends was performed using an ANOVA test with post hoc analysis. Additionally, in order to investigate the incidence of substantial target volume reduction during the course of SMART, the deviation (%) from fraction 5 relative to fraction 1 was calculated for each patient. Results The changes in GTV’s during the course of MRgRT were only modest, and on average (95% CI) for prostate CTV’s +6.7% (6.0 - 7.3%), for pancreatic cancer +1.3% (0.2 - 2.4%), for lung tumors +1.0% (-1.6 - 3.7%), for renal cell cancer +1.0% (-0.7 - 2.8%), and for adrenal metastases +1.8% (-0.8 - 4.4%). Only the average increase in prostate cancer CTV’s as compared to baseline was statistically significant (p < 0.001). Table 1 shows the rate of patients exhibiting a 10% and 20% decrease of