ESTRO 2023 - Abstract Book

S1345

Digital Posters

ESTRO 2023

When models from consortium A were applied to patients from consortium B, PC1 was inside the 90th percentile in 6 out of 10 models in 90% of the patients, highlighting a favourable transferability agreement. In 2 out of 10 models, all patients have a PC1 inside the 90th percentile against 7/10 of consortium A. PC1 for model 6 showed poorest transferability for both consortia. Conclusion Geometrical/anatomical quantitative parameter as PC1 confirms a reasonably good inter-consortium model’s transferability. Variation in models’ transferability will be better investigated looking at national contouring guideline differences, geometrical features, models’ dose prediction and the SD of several DVHs and dose/statistic parameters.

This study is supported by an AIRC grant (IG23150).

PO-1647 Novel dataset validation of deep learning models for autocontouring of head and neck, and prostate

D. Sandys 1 , N. Fersht 2 , A. Thompson 2 , R. Davda 2 , S. Khan 3 , M. Bristow 4 , P. Hessey 4 , A. Schwaighofer 4

1 University College London Hospitals NHS Foundation Trust, Radiotherapy Physics, London, United Kingdom; 2 University College London Hospitals NHS Foundation Trust, Oncology, London, United Kingdom; 3 University College London Hospitals NHS Foundation Trust, Radiotherapy and Proton Beam Therapy, London, United Kingdom; 4 Microsoft Research, Health Futures, Cambridge, United Kingdom Purpose or Objective Manual contouring in radiotherapy is a major time and cost demand. Additionally, interobserver variability (IOV) is a major contributor to treatment variance in radiotherapy. Automated contouring (autocontouring) has the potential to provide consistent contours with substantially reduced demand on resources. Previous work has shown it is possible to train deep learning models to generate autocontours which agree with manual clinician contours within the range of clinician IOV. This work assesses the quality of two previously trained autocontouring models on a novel retrospective patient cohort, from a clinical centre which did not provide training data. Materials and Methods Models (InnerEye, Microsoft Research) were previously trained on head and neck (H&N), and prostate patient cohorts contoured according to EORTC and TROG guidelines. This validation work assessed these models against a retrospective cohort of H&N (n=20) and prostate (n=20) patients treated between 2015-2018 and 2019-2020 respectively. The H&N cohort were treated with EBRT for 65Gy in 30#, or 70Gy in 35#. The prostate cohort were treated with EBRT for 60Gy in 20# to the prostate and seminal vesicles (excluding post-operative or pelvic lemph node radiotherapy). Manual clinician contours were generated according to local contouring guidelines and collected retrospectively. Manual and autocontours were compared using Dice similarity coefficient and Hausdorff distance. Manual contours from the local clinical centre acted as ground truth for comparison. Results Autocontours were generated for each patient (Fig. 1). Here a subset of results is presented (Fig. 2), showing agreement with local clinicians is consistent with previously published results for these models. Statistics which deviate from previously published values by are highlighted. These aggregate statistics indicate results which are broadly consistent with those indistinguishable from IOV.