Advancing offline magnetic resonance-guided prostate radiotherapy through dedicated imaging and deep learning-based automatic contouring of targets and neurovascular structures

Schubert P, May M, Höfler D, Fautz HP, Hutter J, Merten R, Mansoorian S, Weissmann T, Deloch L, Schonath M, Belmas N, Grabenbauer F, Frey B, Gaipl U, Axer BN, Szkitsak J, Uder M, Bert C, Fietkau R, Putz F

Publication Language: English

Publication Type: Journal article

Journal

Physics and Imaging in Radiation Oncology Elsevier

Book Volume: 35

Article Number: 100825

DOI: 10.1016/j.phro.2025.100825

Abstract

Background and Purpose: Erectile dysfunction (ED) affects quality of life following radiotherapy for prostate cancer. Magnetic resonance imaging (MRI) planning provides superior visualization of potency-related anatomical structures compared to computed tomography (CT), enabling improved sparing. However, contouring these structures in clinical practice is time-intensive and requires expertise. Deep learning (DL) auto-contouring with MRI simulation could make neurovascular-sparing radiotherapy more accessible. Material and Methods: High-resolution 3D T2-weighted SPACE MRI sequences (<1 mm3 resolution) were obtained for 50 patients in treatment position. An expert uro-radiation oncologist contoured erectile function-related anatomy (neurovascular bundles [NVB], pudendal arteries [IPA], penile bulb [PB], corpora cavernosa [CC]) and target structures (prostate [PR], seminal vesicles [SV]). Forty datasets trained and ten tested a 3D nnU-net model. DL-generated contours were geometrically evaluated (surface Dice Score [sDSC], mean surface distance [MSD]) and validated by blinded expert review. Results: DL auto-segmentation achieved an average sDSC of 0.82 (IPA: 0.93, NVB: 0.71, PB: 0.84, CC: 0.90, PR: 0.74; SV: 0.79) and average MSD of 0.74 mm (IPA: 0.61 mm; NVB: 0.88 mm; PB: 0.63 mm; CC: 0.47 mm; PR 0.83 mm; SV: 1.01 mm). Blinded ratings showed no significant differences between DL and expert contours, except for pudendal arteries (Mean DL vs. expert; NVB: 82 vs. 85; PB: 86 vs. 88; CC: 83 vs. 88; PR 81 vs. 83; SV 78 vs. 81 all p > 0.05; IPA: 82 vs. 89; p = 0.028). Conclusion: Combining high-resolution MRI simulation with DL postprocessing enables accurate auto-contouring for MR-guided SBRT planning, potentially advancing neurovascular-sparing radiotherapy beyond current standards.

Authors with CRIS profile

Philipp Schubert Department of Radiation Oncology Matthias May Institute of Radiology Daniel Höfler Department of Radiation Oncology Jana Hutter Professur für Smart Imaging and Data Profiling (Heisenberg-Professur) Ricarda Merten Department of Radiation Oncology Sina Mansoorian
Thomas Weissmann Department of Radiation Oncology Lisa Deloch Department of Radiation Oncology Natalia Belmas Department of Radiation Oncology Felix Grabenbauer Department of Radiation Oncology Benjamin Frey Department of Radiation Oncology Udo Gaipl Department of Radiation Oncology Bernd-Niklas Axer Department of Radiation Oncology Juliane Szkitsak Department of Radiation Oncology Michael Uder Lehrstuhl für Diagnostische Radiologie Christoph Bert Professur für Med. Strahlenphysik Rainer Fietkau Lehrstuhl für Strahlentherapie Florian Putz Department of Radiation Oncology

Involved external institutions

Bayerisches Zentrum für Krebsforschung (BZKF) DE Germany (DE)

How to cite

APA:

Schubert, P., May, M., Höfler, D., Fautz, H.P., Hutter, J., Merten, R.,... Putz, F. (2025). Advancing offline magnetic resonance-guided prostate radiotherapy through dedicated imaging and deep learning-based automatic contouring of targets and neurovascular structures. Physics and Imaging in Radiation Oncology, 35. https://doi.org/10.1016/j.phro.2025.100825

MLA:

Schubert, Philipp, et al. "Advancing offline magnetic resonance-guided prostate radiotherapy through dedicated imaging and deep learning-based automatic contouring of targets and neurovascular structures." Physics and Imaging in Radiation Oncology 35 (2025).

BibTeX: Download