ScholarDeck

AI / informaticsPhysics and imaging in radiation oncology · 2026

Deep learning-based generation of direct stopping power ratio maps for MR-only proton therapy of primary brain tumor patients.

Emilie Alvarez-Michael, Nils Peters, Julian Schwengfelder … Esther G C Troost

OphthalmologyAI / informatics

AI-generated proton stopping power maps from magnetic resonance imaging closely matched actual measured values.

Deep learning-based generation of direct stopping power ratio maps for MR-only proton therapy of primary brain tumor patients.

Emilie Alvarez-Michael … Esther G C Troost

Physics and imaging in radiation oncology · 2026

Background

A magnetic resonance-only proton therapy (MRoPT) workflow requires synthetic computed tomography (computed tomography) to derive the stopping power ratio (stopping power ratio) for treatment planning.

Methods

Stopping power ratio maps and heterogenous magnetic resonance imaging data were collected from 140 patients.

n = 140 patients

Results

0.062

how far off the AI map was from the real measurement - lower means more accurate

n = 140 patients

More results

For the clinical target volume, D2% and D98% differences ranged from -0.60% to 0.75%.

Regarding the OAR, the Dmean and Dmax differences, corrected for relative biological effectiveness, were within ±2.70Gy (RBE).

The mean pass rate for the local gamma criterion of 2%/2 mm with 10% dose threshold was 76.19 ± 4.67%.

More results

Averages of absolute range shifts ranged from 0.50 mm to 1.79 mm.

“

Conclusion

Generated sSPR showed minor discrepancies from stopping power ratio. This study is a step closer to an MRoPT workflow for brain tumor patients, albeit the residual errors full impact needs to be more investigated in future work.

Post

Deep learning-based generation of direct stopping power ratio maps for MR-only proton therapy of primary brain tumor patients.

AI-generated proton stopping power maps from magnetic resonance imaging closely matched actual measured values.

Related papers