DKFZ-RadOpt/pyRadPlan-dosecalc-Det-DoseLET-proton-lung

pyRadPlan-dosecalc-Det-DoseLET-proton

Dual-output proton dose + LET predictor (ConvDual architecture, single forward pass). Given a CT cuboid and a bixel energy, predicts a local physical dose cuboid and a dose-weighted LET cuboid.

Loadable via pyRadPlan.ml.load_model and usable directly as the AIBeamletEngine dose-calculation engine in pyRadPlan:

pln.prop_dose_calc = {"engine": "AIBeamlet", "model": "pyRadPlan-dosecalc-Det-DoseLET-proton"}

Repository contents

Follows the pyRadPlan ML model contract (pyRadPlan.ml):

File	Purpose
model.py	ConvDual network definition
preprocessor.py	ConvDoseDualPreprocessor — input assembly + forward pass + channel-split output scaling
weights.safetensors	Trained weights
model_config.json	Declarative model/preprocessing/dose-calc configuration

Outputs

• physical_dose — predicted physical dose
• let_dose — predicted dose-weighted LET

Training assumptions

• Radiation mode: protons
• Energy range: 75–160 MeV
• Trained machine: Generic
• Sampling grid: 2 mm spacing, ±50 mm lateral range, depth range [-220, 100] mm

Predictions outside this range are not guaranteed to be accurate; AIBeamletEngine warns when the plan falls outside the declared range.

Security note: loading this model executes model.py/preprocessor.py shipped in this repository (gated behind trust_remote_code, default True in pyRadPlan.ml).