BEAM3R: Beam's-eye-view architecture with Mamba-3 for implicit dose reconstruction

Chen Cheng; Michael Ferraro; James Grover; David E J Waddington; Emily Hewson

physics.med-phSep 4, 2026

BEAM3R: Beam's-eye-view architecture with Mamba-3 for implicit dose reconstruction

Authors: Chen Cheng, Michael Ferraro, James Grover, David E J Waddington, Emily Hewson

Organizations: Image X Institute, The University of Sydney, Sydney, Australia

Abstract

To enable accurate and rapid photon control point and proton beamlet dose calculation in the DoseRAD2026 challenge, we present BEAM3R, a dose estimation framework operating in beam's-eye-view (BEV). Our core innovation combines a Mamba-3 state-space depth-sequence core with physics-based transport conditioning to model long-range depth transport without expensive 3D convolutions. BEAM3R shares a 2D CNN encoder-decoder architecture for photon and proton dose tasks, processing per-plane BEV slices. Proton beamlets are conditioned on water equivalent thickness and remaining range, encoding the parameters determining Bragg peak position. Photon models use a bidirectional Mamba-3 core to capture dose contributions from materials downstream of the calculation point, while the proton model uses a forward core with learned energy-prefix tokens and a Bragg-peak refinement module. To reduce interpolation artifacts and support high spatial resolution, we introduce axial grid alignment of BEV lattices with CT slices and an implicit super-resolution representation via sub-pixel phase packing, evaluated by a differentiable Triton-accelerated resampler that reconstructs packed cubic B-spline coefficients directly in CT space. For MRI-based tasks, synthetic CTs (sCT) are generated by a patch-based conditional GAN with a SwinUNETR backbone. On the preliminary DoseRAD2026 test set, CT-to-photon and CT-to-proton models achieved 1%/1 mm local gamma pass rates of 96.8% and 96.0%, with stratified plan-level MAEs of 0.0041 and 0.0079. Substituting sCT reduced gamma pass rates to 89.7% for photon and 75.4% proton plan level doses, with stratified plan-level MAEs of 0.0093 and 0.0336. Standardised runtimes were 23.4 s and 18.4 s for CT-to-photon and CT-to-proton prediction, increasing to 39.7 s and 42.8 s for the corresponding MRI-based pipelines.

Figures & tables

Task/modality	Cohort	Training patients	Validation patients	Split seed
Photon	Abdomen	30	1ABB041 , 1ABB070 , 1ABB110 , 1ABB115 , 1ABB155	333
Photon	Thorax	36	1THB017 , 1THB023 , 1THB121	333
Proton	Abdomen	34	1ABB030 , 1ABB083	42
Proton	Thorax	37	1THB095 , 1THB211	42

Table 1: Patient-level training and validation splits. Patient 1ABB102 was excluded for photon model as it had outlier statistics in dose fluence

Figure 1: Left: Shared BEV architecture for photon and proton dose prediction. Green dashed boxes mark proton-only components and the magenta dashed box shows the photon-only components.

Task	MAE	IDD Dist.	Plan MAE	$\gamma$ Index	DVH Score	Runtime
CT to photon	0.0084 $\pm$ 0.0022	0.0069 $\pm$ 0.0105	0.0041 $\pm$ 0.0019	96.7569 $\pm$ 3.0172	0.2173 $\pm$ 0.0663	23.4493
MRI to photon	0.0139 $\pm$ 0.0064	0.0097 $\pm$ 0.0093	0.0093 $\pm$ 0.0056	89.7046 $\pm$ 7.2553	0.9051 $\pm$ 0.1782	39.6613
CT to proton	0.0084 $\pm$ 0.0039	0.0058 $\pm$ 0.0081	0.0079 $\pm$ 0.0039	95.9532 $\pm$ 3.6125	0.5018 $\pm$ 0.3073	18.3807
MRI to proton	0.0326 $\pm$ 0.0216	0.0287 $\pm$ 0.0208	0.0336 $\pm$ 0.0226	75.4368 $\pm$ 18.9888	9.5124 $\pm$ 9.5922	42.8198

Table 2: Evaluation metrics for CT/MRI to photon/proton dose calculated on the preliminary test set for the DoseRAD2026 challenge. MAE: beam-level mean absolute dose error; IDD Dist.: normalised integrated depth-dose curve distance; Plan MAE: stratified plan-level MAE;

\gamma

Index: 3D local gamma pass rate (%) at 1%/1 mm; DVH Score: DVH-based clinical score.

Design decision	Variant	Epoch	MAE	IDD
Thorax cohort, 36 training patients
BEV depth extent b	256 planes	122	0.0117	0.0120
	384 planes	124	0.0107	0.0084
Model capacity b	$1.0\times$	94	0.0116	0.0076
	$\mathbf{2.0\times}$	96	0.0098	0.0070
Sequence core b,c $\ddagger$	mLSTM	115	0.0104	–

Table 3: Experiments motivating design decisions final model. Each pair varies the stated factor; the epoch is the lowest-validation-loss epoch within the range common to both runs. Metric sets differ by experiment generation: a BEV-space masked MAE and BEV IDD distance; b CT-space masked MAE and round-trip BEV IDD distance; c CT-space masked MAE and true-gantry beam IDD distance; d CT-space masked MAE and CT-space MSE (

\times 10^{-6}

), the two runs sharing no IDD variant for the reason given in the text. Rows marked

\dagger

vary one further factor and

\ddagger

several (see text). Lower is better throughout. SR: super-resolution.

Figure 2: Failure modes in photon and proton dose estimates taken from the held out validation set. (a) and (c) show a comparison of ground truth, prediction, difference dose and IDD curve prediction for a CT input to photon beam prediction and a proton beamlet prediction respectively. (b) and (d) show a comparison of the MR input, sCT, real CT, prediction and dose difference for a photon beam and a proton beamlet respectively.

Radiotherapy Dose Prediction

See all 12 papers

Sep 1, 2026

PyDoseRT Proton: A GPU Pencil-Beam Engine with a Convolutional Residual-Correction Network for Fast Proton Dose Calculation

Architecture category. Hybrid method: a physics-based analytical pencil-beam (PB) dose engine followed by a 3-D convolutional residual-correction network (RepVGG-U-Net). We addressed the DoseRAD2026 proton dose-prediction task with PyDoseRT Proton, a GPU-accelerated engine implemented in PyTorch and augmented by a learned residual toward Monte Carlo (MC) accuracy. A double-Gaussian PB kernel was calibrated to GATE/Geant4 integrated depth doses in water in two stages: a classical per-energy curve fit, then a gradient-based fit of the full 3-D dose through the PyTorch physics engine as it retains a differentiable execution path for gradient-based optimization of dose-dependent objectives. The engine computes each beamlet on a beam's-eye-view (BEV) lattice with variance-preserving Gaussian splitting, an analytic nuclear halo, and a Fermi-Eyges heterogeneity term, then rotates the result into the patient frame. Additionally, a compact residual U-Net predicts an additive correction in BEV space. It is conditioned on voxelwise material-label embeddings, a discrete energy embedding and spot size. The same model was used for all anatomical sites (thoracic and abdominal). It was trained with a patient-space L1 objective emphasizing the scored high-dose region and multi-scale BEV deep supervision. The submitted CT configuration obtained preliminary-test beamlet MAE 0.0066, image-z IDD distance 0.0025, plan MAE 0.0049, 98.30% gamma pass rate (1%/1 mm), and DVH error 0.460.

Lukas Zimmermann, Hermann Fuchs, Attila Simkó +1

Department of Radiation Oncology, Medical University of Vienna, Vienna, Austria · Christian Doppler Laboratory for Image and Knowledge Driven Precision Radiation Oncology, Medical University of Vienna, Vienna, Austria · Department of Diagnostics and Intervention, Umeå University, Umeå, Sweden

Healthcare Radiotherapy Dose Prediction

Ask PDF

Apr 14, 2026

DoseRAD2026 Challenge dataset: AI accelerated photon and proton dose calculation for radiotherapy

Purpose: Accurate dose calculation is essential in radiotherapy for precise tumor irradiation while sparing healthy tissue. With the growing adoption of MRI-guided and real-time adaptive radiotherapy, fast and accurate dose calculation on CT and MRI is increasingly needed. The DoseRAD2026 dataset and challenge provide a public benchmark of paired CT and MRI data with beam-level photon and proton Monte Carlo dose distributions for developing and evaluating advanced dose calculation methods. Acquisition and validation methods: The dataset comprises paired CT and MRI from 115 patients (75 training, 40 testing) treated on an MRI-linac for thoracic or abdominal lesions, derived from the SynthRAD2025 dataset. Pre-processing included deformable image registration, air-cavity correction, and resampling. Ground-truth photon (6 MV) and proton dose distributions were computed using open-source Monte Carlo algorithms, yielding 40,500 photon beams and 81,000 proton beamlets. Data format and usage notes: Data are organized into photon and proton subsets with paired CT-MRI images, beam-level dose distributions, and JSON beam configuration files. Files are provided in compressed MetaImage (.mha) format. The dataset is released under CC BY-NC 4.0, with training data available from April 2026 and the test set withheld until March 2030. Potential applications: The dataset supports benchmarking of fast dose calculation methods, including beam-level dose estimation for photon and proton therapy, MRI-based dose calculation in MRI-guided workflows, and real-time adaptive radiotherapy.

Fan Xiao, Nikolaos Delopoulos, Niklas Wahl +13

Department of Radiation Oncology, LMU University Hospital, LMU Munich, Munich, Germany · Division of Medical Physics in Radiation Oncology, German Cancer Research Center (DKFZ), Heidelberg, Germany · National Center for Radiation Research in Oncology (NCRO), Heidelberg Institute for Radiation Oncology (HIRO), Heidelberg, Germany +15

Radiotherapy Dose Prediction Medical Image Benchmarks

Ask PDF

Aug 10, 2026

DoseBridge: Denoising Diffusion Bridge Model for Dose Prediction in Lung Intensity-Modulated Proton Therapy

Most radiotherapy dose-prediction models use only CT images and anatomical structures, although intensity-modulated proton therapy (IMPT) dose also depends strongly on beam geometry and available clinical datasets are often small. We present DoseBridge, a denoising diffusion bridge model that uses the patient CT as a structured bridge endpoint and encodes plan-specific beam geometry in a spatially aligned beam mask. Multiscale fusion combines CT, target, organ-at-risk, and beam-mask representations with 1.95% additional parameters. DoseBridge was retrospectively evaluated on single-institution CT images and treatment plans from 52 patients with advanced-stage lung cancer treated with 60 Gy in 30 fractions; 42 cases were used for training and 10 for testing. Performance was assessed using image-similarity, dose-volume, and Lyman-Kutcher-Burman normal-tissue complication probability (NTCP) metrics and compared with two deep-learning models. On the test cohort, DoseBridge achieved a mean absolute error of 4.170 Gy, peak signal-to-noise ratio of 23.06 dB, and structural similarity index of 0.798, outperforming both comparison models on these metrics. Clinical target volume D95 differed from the reference dose by 0.62 +/- 1.6 Gy; signed organ-at-risk mean-dose differences ranged from -0.32 to 0.24 Gy, and NTCP differences were -0.40 +/- 2.2 and 0.52 +/- 3.4 percentage points for acute esophagitis and radiation pneumonitis, respectively. Changing only the beam mask redirected predicted low-dose entrance regions while preserving the high-dose target region. To our knowledge, DoseBridge is the first denoising diffusion bridge model for radiotherapy dose prediction. These results support its feasibility as a beam-aware planning prior for lung IMPT, pending evaluation in larger external cohorts.

Zerun Zhang, Xiaoda Cong, Xiangkun Xu +2

Corewell Health William Beaumont University Hospital, Royal Oak, Michigan, USA

Healthcare Radiotherapy Dose Prediction Medical Image Synthesis Diffusion Models

Ask PDF

Abstract

Radiotherapy Dose Prediction

Explore similar work

PyDoseRT Proton: A GPU Pencil-Beam Engine with a Convolutional Residual-Correction Network for Fast Proton Dose Calculation

DoseRAD2026 Challenge dataset: AI accelerated photon and proton dose calculation for radiotherapy

DoseBridge: Denoising Diffusion Bridge Model for Dose Prediction in Lung Intensity-Modulated Proton Therapy