Fast Occupational Upper-Limb Radiation Dose Prediction Using Machine Learning and Monte Carlo Simulation

OBJECTIVE: Interventional procedures expose physicians to scattered radiation, particularly to their upper extremities, posing occupational health risks. Existing extremity dosimeters such as TLDs, OSL rings and active personal dosimeters provide limited spatial information, exhibit angular and energy dependence and offer little or no real-time feedback. This study develops machine learning models to estimate radiation dose values at discrete upper-limb locations using Monte Carlo-derived data and procedurespecific parameters. APPROACH: A dataset of 10,000 Monte Carlo dose maps was generated under varied clinical and geometric conditions. After log-transformation and normalization, several machine learning models, including deep neural networks and treebased regressor, were trained and assessed using five-fold cross-validation. Mean absolute error and relative error were evaluated on the original dose scale, and an ensemble of the three best-performing models was constructed to improve robustness. MAIN RESULTS: The ensemble consistently outperformed individual models, achieving an average relative error of 3.69 % and demonstrating stable performance across anatomical regions and dose levels. Highest accuracy was obtained for standard beam geometries, whereas larger discrepancies occurred in extreme configurations with steep dose gradients. Predicted dose patterns were consistent with the expected distributions across the upper-limb regions. SIGNIFICANCE: The findings demonstrate the feasibility of machine learning-based extremity dose estimation in interventional environments. The proposed ensemble provides a rapid (around 10 ms), scalable alternative to full Monte Carlo simulations, enabling near real-time predictions of upper-limb occupational dose and supporting optimization of radiation protection practices in image-guided procedures.