Folder with all data and algorithms

Spatially Offset Raman Spectroscopy (SORS) has emerged as a potential tool for non-invasive biomedical diagnostics, enabling molecularly specific probing of subsurface tissues. Despite its promise, clinical translation remains constrained by competing demands for signal strength and patient safety under laser illumination. In this study, we present an open-source, Python-based computational framework that unifies photon transport modeling, probe geometry optimization, and photothermal safety assessment into a single workflow. Using Monte Carlo simulations coupled with Pennes’ bioheat and Arrhenius/CEM43 damage models, we evaluate four SORS frameworks: conventional puck-point, ring-collector, inverse SORS (iSORS), and a novel reinforced iSORS (riSORS) on a skin model. Results demonstrate that ring-based illumination strategies substantially mitigate thermal loading, extending safe exposure times by over 50-fold compared to point sources, while simultaneously enabling up to 100× greater accumulation of usable Raman energy prior to tissue damage. Among the tested geometries, riSORS achieved the best balance of subsurface selectivity and photon capture efficiency, outperforming conventional approaches in both spectroscopic yield and safety margins. The framework also supports sensitivity analyses across tissue optical properties, illustrating its utility for assessing inter-patient variability. While the current implementation assumes simplified optical and thermal parameters and requires experimental validation, it provides a quantitative roadmap for designing clinically viable SORS probes. By explicitly linking geometric design to damage-limited signal collection, this approach bridges the gap between theoretical optimization and translational feasibility, offering a versatile platform to accelerate the safe deployment of SORS in biomedical applications.