Quantitative model for imaging single fiber reflectance spectroscopy.

Single-fiber reflectance (SFR) spectroscopy enables quantitative retrieval of tissue optical properties from highly localized measurements through a single multimode fiber, but its diagnostic yield can be limited by small sampling volumes and probe-pressure artifacts. Imaging SFR (iSFR) mitigates these issues by enabling contact-free, point-scanning spectral imaging using optics to project the fiber’s illumination–collection cone onto the tissue surface. We present a quantitative model tailored to iSFR that supports inverse retrieval of tissue absorption and scattering properties from reflectance spectra. The model is based on large-scale Monte Carlo simulations using a generalizable reflectance computation that replaces explicit simulation of arbitrary finite source–detector configurations with a distance-based probability weighting, substantially reducing computational cost. Across broad optical and geometric ranges, the iSFR model predicts reflectance with a median error of 6.2% (SFR: 4.2%). In inverse mode, using wavelength parameterizations for optical properties, the approach recovered absorption and scattering coefficients to within approximately 10% over a wide range of added noise for simulated spectra from two simplified tissue models, serving as a proof-of-concept. These results establish a quantitative model and an efficient computational pipeline for (i)SFR spectroscopy.