Application of Transfer Learning for Rapid Calibration of Spatially-resolved Diffuse Reflectance Probes for Extraction of Tissue Optical Properties

Abstract

Significance: Treatment planning for light-based therapies including photodynamic therapy requires tissue optical property knowledge. These are recoverable with spatially-resolved diffuse reflectance spectroscopy (DRS), but requires precise source-detector separation (SDS) determination and time-consuming simulations. Aim: An artificial neural network (ANN) was created to map from DRS at short SDS to optical properties. Transfer learning was used to adapt this trained ANN to fiber-optic probes with varying SDS. Approach: One fiber-optic probe was used to generate an ANN mapping from measurements to Monte Carlo simulation to optical properties. A second probe with different SDS was used for transfer learning algorithm creation. Data from a third were used to test this algorithm. Results: The initial ANN recovered absorber concentration with RMSE = 0.29 uM (7.5% mean error) and us' at 665 nm (us,665') with RMSE = 0.77 1/cm (2.5% mean error). For a second probe, transfer learning significantly improved absorber concentration (RMSE = 0.38 vs. 1.67 uM, p=0.0005) and us,665' (0.71 vs. 1.8 1/cm, p=0.0005) recovery. A third probe also showed improved absorber (0.7 vs. 4.1 uM, p<0.0001) and us,665' (1.68 vs. 2.08 1/cm, p=0.2) recovery. Conclusions: A data-driven approach to optical property extraction can be used to rapidly calibrate new fiber-optic probes with varying SDS, with as few as three calibration spectra.