Soft tissue tumor size prediction using precise fiber-optic Raman probes: in silico investigations

Raman spectroscopy is evolving as a powerful optical tool for tissue discrimination in cancer diagnosis. In addition to diagnosis during biopsy and intraoperative procedures, there is also a compelling need to gain insight on the tumor location and geometry like depth, thickness, and size of subsurface soft tissue tumors for diagnosis and monitoring/surgery or treatment-related decision making. Spatially offset Raman spectroscopy (SORS) offers excellent potential in extracting the depth of the tumor by selectively collecting the photons that have traveled a longer path length using source-detector separation. The current work utilizes the property of SORS to estimate the varying size of an ellipsoid tumor resembling a flat adenoma located at various subsurface depths and tumor thicknesses. The prediction model is based on the continuous incremental movement of the probe position on the surface to analyze the change in relative tumor contribution calculated for the Raman scattered signal. Regression on Monte Carlo (MC) simulated SORS signals predicted the tumor dimension with a coefficient of determination of 0.94 and root mean square error less than 2%. This prediction model will help design probes for optimized application-specific SORS signal extraction and processing.