Abstract
SIGNIFICANCE: Raman spectroscopy has emerged as a promising technique for a variety of biomedical applications. The unique ability to provide molecular specific information offers insight to the underlying biochemical changes that result in disease states such as cancer. However, one of the hurdles to successful clinical translation is a lack of international standards for calibration and performance assessment of modern Raman systems used to interrogate biological tissue. AIM: To facilitate progress in the clinical translation of Raman-based devices and assist the scientific community in reaching a consensus regarding best practices for performance testing. APPROACH: We reviewed the current literature and available standards documents to identify methods commonly used for bench testing of Raman devices (e.g., relative intensity correction, wavenumber calibration, noise, resolution, and sensitivity). Additionally, a novel 3D-printed turbid phantom was used to assess depth sensitivity. These approaches were implemented on three fiberoptic-probe-based Raman systems with different technical specifications. RESULTS: While traditional approaches demonstrated fundamental differences due to detectors, spectrometers, and data processing routines, results from the turbid phantom illustrated the impact of illumination-collection geometry on measurement quality. CONCLUSIONS: Specifications alone are necessary but not sufficient to predict in vivo performance, highlighting the need for phantom-based test methods in the standardized evaluation of Raman devices.