Critical assessment of Raman spectroscopy for clinical diagnostics: progress, limitations, and barriers to translation

Raman spectroscopy (RS) has fascinated as a molecular diagnostic tool because of its potential to deliver label-free biochemical analysis of biological tissues and fluids. A number of proof-of-concept and early clinical studies have been conducted to explore its potential in a variety of medical applications, including cancer, infectious diseases, and metabolic disorders. Nevertheless, despite the amount of research activity, the clinical application of RS is currently very limited. This review critically evaluates the current evidence base for the clinical application of RS in diagnostic applications, focusing on the accuracy, variability, and translational potential of the technology. Rather than offering a general descriptive summary, we will critically assess the current status of the most advanced RS technologies, including spontaneous RS and surface-enhanced Raman spectroscopy (SERS), in relation to clinically relevant applications. Particular emphasis will be placed on the development, sample handling, validation, and use of chemometric and machine learning analysis, with special attention to the risks of bias, overfitting, and inter-laboratory variability. It is also concerned with the technical aspects that are at present impeding its adoption in the clinical setting, such as signal stability, standardization of procedures, cost of the equipment, regulatory aspects, and compatibility with the present clinical practice. If possible, the effectiveness of RS-based methods will be compared with that of the present diagnostic methods to give a clear view of its potential and limitations.