Raman spectroscopy in noninvasive optical diagnostics of skin neoplasms

Abstract

Aim to propose a schematic design of a new prototype having a potential for mass production and use in medical institutions for routine diagnostics of skin neoplasms. Material and methods. The study included more than 600 patients with various skin neoplasms. We studied the samples of various skin neoplasms obtained after surgical treatment in the departments of the Samara Regional Clinical Oncology Center. For this purpose, we used the proposed original experimental device for Raman laser spectroscopy. The received signal was processed with a designed algorithm using artificial intelligence. Results. The proposed diagnostic method reached the 89% sensitivity and 93% specificity proving to be effective. As a result, we have substantiated the technological possibility of creating a portable cost-saving spectroscopic technique with a neural network classifier. The device allows to perform preliminary diagnostics without the involvement of a specialist doctor and is devoid of subjective analysis criteria.