Development and Validation of a Raman Spectroscopic Classification Model for Cervical Intraepithelial Neoplasia (CIN).

Abstract

Simple SummaryThere is an unmet clinical need for new methods to aid clinicians in the early detection of cervical cancer and precancer. Spectroscopic methods such as Raman spectroscopy can provide a rapid, label-free and nondestructive measurement of the biochemical fingerprint of cells, tissues and biofluids. This study aims to demonstrate the clinical utility of Raman spectroscopy for the identification of cervical precancer. Raman spectra were recorded from cervical smear samples (n = 662) and a classifier was developed based on histology. A classification accuracy of 91.3% was achieved in an independent blinded test set (n = 69), demonstrating the potential clinical utility of Raman spectroscopy.The mortality associated with cervical cancer can be reduced if detected at the precancer stage, but current methods are limited in terms of subjectivity, cost and time. Optical spectroscopic methods such as Raman spectroscopy can provide a rapid, label-free and nondestructive measurement of the biochemical fingerprint of a cell, tissue or biofluid. Previous studies have shown the potential of Raman spectroscopy for cervical cancer diagnosis, but most were pilot studies with small sample sizes. The aim of this study is to show the clinical utility of Raman spectroscopy for identifying cervical precancer in a large sample set with validation in an independent test set. Liquid-based cervical cytology samples (n = 662) (326 negative, 200 cervical intraepithelial neoplasia (CIN)1 and 136 CIN2+) were obtained as a training set. Raman spectra were recorded from single-cell nuclei and subjected to a partial least squares discriminant analysis (PLSDA). In addition, the PLSDA classification model was validated using a blinded independent test set (n = 69). A classification accuracy of 91.3% was achieved with only six of the blinded samples misclassified. This study showed the potential clinical utility of Raman spectroscopy with a good classification of negative, CIN1 and CIN2+ achieved in an independent test set.