Identification of different tumor states in nasopharyngeal cancer using surface-enhanced Raman spectroscopy combined with Lasso-PLS-DA algorithm

Abstract

Identification of different states in cancer is of vital importance for cancer treatment and management. A powerful diagnostic algorithm based on Lasso-partial least squares-discriminant analysis (Lasso-PLS-DA) was developed here for improving blood surface-enhanced Raman spectroscopy (SERS) analysis, with the aim to classify different states in nasopharyngeal cancer (NPC). A total of 160 blood plasma samples were collected for this study, obtained from 60 normal volunteers, 25 T1 stage cancer and 75 T2–T4 stages cancer patients. Results show that a diagnostic sensitivity of 68% and a specificity of 84.0% can be achieved for separating T2–T4 stage from T1 stage cancer, which had a 20% improvement in diagnostic specificity compared with the previous work. This exploratory study demonstrates that the Lasso-PLS-DA can be integrated with blood SERS analysis as a promising clinical complement for different T stages detection in NPC.