标题：Detection Method of Falsified Medicines by Using a Low-Cost Raman Scattering Spectrometer Combined with Soft Independent Modeling of Class Analogy and Partial Least Squares Discriminant Analysis
摘要：There are many reports of falsified medicines that may cause harm to patients. A rapid and simple method of identifying falsified medicines that could be used in the field is required. Although Raman scattering spectroscopy has become popular as a non-destructive analysis, few validation experiments on falsified medicines that are actually distributed on the market have been conducted. In this study, we validated a discriminant analysis using an ultra-compact, portable, and low-cost Raman scattering spectrometer combined with multivariate analysis. The medicines were three types of erectile dysfunction therapeutic tablet and one type of antifungal tablet: tadalafil (Cialis), vardenafil hydrochloride (Levitra), sildenafil citrate (Viagra), and fluconazole (Diflucan), which is sometimes advertised as female Viagra. For each medicine, the authentic standard product and products obtained by personal import via the internet (genuine or falsified) were used. Discriminant analyses were performed on the Raman spectra combined with soft independent modeling of class analogy (SIMCA) and partial least squares discriminant analysis (PLS-DA). It was possible to identify all falsified samples by SIMCA using the standard product model for all four products. Using the PLS-DA using the PLS models of the four standard products, falsified Levitra and Diflucan samples were classified correctly, although some falsified Cialis and all Viagra samples also belonged to the standard class. In this study, SIMCA might be more suitable than PLS-DA for identifying falsified medicines. A spectroscopic module that combines the low-cost Raman scattering spectroscopy with SIMCA might contribute to the rapid identification of falsified medicines in the field.
关键词：discriminant analysis;falsified medicine;Raman scattering spectroscopy;low-cost analyzer;soft independent modeling of class analogy;partial least squares discriminant analysis
