FT-IR saliva analysis for the diagnosis of psoriasis: A pilot study

Psoriasis is a chronic, autoimmune disease with multiple interplaying risk factors. Saliva has gained growing interest as an excellent biological fluid exhibiting a strong diagnostic potential in dermopathies. Saliva profiling through Fourier Transform Infrared Spectroscopy in attenuated Total Reflection (FT-IR ATR) was investigated for the diagnosis of psoriasis. Particularly, multivariate analysis was carried out after a suitable pre-processing, applying unsupervised principal component analysis (PCA) for feature extraction in the Amide I/II, Thiocyanate and within Thiocyanate and bio fingerprint bands. Further, linear discriminant analysis (LDA) and support vector machine (SVM) were trained to establish discrimination models between psoriatic subjects and healthy controls. PCA-LDA evidenced a classification performance in the bio fingerprint region (2150-900 cm(-1)) of 93.75% accuracy, and a sensitivity and specificity of 87.5% if compared to SVM (87.5% accuracy, with a sensitivity and specificity of 75%). Saliva profiling and multivariate analysis provide a powerful approach in diagnosis and follow-up of inflammatory dermatopathies. FT-IR saliva profiling, signal processing and machine learning algorithms evidenced the possibility of automatic classification of psoriatic patients, with a potentially interesting insight in mass screening and preliminary diagnosis.

Automated summary

This study investigates the diagnosis of psoriasis through saliva profiling using Fourier Transform Infrared Spectroscopy and machine learning algorithms for automatic classification. The results demonstrate the potential of saliva analysis and multivariate analysis in the diagnosis and follow-up of inflammatory dermatopathies.