High predictive values of RBC membrane-based diagnostics by biophotonics in an integrated approach for Autism Spectrum Disorders

Abstract

Membranes attract attention in medicine, concerning lipidome composition and fatty acid correlation with neurological diseases. Hyperspectral dark field microscopy (HDFM), a biophotonic imaging using reflectance spectra, provides accurate characterization of healthy adult RBC identifying a library of 8 spectral end-members. Here we report hyperspectral RBC imaging in children affected by Autism Spectrum Disorder (ASD) (n = 21) compared to healthy age-matched subjects (n = 20), investigating if statistically significant differences in their HDFM spectra exist, that can comprehensively map a membrane impairment involved in disease. A significant difference concerning one end-member (spectrum 4) was found (P value = 0.0021). A thorough statistical treatment evidenced: i) diagnostic performance by the receiving operators curve (ROC) analysis, with cut-offs and very high predictive values (P value = 0.0008) of spectrum 4 for identifying disease; ii) significant correlations of spectrum 4 with clinical parameters and with the RBC membrane deficit of the omega-3 docosahexaenoic acid (DHA) in ASD patients; iii) by principal component analysis, very high affinity values of spectrum 4 to the factor that combines behavioural parameters and the variable "cc" discriminating cases and controls. These results foresee the use of biophotonic methodologies in ASD diagnostic panels combining with molecular elements for a correct neuronal growth.