Fiche publication


Date publication

septembre 2023

Journal

Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy

Auteurs

Membres identifiés du Cancéropôle Est :
Pr BRONOWICKI Jean-Pierre , Pr SOCKALINGUM Ganesh , Pr THIEFIN Gérard


Tous les auteurs :
Thiéfin G, Bertrand D, Untereiner V, Garnotel R, Bronowicki JP, Sockalingum GD

Résumé

Assessment of liver fibrosis is crucial to guide the therapeutic strategy in patients with chronic liver disease. We investigated the potential of serum Fourier transform infrared (FTIR) spectroscopy for assessing the degree of hepatic fibrosis in patients with chronic hepatitis C (CHC). The study was conducted on dried serum samples from 94 CHC patients at different histological stages of hepatic fibrosis: METAVIR F0 (n = 20), F1 (n = 17), F2 (n = 20), F3 (n = 20) and F4 (n = 17). Transmission FTIR spectra were acquired in the 4000-400 cm range. Wavenumbers were selected by genetic algorithm (GA) according to their diagnostic performance as assessed by a partial least squares discriminant analysis (PLS-DA) model using a training and a validation set to differentiate severe stages of fibrosis from mild or moderate ones. The GA procedure was applied 50 times on randomly selected sets. Furthermore, the best set of wavenumbers was re-tested in 1000 randomly selected validation sets. Wavenumbers selected by GA corresponded to functional groups present in lipids, proteins, and carbohydrates. This model allowed to identify patients with cirrhosis (METAVIR F4), patients with advanced fibrosis (METAVIR F3 and F4), and patients with significant fibrosis (METAVIR F2, F3 and F4), with AUROC (Area Under the Receiver Operating Characteristic) of 0.88, 0.85 and 0.85, respectively. Thus, serum FTIR spectroscopy appears to have a strong potential as a new diagnostic tool for assessing the degree of fibrosis in patients with chronic liver disease.

Mots clés

Genetic algorithm, Humans, Infrared spectroscopy, Liver fibrosis, PLS-DA model, Serum

Référence

Spectrochim Acta A Mol Biomol Spectrosc. 2023 09 25;305:123433