Diagnosis of Rheumatic and Musculoskeletal Diseases from Synovial Fluid Using SERS and Artificial Intelligence

BioSpecT (Translational BioSpectroscopy), EA 7506

Our team is specialized in the development of vibrational spectroscopy techniques (infrared absorption and Raman scattering) for the characterization of biological samples (cell, tissue, biofluid). These techniques allow the identification of new diagnostic biomarkers for the detection of cancer lesions or to highlight molecular alterations associated to tissue aging (skin). Spectral data exploitation is based on the implementation of multivariate statistical algorithms developed by the team. Our team has a real and recognized expertise in biological analysis by vibrational spectrocopy.

Rheumatic and musculoskeletal diseases affect a quarter of the population in the European Union—over 120 million people. They are the leading cause of sick leave and premature retirement worldwide. As a result, these diseases impose a significant economic burden on global healthcare systems, with public spending in Europe exceeding €200 billion per year.

Current diagnostic methods typically fail to detect these diseases until they reach advanced stages, at which point joint damage may already have occurred. Recently, we developed a novel method for the early diagnosis of joint disorders based on the analysis of synovial fluid (SF) using Surface-Enhanced Raman Spectroscopy (SERS). SERS spectra were acquired from the synovial fluid of patients with chondrocalcinosis (n=13), gonarthrosis (n=18), rheumatoid arthritis (n=14), and gout (n=5).

However, the complex multidimensional nature of SERS spectra necessitates the application of artificial intelligence (AI) tools to differentiate between these diseases. The objectives of this internship are twofold:

i) to implement supervised AI classification tools to accurately diagnose diseases based on SERS spectra obtained from synovial fluid;

ii) to expand the dataset by acquiring additional SERS spectra from synovial fluid samples.

This internship is part of a collaboration between the BioSpecT unit (Université de Reims Champagne-Ardenne, Reims, France) and the LCBPT (Université Paris Cité, Paris, France).

The candidate must have strong skills in programming, algorithms, and data science. Proficiency in Python is essential. The internship will take place in Reims, but the candidate must be mobile as trips to Paris will be necessary for the successful completion of the internship. Additionally, the candidate should be curious and open to biophysics, chemistry, and biology, as they will be required to acquire SERS spectra from synovial fluid samples.