Intelligent Bi-Dimensional Skin Biopsies of Rheumatoid Arthritis Based on Raman Spectral Imaging and Machine Learning

Rheumatoid arthritis (RA) is one of the most common autoimmune diseases worldwide, characterized by its progressive and irreversible nature. Early diagnosis is crucial for delaying disease progression and optimizing treatment strategies. Existing diagnostic methods face limitations in asymptomatic screening and often rely on subjective judgment by experienced rheumatologists, restricting their application in early screening and clinical diagnosis. To address these challenges, we developed an innovative approach for intelligent bidimensional skin biopsies, employing Raman spectroscopy for direct spectral scanning and imaging of affected joint skin. This method enables preliminary RA diagnosis after a brief skin surface scan. It generates high-resolution three-dimensional Raman images of the affected skin within 13 min, providing rapid and reliable diagnostic support. Furthermore, Raman data are analyzed and classified using multiple artificial intelligence algorithms, such as naive Bayes, linear discriminant analysis, decision trees, k-nearest neighbors, random forests, and support vector machines, achieving high-accuracy RA differentiation. The design significantly enhances diagnostic precision and speed, enabling nonspecialists to accurately diagnose RA. Extensive experimental data validated the method's 100% diagnostic accuracy. This approach provides a novel and effective tool for early RA screening and demonstrates potential applications in other autoimmune and dermatological diseases.