Predictive Clinical Diagnosis of Rheumatoid Arthritis Flares Using Non-Invasive Infra-red Thermal Imaging and an AI/ML Algorithm

Overview

The hypothesis for this clinical research project is that the severity of RA may be detected and predicted using an optimized ML/AI algorithm that uses infrared thermal images of inflamed joints and standard clinical RA-related markers (i.e., ESR and CRP) by computing DAS-28 ESR scores in real-time. The infrared thermal images coupled with clinical laboratory markers and the ML/AI algorithm are expected to assist a practicing clinician in the RA diagnosis and the prediction of the occurrence of flares in RA patients. Physicians who use this technology, would need minimum training and will be able to accurately and reliably diagnose RA using a cheaper method which does not involve incident radiation emitted by other imaging modalities such a X-RAY, musculoskeletal (MSK) ultrasound, or a magnetic resonance imaging (MRI). The aim would be to have the Infrared thermal imaging devices at remote VA clinics that do not have a rheumatology specialist where veterans can go for their inflammatory arthritis flare and get this image by the local VA RN. These clinical results can then be assessed by and discussed with a Rheumatologist via telehealth visits.

Objective #1: To assess the clinical feasibility of implementing a novel, physician assisting, diagnostic approach for RA when compared to conventional RA examination and diagnostic procedures. This prospective, non-interventional study will assess clinically assess and diagnose the severity of RA in sero-positive RA patients experiencing active flares by using conventional examination and diagnostic methods, and compared those with a physician-assisting, diagnostic approach that involves the use of an infrared thermal imaging device, which detects heat waves to be correlated between RA patients and control subjects (i.e., those who do not have RA or who are in remission). Standard clinical laboratory values will be documented from the EHR system as well and will include ESR (sedimentation rate) and CRP (c-reactive protein). Objective #2: To develop and optimize a ML/Artificial intelligence(AI) algorithm that would process and analyze thermal images and assist in the predictive diagnosis of RA using the DAS-28 ESR score for those thermal images of the inflamed joints of patients. This study will predict the probability of an actual flare occurrence and its severity in RA patients by using an optimized, physician assisting ML/AI algorithm that processes and analyzes thermal images from sero-positive RA patients in pain and experiencing flares and that calculates the DAS-28 scoring system in real-time