AI4RA: AI-powered visual technology to diagnose rheumatoid arthritis activity

The AI4RA project will enable the secondment of Dr Wan Rusli from Imperial College to Arthronica. The objective of the secondment is to employ Dr Rusli's skills in biomechanics and computational modelling to evaluate Arthronica's ability to provide remote monitoring and determine disease activity status for patients with rheumatoid arthritis (RA).

RA is a chronic, disabling autoimmune condition in which the body attacks the cells that line the joints, making the joints swollen, stiff and painful; over time this can also damage the cartilage and nearby bone. The National Audit Office estimates that approximately 580,000 adults in England currently have the disease, with a further 26,000 new cases diagnosed each year.

There is growing evidence to suggest that treatment within 12 weeks is associated with improved response to treatment and patient outcomes. This is further supported by a number of studies that have shown that the best clinical outcomes are achieved through a treat-to-target approach. This requires that patients receive at least bi-monthly follow-ups in order to determine if their disease is active or the treatment has successfully dampened disease activity.

The leading metric to assess disease activity in RA is the Disease Activity Score using 28 joints (DAS-28). The DAS-28 is routinely measured in clinic visits for patients with RA. It involves four domains: a clinician-reported swollen joint count, a clinician-reported tender joint count, a patient global measure of symptoms, and a biomarker of inflammation from a blood test.

In the era of the coronavirus pandemic, it is overtly apparent that we need innovative solutions that enable measurement of disease activity remotely, but still reliably. Biomarker levels can be obtained, as home testing kits are available. Patient symptom severity scores are easily captured.

Joint tenderness can be learnt through patient self-assessment. What is needed is a mechanism for remotely recording the status of the joints, including swelling and range of motion.

To address these needs, Arthronica is engaged in a cross-sectional diagnostic accuracy study with the Rheumatology Department at King's College Hospital to record images of and measure the swollen joints. However, the company is lacking in-depth expertise in biomechanical modelling. Dr Rusli has extensive experience in computational modelling and musculoskeletal biomechanics and will support Arthronica researchers in the validation of the RA biomechanical model to increase the technology readiness for remotely assessing RA disease activity.