NIHR i4i 2: Early Diagnosis of Systemic Sclerosis in the General Rheumatology Clinic - Pilot Image Acquisition Study

Systemic sclerosis (SSc) is a painful, disabling disease affecting around 20,000 people in the UK. Symptoms include painful ulcers on the fingers and toes, sometimes requiring amputation, and life-threatening internal organ damage. The first symptom of SSc is usually Raynaud's phenomenon - white, painful fingers in the cold - though around 5% of otherwise healthy individuals also experience Raynaud's. The National Institute for Health and Care Excellence (NICE) recommends that GPs should refer patients with Raynaud's to a general rheumatology clinic for assessment. When a patient is referred for assessment, using nailfold capillaroscopy to find vessel abnormalities caused by SSc is recognised internationally as key to early diagnosis. Despite this, it is not used in most rheumatology clinics, because it requires expensive equipment and specialised skill. When capillaroscopy is not available, diagnosis is often delayed, which can have serious consequences for patients.

The research team will develop an artificial intelligence (AI) system that helps rheumatology clinics diagnose SSc as early as possible. The system will use nailfold capillaroscopy images of the small blood vessels at the base of the fingernails and digital technology to find changes due to disease. In a previous study has shown that, in expert hands, capillaroscopy images good enough to diagnose SSc can be obtained using a commercially available, hand-held microscope. The investigators have also developed AI image analysis software that finds capillary abnormalities very reliably.

This study will establish if target users can acquire diagnostic-quality images in clinic and provide feedback to inform technology refinement. Images will be acquired using version one of our software from 105 patients across seven centres in England. The investigators will also run focus groups and interviews to gain qualitative input from users and patients to assess usability of the system. The results of this study will contribute to the development of a complete system suitable for use by non-specialists.