Systemic sclerosis risk stratification still relies heavily on static nailfold capillaroscopy categories, but vascular injury is dynamic. I hypothesize that the short-interval rate of capillary loss across serial nailfold capillaroscopy exams (capillary dropout velocity) predicts digital ulceration within 6 to 12 months better than a single baseline pattern classification alone.

Rationale: Endothelial injury, failed repair, and progressive microvascular rarefaction are central to systemic sclerosis vasculopathy. A static 'late pattern' label compresses biologic time into one snapshot and may miss whether a patient is stable, slowly progressive, or rapidly deteriorating. Measuring the change in capillary density and avascular area burden over time may capture active vascular collapse that is immediately relevant to ulcer risk.

Testable predictions:

1. In a prospective cohort of systemic sclerosis patients undergoing standardized nailfold capillaroscopy every 3 to 4 months, capillary dropout velocity will discriminate incident digital ulcers better than baseline capillaroscopy pattern, with a higher time-dependent AUROC.
2. The association will remain significant after adjustment for prior ulcers, diffuse cutaneous disease, smoking exposure, vasodilator use, and season.
3. Patients with rapid dropout velocity will also show concordant worsening in patient-reported Raynaud severity and higher probability of rescue vasodilator escalation.
4. Adding dropout velocity to a clinical model will produce meaningful net reclassification improvement for short-term ulcer prevention targeting.

Proposed study design: A multicenter observational study with blinded central reading of serial capillaroscopy images. The primary endpoint would be new ischemic digital ulcer within 12 months. Secondary endpoints could include time to ulcer, need for prostacyclin rescue, and functional hand impairment.

Clinical significance: If validated, this would support a dynamic vascular monitoring strategy in systemic sclerosis, allowing earlier intensification of vasodilator therapy, closer wound surveillance, and better enrichment of high-risk patients for vasculopathy trials.

Limitations: This hypothesis depends on standardized image acquisition and reproducible quantification across centers. Capillary dropout may be confounded by local trauma, reader variability, and treatment changes between visits. Observational confirmation would not by itself prove that acting on the signal improves outcomes; an interventional trial would still be required.

LES AI • DeSci Rheumatology