Three‐Dimensional Phase‐Contrast MR Angiography at 5 T Assesses Warming‐Associated Changes in the Visible Hand Vascular Network in Systemic Sclerosis

Background Thermal challenge can reveal vascular abnormalities in systemic sclerosis (SSc), but current assessments primarily measure surface temperature or cutaneous perfusion. Warming‐associated changes in the visible three‐dimensional hand vascular network in SSc remain incompletely characterized. Purpose To quantify warming‐associated changes and group differences in hand‐size‐normalized metrics of visible hand vessels using 5‐T three‐dimensional phase‐contrast MR angiography (3D PC‐MRA), and to assess repeatability and reader agreement. Study Type Prospective case–control study. Population Ninety participants (60 SSc, mean age: 44.00 ± 10.73 years, 55 females; 30 healthy controls, mean age: 42.83 ± 13.24 years old, 28 females). Twenty controls (18 females) underwent baseline scan–rescan imaging; reader agreement was assessed in the 60 baseline examinations. Field Strength/Sequence 5 T; noncontrast‐enhanced 3D gradient‐echo PC‐MRA. Assessment Hands were imaged before and after 5‐min immersion in 40°C water. Hand‐size‐normalized metrics were vessel volume ( nVol ), centerline length ( nLen ), and apparent vessel radius ( nRad ); changes (Δ) were post‐immersion minus baseline. Statistical Tests Paired t ‐tests or Wilcoxon signed‐rank tests; Welch t ‐tests; multivariable linear regression adjusted for age, sex, body mass index, and baseline values; intraclass correlation coefficients (ICCs); and Bland–Altman analysis. Significance was q < 0.05 after Benjamini–Hochberg correction and p < 0.05 otherwise. Results Normalized vessel volume and normalized centerline length increased significantly in both groups, with mean changes of 0.76 × 10 −3 and 1.91 in controls and 1.80 × 10 −3 and 4.80 in SSc, respectively. Adjusted increases were significantly greater in SSc for normalized vessel volume (between‐group difference, 1.11 × 10 −3 ; 95% CI, [0.36–1.85] × 10 −3 ) and normalized centerline length (difference, 1.97; 95% CI, 0.36–3.59). Changes in normalized apparent vessel radius did not differ after adjustment (difference, −0.001 × 10 −3 ; q = 0.996). Scan–rescan and reader‐agreement ICCs were 0.948–0.994 and 0.941–0.993, respectively. Data Conclusion At 5 T, 3D PC‐MRA detected greater warming‐associated increases in nVol and nLen in SSc than controls. Baseline measurements showed good‐to‐excellent scan–rescan repeatability in controls and excellent reader agreement. Level of Evidence 2. Technical Efficacy Stage 3.