Quantitative Digital Image Analysis for Assessing Cutaneous Fibrosis Correlates With Histopathological Scoring in Patients With Systemic Sclerosis

Abstract

Systemic sclerosis (SSc) is a multisystem autoimmune disease characterized by progressive fibrosis of the skin, blood vessels, and internal organs. Accurate assessment of skin fibrosis is essential for disease monitoring and treatment evaluation, yet reliable quantitative methods are lacking. We compared computer-assisted digital image analysis with traditional semiquantitative histopathological scoring. Patients with SSc and healthy controls were identified from a hospital database. Clinical data, modified Rodnan skin score, and hematoxylin and eosin–stained skin biopsies were collected. Biopsies were scored using modified Keyser and Farge criteria. Whole-slide images were analyzed using QuPath open source digital pathology software (v0.5.1) for area annotation, normalized cellularity, and collagen density, whereas collagen alignment was evaluated using CurveAlign (MATLAB) software. Seven SSc and 7 healthy controls (no significant demographic differences) were studied. SSc samples showed significant differences in hyalinized collagen (P < .001), dermal fibroblast cellularity (P = .009), and total histopathological score (P = .007). Quantitative analysis confirmed decreased dermal cellularity (P = .024), increased collagen density (P = .009), and higher collagen alignment, with reduced fiber orientation variability, reflecting extracellular matrix restructuring in SSc. Tools such as QuPath and CurveAlign improve objectivity and reproducibility in SSc skin assessment, correlating well with traditional scores. These findings support the integration of computer-aided quantitative analysis into clinical practice and trials, advancing personalized fibrosis evaluation. Larger, longitudinal studies are needed for further validation.