Background The modified Rodnan skin score (mRSS), used to measure dermal thickness in patients with systemic sclerosis (SSc), is agnostic to inflammation and vasculopathy. Previously, we demonstrated the potential of neural network-based digital pathology applied to stained skin biopsies from SSc patients as a quantitative outcome. We leveraged deep learning and histologic analyses of clinical trial biopsies to decipher SSc skin features 'seen' by artificial intelligence (AI). Methods Adults with diffuse cutaneous SSc (disease duration ≤ 6 years) enrolled in an open-label trial evaluating belumosudil underwent serial mRSS assessment and dorsal arm biopsies at week 0, 24 and 52/end of trial. Two blinded dermatopathologists independently scored stained sections [Masson's trichrome, hematoxylin and eosin (H&E), CD3, CD34, CD8, α smooth muscle actin (αSMA)] for 16 published SSc dermal pathological parameters. We applied our previously published deep learning model to generate QIF signatures/biopsy and generated Fibrosis Scores. Associations between Fibrosis Score and mRSS (Spearman correlation); and between Fibrosis Score mRSS versus histologic parameters [odds ratios (OR)] were determined. Results Only ten patients were enrolled because the sponsor terminated the trial early. Median, interquartile range (IQR) for mRSS change (0-52 weeks) for the five participants with paired biopsies was - 2.5 (-11-7.5), and for the ten participants was - 2 (-9-7.5). The correlation between Fibrosis Score and mRSS was R = 0.3; p = 0.674. Per 1-unit mRSS change (0-52W), histologic parameters with the greatest associated changes were (OR, p-value): telangiectasia (2.01, 0.001), perivascular CD3+ (1.03, 0.015), and % of CD8 + among CD3+ (1.08, 0.031). Likewise, per 1-unit Fibrosis Score change, parameters with greatest changes were (OR, p-value): hyalinized collagen (1.1, < 0.001), subcutaneous (SC) fat loss (1.47, < 0.001), thickened intima (1.21, 0.005), and eccrine entrapment (1.14, 0.046). Conclusions Belumosudil was associated with a non-clinically meaningful improvement in mRSS. Fibrosis Score changes correlated with histologic feature changes ( e.g. , hyalinized collagen, SC fat loss) that were distinct from those associated with mRSS changes ( e.g. , telangiectasia, perivascular CD3+, and % of CD8 + among CD3+). These data suggest that AI applied to SSc biopsies may be useful for quantifying pathologic features of SSc beyond skin thickness.