This retrospective two-centre cohort study evaluated whether the CT-based Single Time Point Prediction (STP) score — alongside quantitative CT metrics (QLF, QGG, QILD, QNL) — could predict disease progression in 245 non-IPF fibrosing ILD patients over a median follow-up of 20.4 months.
55.9% of patients (137/245) progressed. After adjustment, QGG ≥10% (HR 1.71, p=0.024), QILD ≥30% (HR 1.73, p=0.006), and STP ≥30% (HR 1.58, p=0.020) were all independent predictors of disease progression; higher QILD/QGG in STP-positive regions increased risk, while higher QNL in STP-negative regions was protective.
Retrospective design limits causal inference; single-institution automation may not generalize; the study excludes IPF, so findings apply only to non-IPF ILD subtypes.
In non-IPF fibrosing ILD, CT-based scores (STP ≥30%, QILD ≥30%, QGG ≥10%) independently flag patients at higher risk for progression — consider integrating these automated quantitative CT metrics into baseline risk stratification alongside FVC and DLCO.
Explore related topics
