A prediction model using four non-radiological variables (age, pack-years, fine crackles, elevated serum surfactant protein-D) was developed and temporally validated to identify fibrotic interstitial lung abnormalities (ILAs) on CT, in a multicenter prospective Japanese cohort (derivation n=164, 2022–2023; validation n=207, 2023–2024), with correlation to 24-month functional decline and treatment initiation within 3 years.
The four-variable model achieved an AUC of 0.79 (derivation) and 0.76 (validation; pooled AUC 0.76, 95% CI 0.68–0.82) with good calibration and minimal between-center heterogeneity; fibrotic ILAs comprised 82% of cases, and ~60% subsequently met ATS-defined ILD criteria. A simplified integer score performed similarly (AUC 0.79/0.75) and correlated with functional decline (%FVC ≥10% or %DLco ≥15%) and treatment initiation.
The cohort is from a single country (Japan), limiting generalizability to other ethnic groups. The validation was temporal rather than fully independent, and the study relied on a referral-based multicenter cohort, which may not reflect community prevalence of ILAs.
Clinicians can use a simple four-variable bedside score (age, smoking pack-years, fine crackles, serum SP-D) to identify patients with fibrotic ILAs at risk of progression, supporting earlier risk stratification without requiring radiological subspecialty input. Patients flagged as high-risk should be considered for closer pulmonary function monitoring and earlier treatment discussion.