ABSTRACT:
Purpose: To compare the diagnostic performance of the American Thoracic Society, Japanese Respiratory Society, and Asociación Latinoamericana del Tórax (ATS/JRS/ALAT) versus the American College of Chest Physicians (ACCP) imaging classifications for hypersensitivity pneumonitis (HP).
Materials and Methods: Patients in the institutional review board–approved Interstitial Lung Disease (ILD) registry referred for multidisciplinary discussion (MDD) at the authors’ institution (January 1, 2006–April 1, 2021) were included in this retrospective study when ILD was diagnosed at MDD. MDD diagnoses included HP, connective tissue disease–ILD, and idiopathic pulmonary fibrosis. Retrospective review of thin-section CT images was performed in consensus by two cardiothoracic radiologists blinded to the diagnosis. Diagnostic patterns were determined for thin-section CT images using both classifications. Discordance rates were determined. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were assessed using MDD diagnosis as the reference standard.
Results: A total of 297 patients were included in the study: 200 (67%) with HP, 49 (16%) with connective tissue disease–ILD, and 48 (16%) with idiopathic pulmonary fibrosis at MDD. The discordance rate between the two classifications was 21%. Assuming low HP prevalence (10%), ATS/JRS/ALAT classification outperformed ACCP classification, with greater accuracy (92.3% vs 87.6%) and greater positive predictive value (60.7% vs 42.9%). Assuming high prevalence (50%), accuracy and negative predictive value were superior using ACCP classification (81.7% vs 79.7% and 77.7% vs 72.6%, respectively), and positive predictive value was superior using ATS/JRS/ALAT classification (93.3% vs 87.1%).
Conclusion: Accuracy of the ATS/JRS/ALAT and ACCP HP classifications was greater in settings with low and high HP prevalence, respectively. Diagnostic performance of both classifications was discordant in a minority of cases.