While smoking-related interstitial fibrosis (SRIF) is known to have a benign clinical course, combined pulmonary fibrosis and emphysema (CPFE) is associated with high morbidity and mortality burden.
The goal of our study is to differentiate SRIF (CPFE...
While smoking-related interstitial fibrosis (SRIF) is known to have a benign clinical course, combined pulmonary fibrosis and emphysema (CPFE) is associated with high morbidity and mortality burden.
The goal of our study is to differentiate SRIF (CPFE-histologic SRIF) from CPFE (CPFE-histologic UIP) using chest computed tomography (CT).
This retrospective study included patients who underwent surgical lung biopsy or resection between January 2015 and January 2025; all had pathologic confirmation of SRIF or UIP. CT features assessed included centrilobular and paraseptal emphysema, multiple thin-walled cysts (MTWC), honeycombing, traction bronchiectasis, subpleural ground-glass opacities (GGO)/reticulation, centrilobular ground-glass nodules, and ratio between the diameter of the pulmonary artery and that of the aorta (PA:Ao). CT features were assessed using univariable and multivariable logistic regression analyses. Receiver operating characteristic (ROC) curve analysis was used to quantify the performance of imaging features.
A total of 106 patients with SRIF (n = 32) and CPFE (n = 74) were included. Among the evaluated CT features, MTWC (75% in SRIF, 25.7% in CPFE) and centrilobular ground-glass nodules (46.9% in SRIF, 32.4% in CPFE) were significantly more prevalent in SRIF than in CPFE, whereas honeycombing (52.7% in CPFE, 6.3% in SRIF) and PA:Ao > 1 (55.4% in CPFE, 25% in SRIF) were more frequently observed in CPFE. Centrilobular and paraseptal emphysema were observed in both groups, with higher frequencies in SRIF. Traction bronchiectasis and subpleural GGO/reticulation were present in all patients in both groups. In both univariable and multivariable analyses comparing SRIF and CPFE, MTWC was a predictive feature of SRIF (p < 0.05), whereas honeycombing and PA:Ao > 1 were predictive features of CPFE (all p < 0.05). ROC analysis of a model including MTWC, honeycombing, and PA:Ao > 1 yielded an area under the curve (AUC) of 0.880 for differentiating SRIF from CPFE.
For differentiating SRIF from CPFE, MTWC, honeycombing, and PA:Ao > 1 may serve as valuable CT features.