Abstract Background Combined pulmonary fibrosis and emphysema (CPFE) is a heterogeneous syndrome with poor outcomes. Artificial intelligence -guided quantitative computed tomography (AIQCT) holds promise for precise phenotyping. Purpose To explore AIQCT-based phenotypes of CPFE and their distinct clinical characteristics as well as prognostic implications. Materials and Methods In this prospective multicenter study, 1688 CPFE patients were classified into emphysema-dominant (CPFE-E, n = 815) and fibrosis-dominant (CPFE-F, n = 873) subtypes based on AIQCT measuring fibrosis volume percentage and emphysema volume percentage of the whole lung (CPFE-E: fibrosisemphysema; CPFE-F: fibrosis≥ emphysema). Clinical features, treatments, and survival outcomes were compared. Multivariable Cox regression models assessed associations between AIQCT metrics and all-cause mortality. Results The CPFE-E group had a higher emphysema proportion (7% vs. 6%, P0.001), smoking history (29.8% vs. 24.4%, P=0.012), lower fibrosis percentage (10% vs. 18%, P0.001), and lower lung cancer history (34.3% vs. 39.3%, P=0.036) compared with the CPFE-F group. With a median follow-up duration of 2.62 years, 542 deaths were observed. CPFE-E showed a lower mortality risk compared with CPFE-F (LogRank: P=0.001). After adjusting demographic factors, lifestyle parameters, pulmonary function, comorbidity, and treatments, CPFE-F was associated with a 134% mortality risk. CPFE subtypes added predictive value to traditional clinical models with increasing C-index and Calibration Conclusion AIQCT identifies two distinct CPFE phenotypes with divergent clinical and prognostic profiles. CPFE-F is a more dangerous status than CPFE-E, with no traditional clinical tools that can assess other than AIQCT. These findings advocate for subtype-specific management strategies in CPFE. This abstract is funded by: the National Key Technologies Research and Development Program Precision Medicine Research (2021YFC2500700 and 2016YFC0901101), the National Natural Science Foundation of China (82370072), and the National High Level Hos pital Clinical Research Funding (2022-NHLHCRF-LX-01-0104)
Wang et al. (Fri,) studied this question.