Smoking reshapes systemic metabolic coordination across lung cancer subtypes: a multicenter total-body ¹⁸F-FDG PET/CT study with tumor habitat imaging.

Abstract

Lung cancer (LC) is increasingly recognized as a systemic disease, and these systemic alterations may differ across histological subtypes. Smoking is a major risk factor for LC, yet its impact on systemic metabolism across LC subtypes remains underexplored. We studied 125 LC patients and 200 healthy controls from three centers who underwent static and dynamic total-body ¹⁸F-FDG PET/CT. Thirty-seven regions of interest spanning the central nervous system and peripheral organs were analyzed. Regional standardized uptake values normalized by lean body mass and time-activity curves were harmonized using ComBat and adjusted for demographic covariates. Population-level metabolic networks were constructed from bootstrapped Pearson correlations of regional uptake to characterize smoking status-associated systemic alterations across LC subtypes. Individual-level metabolic networks were derived from correlations of residual time-activity curves after third-order polynomial fitting, from which organ-specific abnormality strength metrics were calculated. In addition, tumor habitats were defined from PET/CT intensity and entropy features, and their associations with smoking status and organ-level abnormalities were assessed using multivariable regression analyses. Population- and individual-level network analyses showed that smoking status was associated with altered systemic metabolic coordination in a subtype-dependent manner, with the liver consistently emerging as a central hub of metabolic disruption. Tumor habitat analysis further revealed a smoking-associated shift toward hypometabolic tumor microenvironments. These findings provide quantitative insight into smoking status-related systemic metabolic reprogramming in LC and highlight the potential value of imaging-based characterization across histological subtypes.