Whole-body ¹⁸F-FDG PET reveals glycemic state-dependent reorganization of directed inter-organ metabolic networks.

Abstract

Glucose homeostasis depends on coordinated activity across multiple organs and is altered in type 2 diabetes. This study aimed to characterize how whole-body inter-organ metabolic coordination differs across normoglycemia, prediabetes, and diabetes using ¹⁸F-FDG PET/CT. Whole-body ¹⁸F-FDG PET/CT scans from 1,149 adults were stratified into normoglycemic, prediabetic, and diabetic groups. Lean body mass-normalized standardized uptake from 20 regions was adjusted for age, sex, and body mass index. Directed networks were estimated within each glycemic group using a bootstrap-embedded causal discovery framework with stability-weighted consensus. Global topology, between-group differences, cumulative multi-step propagation, and reproducibility were assessed. Reorganization of directed connectivity between glycemic groups was system-wide, encompassing all 20 regions in every pairwise comparison (all family-wise error-corrected P < 0.001). The global architecture shifted while retaining its basic form, with directed global efficiency increasing and local clustering decreasing across the glycemic spectrum, while small-world organization was preserved in all groups. At the level of cumulative multi-step structure, node-level net dominance reorganized across groups, and this profile was internally reliable on repeated split-half resampling with Spearman-Brown reliability 0.83, 0.72, and 0.64. Inter-organ metabolic organization differs across glycemic states in a distributed rather than focal manner, and this reorganization is reproducible within the cohort. These differences describe the inferred networks rather than measured metabolic flux, indicating that routinely acquired whole-body PET carries systemic information beyond organ-by-organ reporting.