Whole-body ¹⁸F-FDG PET/CT identifies subclinical metabolic phenotypes in normoglycemic adults.

Abstract

Early disruptions in systemic glucose regulation may occur even before overt glycemic abnormalities manifest. This study investigated whether patterns of inter-regional glucose uptake coordination, derived from whole-body ¹⁸F-FDG PET/CT, could reveal latent subclinical metabolic phenotypes among normoglycemic adults. We analyzed 536 clinically normoglycemic adults (fasting blood glucose < 6.1 mmol/L) who underwent whole-body ¹⁸F-FDG PET/CT. For each subject, pairwise lean body mass normalized standardized uptake value ratios (SUL_r) across 20 anatomically defined regions of interest (ROIs) were calculated and adjusted for age, sex, and BMI. Nonlinear dimensionality reduction followed by Gaussian Mixture Modeling (GMM) was applied to identify latent phenotypes. Cluster separation and assignment certainty were evaluated using silhouette scores and posterior entropy. Between-phenotype SUL_r patterns were compared, and population-level inter-regional metabolic networks were constructed from covariate-adjusted residuals using bootstrap-stabilized Pearson correlations. To assess biological relevance, individuals with prediabetes (n = 465) and diabetes (n = 148) were embedded into the same analytic framework. Two distinct metabolic phenotypes were identified within the normoglycemic cohort (Phenotype A and Phenotype B), with no significant differences in age, sex, BMI, or fasting blood glucose. Phenotype A was characterized by lower myocardium-to-peripheral ratios (versus adipose, skeletal muscle, and lung) and higher brain-to-myocardium and brain-to-fat ratios compared to Phenotype B. At the network level, Phenotype A showed stronger positive correlations between myocardial and peripheral uptakes, indicating tighter systemic coupling despite reduced relative cardiac allocation. These patterns qualitatively mirrored those observed in prediabetic and diabetic groups. Whole-body ¹⁸F-FDG PET/CT identified metabolically distinct phenotypes in adults with normal fasting glucose, differentiated by both pairwise uptake ratios and systemic network coordination. These findings highlight latent metabolic heterogeneity in clinically healthy individuals and provide a basis for future investigations into early systemic metabolic alterations.