Background: Whole-body FDG-PET captures distributed metabolic information beyond focal lesions, but this systems-level signal is not routinely represented at the individual-patient level. We aimed to develop and validate a deterministic, single-subject organ-network framework capable of recovering this systemic signal from routine FDG-PET. Methods: We transformed FDG-uptake values from seven robust anatomical compartments into a two-dimensional metabolic state space, using directed weighted edges to encode predefined organ-to-organ uptake dominance. Two interpretable graph metrics quantified renal-systemic prominence and global network heterogeneity. The framework was evaluated in 714 scans from 630 individuals spanning healthy, cardiometabolic, post-myocardial-infarction, and oncological cohorts, including test-retest and longitudinal intervention data. Findings: Metabolic-space coordinates were stable in healthy participants rescanned six weeks apart (Hotelling's T², P=0·1) and did not differ significantly by scanner, centre, or sex (all P≥0·2), while distinct disease cohorts separated robustly (P<0·001). Higher BMI and a dyslipidaemic pattern were associated with greater clearance and heterogeneity, whereas advanced tumour stage mapped towards lower values of both metrics (joint-space R²=0·15, P<0·001). Among 446 oncology patients with one-year survival data (98 deaths), a one-dimensional metabolic displacement index predicted one-year mortality (AUC 0·65; odds ratio 1·68 per SD decrease, P<0·001) and, in 410 patients with longitudinal follow-up, stratified overall survival (hazard ratio 0·71 per SD increase, 95% CI 0·60–0·82, P<0·001; one-year survival 80·2%, 84·4%, and 63·3% across tertiles). The index decreased after bariatric surgery (mean 5·8 to 4·8, P=0·002) and distinguished histopathological responders from non-responders in an exploratory lung-cancer cohort (median 4·1 versus 4·9, P=0·005). Interpretation: This transparent, training-free representation recovers a systems biomarker from routine FDG-PET and provides a practical basis for integrating whole-body metabolic organisation into clinical imaging workflows, without requiring additional acquisitions or specialised infrastructure.