Determinants of ordered physiological myocardial 18F-fluorodeoxyglucose uptake patterns on non-cardiac PET/computed tomography: an ordinal regression analysis

OBJECTIVE: On routine non-cardiac PET/computed tomography (CT), myocardial 18F-fluorodeoxyglucose (18F-FDG) uptake demonstrates substantial interindividual variability and may complicate image interpretation. However, the systemic metabolic and clinical determinants associated with different myocardial FDG uptake patterns remain incompletely understood. METHODS: This retrospective single-center study included 292 patients who underwent whole-body 18F-FDG PET/CT following a minimum 6-h fast, without dedicated cardiac preparation. Myocardial FDG uptake was visually categorized into three ordered patterns: no/low, focal, and diffuse uptake. Clinical, laboratory, hemodynamic, and treatment-related variables were analyzed using proportional odds ordinal logistic regression to identify factors associated with increasing uptake categories. Exploratory prediction models for higher myocardial uptake patterns and diffuse uptake were additionally constructed to support clinical interpretability. Model performance was evaluated using receiver operating characteristic curves and calibration analysis. RESULTS: Among the 292 patients, 122 (41.8%) demonstrated no/low uptake, 60 (20.5%) focal uptake, and 110 (37.7%) diffuse uptake. In multivariable ordinal logistic regression, blood pressure control remained strongly associated with myocardial uptake patterns. Compared with poor blood pressure control, no hypertension [odds ratio (OR) 0.155, 95% confidence interval (CI) 0.072-0.334, P < 0.001] and good blood pressure control (OR 0.054, 95% CI 0.019-0.151, P < 0.001) were associated with lower odds of higher uptake categories. Absence of hormone therapy was also independently associated with lower odds of higher uptake patterns (OR 0.216, 95% CI 0.098-0.476, P < 0.001). Fasting glucose, diabetes, blood-pool maximum standardized uptake value, and chemotherapy were not independently significant after adjustment. The higher uptake model achieved an area under the curve of 0.898, while the diffuse uptake model achieved an area under the curve of 0.839, with satisfactory calibration performance. CONCLUSIONS: Myocardial FDG uptake patterns observed on routine non-cardiac PET/CT were associated with a graded spectrum of systemic metabolic and hemodynamic characteristics. These findings may provide a framework for more structured interpretation of myocardial FDG uptake in routine PET/CT imaging. These findings should be interpreted as observational associations rather than evidence of causal relationships.