Combining Clinical Variables with 18F-FDG-PET/CT Metrics Enhances Overall Survival Prediction in Gastric Cancer

FDG PET/CT Prognostic Value in Gastric Cancer

Abstract

Objective: This study aimed to evaluate whether adding a broad set of pre-treatment [18F] fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) metabolic and volumetric parameters to routine clinical variables improves the prediction of overall survival (OS) in patients with gastric cancer (GC). Secondary objectives were to assess the prognostic value of blood- and spleen-normalized metabolic indices and to explore associations between PET metrics and HER2 status.
Methods: In this retrospective cohort, pre-treatment 18F-FDG PET/CT data were analyzed to extract standardized uptake value (SUV)- and volume-based PET metrics, BLR_mean, and SLR_mean. Clinical variables, pathological features, treatment details, HER2 status, and survival outcomes were obtained from institutional records. OS was calculated based on the date of initial management. Prognostic performance was evaluated using Cox models, calibration metrics, time-dependent area under the curve (AUC), and decision curve analysis (DCA). Nested models (clinical-only vs. clinical+PET) were compared to determine the incremental value.
Results: SUV- and volume-based PET metrics showed variable but directionally consistent associations with OS. Metabolic tumor volume (MTV_40) and total lesion glycolysis (TLG_40) demonstrated trends toward worse outcomes, although these effects did not consistently reach statistical significance in multivariable analyses. Blood- and spleen-normalized parameters (BLR_mean and SLR_mean) showed stronger effects in the PET-only model but became attenuated after adjustment for clinical covariates. Incorporating PET parameters into the clinical model modestly improved discrimination and yielded acceptable calibration. HER2-positive tumors exhibited higher metabolic activity; however, no significant interaction was observed between HER2 status and the prognostic effect of PET metrics. Across clinically relevant decision thresholds (10–40%), the combined Clinical+PET model achieved higher net benefit than the clinical model alone. The combined model demonstrated a higher net benefit at 12 and 24 months.
Conclusion: Pretreatment 18F-FDG PET/CT appears to provide additional prognostic information beyond routine clinical variables in GC. The inclusion of SUV-based, volumetric, and normalized metabolic parameters modestly improves risk stratification and is associated with favorable decision-analytic performance. These findings support integrating quantitative PET metrics into prognostic evaluation frameworks for patients undergoing management for GC.