Baseline 18F-FDG PET/CT-derived heterogeneity index predicts poor pathological response after neoadjuvant chemoradiotherapy in locally advanced rectal cancer.

OBJECTIVE: To evaluate whether baseline 18F-FDG PET/CT-derived volumetric and heterogeneity parameters can predict Ryan 3 poor pathological response after neoadjuvant chemoradiotherapy in locally advanced rectal cancer (LARC). METHODS: This retrospective single-center study included 48 patients with histopathologically confirmed LARC who underwent baseline 18F-FDG PET/CT before neoadjuvant chemoradiotherapy followed by curative-intent low anterior resection. PET-derived SUVmax, SUVmean, metabolic tumor volume (MTV), total lesion glycolysis (TLG), and a linear regression-derived heterogeneity index (HI) were calculated from the primary tumor. HI was derived from MTV values measured at 40%, 60%, and 80% of SUVmax thresholds. Patients were classified as Ryan 3 poor/non-response or Ryan 0-2 response according to surgical pathology. ROC analysis, logistic regression, and bootstrap internal validation were performed. RESULTS: Sixteen patients had Ryan 3 poor pathological response and 32 had Ryan 0-2 response. MTV, TLG, HI, and clinical/radiological T stage were significantly higher in Ryan 3 patients, whereas age, sex, CEA, clinical/radiological N stage, SUVmax, and SUVmean did not differ significantly between response groups. SUVmax and SUVmean showed no significant discriminatory value. HI showed the numerically highest AUC (0.760; bootstrapped 95% CI, 0.586-0.899), with predictive performance comparable to MTV (AUC 0.749) and TLG (AUC 0.748). In restricted T-stage-adjusted models, HI, MTV, and TLG remained independently associated with Ryan 3 poor response. Bootstrap optimism-corrected AUCs were 0.847, 0.842, and 0.813 for HI + T stage, MTV + T stage, and TLG + T stage models, respectively. CONCLUSION: Baseline 18F-FDG PET/CT-derived HI, MTV, and TLG may help identify LARC patients at risk for Ryan 3 poor pathological response. The workstation-calculable HI may provide complementary information for pretreatment risk stratification.