The predictive value of 18F-FDG PET/CT metabolic parameters for PD-L1 expression in esophageal squamous cell carcinoma

Abstract Objective This retrospective study aimed to evaluate the predictive value of non-invasive 18 F-fluorodeoxyglucose ( 18 F-FDG) positron emission tomography/computed tomography (PET/CT) metabolic parameters for programmed death-ligand 1 (PD-L1) expression status in esophageal squamous cell carcinoma (ESCC). Methods The associations between 18 F-FDG PET/CT metabolic parameters, PD-L1 expression, and clinicopathological features were assessed in 89 patients with ESCC. The measured metabolic parameters comprised the maximum standardized uptake value (SUVmax), mean standardized uptake value (SUVmean), peak standardized uptake value (SUVpeak), standard deviation of standardized uptake value (SUVsd), metabolic tumor volume (MTV), and total lesion glycolysis (TLG). Clinicopathological characteristics, including age, gender, smoking status, drinking status, tumor location, vertical tumor length, TNM stage, and pathological grade, were recorded. PD-L1 expression in primary tumor tissues was evaluated by immunohistochemistry (IHC) for all patients prior to any anti-tumor treatment. For statistical analysis, patients were stratified into PD-L1-negative (CPS < 10) and PD-L1-positive (CPS ≥ 10) groups. Results SUVmax, SUVmean, SUVpeak, and SUVsd were significantly higher in ESCC patients exhibiting PD-L1-positive expression than in those with PD-L1-negative expression (all P < 0.05). Receiver operating characteristic curve analysis demonstrated that SUVmax (area under the curve [AUC] = 0.636, P = 0.029) and SUVmean (AUC = 0.628, P = 0.044) were predictive of PD-L1 positivity, with optimal cutoff values of 19.74 and 10.01, respectively. Univariate logistic regression analysis revealed that SUVmax ( P = 0.006), SUVmean ( P = 0.011), SUVpeak ( P = 0.022), and SUVsd ( P = 0.024) were significantly associated with PD-L1-positive expression. Owing to severe multicollinearity among these four statistically significant SUV-derived parameters, only SUVmax was incorporated into the final single-predictor logistic model ( P = 0.006). Conclusions 18 F-FDG PET/CT metabolic parameters, including SUVmax, SUVmean, SUVpeak, and SUVsd, were associated with PD-L1-positive expression. Due to severe multicollinearity among these parameters, only SUVmax was selected for the single-predictor logistic model. Nevertheless, given its modest discriminative performance (AUC = 0.636), SUVmax should only serve as an auxiliary imaging reference rather than a reliable noninvasive substitute for pathological PD-L1 detection in patients with ESCC.