Correlation between PD-L1 expression and FDG-PET/CT visual assessments in non-small cell lung cancer resected specimens

PURPOSE: This retrospective study aimed to investigate the validity of fluorodeoxyglucose PET (FDG-PET) visual assessments to predict programmed death-ligand 1 (PD-L1) expression levels in patients with non-small cell lung cancer (NSCLC). MATERIALS AND METHODS: One hundred and seven NSCLC patients who underwent FDG-PET/computed tomography (CT) scans and PD-L1 expression tests were retrospectively identified. Patients were divided into two groups according to PD-L1 expression: PD-L1 high group (PD-L1 tumor proportion score ≥50%) and PD-L1 low group (<50%). We compared clinicopathological characteristics and PET assessments [maximum standardized uptake value (SUVmax) and Deauville score] between the two groups based on PD-L1 expression. RESULTS: High expression of PD-L1 was detected in 25 of 107 cases. In both univariable and multivariable analysis, there were significant differences in PET visual assessments in NSCLC ( P < 0.05). Receiver operating characteristics for the PET visual assessments [area under the curve (AUC) = 0.712, 95% confidence interval (CI) 0.628-0.793] and SUVmax (AUC = 0.753, 95% CI 0.647-0.861) showed equivalent accuracy ( P = 0.227). Based on histopathology, in adenocarcinoma patients, there were significant differences between PET visual assessments and PD-L1 expression ( P < 0.05), while no significant differences were observed in squamous cell carcinoma patients. Based on epidermal growth factor receptor (EGFR) mutation analysis, in patients with EGFR wild type, there were significant differences between PET visual assessments and PD-L1 expression ( P = 0.006), while in patients with EGFR mutations, there were no significant differences between PET visual assessments and PD-L1 expression. CONCLUSION: Results of PET visual assessments correlated with PD-L1 expression in NSCLC.