Abstract
BACKGROUND/AIMS: Programmed death-ligand 1 (PD-L1) expression, a validated predictive biomarker for anti-PD-1/PD-L1 inhibitors, is reported to change over time. This poses challenges during clinical application in non-small cell lung cancer. METHODS: This study included patients with non-small cell lung cancer who underwent surgery or biopsy and evaluation of PD-L1 expression in tumor cells via immunohistochemistry more than twice. We set the threshold of PD-L1 positivity to 10% and categorized patients into four groups according to changes in PD-L1 expression. Clinicopathologic information was collected from medical records. Statistical analyses, including Fisher's exact test and log-rank test, were performed. RESULTS: Of 109 patients, 38 (34.9%) and 45 (41.3%) had PD-L1 positivity in archival and recent samples, respectively. PD-L1 status was maintained in 78 (71.6%) patients, but changed in 31 (28.4%), with 19 (17.4%) from negative to positive. There were no significant differences in characteristics between patients who maintained PD-L1 negativity and whose PD-L1 status changed from negative to positive. Patients harboring PD-L1 positivity in either archival or recent samples achieved better responses (p = 0.129) and showed longer overall survival than those who maintained PD-L1 negativity when they received immune checkpoint inhibitors after platinum failure (median overall survival 14.4 months vs. 4.93 months; hazard ratio, 0.43; 95% confidence interval, 0.20 to 0.93). CONCLUSION: PD-L1 status changed in about one-fourth of patients. PD-L1 positivity in either archival or recent samples was predictive of better responses to immune checkpoint inhibitors. Therefore, archival samples could be used for assessment of PD-L1 status. The need for new biopsies should be decided individually.