The Role of PD-L1 Expression in Prediction and Stratification of Recurrent or Refractory Extranodal Natural Killer/T-Cell Lymphoma

Background and AimsThe clinical outcome of relapsed and refractory (RR) extranodal natural killer/T-cell lymphoma (ENKTL) is poor. It is necessary to identify RR patients in ENKTL and find novel therapeutic targets to improve the prognosis of patients with RR ENKTL. MethodsA total of 189 ENKTL patients with effective clinical characteristics were enrolled. Paraffin specimens were collected for PD-L1 expression identification. Kaplan-Meier curve analysis was performed for survival analysis. Whole exome sequencing (WES) was performed for identifying the mutational characterization of RR and effective treatment (ET) patients. ResultsUnivariate and multivariate Cox proportional hazards regression analysis showed that negative PD-L1 expression (HR = 1.132, 95% CI = 0.739-1.734, P = 0.036) was an independent predictor of poor prognosis in patients with ENKTL. The overall survival (OS) of PD-L1 positive patients was significantly higher than that of PD-L1 negative patients (P = 0.009). Then, we added PD-L1 expression as a risk factor to the model of Prognostic Index of Natural Killer Lymphoma (PINK), and named as PINK+PD-L1. The PINK+PD-L1 model can significantly distinguish RR patients, ET patients, and the whole cohort. Moreover, our data showed that PD-L1 expression was lower than 25% in most RR patients, suggesting that RR subtypes may be associated with low expression of PD-L1 (P = 0.019). According to the whole exome sequencing (WES), we found that the mutation frequencies of JAK-STAT (P = 0.001), PI3K-AKT (P = 0.02) and NF-kappa B (P < 0.001) pathways in RR patients were significantly higher than those in ET patients. ConclusionPatients tend to show RR when PD-L1 expression is lower than 25%. The model of PINK+PD-L1 can stratify the risk of different groups and predict OS in ENKTL patients. The mutational profile of ENKTL patients with RR is different from that of patients with ET.