A new prognostic model for predicting outcomes of patients with recurrent or metastatic nasopharyngeal carcinoma receiving subsequent line (≥2 lines) anti-programmed death-1 monotherapy

Objectives: About 17.7-34.0 % of patients with recurrent or metastatic nasopharyngeal carcinoma (RM-NPC) responded well to anti-PD-1 monotherapy. We sought to establish a nomogram to estimate the progression-free survival (PFS) of RM-NPC patients receiving subsequent-line anti-PD-1 monotherapy. Materials and methods: This cohort study investigated consecutive RM-NPC patients undergoing anti-PD-1 monotherapy. A nomogram was developed in the training cohort (n = 161), using a Cox multivariate model with backward stepwise inclusion, and was validated in the validation cohort (n = 69). Its predictive accuracy was assessed using a concordance index (C-index) and calibration curve. The primary endpoint was PFS. Sec-ondary endpoints included the objective response rate (ORR), disease control rate (DCR), and overall survival (OS). Results: Liver metastasis, albumin, lactate dehydrogenase, monocyte-to-lymphocyte ratio, and plasma Eps-tein-Barr virus DNA were used to develop a nomogram that could separate patients into favourable-and unfavourable-prognosis groups. The C-index in the training and validation cohort were 0.70 and 0.68, respec-tively, which was confirmed by calibration curves. Median PFS (mPFS) was lower for the unfavourable-prognosis than for the favourable-prognosis group (1.80 vs 4.93; hazard ratio 2.49 [95 % confidence interval: 1.78-3.49]; p < 0.001), across all subgroups. OS exhibited the same pattern. The ORR and DCR were markedly lower in the unfavourable-prognosis than in the favourable-prognosis group. All results were confirmed in the validation cohort. Conclusion: Our model is a reliable prognostic indicator of PFS in RM-NPC patients undergoing anti-PD-1 monotherapy, allowing robust estimation of the immunotherapy benefit an individual might derive.