Identifying optimal clinical trial candidates for locoregionally advanced nasopharyngeal carcinoma: Analysis of 9468 real-world cases and validation by two phase 3 multicentre, randomised controlled trial.

BACKGROUND AND PURPOSE: This study aims to identify the optimal high-risk candidates for clinical trials in locoregionally advanced nasopharyngeal carcinoma (NPC). MATERIALS AND METHODS: Non-metastatic NPC patients (n = 9,468) were included. Recursive partitioning analyses (RPA) were performed to generate risk stratification. Receiver operating characteristics curve was used to determine the cut-off value of pre-treatment Epstein-Barr virus (EBV) DNA for progression-free survival (PFS). Individual-level data from two clinical trials were used for validation. RESULTS: Anatomic stratification based on T and N category (eighth edition TNM, TNM-8) classified the N2-3 or T4 as an anatomic high-risk group with 5-year PFS of 69% (95% confidence interval: 68-71%). Prognostic stratification identified patients with pre-treatment EBV DNA >/=4000 copies/mL as a prognostic high-risk group with 5-year PFS of 69% (67-70%). The c-index was significantly higher for anatomic stratification (0.621, p < 0.001) and prognostic stratification (0.585, p < 0.001) compared with existing TNM-8 stage groups (0.562). The validation cohorts based on clinical trials data showed greater PFS benefit than the results of the original trials [Hazard ratio: NCT01245959, 0.64 vs. 0.67; NCT01872962, 0.42 vs. 0.52]. Moreover, detectable post-treatment EBV DNA indicated a high risk of progression with 5-year PFS of 38.7% and was the most adverse independent factor for all endpoints. CONCLUSIONS: N2-3 or T4 NPC patients were ideal candidates for multicenter clinical trials in locoregionally advanced NPC. Patients with detectable post-treatment EBV DNA are suitable candidates for adjuvant trials.