@ARTICLE{10.3389/fonc.2021.689593, AUTHOR={Peng, Hao and Chen, Bin-bin and Wang, Xiao-hui and Mo, Yun-Xian and Han, Fei}, TITLE={Prognostic Value of Regression Rate of Plasma EBV DNA After Induction Chemotherapy for Stage II-IVA Nasopharyngeal Carcinoma}, JOURNAL={Frontiers in Oncology}, VOLUME={11}, YEAR={2021}, URL={https://www.frontiersin.org/articles/10.3389/fonc.2021.689593}, DOI={10.3389/fonc.2021.689593}, ISSN={2234-943X}, ABSTRACT={Background/ObjectiveWe aimed to explore the prognostic value of regression rate (RR) of plasma Epstein–Barr virus (EBV) DNA after induction chemotherapy (IC) in patients with stages II–IVA nasopharyngeal carcinoma (NPC).MethodsEligible patients receiving IC followed by concurrent chemoradiotherapy were included. The cut-off value of pre-treatment EBV DNA (pre-IC DNA) and RR were identified by receiver operating curve (ROC). Recursive partitioning analysis (RPA) was applied to create new staging. Harrell’s c-index and time-independent ROC were employed to compare different RPA staging.ResultsIn total, 1,184 patients were included. The cut-off values of pre-IC DNA and RR were 16,200 copies/ml and 95.127% for patients receiving two cycles, and 5,520 copies/ml and 99.994% for those receiving three cycles. Notably, we only focused on patients receiving two cycles of IC. Patients with a RR >95.127% had significantly better 5-year overall survival (OS) than those with a RR ≤95.127% (86.2% vs. 54.3%, P <0.001). Then, RPA1 (pre-IC DNA + TNM staging + RR) and RPA2 (pre-IC DNA + TNM staging + post-IC EBV DNA [post-IC DNA]) staging systems were created. RPA1 staging achieved stronger power in OS prediction than RPA2 staging and TNM staging (c-index: 0.763 [0.714–0.812] vs. 0.735 [0.684–0.786] vs. 0.677 [0.604–0.749]; AUC: 0.736 vs. 0.714 vs. 0.628), indicating that RR had stronger prognostic power than post-IC DNA. Moreover, patients with stages III–IVRPA1 could benefit from high concurrent cumulative platinum dose (≥160 mg/m2).ConclusionRR in conjunction with current TNM staging could better conduct risk stratification, prognosis prediction and help to guide precise concurrent chemotherapy.} }