Locoregional Recurrence Patterns and Prognostic Outcomes in Early-Stage Natural Killer/T-Cell Lymphoma Treated With Sandwich Chemoradiation Therapy: A Post Hoc Analysis of a Randomized Controlled Trial.

Wang, Shu-Bei; Chen, Jia-Yi; Song, Qi; Xu, Cheng; Cao, Wei-Guo; Han, Yi-Min; Cheng, Shu; Xu, Peng-Peng et al. · Pract Radiat Oncol · 2026

rct · Level II

Where this comes from

Abstract

This study aimed to evaluate the locoregional control in patients with early-stage natural killer/T-cell lymphoma (NKTCL) treated with sandwich chemoradiation therapy. We performed a post hoc analysis of 87 consecutive patients with early-stage NKTCL enrolled in a prospective, multicenter, randomized phase 3 trial. All participants received standardized treatment comprising 4 cycles of asparaginase-based chemotherapy with integrated radiation therapy. With a median follow-up of 37.9 months, the 3-year survival outcomes of the cohort were overall survival (OS) 87.1%, progression-free survival 84.6%, and locoregional recurrence-free survival 90.8%. Six patients experienced locoregional recurrence, with 4 in the nasal cavity (4.6%), 4 in the nasopharynx (4.6%), and 1 in the cervical lymph node. Recurrence is typically in the radiation field (5 of 6), often in the initial high-standardized uptake value (SUV) region (4 of 6). Radiation therapy-chemotherapy interval (RCI) between 32 and 35 days was associated with improved prognosis, including OS (P = .043), progression-free survival (P = .013), and locoregional recurrence-free survival (P = .007) versus shorter RCI (21-31 days). Quantitative imaging analysis demonstrated that higher mean baseline apparent diffusion coefficient values (≥0.1125 × 10<sup>-3</sup> mm²/s) were associated with an improved OS (95.2% vs 71.6%, P = .049). Asparaginase-based sandwich chemoradiation therapy shows favorable efficacy in early-stage NKTCL, with distinct locoregional recurrence patterns. RCI and apparent diffusion coefficient values emerge as critical prognostic factors, potentially informing treatment optimization.