Risk stratification and boost dose optimization for locoregionally advanced nasopharyngeal carcinoma with residual lesions based on clinical stage and post-radiotherapy EBV DNA.

BACKGROUND: Residual locoregionally advanced nasopharyngeal carcinoma (LA-NPC) after definitive chemoradiation leads to poor outcomes, but current treatments lack standardized, evidence-based guidance for individualized dose escalation, causing varied practices. OBJECTIVE: To develop a risk-stratification model to personalize boost doses for these patients. DESIGN: A retrospective study. METHODS: In this retrospective study, 364 LA-NPC patients with post-radiotherapy (post-RT) residual disease were enrolled. Patients were stratified into high-dose (≥600 cGy, n=201) and low-dose (<600 cGy, n=163) boost groups. Independent prognostic factors were identified via Cox regression. A risk stratification model was constructed using these factors, and its utility in guiding boost dose selection was evaluated. RESULTS: Multivariate analysis confirmed that clinical stage (HR=2.012, 95% CI: 1.100-3.682, p=0.023) and detectable post-RT EBV DNA (HR=1.997, 95% CI: 1.129-3.535, p=0.018) were independent prognostic factors for LA-NPC patients with residual disease after definitive chemoradiotherapy. Based on these variables, a three-tiered risk model was established: low-risk (Stage III with undetectable post-RT EBV DNA), middle-risk (Stage III with detectable post-RT EBV DNA or Stage IVA with undetectable post-RT EBV DNA), and high-risk (Stage IVA with detectable post-RT EBV DNA). This model effectively stratified patients for overall survival (OS) (P<0.05). Crucially, a high-dose boost (≥600 cGy) was associated with improved OS in the high-risk group (HR=0.344, 95% CI: 0.122-0.967, P=0.043), whereas no survival advantage was identified in the low- or middle-risk groups. Among high-risk patients, no statistically significant differences in acute or late toxicities were observed between the high-dose and low-dose boost groups. CONCLUSION: We developed a clinically applicable risk stratification model integrating clinical stage and post-RT EBV DNA status. Our findings suggest that only high-risk patients may derive an OS benefit from a boost dose of ≥600 cGy, whereas low- and middle-risk patients do not appear to benefit from further dose escalation. This risk-adapted boost strategy provides a preliminary framework requiring prospective validation. Given the retrospective single-center design of this study, prospective multicenter studies are warranted before routine clinical implementation.