Precision management of radiotherapy interruption in locoregionally advanced nasopharyngeal carcinoma: Induction chemotherapy counteracts the survival impact in high-risk patients defined by a novel model

• First to demonstrate that the survival impact of RTI is confined to a specific high-risk NPC subgroup. • First to provide clinical evidence that induction chemotherapy neutralizes the risk of RTI. • The findings enable precision use of IC to protect vulnerable patients from treatment delays. • Offers a practical strategy to improve survival in real-world radiotherapy practice. To develop a risk stratification model to identify high-risk locoregionally advanced nasopharyngeal carcinoma (NPC) susceptible to radiotherapy interruption (RTI) ≥5 days and evaluate whether induction chemotherapy (IC) mitigated the associated survival detriment. We retrospectively analyzed 820 patients treated with intensity-modulated radiotherapy, randomly allocated into training ( n = 490) and validation ( n = 330) cohorts. Prognostic factors for overall survival (OS) were screened using univariate and multivariate analyses. A Cox regression-based prognostic model was developed and evaluated via C-index and calibration plots. RTI ≥5 days impact on survival endpoints—OS, locoregional recurrence-free survival (LRFS), distant metastasis-free survival (DMFS), and failure-free survival (FFS)—was assessed across risk groups, with IC effect analyzed in high-risk patients using propensity score matching. A prognostic model incorporating age, T stage, N stage, and EBV-DNA >4000 copies/mL was established. The prognostic index (PI) stratified patients into high- and low-risk groups, with high-risk patients exhibiting significantly inferior survival in OS, DMFS, and FFS (all P < 0.05). The model demonstrated C-indices of 0.744 (training) and 0.658 (validation) for OS prediction, outperforming the 8th AJCC staging system. RTI ≥5 days significantly reduced OS in high-risk patients (training: HR 1.816, P = 0.034; validation: HR 2.310, P = 0.024) but not in low-risk patients. Importantly, IC ≥2 cycles eliminated RTI-associated survival differences in high-risk patients following confounding adjustment. The developed model effectively identified high-risk NPC patients vulnerable to RTI ≥5 days, with IC shown to mitigate this risk. These findings support the implementation of risk-stratified treatment strategies and proactive IC use to counter unavoidable treatment delays.