Prognostic Value of Pre‐Treatment Diffusion Kurtosis Imaging for Progression‐Free Survival Prediction in Advanced Nasopharyngeal Carcinoma

Abstract

ABSTRACT Purpose This study aimed to evaluate the value of diffusion kurtosis imaging (DKI) for prognostic value for long‐term PFS in nasopharyngeal carcinoma (NPC). Methods A cohort of 295 NPC patients underwent pretreatment 3.0T MRI with DKI to derive mean kurtosis (MK), mean diffusion (MD), and apparent diffusion coefficient (ADC). Clinical parameters (Tumor stage, EBV‐DNA, neoadjuvant chemotherapy regimens) were recorded. Follow‐up extended to December 2023. Statistical analyses (R software v4.3.0) included univariate/multivariate Cox regression and Kaplan–Meier survival analysis. A prognostic nomogram integrating key predictors was developed. Results Median 10‐year follow‐up revealed 2‐, 5‐, and 10‐year PFS rates of 89%, 79%, and 74%, respectively. Univariate Cox regression analysis demonstrated that T stage, Clinical Stages, NAC regimens, ADC_Group, MK_Group, and MD_Group were significant prognostic factors for PFS in NPC (p < 0.05). Multivariate analysis identified Clinical Stage (HR = 2.230, 95% CI 1.44–3.66, p < 0.001), NAC (neoadjuvant chemotherapy) regimens (HR = 0.56, 95% CI 0.35–0.90, p = 0.017), and MK_Group (HR = 0.52, 95% CI 0.33–0.82, p = 0.003) as independent prognostic factors. The MK_Group high exhibited superior survival rates versus MK_Group low (2‐year: 94% vs. 81%; 5‐year: 85% vs. 66%; 10‐year: 79% vs. 64%; all p < 0.05). The nomogram combining Clinical Stage, NAC, and MK_Group demonstrated moderate predictive accuracy for 2‐, 5‐, and 10‐year PFS (AUC = 0.736, 0.718, 0.697). Conclusion Pretreatment MK serves as a robust noninvasive biomarker for long‐term PFS in NPC. Integration with Clinical Stage and NAC regimens enhances prognostic stratification, supporting personalized therapeutic strategies.

Keywords

• diffusion kurtosis imaging
• magnetic resonance imaging
• nasopharyngeal carcinoma
• progression‐free survival