Value of Diffusion-Weighted Imaging and Dynamic Contrast-Enhanced Magnetic Resonance Imaging for Prediction of Treatment Outcomes in Nasopharyngeal Carcinoma

Objective Magnetic resonance imaging (MRI) parameters that reflect the tumor microenvironment of nasopharyngeal carcinoma (NPC) may predict treatment response and facilitate treatment planning. This study aimed to evaluate the diffusion-weighted imaging and dynamic contrast-enhanced MRI (DCE-MRI) values for predicting the treatment outcomes in NPC patients. Methods Eighty-three patients with NPC underwent pretreatment MRI simulation with diffusion-weighted imaging and dynamic contrast-enhanced MRI. Average values of the apparent diffusion coefficient (ADC), Ktrans, Kep, Ve, Vp, and tumor volume of the primary tumors were measured. Other potential clinical characteristics (age, sex, staging, pathology, pretreatment Epstein-Barr virus level, and treatment type) were analyzed. Patients underwent follow-up imaging 6 months after treatment initiation. Treatment responses were assigned according to the Response Evaluation Criteria in Solid Tumors guideline (version 1.1). Results Fifty-one patients showed complete response (CR), whereas 32 patients did not (non-CR). Univariable logistic regression with variables dichotomized by optimal cutoff values showed that ADC ≥1.45 × 10 −3 mm 2 /s, Vp ≥0.14, tumor volume of ≥14.05 mL, high stage (stages III and IV), and Epstein-Barr virus level of ≥2300 copies/mL were predictors of non-CR ( P = 0.008, 0.05, 0.01, 0.009, and 0.04, respectively). The final multivariable model, consisting of a combination of ADC ≥1.45 × 10 −3 mm 2 /s, Vp ≥0.14, and high stage, could predict non-CR with a good discrimination ability (area under the receiver operating characteristic curve, 0.76 [95% confidence interval, 0.66–0.87]; sensitivity, 62.50%; specificity, 80.39%; and accuracy 73.49%). Conclusions A multivariable prediction model using a combination of ADC ≥1.45 × 10 −3 mm 2 /s, Vp ≥0.14, and high stage can be effective for treatment response prediction in NPC patients.