Quantitative synthetic MRI for pretreatment prediction of chemoradiotherapy response in advanced nasopharyngeal carcinoma

Motivation: About 10%~30% of advanced nasopharyngeal carcinoma (ANPC) remained residual and relapse after chemoradiotherapy (CRT), so it would be valuable to pretreatment prediction of no-response ANPC. Goal(s): Quantitative syMRI parameters for predicting no-response ANPC before CRT. Approach: 95 response and 32 no-response ANPC were enrolled, and the T1, T2, PD values from syMRI and clinicopathological characteristics were compared. Logistic regression was used to construct prediction model. Results: T1, T2, PD values were significantly lower in no-response ANPC. The combination of T1, T2, PD can achieve good prediction performance (AUC=0.897). The syMRI combined with clinicopathological characteristics can improve prediction performance (AUC=0.974). Impact: Our study suggested syMRI can non-invasively identify no-response ANPC patients before CRT. Therefore, the application of syMRI would be beneficial for treatment selection and management of ANPC patients.