Comparative evaluation of MRI Dixon T2-weighted, T2 mapping, and T1 mapping for Graves' ophthalmopathy activity...

This study aimed to compare the diagnostic performance of Dixon T2-weighted imaging (Dixon-T2WI), T2 mapping, and T1 mapping in evaluating orbital activity in Graves' orbitopathy (GO).

Seventy-three GO patients were retrospectively enrolled, with 146 affected eyes. Based on clinical activity score (CAS) criteria, the eyes were divided into inactive (n = 76) and active (n = 70) groups. All patients underwent orbital magnetic resonance imaging (MRI) with Dixon-T2WI, T1 mapping, and T2 mapping sequences. The signal intensity ratio (SIR), T2 relaxation time (T2RT), and T1 relaxation time (T1RT) were measured in the superior rectus (SR), inferior rectus (IR), lateral rectus (LR), and medial rectus (MR). Predicted probabilities were derived from the generalised linear mixed model. Diagnostic performance was evaluated via marginal area under the receiver operating characteristic (ROC) curve (AUC), with pairwise comparisons by DeLong's test.

ROC curve analysis showed the following: The AUC values of the MR-SIR (AUC = 0.742) showed no significant difference from T2RT (P > 0.05), and both were higher than T1RT. The AUC values of the SR-SIR (AUC = 0.770) showed no significant difference from T2RT (P > 0.05), and both were higher than T1RT. The AUC value of the IR-SIR (AUC = 0.823) showed no significant difference from T2RT (P > 0.05), and both were higher than T1RT. The AUC values of the LR-SIR (AUC = 0.700) showed no significant difference among SIR, T2RT, and T1RT (all P > 0.05).

As a routine sequence, Dixon-T2WI is comparable to T2 mapping in diagnostic efficacy and significantly superior to T1 mapping. It is preferred for assessing GO activity in clinical practice.