High Signal Intensity in Dentate Nucleus on Unenhanced T1-weighted MR Images: Association with Linear versus Macrocyclic Gadolinium Chelate Administration.

Abstract

To assess whether an association exists between hyperintensity in the dentate nucleus (DN) on unenhanced T1-weighted magnetic resonance (MR) images and previous administration of gadolinium-based contrast agents (GBCAs) that contain different types of gadolinium chelates. The institutional review board approved this study. Written informed consent was waived because this was a retrospective study. Evaluated were 127 cases among 360 consecutive patients who underwent contrast agent-enhanced brain MR imaging. Two radiologists conducted visual evaluation and quantitative analysis on unenhanced T1-weighted MR images by using regions of interest. DN-to-cerebellum (DN/cerebellum) signal intensity ratios were calculated and the relationship between DN/cerebellum and several factors was evaluated, including the number of previous linear chelate and/or macrocyclic GBCA administrations by using a generalized additive model. The Akaike information criterion was used in model selection. Interobserver correlation was evaluated with paired t tests and the Lin concordance correlation coefficient. The images of nine patients (7.1%) showed hyperintensity in the DN. Twenty-three patients (18.1%) received linear GBCAs (median, two patients; maximum, 11 patients), 36 patients (28.3%) received macrocyclic GBCAs (median, two patients; maximum, 15 patients), 14 patients (11.0%) received both types of GBCA (linear [median, two patients; maximum, five patients] and macrocyclic [median, three patients; maximum, eight patients]), and 54 patients (42.5%) had no history of administration of gadolinium chelate. Interobserver correlation was almost perfect (0.992 [95% confidence interval: 0.990, 0.994]). The DN/cerebellum ratio was associated with linear GBCA (P < .001), but not with macrocyclic GBCA exposure (P = .875). According to the Akaike information criterion, only linear GBCA was selected for the final model, and the DN/cerebellum ratio had strong association only with linear GBCA. Hyperintensity in the DN on unenhanced T1-weighted MR images is associated with previous administration of linear GBCA, while the previous administration of macrocyclic GBCAs showed no such association.