Mesial temporal lobe epilepsy (mTLE) is the most common form of epileptic syndrome in adults, which is characterized by seizures originating primarily from mesial structures of the temporal lobe, and the damage can spread to many other brain structures with anatomic and functional connections to the epileptogenic hippocampus. However, little is known about the biochemical change in vivo whole brain of patients with mTLE. In human brain, T1ρ may be sensitive to both normal and pathologic changes in tissue protein content and has been tested successfully in Alzheimer’s disease, Parkinson’s disease and certain brain tumors[1, 2]. The purpose of our study was to report a systematic investigation of the variations in T1rho values of subcortical structures automatic-drawn by atlas-based segmentation among temporal lobe epilepsy patients, and the underlying relation between the significantly altered T1rho values or volumes of subcortical structures and duration of epilepsy or age at epilepsy onset.

Subjects and data acquisition: MR imaging was conducted on 20 mTLE patients whose postoperative pathological diagnoses were hippocampal sclerosis (9 left mTLE patients, aged 29.1±19.6y; and 11 right mTLE patients, aged 26.3±7.9y) and 10 age-matched controls (age 27.6±13.2y) using a 3.0 Tesla MR scanner (Philips Achieva TX). A 8-channel SENSE head coil was used for signal reception. T1rho weighted images were acquired using a fluid suppressed 3D turbo spin echo technique. Whole brain coverage was achieved using a sagittal aligned acquisition matrix of 240×240, with a spatial resolution of 1.8×1.8×1.8 mm³. TR/TE= 4800ms/229ms, FOV= 250×250mm2, flip angle= 90°, number of slices=100. The frequency of Spin lock pulses was 500 Hz. Total spin lock times (TSL) of 0, 20, 40, 60, 80, and 100 ms were used. For anatomical segmentation a 3D T1-weighted (T1W) scan and an axial T2-weighted (T2W) scan were also performed.

Data analysis: The whole brain T1rho maps were calculated using a mono-exponential decay model on a pixel by pixel basis. For each subject, T1W scan was processed using FIRST (fMRIB Integrated Registration and Segmentation Tool) for subcortical brain segmentation using Bayesian shape and appearance models to localize each subject’s ROIs which included seven subcortical GM structures (hippocampus, caudate nucleus, putamen, globus pallidus, nucleus accumbens, thalamus and amygdala). Then Subject-specific ROI maps were linearly registered to each individual’s T1rho space. For each subject, the corresponding T1rho values and volumes of each ROI were extracted in native T1rho space by software on MATLAB 2013a programs.

Statistical analysis: Statistical analysis of the T1rho values obtained from each region was performed using SPSS 20 (IBM Corp). Using the independent-samples t-test, the volumes and T1rho values in the patient group were compared with those in the control group for ROI-based analysis. To investigate the underlying relation between the significantly altered T1rho values or volumes of subcortical structures and duration of epilepsy or age at epilepsy onset, linear regression analysis was performed. A significant difference was accepted if the p value was less than 0.05.

In the patient-control group comparison of ROI-based analysis, the right caudate nucleus, hippocampus and bilateral amygdala of right mTLE patients exhibited a significantly increase inT1rho values, while the left mTLE patients only showed significantly elevated T1rho values in left amygdala. The volumes of subcortical GM showed apparent atrophy of the right hippocampus in right mTLE patients. However, neither the volumes nor T1rho values of the structures appeared correlation with the duration of disease or age of onset.This is the first time to use a ROI-wise analysis by atlas-based segmentation to investigate the variations in T1rho values of subcortical GM structures among temporal lobe epilepsy patients. In general, reduced volumes and increased T1rho values, although not significant, were universally observed in subcortical structures in mTLE patients. All of these regions show neurodegeneration in previous studies. Ipsilateral hippocampus and amygdala of mTLE patients exhibited significantly increased T1rho values, which were histologically confirmed gliosis and neuronal loss. However, we failed to observe an association between T1rho changes in subcortical GM and disease progression, may be due to limited sample size. These results partly reflect the ability of T1rho imaging to detect early neuronal loss and molecular changes related to epilepsy, and demonstrate the feasibility of ROI-wise analysis by atlas-based segmentation of T1rho imaging among mTLE patients.