Introduction
HIV associated dementia (HAD) leads to cerebral metabolic changes in frontal white matter and subcortical grey matter [1–6]. Within these regions of interest (ROI), N-acetylaspartate (NAA), choline (Cho), and creatine (Cr) have typically been studied using single voxel 1 H MR spectroscopy (MRS). 2D CSI 1 H MRS provides greater localization with smaller voxel sizes of previously studied ratios as well additional metabolites such as lactate (LAC). We evaluated cerebral metabolite levels within the basal ganglia (BG), a subcortical grey matter region, and subcortical white matter (SWM) ROIs in HIV+ patients compared to controls using 2D CSI 1 H MRS.
Methods
Forty-four HIV+ patients (31 males, 13 females, 41 ± 2 yrs) and 10 seronegative controls (4 males, 6 females, 47 ± 2 yrs) underwent standard physical, neurological and neuropsychological evaluation. HIV+ patients were classified by the Memorial Sloan-Kettering system for HIV associated dementia complex (HAD) and grouped as asymptomatic (n=24) and symptomatic (n=21). MR apparatus was a 3 T Siemens scanner equipped with the standard clinical quadrature head coil. Anatomical images were acquired by axial T1 weighted MPRAGE sequence (TR 1620 ms, TE 3.87 ms, 1 mm slice thickness, FOV 25×25 cm). Spectroscopy images were acquired with a SE with 2D phase encoding and outer volume saturation pulses for lipid suppression sequence (TR 2000 ms, TE 135 ms, n=3, 20 mm slice thickness, FOV 20×20 cm, voxel size 12.5 × 12.5 × 20). Voxels overlapping bilateral BG and SWM were analyzed. The area under the peak was measured for quantification of cerebral metabolites and calculated ratios of NAA/Cr and Cho/Cr were determined. Metabolic ratios between HIV+ groups were compared to controls by Student's t-test (p<0.05). LAC was considered present if an observed peak was greater than 1 standard deviation above noise. A paired t-test analysis was performed to determine significance across groups for the presence of LAC (p<0.05).
Results
The percentage of visible lactate peaks in BG was significantly greater in HIV+ patients compared to controls (p<0.05). Although not significant in SWM, visible lactate peaks were more prominent in HIV+ groups. Within the SWM but not BG the NAA/Cr ratio was also significantly reduced in HIV+ patients compared to controls (p<0.05). In contrast, no consistent trend was observed for Cho/Cr ratio for either ROI.
Discussion
Our results demonstrate that high field multi voxel 2D CSI 1H-MRS provides a more accurate understanding of cerebral metabolite changes in HIV+ patients. The reduction of NAA/Cr ratio seen in SWM may be due to neuronal loss from mitochondrial dysregulation. Gliosis, as measured by the Cho/Cr ratio, was not observed within HIV+ patients. The presence of LAC observed within BG in HIV+ patients compared to controls may reflect the presence of macrophages within the brain. These results support converging evidence that HIV is concentrated in certain areas of the brain with neuronal loss and inflammation leading to HAD. Furthermore, 2D CSI 1H-MRS, particularly in detecting LAC within BG, may serve as a biomarker in classifying HIV+ dementia.