There are different MRS approaches to measure N-acetylaspartate (NAA) in the medial temporal lobe (MTL) of patients with Alzheimer's disease (AD). While most studies have reported a reduction of NAA in this brain area, it is unclear which parameter would be the best candidate for a diagnostic tool in the clinical work-up of dementia.

33 AD patients and 23 healthy volunteers were examined by single-voxel MRS. A volume of interest (VOI) of 35x25x20mm was positioned in the left MTL in parallel angulation to include the hippocampus. PRESS-localized 1H spectra with TR/TE 2000/30ms and 2000/272ms were acquired at 1.5T on a clinical whole body MR system (Gyroscan Intera 1.5T, Philips Medical Systems) using a transmit/receive birdcage head coil suited for MRS and MRI. In order to assess the sensitivity, specificity and correlation of different metabolite ratios and absolute metabolite concentrations we measured NAA relative to total creatine (NAA/Cr), the absolute concentration of NAA (NAAabs) referenced to brain tissue water, and NAAabs multiplied with the amount of brain tissue (BT) in the 1H-MRS voxel (NAAabs*BT). Absolute concentrations and percentage of brain tissue were determined applying the unsuppressed water signal of the VOI as an internal reference and correcting for CSF contents by bi-exponential T2 relaxometry.

A comparison of NAA/Cr, NAAabs, and NAAabs*BT was performed. At a sensitivity of 80%, we found a specificity of 44% for NAA/Cr, of 68 % for NAAabs, and of 87% for NAAabs*BT. We observed only a low correlation of the absolute concentration of NAA with NAA/Cr (r=0.24) and no correlation of NAAabs with BT (r=0.14).

The metabolic ratio NAA/Cr of the MTL is only of limited value as a diagnostic tool for the discrimination of AD patients and healthy controls. NAAabs has a moderate specificity, while NAAabs*BT discriminates AD patients from controls sufficiently. NAAabs and brain atrophy are independent measures of brain pathology and supply different information. NAA/Cr does not reflect NAA concentration alone, which should be taken into consideration, when interpreting 1H-MRS data based on metabolic ratios.