Calculated Repositioning Of Single-voxel:a New Method In Longitudinal~1H MRS Studies

Proton magnetic resonance spectroscopy (MRS), allowing individual studies over time non-invasively to evaluate disease progression and response to therapy effect, is frequently used in clinical application. In these longitudinal studies involving repeat scans on the same volunteer, reproducible measurements and minimizing errors of metabolite concentrations in MRS are necessitated to ensure the correctness in assessing the effect of the disease or treatment.Several factors, for instance systematic errors in scanner performance, test-retest effects or repositioning effects, are found to reduce the reproducibility among MRS data. A main interference arises from the imprecise re-localization of VOI. Two conventional methods have been employed to mitigate the VOI repositioning problem. One way is the so-called visual repositioning method in which the volunteer is tried to keep in the same position in scan machine in each follow-up scan as in the first scan. It is feasible for scans performed successively or have short time interval but the scans with long-time interval. The other way generally involves acquiring high-resolution scout images and visually locating VOI. It plays a big part only in scans that subjects are placed in a position without any head rotation. Either of these two ways is of high dependence on operator intervention and time consuming.A new application of the efficient registration algorithm, used to reposition voxels for obtaining in vivo1H MRS data, is presented here. This automatic procedure, based on the prospective registration of brain MRI images, allows the direct transformation (translation and rotation) of the voxel of interest (VOI) and the acquisition of the volume covering nearly identical brain region in longitudinal exams. To study the effect of the repositioning approach, we evaluate the tissue classes' overlap and the reproducibility of metabolite concentrations and concentration ratios in the measurement. Meanwhile, statistical analysis for short-term measurement variability, coefficients of variation (CVs) and reproducibility coefficients are performed when this automatic re-localization method is compared to visual repositioning procedure.