Abstract
The present study was performed to investigate methods of imaging, image evaluation, and quantitative analysis for diagnosis of brain atrophy on brain magnetic resonance imaging (MRI). For quantitative analysis, it is important to obtain optimal MR images by an optimal imaging sequence. Signal-to-noise ratio (SNR) is an excellent index for evaluation of MR image quality. However, it is difficult to measure correct SNR in clinical MR images using parallel imaging (PI) techniques. We developed a method to measure SNR on clinical MR images using the PI technique by the modified identical regions of interest (ROI) method. We developed a semi-automated volumetry method for the cerebrum, cerebellum-brain stem, and temporal lobe on brain MRI using texture and anatomical information on brain MR images. It is possible to evaluate cerebrum volume on images of different qualities obtained by spin echo or gradient echo sequences. Image quality obtained by each imaging sequence was evaluated by our SNR measurement method and the volumetry method consisting of optimal image processing for image quality. It is useful to evaluate MR image quality and to measure volume semi-automatically for diagnosis of brain atrophy in brain MRI.
