Automatic quality assessment protocol for MRI equipment

The authors have developed a protocol and software for the quality assessment of MRT equipment with a commercial test object. Automatic image analysis consists of detecting surfaces and objects, defining regions of interest, acquiring reference point coordinates and establishing gray level profiles. Signal-to-noise ratio, image uniformity, geometrical distortion, slice thickness, slice profile, and spatial resolution are checked. The results are periodically analyzed to evaluate possible drifts with time. The measurements are performed weekly on three MRI scanners made by the Siemens(TM) Company (VISION(R) 1.5T, EXPERT(R) 1.0T, and OPEN(R) 0.2T). The results obtained for the three scanners over approximately 3.5 years are presented, analyzed, and compared. (C) 1999 American Association of Physicists in Medicine. [S0094-2405(99)00812-3].