High-resolution echo-planar imaging of the brain: Is it suitable for routine clinical imaging?

To assess the value of echo-planar imaging (EPI) for ultrafast routine clinical diagnostic MRI, we compared four different EPI sequences with conventional T-2-weighted spin-echo images on a commercial clinical imager. We examined 25 randomly selected patients who posed different clinical questions. The images were interpreted by two experienced neuroradiologists blinded as to the sequence used. Image quality and diagnostic certainty were evaluated and the main diagnosis established from the EPI study was compared to that obtained from the T-2-weighted images. Finally, EPI-and T-2-based diagnoses were compared with the diagnosis resulting from a complete MRI examination. Apart from one sequence that was generally rated low as regards both diagnostic certainty and image quality, the EPI sequences were comparable to each other, but inferior to the T-2-weighted images. However, two EPI sequences gave better diagnostic results than T-2-weighted images compared to the full MRI examination. Gradient-echo EPI was particularly sensitive to haemorrhagic lesions. All normal cases were correctly identified on EPI studies. Only two pathological cases were missed; both had isolated cranial nerve lesions. The absence of false-positive results and the high sensitivity to ischaemic and mass lesions mean that EPI can be used for ultra-fast screening. However, from these initial studies, EPI seems unsuitable for neuroradiological investigation of patients who may have subtle lesions whose detection requires either special sequences or administration of contrast medium. EPI can nevertheless be used in addition to high-resolution T-1-weighted images and may replace T-2-weighted spin-echo sequences for special indications.