OBJECTIVE The purpose of this study was to compare a combined gradient and spin-echo (GRASE) technique, which is a rapid T2-weighted imaging sequence, with conventional spin-echo (SE) sequences for imaging brain lesions. The GRASE sequences would allow increased patient throughput with potential cost savings and be useful in uncooperative patients without requiring echoplanar imaging techniques and specialized hardware. METHODS Conventional SE and GRASE T2-weighted images of 49 consecutive patients (20-86 years old) were reviewed independently by three neurora-diologists for the presence and characterization of lesions (most of which were nonspecific foci of hyperintensity within the white matter), gray-white matter differentiation, conspicuity of lesions, and periventricular signal abnormality. The MR studies were performed on a 1.0-T Siemens Magnetom Impact scanner, with the SE images obtained using a TR/TE of 2400/40 and the GRASE images obtained using a TR/effective TE of 4400/110. RESULTS The number of lesions detected that were 5 mm or larger in maximal diameter did not significantly differ among techniques. For lesions smaller than 5 mm, conventional SE T2-weighted images showed more lesions (p < .01). The SE images were better than the GRASE images for assessing gray-white matter differentiation, conspicuity of lesions, and periventricular signal abnormality. The two hypointense lesions were better assessed on the conventional SE images. CONCLUSIONS Although GRASE imaging may be potentially useful for rapid imaging of the brain, our experience shows it has a markedly diminished sensitivity for detecting lesions smaller than 5 mm in diameter. Currently, GRASE imaging should not replace the routine clinical use of conventional SE sequences.