In this study, we compared a fast fluid-attenuated inversion recovery (fast-FLAIR) sequence to conventional spin-echo (CSE) in the evaluation of brain MRI lesion loads of seven patients with clinically definite multiple sclerosis. Interleaved CSE (3000/20, 5 mm contiguous axial slices) and fast-FLAIR (9000/150/2200, 5 mm contiguous axial slices) sequences were performed on a 1.0 T machine. Lesions were counted consensually by two observers and then segmented independently by two other observers using a local thresholding technique, with subsequent manual editing in the case of poorly defined lesions. Four hundred and two lesions were detected in at least one of the two sequences: 128 were seen only on fast-FLAIR, 17 only on CSE. Forty-one lesions were larger on fast-FLAIR, while no lesion was considered larger on CSE. The numbers of periventricular (P = 0.05), cortical/subcortical (P = 0.02) and discrete (P = 0.05) lesions detected using fast-FLAIR were higher than those detected using CSE. The median lesion load was 7185 mm3 on CSE and 8418 mm3 on the fast-FLAIR, the average being 18% (range = 11.6-29%) higher on the fast-FLAIR images. Lesion contrast ratio was higher for lesions on the fast-FLAIR than on the CSE sequence (P < 0.0001). The percentages of poorly defined lesions which needed manual editing after the local thresholding technique was applied and the total time needed for the measurements were lower (P < 0.001) when fast-FLAIR images were used compared with CSE. This resulted in a reduced inter-rater coefficient of variation in measuring lesion volumes. Our data indicate that fast-FLAIR sequences are more sensitive than CSE in detecting multiple sclerosis lesion burden and that fast-FLAIR is a promising technique for natural history studies and clinical trials in multiple sclerosis.