A comparison of sagittal short T1 inversion recovery and T2-weighted FSE sequences for detection of multiple sclerosis spinal cord lesions

Abstract

Multiple sclerosis (MS) is the most common disabling CNS disease of young adults. MRI is routinely used for the detection of MS plaques in the brain and spinal cord. A significant portion of patients with MS demonstrates spinal cord lesions at the time of initial workup, and these lesions are an important part of the McDonald criteria for diagnosis. However, whereas brain imaging sequences are now fairly standardized, there continues to be debate about the optimal sequences for imaging the spinal cord. The short T1 inversion recovery (STIR) sequence has been shown in the current literature to improve lesion detection with its additive T1/T2 weighting, but current spinal cord imaging protocols from the Consortium on MS Center Consensus Guidelines do not include the STIR sequence. We demonstrate that not only do STIR sequences improve lesion detection when compared directly with conventional T2-weighted sequences, but that they also significantly improve lesion conspicuity, facilitating earlier positive diagnosis and management. Dedicated MR spinal cord imaging of twenty-nine sequential patients with clinically confirmed multiple sclerosis was retrospectively analyzed by two independent neuroradiologists in a novel study design. Sagittal T2-weighted and STIR sequence images from the same study for each patient were examined for MS plaques using a double-blinded review of individual images 'separated in time and space', such that STIR and T2 image pairs were never analyzed simultaneously. Number of lesions and lesion conspicuity for each lesion, using a subjective scale (1-5), were tallied for each sequence. Averages for each observer were compared using a paired t-test analysis for statistical significance, and assessment of inter-rater agreement was assessed using Cohen's kappa index. Significantly, more MS lesions were detected on STIR than on T2-weighted sequences for both observers (P = 0.001 and P = 0.005). In seven patients, the conventional T2 sequence detected no lesions at all, whereas STIR sequences showed significant cord involvement. Lesion conspicuity was also significantly better on STIR for both observers (P < 0.0005). This improved conspicuity leads to more uniform lesion detection. On the conventional T2-weighted sequence, there was a statistically significant difference in the number of lesions detected between the two observers (P = 0.003), but there was no statistically significant difference on STIR (P = 0.43). The kappa index showed greater interobserver agreement in both lesion count and lesion conspicuity on the STIR sequence as compared with T2. Short T1 inversion recovery sequence imaging not only significantly improves detection of MS lesions within the spinal cord, but also provides better contrast and conspicuity of visible lesions, creating a more confident diagnostic measure of MS extent and progression. Short T1 inversion recovery sequences of the spinal cord should be routinely obtained during initial and routine follow-up of MS.