Comparison of MERGE and axial T2-weighted fast spin-echo sequences for detection of multiple sclerosis lesions in the cervical spinal cord

摘要

OBJECTIVE. The purpose of our study was to compare axial multiple-echo recombined gradient echo (MERGE) with axial T2-weighted fast spin-echo (FSE) imaging for the detection of multiple sclerosis (MS) lesions in the cervical spinal cord on MRI. MATERIALS AND METHODS. Twenty-nine cervical spine MRI studies of patients with MS lesions and 29 control cases were reviewed retrospectively. Two blinded neuroradiologists independently assessed randomized axial MERGE and axial T2-weighted FSE sequences from each study, documenting the location and number of cord lesions, the degree of confidence in calling each lesion, and the presence of artifacts. The reference standard was determined by an unblinded consensus review of all sequences performed for each case, with lesions considered present if detected on two or more sequences. Lesion detection rates and conspicuity, false-positive findings, and reader confidence and artifact scores were compared for the sequences, and interreader agreement was assessed. RESULTS. Eighty-three lesions were assessed. The mean true-positive lesion detection rate was 87% (95% CI, 79-93%) with MERGE and 67% (60-75%) with T2-weighted FSE, with interreader positive agreement scores of 74% and 75%, respectively. A greater number of false-positive findings were seen with MERGE for both the MS and control cases. Average confidence and artifact scores were similar for both sequences. Subjectively, lesions were more conspicuous in 21 cases with MERGE and four cases with T2-weighted FSE and were equally conspicuous in four cases. CONCLUSION. MERGE and T2-weighted FSE sequences are complementary. MERGE provided greater sensitivity for cord lesions whereas axial T2-weighted FSE provided improved lesion specificity. Further investigation is required to assess the clinical impact of MERGE in the diagnosis and management of MS.