Cervical spondylotic myelopathy: diagnostic performance of radiologists with varying levels of experience in comparing MR images acquired using field strengths of 1.5 and 3 Tesla

Abstract

Background Although both 1.5-Tesla (T) and 3-T magnetic resonance imaging (MRI) are widely used for the diagnosis of cervical spondylotic myelopathy (CSM), there have been no comparative studies. Purpose This study compared the performance of both imaging modalities for diagnosing CSM among radiologists with varying experience levels. Material and Methods Four independent readers with different levels of experience reviewed 1.5-T and 3-T MR images of 79 patients with cervical spondylopathies. For both field strengths, images were evaluated for the depiction of intramedullary T2 hyperintensity of the cord according to disc level using a 5-point scale: -2 = definitely absent; -1 = probably absent; 0 = equivocally absent or present; 1 = probably present; and 2 = definitely present. The score was transformed into a 3-point certainty scale, which converted to absolute value (0 = equivocal, 1 = probable, 2 = definite lesion). Linear mixed model statistics were used to compare the depiction and certainty scale between 1.5-T and 3-T images. Inter-observer agreement was assessed by using Kendall's W statistics. Results Inter-observer agreements among the four readers were 0.718 for 1.5-T and 0.784 for 3-T MR images. Diagnostic accuracy of each reader was slightly increased using 3-T and varied regardless of level of reader experience. The certainty of CSM diagnosis was significantly improved using 3-T field strength. Equivocal lesions were significantly decreased in 3-T MRI in all readers. Conclusion Compared with 1.5-T MRI, 3-T imaging increased the certainty of the lesion and decreased the number of equivocal lesions in patients with CSM. Diagnostic accuracy was not affected by levels of reader experience.