Comparison Between Diffusion Tensor Imaging and Conventional MR Imaging Sequences in the Detection of Spinal Cord Abnormalities

We read with interest the articles by Renoux et al¹ and Facon et al,² respectively, in the October 2006 and in the June–July 2005 issues of the AJNR.

The authors evaluated the diagnostic accuracy of diffusion tensor imaging (by using the apparent diffusion coefficient and fractional anisotropy) in inflammatory diseases of the spinal cord¹ and in spinal cord compression.² In these 2 articles, diffusion tensor imaging was compared with T2-fast spin-echo (FSE)–weighted sequences. The authors found a higher sensitivity in the detection of spinal cord abnormalities with diffusion tensor imaging than with T2-FSE–weighted sequences in both articles.^1,2

We draw the authors’ attention to the previously published reports about the diagnostic accuracy of spinal cord abnormalities with conventional MR imaging sequences. We cite only 2 of them because of the restriction on the number of references. These reports showed that short τ inversion recovery FSE (STIR-FSE) sequences may have a higher sensitivity than T2-FSE–weighted sequences in the detection of spinal cord lesions.^3,4 Campi et al³ and Rocca et al⁴ concluded that STIR-FSE sequences had a higher sensitivity in the detection of demyelinating lesions. Furthermore, Campi et al showed a better demarcation and maybe a better sensitivity in the detection of spinal cord abnormalities in a group of patients with myelopathy of unknown etiology with STIR-FSE sequences.

We hypothesize that the best conventional sequence, with the highest sensitivity in the detection of spinal cord abnormalities, may be the STIR-FSE sequence. We think that future studies evaluating the diagnostic accuracy of diffusion tensor imaging in the detection of spinal cord abnormalities should include STIR-FSE sequences.