| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
University of California, San Francisco San Francisco, CA 94117
I read with interest the article by Wasserman et al. in the August 2001 issue of the American Journal of Roentgenology [1]. The article discussed the reliability of normal MR imaging findings for excluding the diagnosis of vasculitis of the central nervous system. The authors concluded that angiography might still be necessary to render an accurate diagnosis of central nervous system vasculitis. It is surprising that AJR carried this article, because it has, I believe, several weaknesses.
How do the authors explain their findings when four research groups have reported a very high sensitivity for MR imaging [2,3,4,5]? Sensitivity approaching 100% has been reported, even though MR imaging may not reveal all abnormal brain regions [5].
It appears that the authors did not reach the minimum acceptable quality for MR imaging in the diagnosis of central nervous system vasculitis. Figure 1A, for instance, contains motion artifacts. I cannot be certain that the MR images have normal findings because some foci of hyper signal intensity are visible both in the putamen and around the posterior horn of the left ventricle in figure 1A and in the left insular lobe in figure 2B. In fact, the figures shown in the article are spin-echo T2-weighted images.
Proton density-weighted images are much more sensitive to brain abnormalities because the lesions will show hyper signal foci when the cerebrospinal fluid remains gray or hypointense, depending on the TE value. More recently, use of fluid-attenuated inversion recovery (FLAIR) sequences in conjunction with routine spin-echo sequences has been shown to increase the sensitivity of MR imaging. With FLAIR, signal intensity from the cerebrospinal fluid is suppressed and the gray scale is extended. FLAIR is particularly sensitive for detecting periventricular, cortical, and other supratentorial abnormalities whose bright signal intensity from cerebrospinal fluid on conventional T2-weighted images can lead to partial volume averaging. Recent work has shown that FLAIR is particularly important when findings on routine long-TR and long-TE T2-weighted spin-echo imaging are equivocal or normal [6].
Other imaging techniques, such as perfusion- and diffusion-weighted imaging and spectroscopy, are routinely available at most institutions to enhance lesion detection and define brain regions at risk in central nervous system vasculitis. It is surprising that the authors, working in a hospital that is part of a large university-based program, limited their study to T1- and T2-weighted spin-echo imaging. Does their MR system have the capability of performing FLAIR imaging and perfusion- and diffusion-weighted sequences?
The article would have been better and more useful if it had, in fact, reported on the reliability of normal findings on T1- and T2- weighted spin-echo MR imaging for excluding the diagnosis of vasculitis of the central nervous system.
References
Johns Hopkins University School of Medicine Baltimore, MD 21287-2182
Dr. Guermazi points out several weaknesses in our article on MR imaging of vasculitis of the central nervous system [1]. However, most of these points were addressed clearly in our article. We will respond to each concern.
First, Guermazi asks how we explain our findings when four research groups have reported a very high sensitivity for MR imaging. As discussed in our article, Greenan et al. [2] reported 100% sensitivity for MR imaging in detecting central nervous system vasculitis. Despite this high sensitivity, the authors described 12 of 33 vascular distributions with angiographic evidence of central nervous system vasculitis as having no lesions on the corresponding MR images, suggesting that normal findings on MR studies are possible when angiographic findings are positive. Furthermore, reports of the high sensitivity of MR imaging for detecting central nervous system vasculitis have been in small populations (i.e., n = 11, 7, and 7 [2,3,4, respectively]). The value of our article was that our findings were different from those of other research groups.
Guermazi is concerned that the MR imaging findings are not normal. Given the constraints of the journal, we were able to publish only a limited number of images [1]. We believe that we selected a fair representation of each patient's MR images given those limitations. All MR images with findings deemed normal were interpreted prospectively as such. They were subsequently reviewed by attending neuroradiologists who were not made aware of the patients' clinical data or results of other radiologic studies, as we outlined in the article. Guermazi should realize the danger of interpreting a single MR image without the benefit of the entire study. Some of the questionable lesions the letter points out are areas of partial volume averaging.
Guermazi underscores the absence in our article [1] of proton density—weighted and fluid-attenuated inversion recovery (FLAIR) images, as well as those of other imaging techniques including perfusion- and diffusion-weighted imaging and MR spectroscopy. As we discussed in the materials and methods section, proton density—weighted images were obtained for all six patients who had negative MR imaging findings and positive findings on angiography. We chose to print only T2-weighted images.
As discussed in our article [1], FLAIR imaging was not available because of the retrospective nature of the study. In our discussion, we addressed this limitation, pointing out that, because vasculitic lesions are typically located in deep white matter, FLAIR may not increase the sensitivity of MR imaging for detecting these lesions, and that FLAIR is most advantageous in the detection of peripheral lesions. We agree, however, that the lack of FLAIR images remains a limitation of our study, although its implications are unclear.
We also addressed the lack of advanced imaging techniques in our series [1]. As noted, the lack of diffusion-weighted imaging likely did not affect the outcome of this study because in none of our patients was the MR imaging performed less than 24 hr after the onset of symptoms, making parenchymal changes detectable by diffusion-weighted imaging but not by T2- or proton density—weighted sequences unlikely. We agree that perfusion imaging may play an important role in the evaluation of central nervous system vasculitis, as suggested in our article. Reports of perfusion imaging in central nervous system vasculitis are sparse, and its role, although promising, remains to be tested. We have implemented perfusion imaging as part of our current evaluation of central nervous system vasculitis, but this was not included in our article [1] because of its retrospective nature.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
