Subcortical Low-Intensity Areas on T2-Weighted Images: An Uncommon Finding with a Common Explanation?

Brian Bowen

In this issue of the AJNR, Lee and colleagues describe a finding, referred to as a subcortical low-intensity (SCLI), on T2-weighted fast-spin-echo and fluid-attenuated inversion recovery (FLAIR) images in patients with meningitis, viral encephalitis, or leptomeningeal metastases. For these clinical diagnoses, the frequencies of observation of the finding on retrospective review of the MR images were 8.6% (five of 58 patients with meningitis), 23.7% (nine of 38 patients with viral encephalitis), and 23.8% (five of 21 patients with leptomeningeal metastases). Interestingly, cortical hyperintensity was observed in most patients (73.9%) and leptomeningeal enhancement was observed in all of the 19 patients with SCLI areas; the percentages were approximately double those obtained in the cohort of patients without SCLI areas.

The striking features of SCLI lesions are the following: 1) focal-to-diffuse involvement of the subcortical white matter (unilateral, frontal and parietal lobes), 2) reversibility in patients with meningitis or encephalitis, 3) isointense-to-hypointense signal on isotropic diffusion-weighted images (DWI), and 4) decreased apparent diffusion coefficients (ADCs) (compared with values obtained from the contralateral normal white matter) in half of the lesions evaluated with DWI (mean decrease of approximately 9%). On the basis of these observations, the results of a single biopsy that showed focal myelin pallor, and the results of studies by other investigators that have revealed the uncommon occurrence of SCLI-like areas in a variety of other pathologic conditions (eg, early cortical ischemia, multiple sclerosis, Sturge-Weber syndrome), the authors conclude that SCLI is a nonspecific sign of meningeal and cortical disease (although encephalitis is not an exclusively cortical disease) and that SCLI may be caused by a transient increase in the amount of free radicals.

Why free radicals? Free radicals, which are principally reactive oxygen intermediates (eg, superoxide radical, hydrogen peroxide, hydroxyl radical) are transient and have been implicated in numerous pathologic processes, including brain edema, ischemia and infarction, meningitis, encephalitis, and malignancy. The last includes the processes that Lee and colleagues reported. Furthermore, the oxygen free radicals are paramagnetic because of their unpaired electrons, and paramagnetic species can shorten the T2 relaxation time, resulting in SCLI. The flaw in this reasoning, of course, is that no cause and effect has been demonstrated. The implication of oxygen free radicals seems to be based on “guilt by association,” and the conclusion becomes problematic when the authors attempt to explain why no cortical low-intensity areas are present and how the temporal expression of short-lived oxygen free radicals is related to the time course of SCLI, which may be observed for days to weeks after the onset of symptoms.

What about other potential causes of SCLI? The authors consider two causes in detail and reject both: the accumulation of nonheme iron and a structural change in the subcortical white matter. Abnormal iron accumulation in the extrapyramidal system is generally accepted as the cause of excessive T2 shortening in the deep gray matter nuclei in several neurodegenerative diseases such as Parkinson disease and the Parkinson-Plus syndromes. T2 shortening in the subcortical white matter of patients with cerebral ischemia and infarction has also been attributed to iron deposition (secondary to disruption of axonal transport of iron); however, the link between T2 shortening and iron accumulation in the subcortical white matter is unproven, as Lee and colleagues note. Although iron accumulation is an intriguing possibility from the standpoint of the DWI findings, Lee and colleagues reject it as a cause of SCLI, primarily on the basis of conflicting reports in the literature, the transient nature of SCLI, and the lack of detectable iron on Perls staining of the biopsy sample obtained from the cortex and subcortical white matter in a patient with viral encephalitis.

The second postulated cause of SCLI that the authors rejected may not be so easily dismissed without additional histochemical and ultrastructural information. The statement that “no obvious evidence of structural abnormality” is present on histologic examination is not exactly true because myelin pallor was observed in the only biopsy specimen obtained. Although no histopathologic evidence of myelin catabolism (as it occurs in Wallerian degeneration) was reported, could subtle ultrastructural changes in myelin be occurring in SCLI areas? It is curious that the subcortical white matter signal intensity changes on T2-weighted images in conditions as disparate as Wallerian degeneration and Sturge-Weber syndrome pass through a stage of hypointensity that is attributable to myelin metabolism. These conditions and the SCLI reported here differ, though in the time course of signal intensity changes (longer in Wallerian degeneration and Sturge-Weber syndrome) and in the eventual white matter destruction that occurs in Wallerian degeneration.

With their study, Lee and colleagues raised more questions than they answered. This is often the case in science. The answer to the question of whether the relatively uncommon finding of SCLI in various pathologic conditions can be explained by a common mechanism must await more thorough and hypothesis-driven investigation. Certainly, more tissue specimens should be obtained and analyzed, and conceivably, an animal model may need to be developed to cytochemically demonstrate negative oxygen intermediates by using published techniques. The results of such laboratory investigations are more likely to answer the questions raised in this article than additional clinical MR measurements of parameters such as magnetization transfer ratios, T2 relaxation times, or metabolite concentrations from in vivo proton spectroscopy. Before embarking on extensive investigations, though, one should also ask whether this nonspecific MR sign is of sufficient clinical benefit to be worth the effort.