What COVID-19 Brain Abnormalities Could Mean

May 20, 2020
Whitney J. Palmer

Diagnostic Imaging poses five questions about recent COVID-19 and neurological research to UCSF Radiology Chair Christopher Hess, M.D.

Last week, researchers from Turkey published an article in Radiology that outlined brain abnormalities detected on MRI scans of patients, positive for COVID-19, who are also in the intensive care unit (ICU). While those results garnered significant attention throughout the industry, the findings also prompted speculation from many in the specialty about the strength of the study and what information it actually provided about the neurological impact of COVID-19.

To gain some perspective on how radiologists – and other healthcare providers – should view the study results, Diagnostic Imaging spoke with Christopher Hess, M.D., Ph.D., distinguished professor and chair of radiology and biomedical imaging at the University of California-San Francisco. He offered insights on what the findings of the investigation mean and the potential value they provide.

Related Content: MRI Shows Brain Abnormalities in Some COVID-19 Patients

Diagnostic Imaging: What are your thoughts on this research?

Hess: There has been a lot of interest in whether and how COVID-19 affects the central nervous system, and there have been different ways of going about that – studying clinical neurological symptoms, studying cerebrospinal fluid (CSF), and studying imaging. There have been a lot of studies that are mostly case reports or single isolated events that describe what findings might be present in COVID-19. But, the reality is that, without a systematic study, it is difficult to understand what actually is caused by COVID-19 and what is an epiphenomenon of COVID-19 just from being sick and being in the intensive care unit (ICU) and having a severe illness.

Diagnostic Imaging: Given that, then, what does this study actually tell us that could be potentially actionable?

Hess: This study was unique in the sense that it was a relatively large cohort. They took 749 patients, and apparently 200 of them – about 30 percent – ended up in the ICU. About 50 of those – around 20 percent – had neurological symptoms, and 27 of that group went for an MRI. So, the size of the group is good. But, anytime you have a design like that where you’re not enrolling everyone for an MRI, you’re not studying everyone. You have to be aware of bias in selection to make sure what you’re seeing on the MRI reflects the greater group, not just the selected small group.

We know through all the various existing studies that there is a neurological impact of disease. We don’t know the precise anatomy or physiology of that at all. This study showed several imaging findings, but it didn’t really answer the question of why these findings occur. In particular, the number of patients who had CSF – which is really critical to the evaluation of a lumbar puncture is really critical to the evaluation of infectious illness. And, that number was very small, and there was no infection in the CSF. So, again, we’re stuck with the question of whether this is an epiphenomenon of people being sick or is it directly related to COVID-19? We didn’t really answer that because we were not able to directly show COVID in the CSF of any of the patients.

Among the epiphenomenon, there are several things that could be going on. We’ve all heard about the vascular risk or the vascular thrombotic risk factors that some have articulated. In this study, they did have two patients – one with large vessel stroke and one with dural sinus thrombosis. But, would these findings described in the article be related to some underlying vascular cause? Certainly, they could. Or, it could be due to some outcome related to the physiology of COVID rather than direct COVID. Among the possibilities, there’s direct neural tropism or direct brain invasion by the bug. That’s relatively unlikely. And, in this case, given the CSF, I still think it could be vascular.

And, remember, sick patients have all kinds of metabolic derangements. They often have seizures, and in critically ill patients with COVID-19, in particular, they certainly have respiratory issues. Relatively pure hypoxia could certainly change the physiology of the brain such that gives rise to some of these findings. 

I think this study shows an association between the virus and these findings much in the way that prior clinical studies have shown some association of neurologic symptoms with COVID-19. This shows the association of some imaging findings with COVID. But, there’s no direct causal link. The authors made the statement that prior SARS viruses have been shown to have a neurotropism and have even demonstrated in the autopsy samples of some brains. But, that hasn’t been shown for COVID-19, so we just don’t know.

Diagnostic Imaging: Going forward for radiologists and the providers with whom they collaborate, what’s the value of this study’s findings?

Hess: It’s important to be able to recognize so we don’t make a fallacy of misattribution – of attributing the findings you see to some other disease process. Just knowing that this has been seen in patients with COVID before may provide some diagnostic certainty in terms of understanding what the cause of the imaging finding is. I think there is some scientific value here, as well, and understanding what types of neuro-imaging findings are present in patients with COVID.

I also think, secondarily, this shows the real value of multi-modal evaluation of neurologic disease. It has to be really combined with understanding of symptoms, imaging findings, and CSF together. The CSF especially is a critical piece that is lacking in most of the patients included in this study.

Diagnostic Imaging: Are these findings indicative of a need to scan more patients, particularly with MRI?

Hess: The reasons for scanning a patient are really driven by the clinical situation. A patient in the ICU with neurologic deficits – whether they have COVID-19 or not – should probably have an MRI to evaluate why they’re having neurologic deficits. Will it increase the number of scans or should it increase the number of scans that are being done in COVID patients? I don’t think so because there’s no direct treatment of any of the imaging findings that were shown here. There’s no indication for doing this on a routine screening basis to evaluate them.

For example, if we saw hydrocephalus routinely and unsuspected in some of these patients, then, hydrocephalus would be a treatable disorder. You could make an argument that we should do more imaging so we won’t miss hydrocephalus. But, for most of the imaging findings they described, aside from the large vessel stroke and the sinus thrombosis, the treatment is not driven by imaging findings.

Diagnostic Imaging: What are the take-away messages? What would you say to the neuroradiologists looking at these images, wondering how they should proceed?

Hess: The take-home message here is that there’s still so much uncertainty. They need to continue to be curious – to look for the MRI findings in this disorder. Maybe if we accumulate more data, we’ll be able to understand if they’re directly related, causally related, or if they’re indirectly related and just an epiphenomenon of having severe respiratory illness.

Related Content:

Neurology MRI