Relevance of neuroimaging in publications on COVID-19 and stroke

Dear Editor:

The COVID-19 pandemic is a challenge for healthcare systems worldwide, and there has been great interest in describing and explaining the manifestations of the disease. The volume of scientific articles published as a result has challenged journals, editors, and reviewers alike.¹ In the case of neurological diseases and stroke, the pandemic has also stimulated the medical community to publish their experience with COVID-19, with a view to helping professionals across the globe to provide care as quickly as possible. A recent review suggests that while COVID-19 does not increase the risk of stroke, it does considerably increase mortality.²

However, all parties involved in scientific publication should seek to guarantee the greatest possible accuracy and rigour, with the ultimate aim of providing the best possible care for our patients. With this goal in mind, we would like to comment on 2 studies recently published in this journal.

The valuable study by Barrios-López et al.³ presents a series of 4 patients with acute stroke treated at the authors’ centre, and provides detailed clinical, laboratory, and imaging data. However, we would like to make some remarks on the neuroimages presented. Figure 1 presents a series of images that the authors describe as a “cerebral perfusion study.” However, the images presented are not pre-processed CT images or standard parametric maps, but rather a colour-coded representation resulting from automated post-processing. The authors do not indicate which software or algorithm was used to generate these images; these data are essential for correctly interpreting the images. Furthermore, while the caption refers to “increased mean transit time,” the images themselves indicate Tmax and CBF. Essentially, Tmax is used to quantify the volume of brain tissue at risk of infarction or ischaemic penumbra,^4,5 but it should not be mistaken for mean transit time (MMT), a parameter used by software tools or algorithms other than those that the authors presumably used. Improper use of cerebral perfusion terminology may result in incorrect interpretation and application of the data presented. Furthermore, the description of figure 2 is, to say the least, unconventional from a radiological viewpoint: the authors describe image A as an “infratentorial section,” but the image shows a large part of the temporal and occipital lobes, which may be confusing for less experienced readers.

Aguirre et al.⁶ present an excellent description of an interesting case of contrast-induced encephalopathy following mechanical thrombectomy to treat occlusion of the left middle cerebral artery. In this case, figure 1 provides CT perfusion parametric maps with time to peak and cerebral blood volume data, enabling readers to correctly interpret and understand the case. However, including a key explaining the colours used and their correlation with quantitative volume and time data would have enabled more precise interpretation of the images, since the colour space may be modified during post-processing; assessment and quantification is therefore difficult without a colour scale. Regarding image H (24-h CT scan), the authors note the presence of “signs of oedema,” which are not evident, in our opinion. The image does not show white matter hypodensity, loss of grey-white differentiation, or mass effect,⁷ which constitute the main signs of cerebral oedema on CT images. A more detailed description of the localisation of contrast extravasation (probably the basal ganglia and cerebral cortex) would have also increased the precision of the radiological data presented and provided more complete data on the pathophysiology of the patient’s symptoms.

Our comments are intended to emphasise the importance of radiological images in scientific articles and to underscore the role of neuroradiologists in providing precise, expert descriptions and analyses of these images. Research should ideally be multidisciplinary, and radiologists should therefore participate in studies presenting radiological images.^8,9