Imaging of Stroke Pathology without Predefined Gold Standard

Imaging of patients with acute ischemic stroke is fascinating, because is offers the opportunity to better understand and treat a disabling disease with variable causes and pathophysiology. We appreciated computed tomography (CT) in the early 1970s as the first method demonstrating the diseased brain in vivo. In stroke patients, CT detects hyperattenuating masses and arteries, hypoattenuating brain tissue, and swelling of brain tissue without a change in x-ray attenuation, and we diagnose hemorrhage, intra-arterial thrombus, ischemic brain edema, and tissue swelling due to vasodilatation Because stroke is fortunately often not a fatal disease, autopsies are rarely done and direct comparison between the image and the stroke pathology is seldom possible. Therefore, we cannot be quite sure that stroke patients without hyperattenuating substrates have really no intracranial hemorrhage. Subtle hemorrhagic tissue transformation of tiny amounts of blood in the cerebrospinal fluid may remain undetected by CT. Stroke patients without hypoattenuating tissue on CT may have no ischemic edema and no irreversible tissue damage, but it is hard to prove. We are missing a gold standard to determine the sensitivity and specificity of brain tissue imaging for certain pathologies. The use of imaging modalities, e.g. the comparison between CT and magnetic resonance imaging (MRI), is of limited value as long as the specificity of findings is undetermined. The recent paper of Jaillard et al. [2] is an example. Jaillard et al. intended 'to study the pathophysiology of early CT signs'. They compared CT and MRI findings in 16 patients within 29 h of stroke onset. The CT was obtained 30 min to 24 h earlier than MRI (mean 4:48 h) -interestingly, a common finding in papers comparing both modalities Why is MRI delayed in most studies compared to CT? Based on this comparison with different intervals between both assessments and at different timepoints of the evolving ischemic tissue changes, the authors conclude that early CT signs may show either preinfarction or ischemic penumbra. In particular, they base this conclusion on the observation of patients who had a follow-up T2-weighted MRI (T2WI) that matched the diffusionweighted image (DWI), but was smaller than the early CT findings. The authors do not discuss whether T2WI is really the gold standard for irreversible tissue damage and they do not describe the methods of their follow-up imaging. Moreover, they do not describe which 'CT sign' appeared reversible. It would make sense to find that brain tissue swelling due to compensatory arterial dilatation is reversible if perfusion pressure increases, whereas hypoattenuating tissue represents ischemic edema and irreversible injury Unfortunately, the