Schaefer PW, Souza L, Kamalian S, Hirsch JA, Yoo AJ, Kamalian S, et al. Limited Reliability of Computed Tomographic Perfusion Acute Infarct Volume Measurements Compared With Diffusion-Weighted Imaging in Anterior Circulation Stroke. Stroke. 2015

For anyone who was worried that acute stroke evaluation and management was getting too easy, square this circle: should we be making major decisions based on CT-based perfusion mapping or MRI-based restricted diffusion?



This question has major implications for clinical practice and acute stroke research. The broad availability and rapid acquisition of CT makes it an attractive means of inferring “physiologic data” during hyperacute stroke, and is the dominant mode of advanced imaging in acute stroke practice at large. That being the case, many patients are likely triaged for care, including endovascular reperfusion, based on the clinical scenario and CT perfusion data. Furthermore, CT perfusion thresholds have been used in several large-scale acute stroke treatment trials.

This contribution from Dr. Schaefer, et al calls these practices into question. The investigators designed their study around the practical questions, in their words: “Can CTP substitute for DWI in individual patient triage? And, can CTP replace DWI for stroke treatment trials?” They sought to answer these questions by recruiting consecutive acute stroke patients who were able to undergo CTA, CTP and MRI within hours of one another, a demonstrable proximal anterior circulation occlusion and a 3-month mRS score. They were able to analyze 55 cases meeting these criteria and scrutinized the correlation between CTP (more specifically, CBF and CBV maps) and DWI in terms of size and signal-to-noise ratio. Their results were clear if not surprising: visual inspection of perfusion abnormality (CBF better than CBV) correlated with DWI overall but not perfectly so, and the signal-to-noise was much higher with DWI, allowing for better delineation of the infarct volume.

As far as I’m concerned, these findings beg more questions than they answer, making this a very successful contribution! This puts data to that sense we all get when trying to digest Stroke and the Amazing Technicolor DreamCTP, but in practice and research design we have to seriously consider at what we are looking. The availability and rapidity of CT makes us hope and wish it is an acceptable surrogate for stroke lesion size, but these data suggest it is too imprecise a tool to make it the definitive approach as compared to MRI. I anticipate these data will positively inform future diagnostic and treatment trials for hyperacute stroke.