Perfusion CT in acute stroke: effectiveness of automatically-generated colour maps

Abstract

Objective: To evaluate the accuracy of perfusion CT (pCT) in the definition of the infarcted core and the penumbra, comparing the data obtained from the evaluation of parametric maps [cerebral blood volume (CBV), cerebral blood flow (CBF) and mean transit time (MTT)] with software-generated colour maps.

Methods: A retrospective analysis was performed to identify patients with suspected acute ischaemic strokes and who had undergone unenhanced CT and pCT carried out within 4.5 h from the onset of the symptoms. A qualitative evaluation of the CBV, CBF and MTT maps was performed, followed by an analysis of the colour maps automatically generated by the software.

Results: 26 patients were identified, but a direct CT follow-up was performed only on 19 patients after 24-48 h. In the qualitative analysis, 14 patients showed perfusion abnormalities. Specifically, 29 perfusion deficit areas were detected, of which 15 areas suggested the penumbra and the remaining 14 areas suggested the infarct. As for automatically software-generated maps, 12 patients showed perfusion abnormalities. 25 perfusion deficit areas were identified, 15 areas of which suggested the penumbra and the other 10 areas the infarct. The McNemar's test showed no statistically significant difference between the two methods of evaluation in highlighting infarcted areas proved later at CT follow-up.

Conclusion: We demonstrated how pCT provides good diagnostic accuracy in the identification of acute ischaemic lesions. The limits of identification of the lesions mainly lie at the pons level and in the basal ganglia area. Qualitative analysis has proven to be more efficient in identification of perfusion lesions in comparison with software-generated maps. However, software-generated maps have proven to be very useful in the emergency setting. Advances in knowledge: The use of CT perfusion is requested in increasingly more patients in order to optimize the treatment, thanks also to the technological evolution of CT, which now allows a whole-brain study. The need for performing CT perfusion study also in the emergency setting could represent a problem for physicians who are not used to interpreting the parametric maps (CBV, MTT etc.). The software-generated maps could be of value in these settings, helping the less expert physician in the differentiation between different areas.

Figure 1.

(a–d) Perfusion maps (a) demonstrating abnormalities on all four parameters in the head of caudate nucleus and parietal lobe on the left side: the mean transit time (MTT) map shows a wider alteration consistent with an area of ischaemic penumbra around an already infarcted area. The software-generated colour maps (b) show a good concordance, highlighting the already infarcted areas in red (head of caudate nucleus and parietal lobe) and the area of ischaemic penumbra in green. No alterations suggesting ischaemic lesions are present at initial transverse head CT image (c), while the CT follow-up image (d) shows two hypodense ischaemic lesions that correlate well with the perfusion maps. CBF, cerebral blood flow; CBV, cerebral blood volume.

Figure 2.

(a–c) The maps (a) show decreased cerebral blood volume (CBV) and cerebral blood flow (CBF) with elevated mean transit time (MTT) and TTP in pons Varolii on the left side, consistent with infarction. The software-generated colour maps (b) do not show any alterations in pons Varolii. CT follow-up image (c) confirms infarct in the pons Varolii on the left side.

Figure 3.

(a–d) The perfusion maps (a) demonstrating a small perfusion deficit in the basal ganglia on the left side; the change, however, correlates with a chronic stroke on direct scan. The software-generated colour maps (b) show a small green area which, in this case too, correlates with the chronic ischaemic area, thus not meeting the inclusion criteria. The initial CT image (c) shows the chronic ischaemic area in the basal ganglia on the left side (already visible in previous examinations). However, the CT follow-up image (d) shows the presence of a new lesion next to the chronic lesion. This was a false-negative perfusion CT study. CBF, cerebral blood flow; CBV, cerebral blood volume; MTT, mean transit time.