Is there a simple and accessible solution to improve acute infarct core imaging? The utility of steady-state CT angiographic source images obtained from a delayed phase acquisition

Abstract

Background: Early identification and quantification of core infarct is of importance in stroke management for treatment selection, prognostication, and complication prediction. Non-contrast computed tomography (CT) (NCCT) remains the primary tool, but it suffers from limited sensitivity and inter-rater variability; CT perfusion is inconsistently available and commonly blighted by movement artefact. We assessed the performance of a standardised form of CT angiographic source imaging (CTASI) obtained through addition of a delayed phase at 40 seconds post-contrast injection (DP40) following fast-acquisition CT angiography.

Methods: Contrast resolution between ischaemic and normal grey matter (GM) was compared qualitatively and quantitatively to NCCT. Using Alberta Stroke Program Early CT Score (ASPECTS), DP40 low density was compared to NCCT and venous phase CT perfusion source images (CTPSI) and to 24-hour NCCT ASPECTS in patients with timely endovascular recanalisation (Thrombolysis In Cerebral Infarction 2C/3).

Results: Seventy-four patients with a proximal middle cerebral artery or terminal internal carotid artery occlusion were included. The mean attenuation difference between ischaemic and normal GM increased from 4.86+/-3.12 HU (NCCT) to 9.30+/-3.14 HU (DP40) (p < 0.0001). Subjective assessment by two raters revealed that DP40 improved ischaemic tissue conspicuity in 39 to 41 (78-82%) of cases (kappa 0.805, standard error 0.108, 95% confidence interval: 0.593-1.000). The correlation between ASPECTS on baseline imaging and eventual 24-hour ASPECTS improved from R = 0.7197 for NCCT to R = 0.9875 for DP40 (z = 7.89, p < 0.0001). The correlation between DP40 and venous phase CTPSI ASPECTS was 0.9681, p < 0.0001.

Conclusion: DP40 CTASI represent a simple technique for improving detection and estimation of extent of ischaemia over NCCT and show close correlation with surrogate measures of infarct core.

Keywords: CT angiographic source images; CT angiography; Stroke; infarct; thrombectomy.

Figure 1.

(a) Demonstration of region of interest attenuation measurement in Hounsfield Units in the right lentiform nucleus within an area of striatal ischaemic low density (in a patient presenting at 1-hour post-stroke onset with right M1 occlusion) on NCCT (left) and DP40 (centre) with comparison to mirrored contralateral grey-matter and 24-hour CT after successful thrombectomy (right) with striatal infarct. (b) Baseline DP40 (upper row) and NCCT (lower row) images in another patient with a left M1 occlusion and confluent leucoariosis and an old left frontal infarct showing how ASPECTS can differ (acute ischaemic low density highlighted with arrowheads). ASPECTS: Alberta Stroke Program Early CT Score; DP40: delayed phase at 40 seconds after contrast injection; NCCT: non-contrast computed tomography.

Figure 2.

(a) Relationship between normal and ischaemic grey matter attenuation (Hounsfield Units, HU) and the difference between these measurements on NCCT and DP40 imaging for each of the consecutive 50 patients. The DP40 scans show a much larger increase in normal grey matter attenuation resulting in a larger difference in attenuation between ischaemic and normal grey matter on DP40 imaging compared to NCCT. (b) Relationship between the ischaemic grey matter and normal grey matter attenuation difference (HU) for DP40 and NCCT imaging relative to time from stroke onset (minutes). Improved contrast between normal and ischaemic grey matter was evident for DP40 imaging at early time points (<6 hours) following stroke onset. (c) Correlation between NCCT and DP40 ASPECTS with 24-hour ASPECTS. Baseline DP40 imaging showed improved correlation with eventual 24-hour ASPECTS compared to NCCT. (d) Scatter plot displaying differences in ASPECTS between NCCT and DP40 scans relative to time from stroke onset (minutes). This difference was notably evident at early time points post-stroke onset (<6rs). ASPECTS: Alberta Stroke Program Early CT Score; DP40: delayed phase at 40 seconds after contrast injection; NCCT: non-contrast computed tomography.

Figure 3.

NCCT (left), DP40 (centre), and 24-hour NCCT post-successful recanalisation for comparison (right) in 5 different patients initially imaged at early time points (stated in figure) post-stroke onset showing improved conspicuity and display of ischaemic change using DP40 imaging relative to NCCT (ringed). DP40 scans displayed with window settings of WW45 WL45 which we have found of value in accentuating tissue contrast between ischaemic and normal grey matter. ASPECTS: Alberta Stroke Program Early CT Score; DP40: delayed phase at 40 seconds after contrast injection; NCCT: non-contrast computed tomography.

Figure 4.

Plot demonstrating correlation between DP40 and venous phase CTPSI ASPECTS. ASPECTS: Alberta Stroke Program Early CT Score; CT: computed tomography; CTPSI: CT perfusion source images; DP40: delayed phase at 40 seconds after contrast injection;

Figure 5.

Representative images from 5 patients imaged within 12 hours of stroke onset. NCCT (left), DP40 (centre) and CTPSI at peak venous phase (right) showing improved identification and demarcation of ischaemic change (ringed) on DP40 images relative to NCCT and concordance with venous phase CTPSI.CT: computed tomography; CTPSI: CT perfusion source images; DP40: delayed phase at 40 seconds after contrast injection; NCCT: non-contrast CT.