Comparison of Noncontrast Computed Tomography, Multiphase Computed Tomography Angiography, and Computed Tomography Perfusion to Assess Infarct Growth Rate in Acute Stroke

BACKGROUND: Infarct growth rate is remarkably heterogeneous in acute ischemic stroke, reflecting diverse clinical-physiological phenotypes. We compared different methods of estimating infarct growth rate in patients with acute ischemic stroke undergoing thrombectomy using multimodal computed tomography (CT) stroke imaging. METHODS: Secondary analysis of the international ESCAPE-NA1 trial (Efficacy and Safety of Nerinetide for the Treatment of Acute Ischemic Stroke) which evaluated the effect of nerinetide in patients with large vessel occlusion undergoing thrombectomy. Infarct growth rate was estimated leveraging each component of multimodal stroke CT imaging: (1) 10 minus baseline Alberta Stroke Program Early CT Score (ASPECTS) divided by hours elapsed from symptom onset on noncontrast CT (ASPECTS decay per hour); (2) collateral status on multiphase CT angiography (mCTA), and (3) hypoperfusion intensity ratio on CT perfusion. Patients were dichotomized into intermediate and slow progressors (since fast progressors were likely to be excluded from ESCAPE-NA1 based on trial enrollment criteria) according to median ASPECTS decay, presence of good versus moderate/poor mCTA collaterals, and median hypoperfusion intensity ratio, respectively. Associations between progressor phenotypes and 90-day modified Rankin Scale score were assessed across neuroimaging modalities using adjusted logistic regression analyses. RESULTS: Among 1105 patients enrolled in ESCAPE-NA1 between 2017 and 2019, 619 (56.0%) were assessed for progressor phenotypes using noncontrast CT, 1084 (98.1%) with mCTA, and 415 (37.6%) with CT perfusion. Median ASPECTS decay per hour was 1.05 (interquartile range, 0.05–1.85), 188/1084 (17%) patients had good collateral status on mCTA, and the median hypoperfusion intensity ratio was 0.44 (interquartile range, 0.28–0.59). Intermediate progressors showed worse functional outcomes compared with slow progressors only in CT perfusion strata: adjusted common odds ratio for modified Rankin Scale ordinal shift analysis of 1.69 (95% CI, 1.14–2.49). No significant association between progressor phenotypes and 90-day modified Rankin Scale was seen when the noncontrast CT and the mCTA approaches were used. CONCLUSIONS: Stroke progressor phenotypes based on CT perfusion criteria (using the hypoperfusion intensity ratio approach) were associated with clinical outcomes, while stroke progressor phenotypes based on noncontrast CT (ASPECTS decay) and mCTA (collateral status) criteria were not.