Comparison of methods between CT perfusion source images and CT angiography in collateral flow assessment

The methods used for grading leptomeningeal collateral flow (LMF) on single-phase computed tomography angiography (CTA) are heterogeneous and limited by temporal resolution.

To compare the reliability of relative filling time delay (rFTD) on CT perfusion source images (CTP-SI) and the currently used single-phase CTA collateral assessment methods and evaluate their ability to predict clinical outcomes in acute ischemic stroke patients.

We analyzed consecutive middle cerebral artery or internal carotid artery occlusion patients who received multimodal CT before treatment and within 12 h of stroke symptom onset from October 2015 to December 2018. Patients were dichotomized using the 90-day mRS into good (0-1) versus adverse (2-6) outcomes. CTP-SI was used to identify the rFTD score. CTA images were reconstructed to assess collateral status using the collateral score (Cs) and region leptomeningeal collateral score (rLMCs). Two observers independently assessed images.

The baseline characteristics (n = 54) were median age of 67 years and 68.5% of the participants were men. The baseline median NIHSS was 14. Good clinical outcomes were observed in 19 (35.2%) patients. The k value was higher for rFTDs (k = 0.779, P < 0.001) than Cs (k = 0.666, P < 0.001) and rLMCs (k = 0.763, P < 0.001). Higher rFTDs were correlated with lower rLMCs (Spearman's rho -0.68, P < 0.001) and Cs (rho -0.66, P < 0.001). In multivariate logistic regression, rFTD was associated with functional outcomes (P = 0.044).

The rFTDs method is comparable to single-phase CTA-based assessments in assessing LMFs in acute ischemic stroke patients. Higher rFTDs is independently associated with adverse long-term functional outcomes.