Return To Abstract Listing

Scientific Session 19 — SS19: Neuroradiology - Brain Tumor, Vascular and Neurodegenerative

Thursday, May 9, 2019

Abstracts 1879-2985

2063. Collaterals and Perfusion Parameters in Large Vessel Occlusion Ischemic Strokes: A Stroke Registry Real World Experience

Voleti S*,  Vidovich J,  Corcoran B,  Mistry E,  Vagal A. University of Cincinnati Medical Center, Cincinnati, OH

Address correspondence to S. Voleti (

Objective: The recent endovascular trials have produced major changes in the treatment of patients with large vessel occlusion (LVO) ischemic strokes with a paradigm shift for use of CT perfusion (CTP) to make endovascular treatment decisions in an extended time window. Although better collaterals lead to better outcomes, the relationship between collaterals and CTP parameters is unclear. Our objective was to explore the relationship between CTP and CT angiogram (CTA) collaterals in LVO strokes. We hypothesized that better collaterals are associated with smaller CTP core and larger mismatch.

Materials and Methods: We used a single-institution, retrospective registry of consecutive patients with acute ischemic strokes with LVO (internal carotid artery, M1, M2 occlusion) between May 2016 and May 2018. We specifically chose this timeframe to provide us real-world data after the new endovascular trial evidence. We graded baseline CTA collaterals (0 = absent; 1 = filling <=50%; 2 = filling >50% & <100%; 3 = filling 100% of the occluded vascular territory) and dichotomized into good (2-3) and poor (0-1) collaterals. We recorded automated CTP parameters by using RAPID software including core (relative cerebral blood flow (CBF)<30%), hypoperfusion (Tmax>6 sec), and mismatch ratios. We also recorded additional CTP thresholds (CBF<34% and <38%), Tmax (>4 sec, >8 sec, and >10 sec), and hypoperfusion index (Tmax>10 sec/Tmax>6 sec). We collected demographic and clinical data from electronic medical records. We performed Pearson’s correlation and logistic regressions to assess the association between collaterals and CTP parameters.

Results: Among 158 patients who received CTP, 53 had LVO. Among the 53 included patients (mean age: 62.2 ± 16.6; 39.6% female), the median ASPECTS was 7, the median CTP core was 15 ml, and the median hypoperfusion was 123 ml. Patients with good collaterals had smaller core volumes (14.92 ± 22.76 ml vs 50.48 ± 52.86 ml; p=0.003), smaller volumes of severe hypoperfusion (Tmax>10 sec) (39.29 ± 41.86 ml vs 83.81 ± 62.62 ml; p=0.005), larger volumes of moderate hypoperfusion (Tmax>6 sec) (99.72 ± 62.88 ml vs 137.13 ± 72.08 ml; p=0.051), and lower hypoperfusion index (0.28 ± 0.23 vs 0.54 ± 0.16; p<0.0001) than patients with poor collaterals. The correlation coefficient between collaterals and CTP core was -0.56 (p<0.0001), between collaterals and CTP hypoperfusion was -0.42 (p=0.002), and between collaterals and mismatch ratio was 0.389 (p=0.004). There were significant associations between collaterals and core (OR=1.05, 95% Cl: 1.02-1.07; p=0.0002), hypoperfusion (OR=1.014, 95% Cl: 1.005-1.022, p=0.001), and mismatch (OR=0.98, 95% Cl: 0.96-0.99, p=0.01).

Conclusion: Better collaterals are associated with smaller CTP core, larger hypoperfusion, and larger mismatch in LVO ischemic stroke patients. This finding is clinically relevant because CTA collateral assessment may be an alternative to CTP, particularly in centers where CTP is not readily available for stroke workup.