Abstract T P7: CT Perfusion Accurately Identifies Core and Penumbral Tissue in Large Vessel Occlusion Strokes: a Paradigm Validation Using a Large Cohort of Endovascularly-Reperfused and Medically-Treated Non-Reperfused Strokes
Introduction: MRI core and penumbral imaging paradigms show promise in tissue fate prediction in large vessel occlusion strokes (LVOS). The comparability of CT perfusion (CTP) in this context remains to be established.
Hypothesis: We hypothesized that initial CTP “core estimates” would strongly correlate with final infarct volumes (FIV) in Endovascularly-Reperfused patients, while initial CTP “hypoperfusion estimates” would correlate with FIV in Medically-Treated Non-Reperfused patients. We aimed to ascertain which CTP parameters best predict initial core and at-risk tissue volumes in reperfused vs. medically-treated patients.
Methods: Retrospective analysis of AIS patients presenting within 12 hours was performed. Subjects were dichotomized into either successfully reperfused post-thrombectomy (TICI 2b/3) or non-reperfused (no IV/IA therapy or confirmed persistent occlusion post IV tpa) cohorts. CTP predictors of infarct core included cerebral blood flow and blood volume (rCBV, rCBF). Total hypoperfused tissue volumes were estimated at 2s increments of time-to-maximum of the residue function (Tmax) from 4-10s. Perfusion analysis was performed on a custom computational tool (RAPID). Imaging outcome measure included final infarct volumes (FIV) on follow-up imaging.
Results: Among 92 total patients, 57 were reperfused (21 male, mean 64±14years) with median (IQR) NIHSS=18 (7); time (min) from symptom onset to reperfusion=370 (193); ICA lesions=10 (18%); MCA-M1=37 (64%); MCA-M2=11 (19%). Among 35 non-reperfused patients (12 male, mean 68±17 years) NIHSS=12 (IQR 13); ICA lesions=8 (23%); MCA-M1=19 (54%); MCA-M2=8 (23%). In reperfused patients, significant correlations with FIV were observed only with rCBF (Pearson’s Correlation 0.42, p=0.005) (median predicted core=10cc; FIV=13.9cc) and rCBV (0.30, p=0.05) core estimates. Among non-reperfused controls, strongest correlations were again observed with rCBF (0.72) (median rCBF=2cc; FIV=16cc, IQR=62cc) and rCBV (0.74) estimates (P<0.01); however, significant correlations in this cohort were present for all hypoperfusion estimates (p<0.01).
Conclusion: Baseline CTP profiles correlate variably with imaging outcome, depending upon reperfusion status and parameter selection.
Author Disclosures: S. Dehkharghani: None. A.B. Prater: None. D. Qiu: None. R. Bammer: Ownership Interest; Significant; Ischemaview, Inc. F. Nahab: None. M. Straka: Employment; Significant; Ischemiaview, Inc. Ownership Interest; Significant; Ischemaview, Inc. M. Bowen: None. R.G. Nogueira: Other Research Support; Modest; Stryker: PI for TREVO 2 and DAWN trials, Penumbra: executive committee for the 3D Separator Trial. Other Research Support; Significant; Covidien: Steering Committee for SWIFT and SWIFT PRIME trials and core lab, STAR trial.
- © 2015 by American Heart Association, Inc.