Article Text

Download PDFPDF

E-039 Percent insular ribbon infarction on non-contrast computed tomography and computed tomography angiography for predicting infarct growth rate and 90-day outcomes in large-vessel occlusive stroke: a secondary analysis of a prospective clinical trial
Free
  1. R Regenhardt1,
  2. N Jarunnarumol2,
  3. A Kam2,
  4. J He2,
  5. R Gonzalez2,
  6. A Singhal3,
  7. M Lev2
  1. 1Neurosurgery, Neurology, Massachusetts General Hospital, Boston, MA
  2. 2Radiology, Massachusetts General Hospital, Boston, MA
  3. 3Neurology, Massachusetts General Hospital, Boston, MA

Abstract

Introduction Understanding infarct growth progression can aid in identifying stroke patients with slow evolution who might benefit from delayed endovascular thrombectomy (EVT). We sought to evaluate associations between percent insular ribbon infarction on non-contrast CT (NCCT-PIRI) and CT angiography (CTA-PIRI), collateral pattern on CTA, ASPECTS (NCCT-ASPECTS), infarct growth rate (IGR), and 90-day outcomes in large-vessel occlusion (LVO) stroke.

Methods We conducted a secondary analysis of clinical trial data of acute stroke patients not receiving reperfusion therapies between January 2007 - June 2009. A total of 31 participants with anterior circulation LVO from the original trial who underwent initial CT followed by sequential MRI scans were included. Two neuroradiologists independently rated NCCT-PIRI and CTA-PIRI at presentation according to the following 0–4 scale: mild (PIRI 0–1, <25%), moderate (PIRI 2, 25–50%), or severe (PIRI 3–4, >50%); differences between readers were reconciled by consensus. Collateral patterns were classified as symmetric, malignant, or other based on CTA. Serial infarct volumes were measured using DWI within 48 hours and FLAIR at 90 days. Infarct growth rate (IGR) was calculated as change in infarct volume per time. Clinical outcomes were assessed using the 90-day Barthel Index (BI) and modified Rankin Scale (mRS).

Results The median age was 71 years, and 39% were women. The inter-rater agreement for NCCT-PIRI was excellent (κ=0.83, p<0.001). Similarly, the inter-rater agreement for CTA-PIRI was excellent (κ=0.83, p<0.001). Multivariable models, controlling for age, occlusion location, and collateral pattern, showed NCCT-PIRI category independently predicted infarct volume at presentation (β=1.19, 95%CI=0.16,2.22, p=0.02), 90-day mRS≤2 (OR=0.16, 95%CI=0.03,0.84, p=0.03), and 90-day BI (β=-1.92, 95%CI=-3.22,-0.61, p=0.004). Furthermore, CTA-PIRI category independently predicted infarct volume at presentation (β=1.80, 95%CI=0.68,2.93, p=0.002), onset-to-presentation IGR (β=1.28, 95%CI=0.22,2.35, p=0.02), and 90-day BI (β=-1.28, 95%CI=-2.52,-0.04, p=0.04). For predicting 90-day favourable mRS ≤2, the sensitivity and specificity were 90.0% (9/10) and 73.7% (14/19) for mild-to-moderate NCCT-PIRI, and 70.0% (7/10) and 84.2% (16/19) for mild-to-moderate CTA-PIRI, as compared to 100.0% (10/10) and 31.6% (6/19) for ASPECTS ≥6, 60.0% (6/10) and 78.9% (15/19) for presentation DWI infarct volume <30 cc, and 70.0% (7/10) and a specificity of 73.7% (14/19) for symmetric collateral pattern.

Conclusion In this prospective cohort of patients with untreated anterior LVO stroke, percentage insular ribbon infarction on admission head CT and CTA were independently associated with IGR and 90-day clinical outcome. A less than 50% insula infarct at presentation (mild-to-moderate PIRI <2) outperformed ASPECTS and CTA collateral pattern for prediction of 24 hr infarct volume and compared highly favourably with DWI. NCCT-PIRI outperformed ASPECTS, CTA collateral pattern and DWI infarct volume for 90 day mRS≤2 outcome prediction. The widely available, simple to obtain CT/CTA PIRI metrics can be used to identify patients who may benefit from late-window EVT and aid in outcome prognostication at the time of stroke presentation.

Disclosures R. Regenhardt: 1; C; Heitman Foundation for Stroke. 2; C; DSMB for trial sponsored by Rapid Medical. N. Jarunnarumol: None. A. Kam: None. J. He: None. R. Gonzalez: None. A. Singhal: None. M. Lev: None.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.