7 COMPUTED TOMOGRAPHY ANGIOGRAPHY SOURCE IMAGES ARE MORE ACCURATE THAN UNENHANCED COMPUTED TOMOGRAPHY FOR THE DETECTION AND DELINEATION OF ACUTE ISCHEMIC LESIONS: RECEIVER OPERATOR CHARACTERISTIC CURVE ANALYSIS USING A MODIFIED ASPECTS RATING SCALE. (1st March 2006)

• Record Type: Journal Article
• Authors: Camargo, E. C.; Furie, K. L.; Singhal, A. B.; Roccatagliata, L.; Cunnane, M. E.; Halpern, E. F.; Harris, G. J.; Smith, W. S.; Gonzalez, R. G.; Koroshetz, W. J.; Lev, M. H.
• Is Part Of: Journal of investigative medicine. Volume 54:Number 2(2006)
• Page Start: S374
• Page End: S374
• Publication Date: 2006-03-01
• Subjects: Clinical medicine -- Periodicals; Medicine -- Research -- Periodicals; Medicine; Research -- United States; Clinical medicine; Medicine -- Research; Periodicals; 616.075
• DOI: 10.2310/6650.2005.x0015.85
• Languages: English
• ISSNs: 1081-5589

Abstract

Abstract : Purpose: ASPECTS is a robust method for quantifying acute stroke hypodensity on noncontrast CT (NCCT) and helps predict clinical outcome. Whole brain source images obtained with CT angiography (CTA-SI) may increase ischemic brain conspicuity. We sought to determine if CTA-SI is more sensitive than NCCT in detecting acute middle cerebral artery (MCA) infarction and in predicting final infarct size using a modified ASPECT score for lesion delineation. Materials and Methods: After informed consent, we reviewed NCCT and CTA-SI scans of 51 patients with suspected MCA stroke, imaged within 12 hours of symptoms onset, enrolled in a dual-center cohort study. Two blinded neuroimagers rated presence and extent of NCCT and CTA-SI hypodensity using modified ASPECTS. Level of certainty for hypodensity detection was graded on admission and follow-up images using a 5-point scale (5 = definitely present; 1 = definitely absent). Linear regression and receiver operator characteristic (ROC) analyses were performed. Results: Thirty-three patients had confirmed infarction at follow-up (mean 6.5 days). Using certainty levels $ 4 (probable/definite) for ischemic hypodensity, sensitivity for acute stroke detection was 49% with NCCT and 70% with CTA-SI ( p = .04, ROC analysis); specificity was 100% for both. Linear regression showed R 2 = .42 ( p < .0001) for the correlation between acute NCCT modified ASPECTS and follow-up modified ASPECTS, and .73 (p < .0001) for the correlation between CTA-SI and follow-up modified ASPECTS (p < .0001 for the NCCT/CTA-SI comparison). Conclusion: CTA-SI, compared to NCCT, is more sensitive in detecting early irreversible ischemia and more accurate for final infarct volume prediction using modified ASPECTS.