J Korean Neurol Assoc > Volume 8(1); 1990 > Article
Journal of the Korean Neurological Association 1990;8(1): 29-34.
뇌실내 출혈을 동반한 기저핵 및 시상부 출혈환자에 있어 CT계측을 이용한 예후판별
한문강, 정진상,김재문,위봉애
충남대학교 신경과. 장홍백병원 신경과. 서울대학교 신경과.
Survival Prediction of Ganglio-thalamic Hemorrhage Accompanying Intraventricular Hemorrhage Using CT Scan Indices
Moon Gang Han, M.D., Jin Sang Jung, M.D., Jae Moon Kim, M.D.,Bong Ae Wie,M.D.,
Department of Neurology, Chung Nam University, College of Medicine Department of Neurology, Chang- heung Paik Hospital Department of Neurology, Seoul National University Hospital
Abstract
To develop a survival prediction model and to use it as a therapeutic guideline a series of 66 cases with ganglio-thalamic hemorrhage accompanying intraventricular hemorrhage were analyzed retrospectively. They were classified into the benign and fatal groups according to their final outcomes. Fourteen indices obtained from the initial CT scans were final outcomes. Fourteen indices obtained from the initial CT scans were subjected to multivariate discriminant analysis. The resultant discriminant function(Z) included the amount of hematoma in the parenchyme(AH), bicaudate cerebroventricular index(BCCI) maximum fourth ventricular width(FVW), and third cerebroventricular ratio(TCR) in an order of decreasing discriminating power and was as follows :Z=-3.2639+0.3508x10-1xAH+6.8816xBCCI+0.1139xFVW-5.7794xTCR. This function predicted survivability with accuracy of 84.9% when it was applied to the original subjects. The conclusion is that AH, BCCI, FVW, and TCR are the potent predictors of the survival of patients with ganglio-thalamic hemorrhage accompanying intraventricular hemorrhage.
TOOLS
METRICS Graph View
  • 1,204 View
  • 6 Download
Related articles


ABOUT
BROWSE ARTICLES
EDITORIAL POLICY
FOR CONTRIBUTORS
Editorial Office
(ZIP 03163) #1111, Daeil Bldg, 12, Insadong-gil, Jongno-gu, Seoul, Korea
Tel: +82-2-737-6530    Fax: +82-2-737-6531    E-mail: jkna@neuro.or.kr                

Copyright © 2024 by Korean Neurological Association.

Developed in M2PI

Close layer