Version 2.5.2.0 CRISP Logo CRISP Homepage Help for CRISP Email Us

Abstract

Grant Number: 5R01LM006726-04
Project Title: KNOWLEDGE DISCOVERY IN DISTRIBUTED CARDIAC IMAGE BASES
PI Information:NameEmailTitle
EZQUERRA, NORBERTO F. norberto.ezquerra@cc.gatech.edu ASSOCIATE PROFESSOR

Abstract: DESCRIPTION (taken from application abstract): Heart disease continues to be the primary cause of death in the U.S., with 25% of all deaths related to coronary artery disease (CAD). In addition to the loss of irreplaceable human life, there are also staggering health care costs and losses in productivity associated with the 1.5 million myocardial infarctions suffered in the U.S. every year. The present competing renewal application seeks to make a contribution toward this vital health care problem by exploring frontier computing methods to support and facilitate CAD assessment. The objective of the proposed research is to develop and evaluate a methodology to accomplish the following specific aims: (1)Knowledge Discovery: To design, implement and test novel database (DB) "mining" algorithms to uncover associations and inferences imbedded in clinical DBs and which can improve diagnostic performance. (2)Knowledge Base Enrichment: To use the knowledge resulting from DB mining as well as conventional knowledge-acquisition methods to create and evaluate a robust knowledge base (KB) with which to interpret cardiovascular SPECT imagery and other types of relevant, patient- specific information. (3)Distributed Knowledge Discovery and Processing: To extend both the Knowledge-discovery and knowledge-based processing methods to distributed, Internet-based setting for a twofold purpose: (I) to provide users with widespread access to the resulting KB, and (ii) to access and mine remote multi center DBs to further improve our knowledge regarding the assessment of CAD. The proposed work represents pioneering research in several ways, especially: (I) the creation of innovative algorithms to mine image DBs, (II) the application of these algorithms to the clinical assessment of CAD, and (III) the creation of distributed DB mining and knowledge-based processing methods to link geographically dispersed users and clinical DBs. The proposed research builds on our previous work on knowledge- guided image interpretation, and represents an interinstitutional and interdisciplinary effort between Georgia Tech and Emory University, a longstanding collaboration that has previously resulted in numerous joint publications and valuable insights centering on diagnostic imaging, and which has also supported several academic degrees.

Public Health Relevance:
This Public Health Relevance is not available.

Thesaurus Terms:
cardiovascular visualization, computer assisted diagnosis, computer data analysis, computer processing of clinical data, coronary disorder, diagnosis design /evaluation, heart disorder diagnosis
Internet, artificial intelligence, computer assisted medical decision making, information system analysis, mathematical model, myocardium, perfusion
angiography, bioimaging /biomedical imaging, human data, single photon emission computed tomography

Institution: GEORGIA INSTITUTE OF TECHNOLOGY
505 10TH ST NW
ATLANTA, GA 303320420
Fiscal Year: 2001
Department: NONE
Project Start: 01-FEB-1998
Project End: 31-JAN-2003
ICD: NATIONAL LIBRARY OF MEDICINE
IRG: BLR


CRISP Homepage Help for CRISP Email Us