Vascular Plaque Determination for Stroke Risk Assessment

Abstract

Each year, about 800,000 Americans experience a new or recurrent stroke. The long-term goal of this research program is to create non-invasive methods utilizing ultrasound to identify plaques at high risk for initiating a cerebrovascular accident. The core of the current research project is a pilot clinical study to enroll 100 subjects who are scheduled for carotid endarterectomy (CEA). From each subject, the research effort obtains ultrasound data from the carotid plaque (or carotid artery for Normal subjects) prior to surgery and then creates histology slides of the removed plaque tissue which are used to train a statistical classifier for determining plaque composition. The main accomplishments from the prior year include: increasing enrollment to 53 CEA subjects, determination of attenuation compensation approach, testing of the use of the nonlinearly generated harmonic in addition to the fundamental for plaque characterization, and the production of a course machine learning algorithm based on data from the first 32 subjects enrolled in year 1 with accuracy of 77%.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Oct 01, 2018
Accession Number
AD1063856

Entities

People

  • David G. Vince
  • Russell J. Fedewa

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Accidents
  • Attenuation
  • Biomedical Engineering
  • Compensation
  • Engineering
  • Frequency
  • Frequency Shift
  • Histology
  • Institutional Review Board
  • Machine Learning
  • Physicians
  • Power Spectra
  • Risk
  • Risk Analysis
  • Signal Processing
  • Supervised Machine Learning
  • Ultrasounds

Fields of Study

  • Medicine

Readers

  • Cardiovascular Physiology
  • Clinical Trial Research.
  • Molecular Genetics

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference