Improving Detection of Axillary Lymph Nodes by Computer-Aided Kinetic Feature Identification in Positron Emission Tomography

Abstract

This project seeks to improve detection of metastatic axillary breast cancer and decrease a number of unnecessary breast biopsies through developing a sophisticated statistical signal processing system to aid identification of metabolic information within dynamic PET data. This Year 1 effort has been concentrated on mathematical modeling, validity analysis and algorithm development. We have found some interesting, natural relationship of conventional PET-FDG kinetic modeling to subspace representation in linear algebra. By representing time-activity curves of normal and malignant tissues with distinct subspaces, the distance between the subspaces can quantitatively characterize the difference between normal and malignant tissues as well as the similarity between primary and metastatic lesions. It turns out that OSEM outperforms FBP in revealing such difference and similarity. To accelerate the subspace estimation performed on a fully reconstructed dynamic image sequence, the validity of estimating parameters directly from projection data prior to the image reconstruction has been studied. With such estimations, a data compression algorithm has been developed for a fast dynamic image reconstruction, which maximizes the signal-to-noise ratio (SNR) .

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 2000
Accession Number
ADA385393

Entities

People

  • Xiaoli Yu

Organizations

  • University of Southern California

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Algorithms
  • Chemical Reactions
  • Data Compression
  • Detection
  • Detectors
  • Diagnostic Imaging
  • Electrical Engineering
  • Health Services
  • Identification
  • Image Processing
  • Image Reconstruction
  • Lymph Nodes
  • Metabolism
  • Positron Emissions
  • Signal Processing
  • Simulations
  • Tomography

Fields of Study

  • Physics

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Computational Modeling and Simulation
  • Medical Imaging.