Advertisement

Ultrasonic Tissue Characterization Imaging

  • Frederic L. Lizzi
  • Ernest J. Feleppa
  • Mykola M. Yaremko
Part of the Acoustical Imaging book series (ACIM, volume 14)

Abstract

Clinical tissue characterization research programs are being conducted in collaboration with the College of Physicians and Surgeons at Columbia University and the Cornell University Medical College. These programs use ultrasonic tissue characterization techniques to analyze radio-frequency echo signals from the liver and eye, respectively.

In both programs, a variety of frequency domain analysis techniques have been employed to study normal and diseased tissues. In the eye, a data base comprising over 700 cases has been established and discriminant analysis has identified combinations of diagnostically useful spectral features. These features can identify metastatic carcinoma, spindle-B malignant melanoma, and mixed/epitheliod melanoma. In the liver, a smaller data base exists; this has shown that different parameters can be useful in differentiating normal livers from mildly cirrhotic or severely cirrhotic livers.

The diagnostic utility of these parameters has motivated us to investigate new types of cross-sectional images depicting specific spectral or cepstral features. Several types of computer-generated images have been produced. “Stained” images of the eye depict discriminant functions based on several spectral parameters and use color-coding to indicate correlations with specific tissue types. Spectral parameter images depict well-defined spectral features and are useful for monitoring the treatment of ocular tumors. Scatterer-size images compare spectral features with a theoretical model to estimate sub-resolution scatterer sizes. Cepstral images of the liver are proving very interesting; these images indicate the relative scattering strength of tissue elements with selectable periodicities.

This presentation will provide clinical examples of these types of images and indicate the factors that govern the spatial resolution they can provide. Of special interest are the trade-offs that must be considered among spatial and spectral resolutions and statistical uncertainty.

Keywords

Spectral Feature Discriminant Function Spectral Resolution Spectral Parameter Statistical Uncertainty 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Copyright information

© Plenum Press, New York 1985

Authors and Affiliations

  • Frederic L. Lizzi
    • 1
  • Ernest J. Feleppa
    • 1
  • Mykola M. Yaremko
    • 1
  1. 1.Riverside Research InstituteNew YorkUSA

Personalised recommendations