Few-View Tomographic Reconstruction of Technetium-99m- sestamibi Distribution for the Detection and Diagnosis of Breast Lesions.

Abstract

Scintimammography (SMM) is a nuclear medicine test with the potential to provide relatively low-cost, minimally invasive differentiation of breast abnormalities detected by physical examination or mammography. While the most widely used clinical protocol involves acquiring one or two planar views, occasionally supplemented by conventional SPECT, we have used the ideal-observer SNR framework to establish that a dedicated breast SPECT geometry, in which the camera revolves around the breast alone, would provide superior lesion detectability to these other two approaches. Because the time required by tomographic studies to acquire data useful to popular reconstruction algorithms might be excessive for SPECT SMM, we have worked to develop reconstruction algorithms that can generate diagnostically useful SMM images from a smaller number of views than is usually used. In particular, we have developed a technique in which the few-view sinogram is first smoothed using an effectively two-dimensional, adaptive smoothing algorithm and additional views are then interpolated using periodic spline interpolation. The spline interpolation approach was chosen after extensive investigation of the accuracy and noise properties of various periodic interpolation approaches. We find that using this technique, meaningful image can be reconstructed from as few as 16 Angular views.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1998
Accession Number
ADA358869

Entities

People

  • Patrick J. Lariviere

Organizations

  • University of Chicago

Tags

Communities of Interest

  • Biomedical
  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Accuracy
  • Algorithms
  • Biomedical Research
  • Cancer
  • Data Acquisition
  • Data Sets
  • Detection
  • Frequency Domain
  • Geometry
  • Image Reconstruction
  • Imaging Techniques
  • Nuclear Medicine
  • Periodic Functions
  • Power Spectra
  • Three Dimensional
  • Tomography
  • Two Dimensional

Fields of Study

  • Physics

Readers

  • Computer Vision.
  • Medical Imaging.
  • Theoretical Analysis.