Abstract

The long term objective of this project is to implement Inverse Monte Carlo (IMOC) as a unified reconstruction algorithm to facilitate quantitative Single Phonton Emission Computed Tomography (SPECT). The most serious obstacles to clinical quantitative SPECT imaging are the degrading effects due to photon scatter and attenuation, the spatial variance of the collimation, and the deficiency of counts in the acquired projections. IMOC will be developed as a unified reconstruction algorithm providing simultaneous compensation for these degrading effects. IMOC combines detailed photon transport modeling (using Monte Carlo techniques) with iterative statistical estimators. Three statistical estimators which have been suggested for emission computed tomography (maximum likelihood, minimum variance, and maximum entropy) will be evaluated for convergence and stability as a function of iteration, accuracy of the system model, and photon count density. Experimental projection data from phantoms will be reconstructed to evaluate IMOC for improvement in lesion detectability, noise propagation, resolution recovery, and quantitation of uptake ratios for source distributions which reflect clinical imaging geometries. Algorithm acceleration techniques will be investigated (through estimator modification and Monte Carlo approximation) to evaluate computation burden reduction. IMOC will be implemented and performance evaluated on vector and parallel computer architectures to assess the viability of these new computation environments for tomographic reconstruction. Results from preliminary investigations are encouraging, indicating that IMOC can provide quantitative compensation in reconstructed SPECT images. Presently, IMOC is the only reconstruction algorithm for SPECT which has demonstrated this simultaneous compensation. This proposed evaluation of alternatives for IMOC development is intended to help establish a strategy for implementing IMOC as a practical reconstruction algorithm enhancing the use of SPECT routinely as a quantitative functional imaging modality.

  Funding Agency

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Research Project (R01)
Project #
5R01CA046856-02
Application #
3190321
Study Section
Diagnostic Radiology Study Section (RNM)
Project Start
1988-03-01
Project End
1991-02-28
Budget Start
1989-03-15
Budget End
1990-02-28
Support Year
2
Fiscal Year
1989
Total Cost
Indirect Cost

  Institution

Name
Duke University
Department
Type
Schools of Medicine
DUNS #
071723621
City
Durham
State
NC
Country
United States
Zip Code
27705