Statistical approaches in quantitative positron emissiontomography

  • Authors:

  • Richard M. Leahy;Jinyi Qi

  • Affiliations:

  • Signal and Image Processing Institute, Department of Electrical Engineering, University of Southern California, Los Angeles, CA 90089-2564, USA. leahy@sipi.usc.edu;Signal and Image Processing Institute, Department of Electrical Engineering, University of Southern California, Los Angeles, CA 90089-2564, USA

  • Venue:
  • Statistics and Computing
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

Positron emission tomography is a medical imagingmodality for producing 3D images of the spatial distribution ofbiochemical tracers within the human body. The images arereconstructed from data formed through detection of radiationresulting from the emission of positrons from radioisotopes taggedonto the tracer of interest. These measurements are approximate lineintegrals from which the image can be reconstructed using analyticalinversion formulae. However these direct methods do not allowaccurate modeling either of the detector system or of the inherentstatistical fluctuations in the data. Here we review recent progressin developing statistical approaches to image estimation that canovercome these limitations. We describe the various components of thephysical model and review different formulations of the inverseproblem. The wide range of numerical procedures for solving theseproblems are then reviewed. Finally, we describe recent work aimed atquantifying the quality of the resulting images, both in terms ofclassical measures of estimator bias and variance, and also usingmeasures that are of more direct clinical relevance.