Computer simulation and analysis of nuclear medicine studies

Research output: ThesisDoctoral Thesis

56 Downloads (Pure)

Abstract

Problem investigated: There is no substantive literature on the three dimensional segmental anatomy of the lungs. Extant literature has primarily been concerned with the surface markings of the segments. There is clearly a need for the establishment of such a model of the lungs to permit investigation of basic issues germane to planar lung scintigraphy in the diagnosis of pulmonary embolic disease. Issues as basic as which views of the lungs offer the highest information density, variability of perceived defect size, the utility of lung charts in reporting, how small a lesion must be for diagnosis and the investigations of artifacts in lung scintigraphy require exploration. These problems were identified by systematic review of clinical reporting in a large database of cases in a typical hospital practice.
Procedures followed: A three dimensional model of the scintigraphic anatomy of the lungs was constructed based on computed tomography, cadaveric dissection and extant literature. A Monte Carlo simulation technique was then applied with generation of photons within the segmented model with the subsequent emission, scatter, attenuation and eventual collimation and detection to produce clinically acceptable images in standard projections.

General results obtained: 1. Information density about the lung segments is highest in the anterior and posterior oblique projections. 2. Classification of lesion size by clinicians has poor accuracy and reproducibility. This is improved with the use of a guidechart of segmental anatomy. 3. The smallest detectable lesion is 3% of the volume of a lung with reduced detectability in the right lung base. 4. The “stripe sign” can be readily produced in both lungs with the major mechanism being scatter from adjacent segments in the same lung rather than shine through from the other lung.

Major conclusions: The model developed is clinically acceptable and capable of providing solutions to major problems involved in planar lung scintigraphy. It also has the potential to provide a framework for more complex problem-solving in tomographic lung imaging.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • The University of New South Wales
Supervisors/Advisors
  • Bertram, Chris, Supervisor, External person
  • Van der Wall, Hans, Supervisor, External person
Award date20 Apr 1999
Publication statusPublished - 1999
Externally publishedYes

    Fingerprint

Bibliographical note

Copyright the Author 1998. Version archived for private and non-commercial use with the permission of the author/s and according to publisher conditions. For further rights please contact the publisher.

Keywords

  • Biomedical Research
  • Nuclear medicine
  • lung perfusion
  • Lung segmental reference chart
  • Monte Carlo simulation

Cite this