Advanced

Quantitative tissue classication with dual-energy computed tomography for radiation treatment planning evaluation and optimization of a new method

Karlsson, Mattias (2010)
Medical Physics Programme
Abstract (Swedish)
A new method to extract mass density and elemental compositions of tissues from dual-energy computed tomography (DECT) scans is described and investigated. The method decomposes a tissue into proportions of two or three predened base materials and was proposed in 2008 by a research team at Linkoping University. Informations of tissue parameters may be important in radiation treatment planning to calculate the distribution of absorbed dose in the patient. The study was performed with tissue compositions taken from scientic publications and the simulated response of a DECT scanner. The derived tissue parameters were used as input for Monte Carlo computer simulations of dierent types of radiation treatments to evaluate the eects on... (More)
A new method to extract mass density and elemental compositions of tissues from dual-energy computed tomography (DECT) scans is described and investigated. The method decomposes a tissue into proportions of two or three predened base materials and was proposed in 2008 by a research team at Linkoping University. Informations of tissue parameters may be important in radiation treatment planning to calculate the distribution of absorbed dose in the patient. The study was performed with tissue compositions taken from scientic publications and the simulated response of a DECT scanner. The derived tissue parameters were used as input for Monte Carlo computer simulations of dierent types of radiation treatments to evaluate the eects on distributions of absorbed dose.It was found that the accuracy of the extracted elemental compositions depended on the selected base materials. An inappropriate choice of base materials resulted in negative elemental weights, which are unphysical. Skeletal tissues were decomposed into proportions of cortical bone, yellow bone marrow and red bone marrow. 12 of 19 investigated skeletal tissues could be decomposed into positive elemental weights with the largest error of 3.0 percentage points. Properties for soft tissues were more dicult to determine with only one set of base materials. 29 of 51 soft tissues resulted in positive elemental weights when decomposed into proportions of water, lipid and protein. Errors up to 19.2 percentage points were seen in the elemental weights of the soft tissues. The simulated absorbed dose distributions of 192Ir brachytherapy did not show any substantial dependence on dierent tissue compositions. The results from simulations with the lower energy photons emitted by 125I were more aected by variations in tissue compositions. In both proton and 12C-ion therapy simulations, the error in Bragg peak position introduced by the extracted tissue compositions were less than 0.3 mm.This study indicates that an appropriate choice of base materials is crucial in order to determine accurate tissue compositions. The Monte Carlo simulations showed that elemental compositions may be important to consider in radiation treatment planning, especially for low-energy brachytherapy. Further evaluation with real DECT data is needed to see how the performance of the method is aected by for example image artifacts and individual variations in tissue compositions. (Less)
Please use this url to cite or link to this publication:
author
Karlsson, Mattias
supervisor
organization
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Röntgen
language
English
id
2157127
date added to LUP
2011-09-13 15:19:41
date last changed
2016-02-04 03:52:04
@misc{2157127,
  abstract     = {A new method to extract mass density and elemental compositions of tissues from dual-energy computed tomography (DECT) scans is described and investigated. The method decomposes a tissue into proportions of two or three predened base materials and was proposed in 2008 by a research team at Linkoping University. Informations of tissue parameters may be important in radiation treatment planning to calculate the distribution of absorbed dose in the patient. The study was performed with tissue compositions taken from scientic publications and the simulated response of a DECT scanner. The derived tissue parameters were used as input for Monte Carlo computer simulations of dierent types of radiation treatments to evaluate the eects on distributions of absorbed dose.It was found that the accuracy of the extracted elemental compositions depended on the selected base materials. An inappropriate choice of base materials resulted in negative elemental weights, which are unphysical. Skeletal tissues were decomposed into proportions of cortical bone, yellow bone marrow and red bone marrow. 12 of 19 investigated skeletal tissues could be decomposed into positive elemental weights with the largest error of 3.0 percentage points. Properties for soft tissues were more dicult to determine with only one set of base materials. 29 of 51 soft tissues resulted in positive elemental weights when decomposed into proportions of water, lipid and protein. Errors up to 19.2 percentage points were seen in the elemental weights of the soft tissues. The simulated absorbed dose distributions of 192Ir brachytherapy did not show any substantial dependence on dierent tissue compositions. The results from simulations with the lower energy photons emitted by 125I were more aected by variations in tissue compositions. In both proton and 12C-ion therapy simulations, the error in Bragg peak position introduced by the extracted tissue compositions were less than 0.3 mm.This study indicates that an appropriate choice of base materials is crucial in order to determine accurate tissue compositions. The Monte Carlo simulations showed that elemental compositions may be important to consider in radiation treatment planning, especially for low-energy brachytherapy. Further evaluation with real DECT data is needed to see how the performance of the method is aected by for example image artifacts and individual variations in tissue compositions.},
  author       = {Karlsson, Mattias},
  keyword      = {Röntgen},
  language     = {eng},
  note         = {Student Paper},
  title        = {Quantitative tissue classication with dual-energy computed tomography for radiation treatment planning evaluation and optimization of a new method},
  year         = {2010},
}