Advanced

Evaluation of a decision support system for interpretation of myocardial perfusion gated SPECT

Lomsky, Milan; Gjertsson, Peter; Johansson, Lena; Richter, Jens; Ohlsson, Mattias LU ; Tout, Deborah; van Aswegen, Andries; Underwood, S. Richard and Edenbrandt, Lars LU (2008) In European Journal of Nuclear Medicine and Molecular Imaging 35(8). p.1523-1529
Abstract
Purpose We have recently presented a decision support system for interpreting myocardial perfusion scintigraphy (MPS). In this study, we wanted to evaluate the system in a separate hospital from where it was trained and to compare it with a quantification software package. Methods A completely automated method based on neural networks was trained for the interpretation of MPS regarding myocardial ischaemia and infarction using 418 MPS from one hospital. Features from each examination describing rest and stress perfusion, regional and global function were used as inputs to different neural networks. After the training session, the system was evaluated using 532 MPS from another hospital. The test images were also processed with the... (More)
Purpose We have recently presented a decision support system for interpreting myocardial perfusion scintigraphy (MPS). In this study, we wanted to evaluate the system in a separate hospital from where it was trained and to compare it with a quantification software package. Methods A completely automated method based on neural networks was trained for the interpretation of MPS regarding myocardial ischaemia and infarction using 418 MPS from one hospital. Features from each examination describing rest and stress perfusion, regional and global function were used as inputs to different neural networks. After the training session, the system was evaluated using 532 MPS from another hospital. The test images were also processed with the quantification software package Emory Cardiac Toolbox (ECTb). The images were interpreted by experienced clinicians at both the training and the test hospital, regarding the presence or absence of myocardial ischaemia and/or infarction and these interpretations were used as gold standard. Results The neural network showed a sensitivity of 90% and a specificity of 85% for myocardial ischaemia. The specificity for the ECTb was 46% (p < 0.001), measured at the same sensitivity. The neural network sensitivity for myocardial infarction was 89% and the specificity 96%. The corresponding specificity for the ECTb was 54% (p < 0.001). Conclusions A decision support system based on neural networks presents interpretations more similar to experienced clinicians compared to a conventional automated quantification software package. This study shows the feasibility of disseminating the expertise of experienced clinicians to less experienced physicians by the use of neural networks. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
radionuclide imaging, neural networks (computer), image interpretation, computer assisted, heart function tests, heart disease
in
European Journal of Nuclear Medicine and Molecular Imaging
volume
35
issue
8
pages
1523 - 1529
publisher
Springer
external identifiers
  • wos:000257922400015
  • scopus:48149085300
ISSN
1619-7070
DOI
10.1007/s00259-008-0746-9
language
English
LU publication?
yes
id
b9ff6788-e4a1-4bae-830f-123685d71d70 (old id 1253727)
date added to LUP
2008-11-10 11:54:07
date last changed
2017-01-01 04:30:55
@article{b9ff6788-e4a1-4bae-830f-123685d71d70,
  abstract     = {Purpose We have recently presented a decision support system for interpreting myocardial perfusion scintigraphy (MPS). In this study, we wanted to evaluate the system in a separate hospital from where it was trained and to compare it with a quantification software package. Methods A completely automated method based on neural networks was trained for the interpretation of MPS regarding myocardial ischaemia and infarction using 418 MPS from one hospital. Features from each examination describing rest and stress perfusion, regional and global function were used as inputs to different neural networks. After the training session, the system was evaluated using 532 MPS from another hospital. The test images were also processed with the quantification software package Emory Cardiac Toolbox (ECTb). The images were interpreted by experienced clinicians at both the training and the test hospital, regarding the presence or absence of myocardial ischaemia and/or infarction and these interpretations were used as gold standard. Results The neural network showed a sensitivity of 90% and a specificity of 85% for myocardial ischaemia. The specificity for the ECTb was 46% (p &lt; 0.001), measured at the same sensitivity. The neural network sensitivity for myocardial infarction was 89% and the specificity 96%. The corresponding specificity for the ECTb was 54% (p &lt; 0.001). Conclusions A decision support system based on neural networks presents interpretations more similar to experienced clinicians compared to a conventional automated quantification software package. This study shows the feasibility of disseminating the expertise of experienced clinicians to less experienced physicians by the use of neural networks.},
  author       = {Lomsky, Milan and Gjertsson, Peter and Johansson, Lena and Richter, Jens and Ohlsson, Mattias and Tout, Deborah and van Aswegen, Andries and Underwood, S. Richard and Edenbrandt, Lars},
  issn         = {1619-7070},
  keyword      = {radionuclide imaging,neural networks (computer),image interpretation,computer assisted,heart function tests,heart disease},
  language     = {eng},
  number       = {8},
  pages        = {1523--1529},
  publisher    = {Springer},
  series       = {European Journal of Nuclear Medicine and Molecular Imaging},
  title        = {Evaluation of a decision support system for interpretation of myocardial perfusion gated SPECT},
  url          = {http://dx.doi.org/10.1007/s00259-008-0746-9},
  volume       = {35},
  year         = {2008},
}