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). CONCLUSION: 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.
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). CONCLUSION: 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.
Authors: X Kang; D S Berman; K F Van Train; A M Amanullah; J Areeda; J D Friedman; H Kiat; G Germano Journal: J Nucl Med Date: 1997-09 Impact factor: 10.057
Authors: K F Van Train; J Areeda; E V Garcia; C D Cooke; J Maddahi; H Kiat; G Germano; G Silagan; R Folks; D S Berman Journal: J Nucl Med Date: 1993-09 Impact factor: 10.057
Authors: R Hachamovitch; D S Berman; L J Shaw; H Kiat; I Cohen; J A Cabico; J Friedman; G A Diamond Journal: Circulation Date: 1998-02-17 Impact factor: 29.690
Authors: Milan Lomsky; Jens Richter; Lena Johansson; Henrik El-Ali; Karl Aström; Michael Ljungberg; Lars Edenbrandt Journal: Clin Physiol Funct Imaging Date: 2005-07 Impact factor: 2.273
Authors: K F Van Train; E V Garcia; J Maddahi; J Areeda; C D Cooke; H Kiat; G Silagan; R Folks; J Friedman; L Matzer Journal: J Nucl Med Date: 1994-04 Impact factor: 10.057
Authors: D S Berman; X Kang; K F Van Train; H C Lewin; I Cohen; J Areeda; J D Friedman; G Germano; L J Shaw; R Hachamovitch Journal: J Am Coll Cardiol Date: 1998-12 Impact factor: 24.094
Authors: Levent A Guner; Nese Ilgin Karabacak; Ozgur U Akdemir; Pinar Senkul Karagoz; Sinan A Kocaman; Atiye Cengel; Mustafa Unlu Journal: J Nucl Cardiol Date: 2010-03-04 Impact factor: 5.952
Authors: Helen Fransson; Michael Ljungberg; Marcus Carlsson; Henrik Engblom; Håkan Arheden; Einar Heiberg Journal: J Nucl Cardiol Date: 2014-02-15 Impact factor: 5.952
Authors: A Chiba; T Kudo; R Ideguchi; M Altay; S Koga; T Yonekura; A Tsuneto; M Morikawa; S Ikeda; H Kawano; Y Koide; M Uetani; K Maemura Journal: Int J Cardiovasc Imaging Date: 2021-03-11 Impact factor: 2.357
Authors: Lars Edenbrandt; Peter Höglund; Sophia Frantz; Philip Hasbak; Allan Johansen; Lena Johansson; Annett Kammeier; Oliver Lindner; Milan Lomsky; Shinro Matsuo; Kenichi Nakajima; Karin Nyström; Eva Olsson; Karl Sjöstrand; Sven-Eric Svensson; Hiroshi Wakabayashi; Elin Trägårdh Journal: BMC Med Imaging Date: 2014-01-31 Impact factor: 1.930