Comparison between radiological and artificial neural network diagnosis and screening.

A. Degenhard, C. Tanner, C. Hayes, D. J. Hawkes, M. O. Leach, Fiona Jane Gilbert, MARIBS Study Group

Research output: Contribution to journalArticle

20 Citations (Scopus)

Abstract

The imaging protocol of the UK multicentre magnetic resonance imaging study for screening in women at genetic risk of breast cancer aims to assist in detecting and diagnosing malignant breast lesions. In this paper, we evaluate a three-layer, feed-forward, backpropagation neural network as an artificial radiological classifier using receiver operating characteristic (ROC) curve analysis and compare the results with those obtained using a proposed radiological scoring system for the study which currently supplements the radiologist's clinical opinion, in comparison with histological diagnosis. Based on the 76 symptomatic cases evaluated, descriptive features scored by radiologists showed considerable overlap between benign and malignant, although some features such as irregular contours and heterogeneous enhancement were more often associated with malignant pathology. In this preliminary evaluation, ROC analysis showed that the proposed scoring scheme did not perform well, indicating further refinement is required. When all 23 features were used in the neural network, its performance was poorer than that of the scoring scheme. When only ten features were used, limited to descriptors of enhancement characteristics, the neural network performed similar to the scoring scheme. This comparison shows that the neural network approach to clinical diagnosis has considerable potential and warrants further development.

Original languageEnglish
Pages (from-to)727-739
Number of pages12
JournalPhysiological Measurement
Volume3
Publication statusPublished - 2002

Keywords

  • magnetic resonance imaging
  • clinical imaging studies
  • dynamic contrast enhanced imaging
  • radiological diagnosis of breast cancer
  • artificial neural networks
  • receiver operating characteristic curves
  • BREAST-LESIONS
  • MAMMOGRAPHY
  • CLASSIFICATION
  • CANCER

Cite this

Degenhard, A., Tanner, C., Hayes, C., Hawkes, D. J., Leach, M. O., Gilbert, F. J., & MARIBS Study Group (2002). Comparison between radiological and artificial neural network diagnosis and screening. Physiological Measurement, 3, 727-739.

Comparison between radiological and artificial neural network diagnosis and screening. / Degenhard, A.; Tanner, C.; Hayes, C.; Hawkes, D. J.; Leach, M. O.; Gilbert, Fiona Jane; MARIBS Study Group.

In: Physiological Measurement, Vol. 3, 2002, p. 727-739.

Research output: Contribution to journalArticle

Degenhard, A, Tanner, C, Hayes, C, Hawkes, DJ, Leach, MO, Gilbert, FJ & MARIBS Study Group 2002, 'Comparison between radiological and artificial neural network diagnosis and screening.', Physiological Measurement, vol. 3, pp. 727-739.
Degenhard A, Tanner C, Hayes C, Hawkes DJ, Leach MO, Gilbert FJ et al. Comparison between radiological and artificial neural network diagnosis and screening. Physiological Measurement. 2002;3:727-739.
Degenhard, A. ; Tanner, C. ; Hayes, C. ; Hawkes, D. J. ; Leach, M. O. ; Gilbert, Fiona Jane ; MARIBS Study Group. / Comparison between radiological and artificial neural network diagnosis and screening. In: Physiological Measurement. 2002 ; Vol. 3. pp. 727-739.
@article{daaf6de9c9db455888cedc581fedcfcc,
title = "Comparison between radiological and artificial neural network diagnosis and screening.",
abstract = "The imaging protocol of the UK multicentre magnetic resonance imaging study for screening in women at genetic risk of breast cancer aims to assist in detecting and diagnosing malignant breast lesions. In this paper, we evaluate a three-layer, feed-forward, backpropagation neural network as an artificial radiological classifier using receiver operating characteristic (ROC) curve analysis and compare the results with those obtained using a proposed radiological scoring system for the study which currently supplements the radiologist's clinical opinion, in comparison with histological diagnosis. Based on the 76 symptomatic cases evaluated, descriptive features scored by radiologists showed considerable overlap between benign and malignant, although some features such as irregular contours and heterogeneous enhancement were more often associated with malignant pathology. In this preliminary evaluation, ROC analysis showed that the proposed scoring scheme did not perform well, indicating further refinement is required. When all 23 features were used in the neural network, its performance was poorer than that of the scoring scheme. When only ten features were used, limited to descriptors of enhancement characteristics, the neural network performed similar to the scoring scheme. This comparison shows that the neural network approach to clinical diagnosis has considerable potential and warrants further development.",
keywords = "magnetic resonance imaging, clinical imaging studies, dynamic contrast enhanced imaging, radiological diagnosis of breast cancer, artificial neural networks, receiver operating characteristic curves, BREAST-LESIONS, MAMMOGRAPHY, CLASSIFICATION, CANCER",
author = "A. Degenhard and C. Tanner and C. Hayes and Hawkes, {D. J.} and Leach, {M. O.} and Gilbert, {Fiona Jane} and {MARIBS Study Group}",
year = "2002",
language = "English",
volume = "3",
pages = "727--739",
journal = "Physiological Measurement",
issn = "0967-3334",
publisher = "IOP Publishing Ltd.",

}

TY - JOUR

T1 - Comparison between radiological and artificial neural network diagnosis and screening.

AU - Degenhard, A.

AU - Tanner, C.

AU - Hayes, C.

AU - Hawkes, D. J.

AU - Leach, M. O.

AU - Gilbert, Fiona Jane

AU - MARIBS Study Group

PY - 2002

Y1 - 2002

N2 - The imaging protocol of the UK multicentre magnetic resonance imaging study for screening in women at genetic risk of breast cancer aims to assist in detecting and diagnosing malignant breast lesions. In this paper, we evaluate a three-layer, feed-forward, backpropagation neural network as an artificial radiological classifier using receiver operating characteristic (ROC) curve analysis and compare the results with those obtained using a proposed radiological scoring system for the study which currently supplements the radiologist's clinical opinion, in comparison with histological diagnosis. Based on the 76 symptomatic cases evaluated, descriptive features scored by radiologists showed considerable overlap between benign and malignant, although some features such as irregular contours and heterogeneous enhancement were more often associated with malignant pathology. In this preliminary evaluation, ROC analysis showed that the proposed scoring scheme did not perform well, indicating further refinement is required. When all 23 features were used in the neural network, its performance was poorer than that of the scoring scheme. When only ten features were used, limited to descriptors of enhancement characteristics, the neural network performed similar to the scoring scheme. This comparison shows that the neural network approach to clinical diagnosis has considerable potential and warrants further development.

AB - The imaging protocol of the UK multicentre magnetic resonance imaging study for screening in women at genetic risk of breast cancer aims to assist in detecting and diagnosing malignant breast lesions. In this paper, we evaluate a three-layer, feed-forward, backpropagation neural network as an artificial radiological classifier using receiver operating characteristic (ROC) curve analysis and compare the results with those obtained using a proposed radiological scoring system for the study which currently supplements the radiologist's clinical opinion, in comparison with histological diagnosis. Based on the 76 symptomatic cases evaluated, descriptive features scored by radiologists showed considerable overlap between benign and malignant, although some features such as irregular contours and heterogeneous enhancement were more often associated with malignant pathology. In this preliminary evaluation, ROC analysis showed that the proposed scoring scheme did not perform well, indicating further refinement is required. When all 23 features were used in the neural network, its performance was poorer than that of the scoring scheme. When only ten features were used, limited to descriptors of enhancement characteristics, the neural network performed similar to the scoring scheme. This comparison shows that the neural network approach to clinical diagnosis has considerable potential and warrants further development.

KW - magnetic resonance imaging

KW - clinical imaging studies

KW - dynamic contrast enhanced imaging

KW - radiological diagnosis of breast cancer

KW - artificial neural networks

KW - receiver operating characteristic curves

KW - BREAST-LESIONS

KW - MAMMOGRAPHY

KW - CLASSIFICATION

KW - CANCER

M3 - Article

VL - 3

SP - 727

EP - 739

JO - Physiological Measurement

JF - Physiological Measurement

SN - 0967-3334

ER -