Solitary pulmonary nodules: Determining the likelihood of malignancy with neural network analysis

J. W. Gurney, S. J. Swensen

Research output: Contribution to journalArticle

65 Citations (Scopus)

Abstract

PURPOSE: To test a neural network in differentiation of benign from malignant solitary pulmonary nodules. MATERIALS AND METHODS: Neural networks were trained and tested on the characteristics of 318 nodules. Predictive accuracy of the network was judged for calibration and discrimination. Network results were compared with those with a simpler Bayesian method. RESULTS: The Brier score was 0.142 (calibration, 0.003; discrimination, 0.139) for the neural network and 0.133 for the Bayesian analysis (calibration, 0.012; discrimination, 0.121). Analysis of the calibration curve revealed no significant difference (P < .05) between the slope (b = 1.09) and the line of identity (b = 1) for the neural network or the Bayesian analysis. The area under the receiver operating characteristic curve was 0.871 for the neural network and 0.894 for the Bayesian analysis (P < .05). There were 23 and 21 false-positive predictions and 18 and six false-negative predictions for the neural network and Bayesian analysis, respectively. CONCLUSION: The Bayesian method was better than the neural network in prediction of probability of malignancy in solitary pulmonary nodules.

Original languageEnglish (US)
Pages (from-to)823-829
Number of pages7
JournalRadiology
Volume196
Issue number3
StatePublished - 1995

Fingerprint

Solitary Pulmonary Nodule
Bayes Theorem
Calibration
Neoplasms
ROC Curve

Keywords

  • Computers, diagnostic aid
  • Computers, neural network
  • Lung neoplasms, diagnosis
  • Lung, nodule

ASJC Scopus subject areas

  • Radiological and Ultrasound Technology

Cite this

Solitary pulmonary nodules : Determining the likelihood of malignancy with neural network analysis. / Gurney, J. W.; Swensen, S. J.

In: Radiology, Vol. 196, No. 3, 1995, p. 823-829.

Research output: Contribution to journalArticle

Gurney, J. W. ; Swensen, S. J. / Solitary pulmonary nodules : Determining the likelihood of malignancy with neural network analysis. In: Radiology. 1995 ; Vol. 196, No. 3. pp. 823-829.
@article{5a32ab91eabf4a5eb9139210d50ba14b,
title = "Solitary pulmonary nodules: Determining the likelihood of malignancy with neural network analysis",
abstract = "PURPOSE: To test a neural network in differentiation of benign from malignant solitary pulmonary nodules. MATERIALS AND METHODS: Neural networks were trained and tested on the characteristics of 318 nodules. Predictive accuracy of the network was judged for calibration and discrimination. Network results were compared with those with a simpler Bayesian method. RESULTS: The Brier score was 0.142 (calibration, 0.003; discrimination, 0.139) for the neural network and 0.133 for the Bayesian analysis (calibration, 0.012; discrimination, 0.121). Analysis of the calibration curve revealed no significant difference (P < .05) between the slope (b = 1.09) and the line of identity (b = 1) for the neural network or the Bayesian analysis. The area under the receiver operating characteristic curve was 0.871 for the neural network and 0.894 for the Bayesian analysis (P < .05). There were 23 and 21 false-positive predictions and 18 and six false-negative predictions for the neural network and Bayesian analysis, respectively. CONCLUSION: The Bayesian method was better than the neural network in prediction of probability of malignancy in solitary pulmonary nodules.",
keywords = "Computers, diagnostic aid, Computers, neural network, Lung neoplasms, diagnosis, Lung, nodule",
author = "Gurney, {J. W.} and Swensen, {S. J.}",
year = "1995",
language = "English (US)",
volume = "196",
pages = "823--829",
journal = "Radiology",
issn = "0033-8419",
publisher = "Radiological Society of North America Inc.",
number = "3",

}

TY - JOUR

T1 - Solitary pulmonary nodules

T2 - Determining the likelihood of malignancy with neural network analysis

AU - Gurney, J. W.

AU - Swensen, S. J.

PY - 1995

Y1 - 1995

N2 - PURPOSE: To test a neural network in differentiation of benign from malignant solitary pulmonary nodules. MATERIALS AND METHODS: Neural networks were trained and tested on the characteristics of 318 nodules. Predictive accuracy of the network was judged for calibration and discrimination. Network results were compared with those with a simpler Bayesian method. RESULTS: The Brier score was 0.142 (calibration, 0.003; discrimination, 0.139) for the neural network and 0.133 for the Bayesian analysis (calibration, 0.012; discrimination, 0.121). Analysis of the calibration curve revealed no significant difference (P < .05) between the slope (b = 1.09) and the line of identity (b = 1) for the neural network or the Bayesian analysis. The area under the receiver operating characteristic curve was 0.871 for the neural network and 0.894 for the Bayesian analysis (P < .05). There were 23 and 21 false-positive predictions and 18 and six false-negative predictions for the neural network and Bayesian analysis, respectively. CONCLUSION: The Bayesian method was better than the neural network in prediction of probability of malignancy in solitary pulmonary nodules.

AB - PURPOSE: To test a neural network in differentiation of benign from malignant solitary pulmonary nodules. MATERIALS AND METHODS: Neural networks were trained and tested on the characteristics of 318 nodules. Predictive accuracy of the network was judged for calibration and discrimination. Network results were compared with those with a simpler Bayesian method. RESULTS: The Brier score was 0.142 (calibration, 0.003; discrimination, 0.139) for the neural network and 0.133 for the Bayesian analysis (calibration, 0.012; discrimination, 0.121). Analysis of the calibration curve revealed no significant difference (P < .05) between the slope (b = 1.09) and the line of identity (b = 1) for the neural network or the Bayesian analysis. The area under the receiver operating characteristic curve was 0.871 for the neural network and 0.894 for the Bayesian analysis (P < .05). There were 23 and 21 false-positive predictions and 18 and six false-negative predictions for the neural network and Bayesian analysis, respectively. CONCLUSION: The Bayesian method was better than the neural network in prediction of probability of malignancy in solitary pulmonary nodules.

KW - Computers, diagnostic aid

KW - Computers, neural network

KW - Lung neoplasms, diagnosis

KW - Lung, nodule

UR - http://www.scopus.com/inward/record.url?scp=0029102693&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0029102693&partnerID=8YFLogxK

M3 - Article

C2 - 7644650

AN - SCOPUS:0029102693

VL - 196

SP - 823

EP - 829

JO - Radiology

JF - Radiology

SN - 0033-8419

IS - 3

ER -