Advertisement

Journal of Digital Imaging

, Volume 7, Issue 4, pp 196–207 | Cite as

Reduction of false positives in computerized detection of lung nodules in chest radiographs using artificial neural networks, discriminant analysis, and a rule-based scheme

  • Yuzheng C. Wu
  • Kunio Doi
  • Maryellen L. Giger
  • Charles E. Metz
  • Wei Zhang
Article

Abstract

A computer-aided diagnosis (CAD) scheme is being developed to identify image regions considered suspicious for lung nodules in chest radiographs to assist radiologists in making correct diagnoses. Automated classifiers—an artificial neural network, discriminant analysis, and a rule-based scheme—are used to reduce the number of false-positive detections of the CAD scheme. The CAD scheme first detects nodule candidates from chest radiographs based on a difference image technique. Nine image features characterizing nodules are extracted automatically for each of the nodule candidates. The extracted image features are then used as input data to the classifiers for distinguishing actual nodules from the false-positive detections. The performances of the classifiers are evaluated by receiver-operating characteristic analysis. On the basis of the database of 30 normal and 30 abnormal chest images, the neural network achieves an AZ value (area under the receiver-operating-characteristic curve) of 0.79 in detecting lung nodules, as tested by the round-robin method. The neural network, after being trained with a training database, is able to eliminate more than 83% of the false-positive detections reported by the CAD scheme. Moreover, the combination of the trained neural network and a rule-based scheme eliminates 96% of the false-positive detections of the CAD scheme.

Key Words

Artificial neural networks discriminant analysis rule-based schemme lung nodules receiver-operating characteristic (ROC) analysis 

References

  1. 1.
    Rumelhart DE, McClellend JL (eds): Parallel Distributed Processing, Cambridge, MA, MIT, 1986Google Scholar
  2. 2.
    Hecht-Nielsen R: Neurocomputing. Reading, MA, Addison-Wesley, 1990Google Scholar
  3. 3.
    Boone JM, Sigilitto VG, Shaber GS: Neural networks in radiology: An introduction and evaluation in a signal detection task. Med Phys 17:234–241, 1990CrossRefPubMedGoogle Scholar
  4. 4.
    Wu Y, Doi K, Metz CE, et al: Simulation studies of data classification by artificial neural networks: Potential applications in medical imaging and decision making. J Digit Imaging 6:117–125, 1993PubMedGoogle Scholar
  5. 5.
    Wu Y, Doi K, Giger ML, et al: Computerized detection of clustered microcalcifications in digital mammograms: Applications of artificial neural networks. Med Phys 19:555–560, 1992CrossRefPubMedGoogle Scholar
  6. 6.
    Wu Y, Doi K, Giger ML, et al: Application of neural networks in mammography for the diagnosis of breast cancer. Proc SPIE Imaging Technol Appl 1778:19–27, 1992Google Scholar
  7. 7.
    Wu Y, Giger ML, Doi K, et al: Artificial neural networks in mammography: Application to decision making in the diagnosis of breast cancer. Radiology 187:81–87, 1993PubMedGoogle Scholar
  8. 8.
    Asada N, Doi K, MacMahon H, et al: Potential usefulness of an artificial neural network for differential diagnosis of interstitial lung diseases: Pilot study. Radiology 177:857–860, 1990PubMedGoogle Scholar
  9. 9.
    Gross GW, Boone JM, Greco-Hunt V, et al: Neural networks in radiologic diagnosis: II. Interpretation of neonatal chest radiographs. Invest Radiol 25:1017–1023, 1990CrossRefPubMedGoogle Scholar
  10. 10.
    Wu Y, Doi K, Giger ML, et al: Detection of lung nodules in digital chest radiographs: Artificial neural networks and discriminant analysis. Presented at the 34th annual meeting of AAPM, Calgary, Canada, August 23–27, 1992Google Scholar
  11. 11.
    Wu Y, Doi K, Giger ML, et al: Detection of lung nodules in digital chest radiographs: Comparison of artificial neural networks and discriminant analysis. Presented at the 78th scientific assembly and annual meeting of RSNA, Radiology 185 (P):156, 1992 (suppl, abstr)Google Scholar
  12. 12.
    Lo SCB, Freedman MT, Lin JS, et al: Profile-matching techniques and neural classifier for automatic nodule detections in pulmonary radiographs. Presented at the 78th scientific assembly and annual meeting of RSNA, Radiology 185 (P):196, 1992 (suppl, abstr)Google Scholar
  13. 13.
    Garg S, Floyd CE: Neural network localization of pulmonary nodules on digital chest radiographs. Presented at the 78th scientific assembly and annual meeting of RSNA, Radiology 185 (P):157, 1992 (suppl, abstr)Google Scholar
  14. 14.
    Kim JH, Min BG, Han MC, et al: Computer-assisted detection of lung nodules by using artificial neural net. Presented at the 78th scientific assembly and annual meeting of RSNA. Radiology 185 (P):156, 1992 (suppl, abstr)Google Scholar
  15. 15.
    Lo SCB, Freedman MT, Lin JS, et al: Automatic lung nodule detection using profile matching and back-propagation neural network techniques. J Digit Imaging 6:48–54, 1993PubMedCrossRefGoogle Scholar
  16. 16.
    Chiou YS, Lin JS, Fleming Lure YM, et al: Shape feature analysis using artificial neural networks for improvements of hybrid lung nodule detection (HLND) system. Image Processing, Medical Imaging VII, SPIE Medical Imaging 1993 Symposium, Newport Beach, CA, February 13–19, 1993Google Scholar
  17. 17.
    Chiou YS, Lin JS, Fleming Lure YM, et al: Neural network based hybrid lung nodule detection (HLND) system. 1993 IEEE International Conference on Neural Network, San Francisco, CA, March 28-April 1993.Google Scholar
  18. 18.
    Lo JY, Floyd CE, Bowsher JE, et al: Spatially varying scatter estimation in portable chest radiography with an artificial neural network. Presented at the 78th scientific assembly and annual meeting of RSNA, Radiology 185 (P):300, 1992Google Scholar
  19. 19.
    Floyd CE, Tourassi GD: An artificial neural network for lesion detection on single-photon emission computed tomographic images. Invest Radiol 28:667–672, 1992CrossRefGoogle Scholar
  20. 20.
    Chan KK, Hayrapetian AS, Lau CC, et al: Neural network segmentation of double-echo MR images. Presented at the 78th scientific assembly and annual meeting of RSNA. Radiology 185 (P):157, 1992Google Scholar
  21. 21.
    Forrest JV, Friedman PJ: Radiologic errors in patients with lung cancer. West J Med 134:485–490, 1981PubMedGoogle Scholar
  22. 22.
    Giger ML, Doi K, MacMahon H: Image feature analysis and computer-aided diagnosis in digital radiography, III: Automated detection of nodules in peripheral lung fields. Med Phys 15:158–166, 1988CrossRefPubMedGoogle Scholar
  23. 23.
    Giger ML, Doi K, MacMahon H, et al: Pulmonary nodules: Computer-aided detection in digital chest images. RadioGraphics 10:41–51, 1990PubMedGoogle Scholar
  24. 24.
    Giger ML, Ahn K, Doi K, et al: Computerized detection of pulmonary nodules in digital chest images: Use of morphological filters in reducing false-positive detections. Med Phys 17:861–865, 1990CrossRefPubMedGoogle Scholar
  25. 25.
    Yoshimura H, Giger ML, Doi K, et al: Computerized scheme for the detection of pulmonary nodules A nonlinear filtering technique. Invest Radiol 27:124–129, 1992CrossRefPubMedGoogle Scholar
  26. 26.
    Katsuragawa S, Doi K, MacMahon H: Image feature analysis and computer-aided diagnosis in digital radiography: Detection and characterization of interstitial disease in digital chest radiography. Med Phys 15:311–319, 1988CrossRefPubMedGoogle Scholar
  27. 27.
    Ishida M, Kato H, Doi K, et al: Development of a new digital radiographic image processing. Proc SPIE 347:42–48, 1982Google Scholar
  28. 28.
    Metz CE: ROC methodology in radiologic imaging. Invest Radiol 21:720–733, 1986CrossRefPubMedGoogle Scholar
  29. 29.
    Metz CE, Shen JH, Herman BA: New methods for estimating a binormal ROC curve from continuously-distributed test results. Presented at the 1990 Joint Meetings of the American Statistical Society and the Biometric Society, Anaheim, CA, August 1990Google Scholar
  30. 30.
    Metz CE: Some practical issues of experimental design and data analysis in radiological ROC studies. Invest Radiol 24:234–245, 1989CrossRefPubMedGoogle Scholar
  31. 31.
    Metz CE, Wang PL, Kronman HB: A new approach for testing the significance of differences between ROC curves measured from correlated data, in Deconinck F (eds): Information Processing in Medical Imaging. Nijhoff, The Hague, The Netherlands, 1989, pp 432–445Google Scholar
  32. 32.
    Metz CE: Quantification of failure to demonstrate statistical significance: The usefulness of confidence intervals. Invest Radiol 28:59–63, 1993CrossRefPubMedGoogle Scholar
  33. 33.
    Pratt WK: Digital Image Processing. New York, NY, Wiley, 1978Google Scholar
  34. 34.
    Sanada S, Doi K, MacMahon H: Image feature analysis and computer-aided diagnosis in digital chest radiography: Automated delineation of posterior ribs in chest images. Med Phys 18:964–971, 1991CrossRefPubMedGoogle Scholar
  35. 35.
    Rumelhart DE, Hinton GE, Williams RJ: Learning internal representations by error propagation, in Rumelhart DE, McClellend JL (eds): Parallel Distributed Processing. Cambridge, MA, MIT, 1986, pp 318–362Google Scholar
  36. 36.
    Lachenbruch PA, Discriminant Analysis. New York, NY, Hafner, 1975Google Scholar
  37. 37.
    Matsumoto T, Yoshimura H, Giger ML, et al: Potential usefulness of computerized nodule detection on screening programs for lung cancer. Invest Radiol 27:471–475, 1992CrossRefPubMedGoogle Scholar
  38. 38.
    Matsumoto T, Yoshimura H, Doi K, et al: Image feature analysis of false-positive diagnoses produced by automated detection of lung nodules. Invest Radiol 27:587–597, 1992CrossRefPubMedGoogle Scholar

Copyright information

© Society for Imaging Informatics in Medicine 1994

Authors and Affiliations

  • Yuzheng C. Wu
    • 1
    • 2
  • Kunio Doi
    • 1
    • 2
  • Maryellen L. Giger
    • 1
    • 2
  • Charles E. Metz
    • 1
    • 2
  • Wei Zhang
    • 1
    • 2
  1. 1.Kurt Rossmann Laboratories for Radiologic Image Research, Department of RadiologyThe University of ChicagoChicago
  2. 2.Center of Information Sciences and Imaging Systems, Department of RadiologyGeorgetown UniversityWashington, DC

Personalised recommendations