Special Section on Pioneers in Medical Imaging: Honoring the Memory of Robert F. Wagner

Impact of lesion segmentation metrics on computer-aided diagnosis/detection in breast computed tomography

[+] Author Affiliations
Hsien-Chi Kuo

University of Chicago, Department of Radiology, 5841 S. Maryland Avenue, Chicago 60637, Illinois, United States

Maryellen L. Giger

University of Chicago, Department of Radiology, 5841 S. Maryland Avenue, Chicago 60637, Illinois, United States

Ingrid Reiser

University of Chicago, Department of Radiology, 5841 S. Maryland Avenue, Chicago 60637, Illinois, United States

Karen Drukker

University of Chicago, Department of Radiology, 5841 S. Maryland Avenue, Chicago 60637, Illinois, United States

John M. Boone

University of California at Davis, Department of Radiology, 4860 Y Street, Suite 3100, Sacramento 95817, California, United States

Karen K. Lindfors

University of California at Davis, Department of Radiology, 4860 Y Street, Suite 3100, Sacramento 95817, California, United States

Kai Yang

University of Oklahoma Health Sciences Center, Department of Radiological Sciences, 940 N.E. 13th Street, Oklahoma City 73104, Oklahoma, United States

Alexandra Edwards

University of Chicago, Department of Radiology, 5841 S. Maryland Avenue, Chicago 60637, Illinois, United States

J. Med. Imag. 1(3), 031012 (Dec 24, 2014). doi:10.1117/1.JMI.1.3.031012
History: Received May 19, 2014; Accepted December 2, 2014
Text Size: A A A

Abstract.  Evaluation of segmentation algorithms usually involves comparisons of segmentations to gold-standard delineations without regard to the ultimate medical decision-making task. We compare two segmentation evaluations methods—a Dice similarity coefficient (DSC) evaluation and a diagnostic classification task–based evaluation method using lesions from breast computed tomography. In our investigation, we use results from two previously developed lesion-segmentation algorithms [a global active contour model (GAC) and a global with local aspects active contour model]. Although similar DSC values were obtained (0.80 versus 0.77), we show that the global + local active contour (GLAC) model, as compared with the GAC model, is able to yield significantly improved classification performance in terms of area under the receivers operating characteristic (ROC) curve in the task of distinguishing malignant from benign lesions. [Area under the ROC curve(AUC)=0.78 compared to 0.63, p0.001]. This is mainly because the GLAC model yields better detailed information required in the calculation of morphological features. Based on our findings, we conclude that the DSC metric alone is not sufficient for evaluating segmentation lesions in computer-aided diagnosis tasks.

Figures in this Article
© 2014 Society of Photo-Optical Instrumentation Engineers

Citation

Hsien-Chi Kuo ; Maryellen L. Giger ; Ingrid Reiser ; Karen Drukker ; John M. Boone, et al.
"Impact of lesion segmentation metrics on computer-aided diagnosis/detection in breast computed tomography", J. Med. Imag. 1(3), 031012 (Dec 24, 2014). ; http://dx.doi.org/10.1117/1.JMI.1.3.031012


Access This Article
Sign in or Create a personal account to Buy this article ($20 for members, $25 for non-members).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Related Book Chapters

Topic Collections

PubMed Articles
Advertisement
  • Don't have an account?
  • Subscribe to the SPIE Digital Library
  • Create a FREE account to sign up for Digital Library content alerts and gain access to institutional subscriptions remotely.
Access This Article
Sign in or Create a personal account to Buy this article ($20 for members, $25 for non-members).
Access This Proceeding
Sign in or Create a personal account to Buy this article ($15 for members, $18 for non-members).
Access This Chapter

Access to SPIE eBooks is limited to subscribing institutions and is not available as part of a personal subscription. Print or electronic versions of individual SPIE books may be purchased via SPIE.org.