The purpose of this study is to increase the specificity of breast biopsy by building computer models which combine both mammography and breast ultrasound (US) findings to identify probably benign breast masses. In current clinical practice, breast US is used only to distinguish between fluid-filled cysts vs. solid masses. The proposed artificial neural network (ANN) model would go one step further and quantitatively identify probably benign cases which may undergo short-term follow-up in lieu of biopsy. The hypothesis is that by combining information from both modalities, the model will be more robust and more accurate than those based upon either modality alone, and be able to improve upon the performance of the radiologists. In preliminary studies, ANN models successfully identified probably benign breast masses using just mammographic findings or just US findings.
The specific aims of the proposed study are to: 1. Prospectively collect data for 300 cases of biopsy-proven breast lesions for which mammography and ultrasound (US) data are both available. 2. Optimize artificial neural network (ANN) models to identify probably lesions based on US findings only. 3. Develop unified models to identify probably benign lesions using both US and mammography findings. 4. Perform statistical analysis to evaluate contribution of US findings to the diagnostic performances of radiologists and ANN models. The immediate benefit of this proposal is a computer-based decision aid to improve the specificity of breast biopsy and thus reduce the cost associated with benign biopsies. This proposal has the potential to reduce significantly the number of unnecessary breast biopsies and their associated cost, physical pain, and emotional distress to the patient.

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Exploratory/Developmental Grants (R21)
Project #
5R21CA093461-02
Application #
6620433
Study Section
Diagnostic Imaging Study Section (DMG)
Program Officer
Torres-Anjel, Manuel J
Project Start
2002-03-01
Project End
2005-02-28
Budget Start
2003-03-17
Budget End
2005-02-28
Support Year
2
Fiscal Year
2003
Total Cost
$154,000
Indirect Cost
Name
Duke University
Department
Radiation-Diagnostic/Oncology
Type
Schools of Medicine
DUNS #
044387793
City
Durham
State
NC
Country
United States
Zip Code
27705
Jesneck, Jonathan L; Lo, Joseph Y; Baker, Jay A (2007) Breast mass lesions: computer-aided diagnosis models with mammographic and sonographic descriptors. Radiology 244:390-8
Jesneck, Jonathan L; Nolte, Loren W; Baker, Jay A et al. (2006) Optimized approach to decision fusion of heterogeneous data for breast cancer diagnosis. Med Phys 33:2945-54