Breast cancer is one of the leading causes of death in women. Early detection, diagnosis and treatment of breast cancer can significantly reduce breast cancer morbidity and mortality. Accurate mammogram interpretation by radiologists is the key to the early detection of breast cancer. But, it is still challenging for radiologists to maintain accurate and efficient interpretation when given a massive volume of mammograms. With the advances of digital radiography and computer technology, computer- aided detection (CAD) schemes have been investigated with the hope to provide a """"""""second opinion"""""""" to radiologists so that their diagnostic accuracy and efficiency can be significantly improved. In this project, the applicants proposed to develop a knowledge-based CAD scheme, to optimize the scheme, and to evaluate scheme's performance for the identification of breast abnormalities on digitized mammograms. The design of the scheme includes a learning process to establish a knowledge base and an identification process to identify breast abnormalities on digitized mammograms. The main idea is to investigate a CAD scheme using knowledge based approaches so that the scheme is largely independent from other rule-based approaches. The methods include quantitative characterization of a suspicious region, and """"""""similarity"""""""" measures among the suspicious region and regions with """"""""known"""""""" truth (""""""""positive"""""""" or """"""""negative""""""""). In addition, a classification scheme will be investigated to classify the suspicious region as either """"""""positive"""""""" or """"""""negative"""""""" using a """"""""likelihood"""""""" measure. Efforts will be undertaken to optimize and to evaluate the knowledge-based CAD scheme in terms of sensitivity and specificity using patient cases that are reviewed and verified by radiologists retrospectively. Receiver- operating characteristics (ROC) analysis will also be performed to evaluate the scheme performance using a clinical database that was never used during the scheme development and optimization.

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Exploratory/Developmental Grants (R21)
Project #
5R21CA082912-02
Application #
6150405
Study Section
Diagnostic Imaging Study Section (DMG)
Program Officer
Menkens, Anne E
Project Start
1999-02-01
Project End
2002-01-31
Budget Start
2000-02-01
Budget End
2002-01-31
Support Year
2
Fiscal Year
2000
Total Cost
$150,000
Indirect Cost
Name
University of Pittsburgh
Department
Radiation-Diagnostic/Oncology
Type
Schools of Medicine
DUNS #
004514360
City
Pittsburgh
State
PA
Country
United States
Zip Code
15213
Chang, Yuan-Hsiang; Wang, Xiao-Hui; Hardesty, Lara A et al. (2002) Computerized assessment of tissue composition on digitized mammograms. Acad Radiol 9:899-905
Chang, Y H; Hardesty, L A; Hakim, C M et al. (2001) Knowledge-based computer-aided detection of masses on digitized mammograms: a preliminary assessment. Med Phys 28:455-61
Zheng, B; Sumkin, J H; Good, W F et al. (2000) Applying computer-assisted detection schemes to digitized mammograms after JPEG data compression: an assessment. Acad Radiol 7:595-602
Chang, Y H; Good, W F; Sumkin, J H et al. (1999) Computerized localization of breast lesions from two views. An experimental comparison of two methods. Invest Radiol 34:585-8