Specificity/Sensitivity in Mammographic Screening

Summary

Principal Investigator: Daniel B Kopans
Affiliation: Massachusetts General Hospital
Country: USA
Abstract: The early detection of breast cancer saves lives. Mammography is the best method of detecting early stage cancer. However, a significant fraction of breast cancers are missed on mammographic screening, primarily because the complex background of normal, radiographically dense breast structures may mask the presence of breast cancer. The goal of this project is to utilize digital tomosynthesis to "see through" the dense breast tissues, to improve the visibility of breast lesions and increase early cancer detection. In conventional tomography, the x-ray source is moved above the patient and one tomography plane is obtained for each movement of the source. In tomosynthesis, multiple tomography planes can be reconstructed from a single movement of the x-ray source. In the proposed tomosynthesis method, multiple images are acquired as the x-ray tube is moved in an arc above the stationary breast and digital detector. A step-and-expose method of imaging is used. The image obtained at each angle is of low radiation dose, with the total radiation dose required for imaging the entire breast being approximately equal to the dose used for a single film-screen mammogram. By shifting and adding the images, it is possible to bring any plane of the breast into sharp focus. Breast structures outside the plane are blurred. Thus, lesions within a selected plane can be detected more easily. The MGH-General Electric Prototype Digital Breast Tomosynthesis (DBT) system installed at our institution is ideally suited to explore DBT because of its properties: 1) low noise, 2) high resolution, 3) large flat detector area with minimal image distortion, and 4) rapid image readout times. We have developed the initial algorithms and techniques for DBT. We have patented and licensed the Tomosynthesis system to General Electric Medical Systems Base don our initial experience, we propose to perform a 3000 subject clinical screening trial of breast tomosynthesis to determine specificity improvement (reduced callbacks) and estimate sensitivity improvement in order to design a multi-site screening trial. We expect to detect otherwise occult breast cancers in the screening setting. Tomosynthesis and conventional digital imaging will be evaluated on their ability to predict breast cancer by comparing the recommendations from each modality with clinical truth. MGH will coordinate the clinical study, determine the specific technical parameters to be used, perform quality control and data analysis. The clinical effectiveness of breast tomosynthesis will be determined. This program will accelerate an innovative technique for breast imaging from the pre-clinical stages into clinical practice. It is expected that tomosynthesis will have major impact on the practice of mammography, providing the radiologist with an important tool for detecting cancer lesions which are difficult to see with current imaging technology. Tomosynthesis may result in improved detection, characterization and visualization of breast lesions which may ultimately reduce mortality from breast cancer, while dramatically reducing callback imaging studies.
Funding Period: 2004-05-01 - 2007-04-30
more information: NIH RePORT

Top Publications

  1. pmc Enhanced imaging of microcalcifications in digital breast tomosynthesis through improved image-reconstruction algorithms
    Emil Y Sidky
    Department of Radiology, University of Chicago, 5841 S Maryland Avenue Chicago, Illinois 60637, USA
    Med Phys 36:4920-32. 2009

Scientific Experts

Detail Information

Publications1

  1. pmc Enhanced imaging of microcalcifications in digital breast tomosynthesis through improved image-reconstruction algorithms
    Emil Y Sidky
    Department of Radiology, University of Chicago, 5841 S Maryland Avenue Chicago, Illinois 60637, USA
    Med Phys 36:4920-32. 2009
    ..The authors develop a practical, iterative algorithm for image-reconstruction in undersampled tomographic systems, such as digital breast tomosynthesis (DBT)...