 |
Department: | Biomedical Engineering Deaprtment of Radiology |
| Division: | Biomedical Engineering/ Radiology | |
| Address: | 2424 Erwin Road Suite 302 Durham, NC 27705 |
|
| Phone: | (919) 684-1459/ (919) 358-0852 | |
| Fax: | (919) 684-1491 | |
| Email: | swatee.singh@duke.edu | |
| Web: |
Area of expertise
My work involves digital medical imaging, specifically the processing, and manupilation of medical images to help radiologists for the task of breast cancer detection with the aid of a computer-aided detection (CADe) tool. I work with both mammographic images as well as digital tomosynthesis images and combine tools from the fields of digital signal processing, computer vision, and machine learning to accomplish this task.
Why CADe?
For women in the United States, breast cancer is the second-most deadly type of cancer. Survival rates are significantly higher when the cancer is detected at an early stage. At present, the most common, and effective early-detection tool currently available to clinicians is screening mammography, which still misses 10-30% of breast cancers. Mammograms are read more accurately when read by more than one mammographer. Unfortunately, double reading is not practical in US. CADe systems have been demonstrated to serve as a reliable, accurate, and efficient second-reader to aid mammographers. One study indicated that a commercial CADe system can successfully identify 77% of overlooked breast malignancies, while another demonstrated that routine use of a CADe system may increase the number of cancers detected at screening mammography up to 20%.
Digital Tomosynthesis and CADe
Digital tomosynthesis is under active development at this institution and others. These systems provide 3D slice images from a modified digital mammography system. It is believed that tomosynthesis will show improvement in characterization, and detection of breast masses by removing overlapping dense fibroglandular tissue. The goal is to provide 3D information at high resolution, comparable dose to mammography, and with lower cost and hardware requirements compared to alternatives such as breast Computed Tomography or breast Magnetic Resonance Imaging.
In a screening setting, radiologists typically look at 4 views per patient for mammography. However, if tomosynthesis were to replace mammography as a screening tool, then a radiologist would potentially have to look at 50 to 80 reconstructed slices per exam. This increase in the number of images will likely affect workflow dramatically. The role of CADe in such a setting becomes even more important as not just a second reader, but also to possibly identify initial suspicious breast volumes for the radiologist to focus their attention. As current investigators in CT colonography have suggested, CADe can potentially ease radiologist workflow when working with large 3D data sets.
Peer Reviewed Publications
G. D. Tourassi, B. Harrawood, S. Singh, J. Y. Lo, C. E. Floyd, Jr, “Evaluation of Information-Theoretic Similarity Measures for Content Based Retrieval and Detection of Masses in Mammograms,” Medical Physics 2007
G. D. Tourassi, B. Harrawood, S. Singh, J. Y. Lo, “Information- Theoretic CAD System in Mammography: Entropy- Based Indexing for Computational Efficiency and Robust Performance,” (accepted for publication with minor revisions in Medical Physics)
Conference Proceedings
S. Singh, G. Tourassi, “Breast Mass Detection in Tomosynthesis Projection Images Using Information Theoretic Similarity Measures.” Medical Imaging: CAD, 2007
A. Jerebko, Y. Quan, N. Merlet, E. Ratner, S. Singh, J. Y. Lo, Arun Krishnan, “Feasibility study of breast tomosynthesis CAD system.” Medical Imaging: CAD, 2007
J. Y. Lo, J. A. Baker, J. Orman, T. Mertelmeier, S. Singh, “Breast Tomosynthesis: Initial Clinical Experience with 100 Human Subjects.” RSNA 2006
S. Singh, A. Baydush, B. Harrawood, J. Y. Lo, “Mass Detection in Mammographic ROIs using Watson Filters.” Medical Imaging: Medical Perception, 2006
S. Singh, J. Y. Lo, D. Catarious, C. E. Floyd, “Design and Clinical Efficacy of a Computer-Aided Detection Tool For Masses In Mammograms.” Era of Hope, US Army, 2005
M. Bissonnette, M. Hansroul, E. Masson, S. Savard, S. Cadieux, P. Warmoes, D. Gravel, J. Agopyan, B. T. Polischuk, W. H. Haerer, T. Mertelmeier, J. Y. Lo, Y. Chen, J. T. Dobbins III, J. L. Jesneck, S. Singh, “Digital breast tomosynthesis using an amorphous selenium flat panel detector”, Medical Imaging: Physics of Medical Imaging, 2005
J. Y. Lo, E. Samei. J. L. Jesneck, J. T. Dobbins, J. A. Baker, R. Saunders, S. Singh, C. E. Floyd, Jr, “Radiographic technique optimization for amorphous selenium FFSM system: phantom and patient results.” International Workshop for Digital Mammography (IWDM), 2004