This proposal addresses the need for and potential solution to subspeckle visibility of small hypoechogenic tumors in the breast. The work is a collaboration between the Valley Forge Research Center (VFRC), Dept. of Electrical Engineering, UPenn, and the Breast Imaging Section, Dept. of Radiology, Hospital of the Univ. of Pennsylvania (HUP). In vivo measurements on mammography patients at HUP will be the primary source of data for developing and testing algorithms and procedures at VFRC. Breast cancer is a significant killer of women (approximately 40,000 Americans per yr.), many young and in their child rearing years. Molecular biology and genetic engineering hold hopes for eventual means of prevention; until that time, however, the best route is early detection of tumors, distinction between benign processes and actively growing masses, and elimination of malignant lesions by clinical practices. This proposal addresses the first two problems (early detection and distinction between benign and malignant lesions), with primary emphasis upon the first. Current clinical practice (particularly needle biopsy) would be largely sufficient if the first two problems were solved. Thus early detection of tumors is the bottleneck and therefore the key to saving lives. The preponderance of breast tumors originate in terminal ducts (approximately 90 percent, as inferred from data in Page et al. 1987), which are the smallest branches of the ductal system that transports milk from the mammary glands to the nipple. """"""""[I]f breast cancer can be reliably detected in the in situ stage, and successfully removed, survival should be significantly increased"""""""" (Sullivan 1997). The problem is difficult because terminal ductal crossection is a fraction of a mm. To eliminate intraductal tumors by existing clinical practices and thereby provide a high probability of a long term survival to patients, tumor detection must be accomplished before there is danger of penetration of duct wall. Ducts expand as carcinogenic cells multiply in the ductal interior and ductal crossection can expand to two or more mm before there is significant danger of penetration (Alcorn 1990). Thus tumors should be discovered and evaluated at approximately 2 mm. A further difficulty is that most breast tumors are hypoechogenic, causing small tumors to easily disappear into the surrounding soft-tissue speckle background when imaged in a B-scan, making them difficult to detect. """"""""Speckle interference can easily obscure the normal duct structure, so that an adequate algorithm is necessary to suppress this problem as much as possible"""""""" (Kelley 1996). Disparity mapping, the technique proposed, directly addresses Dr. Kelley's concern by providing high contrast, subspeckle visibility of small hypoechogenic mammary tumors. Evidence is given in Figure 11b. Elimination of intraductal tumors, based on detection at this early stage, would be a major advance toward the ultimate goal of total eradication of the treat imposed by breast cancer.
| Steinberg, B D; Carlson, D L; Birnbaum, J A (2001) Sonographic discrimination between benign and malignant breast lesions with use of disparity processing. Acad Radiol 8:705-12 |
