Breast cancer accounts for nearly 1 in 3 cancers diagnosed in US women. X-ray mammography is the gold standard method for the detection of micro-calcifications (MC) that are preliminary indicators of breast cancer. Every year millions of women undergo biopsy for diagnosis of breast cancer after suspicious mammograms. However, mammography cannot be used as real time image guidance for biopsy because of the relatively large dose of ionizing radiation. Ultrasound imaging does not use ionizing radiation, provides real time images, and can be used in the dense breast of younger women. Clearly, a system that integrates a MC sensitive ultrasound scanner that rivals the capabilities of mammography and a haptic assisted biopsy robot that can help the interventional radiologist in reaching the biopsy area is highly desirable.
However, detection of the smallest micro-calcifications is frustrating problem for clinical ultrasound since tissue attenuation limits the frequency and thus the resolution. In order to develop an ultrasound system able to detect MC in real time one has to develop a system capable of localizing objects smaller than the wave length (super-resolution). The Colton-Kirsch Linear Sampling (CK-LS) is a scalar inverse scattering algorithm that has been shown to provide super-resolution images in passive radar applications, but has not been applied in the field of ultrasound imaging. The Department of Defense Breast Cancer Research Program has recently funded a research project, together with Dr. M A Lewis of UTSWMED and Dr. T Aktosun of UTA, aimed to determine the feasibility of this approach in a laboratory models.
The CK-LS algorithm can be implemented in an imaging system with 6 ultrasound transmitters and 6 receiver array elements forming a hexagonal cup around the breast. This configuration can effectively sample the scattering field by electronically multiplexing the transmitters and the six readout arrays (see Figure 5), without painful compression of the breast. The theoretical development required to apply the CK-LS method in a medium with heterogeneous background (as is the case of real breast tissue), the design, construction, and characterization of the imager will constitute the subject of a new proposal that it will be submitted to the same funding agency.
The proposed ultrasound imager will be capable to provide real time images that can be used to create a 3D model of the breast. The interventional radiologist would interact with the model to identify the MC and plan the biopsy procedure during the same appointment, reducing the time interval between identification of a suspicious area in the breast and the diagnosis by pathology, and thus reducing the period of uncertainty for the patient. Software able to track the biopsy target area in real time and calculate the path for the biopsy needle accounting for patient motion and tissue deformation can then be developed. A haptic system that will provide a virtual envelope around the optimal needle trajectory and provide force feedback that will accurately guide the needle towards the biopsy target area and a 6 DOF robotic arm used to control the biopsy probe would complete the clinical system.
A research proposal aimed to develop a clinical system that integrates the MC sensitive ultrasound scanner with the 3D modeling of the breast tissue, and the haptic assisted biopsy robot (as seen in Figure 6) will be pursued with the appropriate funding agencies.
Far field simulation showing the localization of 3 sound-hard scatterers (left) and the proposed CK-LS ultrasound imager schematic (right).