Font Size: a A A

Electrical impedance tomography for breast imaging

Posted on:2002-12-19Degree:Ph.DType:Dissertation
University:Dartmouth CollegeCandidate:Kerner, Todd EricFull Text:PDF
GTID:1464390014450028Subject:Engineering
Abstract/Summary:
Electrical impedance tomography (EIT) is a potential, non-invasive technique to image women for breast cancer. Ex vivo studies have shown characteristic frequency dispersions of the electrical conductivity (σ) and permittivity (&egr;) of cancerous versus normal tissue. By applying alternating current around the breast and measuring the voltage distribution that results, one can use computational model-based methods to reconstruct an image of σ and &egr; inside the breast. The σ and &egr;, correspond to both the anatomy and pathology within the imaging plane. A 32 channel EIT capable of operating between DC and 1 MHz was designed and constructed. The system was validated by generating physically meaningful images from a saline tank and agar gels with embedded objects. With a 10 cm diameter agar gel and 4:1 contrast inclusions, the minimum discernable width was 0.5 cm at 1 cm depth and 1.2 cm in the center. Our EIT should be able to detect similar size lesions in the breast. Over 50 women have been imaged with this device. A radial translatable circular array of 16 Ag/AgCl electrodes was placed around the breast while the patient lay prone. We applied trigonometric voltage patterns at 10 frequencies between 10 kHz and 950 kHz. Although there were significant surface electrode artifacts, the results were promising. A characteristic homogeneous image emerged for mammographically normal cases. A variety of improvements were made from building a new electrode mechanical interface to modeling the finite electrode length in the reconstruction algorithm. In a recent study, we demonstrated the reproducibility of our breast images, and we produced several striking images that correctly localized malignancies. Using a newly defined visual criterion, the EIT identified 83% of the ACR 4–5 lesions, and with a numerical criterion, the EIT identified 67% of these lesions. Overall, multi-frequency EIT appears to be promising to image breasts for malignancies, but improvements must be made before the system reaches its full potential.
Keywords/Search Tags:Breast, EIT, Image
Related items