The objective of this study is to improve the diagnosis and management of patients with breast cancer through development and evaluation of more effective techniques for ultrasound color flow imaging and quantification. This medical objective is to be accomplished by: 1) further developing methods for acquisition, display and quantification of vascularity and pulse echo information from breast ultrasound; 2) evaluating these methods and the potential of ultrasound in discriminating breast cancer; 3) making more long range improvements in the acquisition flexibility and information content of ultrasound imaging of the breast. Significant advances were made toward each of these objectives in Years 01-03. Methods have been developed for displaying complex vascular and gray scale information to show their relationships in 3D and for quantifying what is offered to the physician in the Doppler images. The introduction of power mode imaging as a medical product was motivated by our proposal and work on blood volume and perfusion measurement as part of this project, which was proposed initially February 1, 1991. Data acquisition for initial human studies of the developed 3-D color flow scanning, display and quantification is nearly complete. Preliminary analyses indicate that 3-D visualization is quite promising for displaying the vascular architecture of the region of the breast mass and its relation to the pulse echo lesion when visible. 3D-based visual ratings of vascularity contributed to the discrimination of breast cancer provided by gray scale information and normalized, power-weighted, Doppler signal pixel density discriminated every better. It is proposed to better characterize the measurements, make the imaging more versatile and extend the human studies for an additional 1.5 years to increase the statistics. A small number of malignant and benign lesions will be studied with and without microbubble contrast as will a group with locally advanced disease to assess the predictive value of 3-D color flow ultrasound for response to systemic chemotherapy. Very promising preliminary demonstrations of multiview (compound) 3-D imaging of breast blood flow and other structures have been accomplished by 3-D image set alignment and unwarping from coregistration of points and lines found manually to be in common in the various views. It is proposed to bring this exciting long range development of technology to a practical level in years 04-06 by computerized identification and registration of homologous tissue volumes in the different views. To facilitate acquisition of 3-D image sets anywhere round the breast with many scanheads, our recently-demonstrated image-based scan position registration approach will be developed.
Showing the most recent 10 out of 14 publications
