Combining Clinical, Sonographic, and Elastigraphic Features to Improve the Detection of Prostate Cancer
Abstract
The ultimate goal of this project is to combine features derived from ultrasound (US) images, US radio-frequency (RE) data, tissue elasticity imaging, and clinical data such as PSA into a computerized system for displaying prostate images that indicate probable location(s) of cancer. This project proposed to begin by gathering RE data from in-vitro prostatectomy specimens in cross sectional planes 2mm apart. These data are used to calculate RE features such as scatterer size, and backscatter coefficient at each location in the gland. The data are also used to generate images and elastograms from which image texture features and tissue hardness features are computed. The features are then correlated with histology taken at the same tissue planes to determine which features and feature combinations most accurately predict the presence of cancer. Despite administrative problems that delayed the start of the project by six months, we have brought the acquisition equipment on line, acquiring both RE and pathology data from the prostate glands of 26 patients. From the RE we produced the first elastograms ever from a gland NOT embedded in a gel block proving that acquisition of reliable strain data from a gland immersed in saline is possible. We also developed new standoff devices to improve data quality from the peripheral zone of the prostate and developed improved software for elastogram and RE feature calculation. The assembly of the pathology images into whole mount images corresponding to the scan planes obtained with ultrasound is well underway. Development of software for texture analysis and integration of existing software components awaits the arrival of a replacement postdoctoral fellow in January.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1999
- Accession Number
- ADA386633
Entities
People
- Brian S. Garra
Organizations
- University of Vermont