Temporal Subtraction of Digital Breast Tomosynthesis Images for Improved Mass Detection
Abstract
The purpose of this project is to determine the feasibility of using temporal subtraction on Digital Breast Tomosynthesis (DBT) phantom images to allow for easier and earlier detection of breast cancer than with either technique alone. The ultimate goal of project would be applied to human subject data. Therefore, initial feasibility was tested on temporally spaced DBT images of human subjects. The investigator implemented a temporal subtraction algorithm using a rigid registration technique. However, the resulting subtraction showed many misregistration errors likely due to differences in positioning and compression forces used during the temporally spaced acquisitions. This work has demonstrated that temporal subtraction of DBT images is most likely infeasible, although may be possible with fully 3D images such as dedicated breast CT data. The investigator is currently developing a 3D computer simulated breast phantom, a specific aim of proposal. Dedicated breast CT data was used as the basis for the phantom. In order to create the phantom, several methods are under development: an automated segmentation algorithm, a compression algorithm, and a simulated image acquisition algorithm. The phantom resulting from this work will combine the realism of empirical data with the flexibility of mathematical models. The phantom will incorporate information from breast models of several different human subjects and include the ability to change the breasts' size, composition, and compressibility.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 2008
- Accession Number
- ADA500912
Entities
People
- Christina M. Li
- James T. Dobbins Iii
Organizations
- Duke University