Digital breast tomosynthsis (DBT) has been growing rapidly in its application to mammographic cancer screening. While evidence exists suggesting that iterative image reconstruction (IIR) algorithms may improve DBT image quality in terms of visualizing tumor spiculations and microcalcifications in the breast without any adjustment to the DBT hardware, there remains a large gap between development of advanced IIR and its translation to the clinic. This project, building upon our previous success on IIR development for DBT, focuses on filling in this gap through development and integration of novel IIR algorithms into DBT systems with the parameter selection in an automated fashion, thus realizing the potential of IIR for improving DBT-image quality. The project has available a database of hundreds of normal/abnormal DBT cases with clinical DBT systems, and the assistance of our in-house imaging physicists and radiologists.
The specific aims of the research are: 1: Investigate novel advanced IIR algorithms; 2A: Design image quality metrics specific to DBT volume characterization; 2B: Determine of IIR algorithms parameters from simulation-based IQ metrics; 3: Quantitatively evaluate the performance of automated advanced IIR on DBT imaging. The benefit of the resulting automated IIR algorithms from Aims 1-2 will be evaluated quantitatively in Aim 3 by expert observers against the clinical processing with respect to imaging tasks relevant for DBT. The proposed project has high clinical and technical significance, because the use of DBT for mammography screening is becoming the standard in the US and because the research proposed enables the translation of advanced IIR to impact DBT clinic applications. We will directly develop the automated IIR algorithms on the industrial leading scanner, the Hologic Selenia Dimensions, employed in our clinic, and thus improvements gained in this project may have an immediate impact for mammographic screening in terms of increasing sensitivity and reducing call-back rates. The team assembled for this project includes leading imaging scientists, physicists, and breast-imaging radiologists, along with industrial consultants.

Public Health Relevance

Digital breast tomosynthesis (DBT) plays an increasingly important role in screening and assessment of breast cancer. The objective of the project is to improve current DBT performance by increasing sensitivity to malignant tumors and reducing unnecessary patient call backs through the development and translation of innovative iterative image reconstruction (IIR) in DBT, leading to an enhanced breast-cancer detection in DBT- based breast cancer screening. Our team of academic and industrial investigators possesses complementary expertise in hardware, algorithm, and clinics necessarily ensuring that the project is carried out successfully in a timely manner.

Agency
National Institute of Health (NIH)
Institute
National Institute of Biomedical Imaging and Bioengineering (NIBIB)
Type
Research Project (R01)
Project #
1R01EB026282-01A1
Application #
9669830
Study Section
Biomedical Imaging Technology Study Section (BMIT)
Program Officer
Shabestari, Behrouz
Project Start
2018-09-14
Project End
2022-05-31
Budget Start
2018-09-14
Budget End
2019-05-31
Support Year
1
Fiscal Year
2018
Total Cost
Indirect Cost
Name
University of Chicago
Department
Radiation-Diagnostic/Oncology
Type
Schools of Medicine
DUNS #
005421136
City
Chicago
State
IL
Country
United States
Zip Code
60637