Jan. 16, 2026 — Vega Imaging Informatics has announced the successful curation of the world’s largest digital breast tomosynthesis (DBT) dataset with paired histology outcomes. The dataset was prepared for a leading developer of AI-based radiology software focused on breast imaging and will support the development of next-generation AI solutions designed to assist in the interpretation of 3D mammography exams. The dataset includes over 1 million DBT studies across three DBT manufacturers, which are paired with biopsy outcome data for more than 22,000 patients, including over 7,000 cancer cases. All data was de-identified according to 45 C.F.R. § 164.514(b) of the HIPAA Privacy Rule.
Strengthening AI Performance
DBT, or 3D mammography, has become an increasingly important technique in breast cancer screening, because it reduces overlapping breast tissue and improves lesion visibility as opposed to traditional 2D mammography techniques. At the same time, DBT exams generate hundreds of image slices per study, increasing interpretation time and cognitive load for radiologists.
By leveraging Vega’s large-scale imaging resources for this project, the AI developer was able to attain superior diversity in its data, which is a key factor in the development of models that are effective across a wide range of breast densities, anatomical variations, and real-world clinical conditions.
Large, Multi-Modal Imaging Datasets
“High-quality, clinically representative imaging data is essential for building AI tools that perform reliably in real-world care settings,” said Luke Bideaux, President and CEO of Vega Imaging Informatics. “This multi-modal dataset demonstrates Vega’s ability to deliver imaging data at scale while meeting the stringent quality, privacy, and compliance standards required for medical AI development. With a single DBT study reaching file sizes over 50 times larger than many other types of imaging studies, such as most chest x-rays, the sheer file size of this dataset demonstrates the scale Vega can achieve in its curation initiatives.”
Next Generation of Breast Imaging Workflows
AI-enabled DBT interpretation tools are increasingly being adopted to support radiologists through improved workflow efficiency, reducing reading time, and enhancing diagnostic confidence. Increased adoption results in higher demand and competition for AI solutions that can deliver value to breast imaging providers. As the demand for large, high-quality imaging datasets continues to rise, Vega has proven that it is well positioned to meet the needs of the AI development community.
January 20, 2026 
