In a recent interview, Manisha Bahl, M.D., discussed key findings from a new study on AI and digital breast tomosynthesis that she presented at the Society for Breast Imaging (SBI) conference.
While there has been a fair amount of research examining the impact of artificial intelligence (AI) with two-dimensional mammography, there has been relatively limited data on the use of AI with digital breast tomosynthesis (DBT), noted Manisha Bahl, M.D., MPH, FSBI, in a recent interview.
Dr. Bahl also noted there has been little data in the AI computer-assisted diagnosis (CAD) breast imaging research with respect to what specific breast cancers are detected or not detected with AI.
“The purpose of our (new) study was to fill that gap in knowledge or understanding by providing an analysis of the types of breast cancers that are missed versus detected by a commercial AI based CAD algorithm for tomosynthesis,” explained Dr. Bahl, an associate professor of radiology at Harvard Medical School.
For the study, which was recently presented at the Society for Breast Imaging (SBI) conference, Dr. Bahl and colleagues assessed the Genius AI Detection 2.0 software (Hologic) in 5,000 patients who had DBT screening exams. They found that the AI software detected and correctly localized 90 percent of true-positive cases.
Dr. Bahl noted the AI software was more likely to detect invasive ductal carcinoma and lymph node positive carcinoma. The AI software was less likely to detect true positive cancers that presented as asymmetries on mammography as well as invasive lobular carcinoma, according to Dr. Bahl.
“These results suggest that the AI algorithm is optimized to detect clinically significant lymph node positive cancers (and) may require further training to identify invasive lobular cancers, which are known to be difficult to detect because of their unique infiltrative growth pattern,” posited Dr. Bahl, the director of the breast imaging fellowship program at Massachusetts General Hospital.
Dr. Bahl also said she was impressed by the ability of the AI software to detect 32 percent of the false negative cases in the study.
“I think those results suggest that integrating AI into our practice could potentially help decrease the false negative rate of screening tomosynthesis, which is an important metric, an important surrogate marker for long term outcomes. Those results are very promising to me and show that AI has the potential to improve our quality in addition to our efficiency,” noted Dr. Bahl.
(Editor’s note: For related content, see “Multicenter Mammography Study Shows Greater than 10 Percent Increase in Breast Cancer Detection with Adjunctive AI,” “New Mammography Studies Assess Image-Based AI Risk Models and Breast Arterial Calcification Detection” and “Can AI Bolster Breast Cancer Detection in DBT Screening?”)
For more insights from Dr. Bahl, watch the video below.
Stay at the forefront of radiology with the Diagnostic Imaging newsletter, delivering the latest news, clinical insights, and imaging advancements for today’s radiologists.
The Reading Room Podcast: Current and Emerging Insights on Abbreviated Breast MRI, Part 2
July 23rd 2025In the second part of a multi-part podcast episode, Stamatia Destounis, MD, Emily Conant, MD and Habib Rahbar, MD, discuss key sequences for abbreviated breast MRI and how it stacks up to other breast cancer screening modalities.
FDA Clears Point-Of-Care Ultrasound Platform and AI Software for Neuraxial Procedures
July 17th 2025The dual FDA clearances for the Accuro 3S point-of-care ultrasound device and the SpineNav-AI machine learning-based software may enhance precision and safety with ultrasound-guided neuraxial procedures.
Stroke MRI Study Assesses Impact of Motion Artifacts Upon AI and Radiologist Lesion Detection
July 16th 2025Noting a 7.4 percent incidence of motion artifacts on brain MRI scans for suspected stroke patients, the authors of a new study found that motion artifacts can reduce radiologist and AI accuracy for detecting hemorrhagic lesions.