Can deep learning help doctors find Breast Cancer on tomosynthesis scans?
Deep learning (DL) is a type of artificial intelligence that can analyze medical images. For digital breast tomosynthesis (DBT) — a 3D mammogram — DL algorithms show strong potential to help radiologists detect breast cancer. Recent studies find that DL can match the accuracy of experienced radiologists and may even catch more cancers in certain situations.
What the research says
A 2026 meta-analysis of 13 studies with over 38,000 patients found that stand-alone DL algorithms achieved a pooled sensitivity of 0.88 (catching 88% of cancers) and specificity of 0.74 (correctly ruling out 74% of non-cancers), with an overall accuracy (AUC) of 0.89 410. This performance was not statistically different from all radiologists (AUC 0.89 vs 0.88) or senior radiologists (AUC 0.89 vs 0.90) 410. However, DL showed significantly better sensitivity compared to radiologists in some analyses, meaning it may detect more cancers 410. Another study used DL to estimate breast density from DBT images, achieving high accuracy (Dice score 0.88) and linking density measurements to cancer risk 9. These findings suggest DL can serve as a reliable second reader or assist less experienced radiologists, potentially reducing missed cancers and false positives.
What to ask your doctor
- Is deep learning used to read my tomosynthesis scans at this facility?
- How does the AI tool compare with radiologist readings in terms of accuracy?
- Could AI help detect cancers that might otherwise be missed on my scan?
- Are there any limitations or false positives I should be aware of with AI-assisted reading?
- How is the AI integrated into the radiologist's workflow here?
This question is drawn from common patient questions about Oncology and answered using cited medical research. We do not provide individualized advice.