Algorithms from the 2023 RSNA Screening Mammography AI Challenge demonstrated strong performance, with leading models achieving high sensitivity and specificity in breast cancer detection.
Key Details
- 1The 2023 RSNA AI Challenge evaluated 1,537 algorithms on an independent dataset of 5,415 women from the US and Australia.
- 2Median recall rate of all algorithms was 1.7%, with the top-performing algorithm at 1.5%.
- 3Top algorithm sensitivity was 48.6% vs. median of 27.6%; specificity was 99.5% vs. 98.7% median.
- 4Ensemble models of the top 3 and top 10 algorithms achieved sensitivities of 60.7% and 67.8%, with corresponding recall rates of 2.4% and 3.5%.
- 5Sensitivity was higher in the Australian evaluation set (68.1%) than the US set (52.0%).
- 6The top models had higher sensitivity for invasive (68.0%) over noninvasive cancers (43.8%).
Why It Matters
These results highlight the continued advancement and clinical promise of AI tools for breast cancer screening in mammography, especially with ensemble and robust model approaches. The focus on real-world variability, dataset diversity, and robust benchmarking is crucial for safe, scalable adoption in radiology practice.
Source
AuntMinnie
Related News
•AuntMinnie
Toronto Study: LLMs Must Cite Sources for Radiology Decision Support
University of Toronto researchers found that large language models (LLMs) such as DeepSeek V3 and GPT-4o offer promising support for radiology decision-making in pancreatic cancer when their recommendations cite guideline sources.
•AuntMinnie
AI Model Using Mammograms Enhances Five-Year Breast Cancer Risk Assessment
A new image-only AI model more accurately predicts five-year breast cancer risk than breast density alone, according to multinational research presented at RSNA 2025.
•Radiology Business
AI Model Uses CT Scans to Reveal Biomarker for Chronic Stress
Researchers developed an AI model to measure chronic stress using adrenal gland volume on routine CT scans.