AI matches human radiologists in interpreting breast cancer scans.
AI matches human radiologists in interpreting breast cancer scans.
AI Shows Promise in Breast Cancer Detection, but Human Expertise Still Vital
Artificial intelligence (AI) continues to show promise in the field of healthcare, with a recent study suggesting that AI can perform as well as human specialists in detecting breast cancer on mammograms. However, experts emphasize that this technology should not replace human expertise, but rather complement it by enhancing efficiency and accuracy.
The study compared the mammography-reading skills of an AI tool with those of over 500 medical professionals and found that both groups achieved a similar level of accuracy. On average, both humans and AI successfully detected around 90% of breast tumors and correctly identified over three-quarters of cancer-free mammograms. While this is a positive outcome, it is important to note that there is still room for improvement.
Mammography is a widely-used method for breast cancer screening, but it poses challenges even for the most skilled professionals. Dr. Liane Philpotts, a radiology professor at Yale School of Medicine, explains that mammograms rely on detecting subtle patterns, which can be difficult for both humans and AI. However, the fact that AI performed on par with radiologists is encouraging.
Dr. Mozziyar Etemadi, an assistant professor at Northwestern University Feinberg School of Medicine, highlights the potential of AI in addressing the limitations of human expertise. Humans are prone to missing certain details, and AI could assist in prioritizing suspicious cases and identifying areas that may require further examination.
The study focused on a commercially available AI tool called Lunit’s INSIGHT MMG. It compared the performance of the AI tool against 552 radiologists, radiographers, and breast clinicians from the United Kingdom who evaluated a set of challenging mammograms. The results showed strikingly similar rates of sensitivity and specificity between humans and AI. Lead researcher Yan Chen, a professor at the University of Nottingham School of Medicine, views this as strong evidence that AI can perform as well as human readers in breast cancer screening.
- Plastic chemical may harm boys’ development.
- Prostate Cancer Expert Answers Your Questions
- Smoking boys may harm future children’s health.
However, it is important to note that the study’s findings may not be directly applicable to the United States. In the U.S., 3D mammography is becoming increasingly common, which presents different challenges compared to the 2D mammography used in the study. Additionally, the recommended age for breast cancer screening differs between countries, potentially affecting the density of breast tissue and the difficulty of mammogram interpretation.
To further enhance the effectiveness of AI in mammography, experts emphasize the need for AI tools that work seamlessly with 3D mammography. Moreover, the ability of AI to analyze past mammograms and detect changes over time could be incredibly valuable. However, integrating AI into real-world workflows and determining its overall impact on women undergoing screening are still areas that require exploration.
Ongoing clinical trials, such as the one being conducted in Sweden, directly test the effectiveness of AI-supported mammography. Early findings indicate that radiologists using AI assistance detect an additional 20% of breast cancers. These studies are crucial in providing concrete evidence of AI’s benefits and should include diverse populations to reflect real-world conditions.
While AI undoubtedly holds promise for breast cancer detection, it should be seen as a tool to enhance human expertise rather than replace it. The collaboration between AI and healthcare professionals has the potential to optimize screening processes, improve accuracy, and ultimately save lives.
