The Impact of Deep Learning AI on Mammography

A new study by Dr. Shepherd and colleagues has recently shown that deep learning, a type of artificial intelligence, has a higher precision rate of detecting cancer risks in mammograms compared to conventional methods. Screening mammograms can detect early signs of breast cancer, which improves the chances of surviving for the patient, but screening mammograms usually miss about 20% of breast cancers. On the flip side, a false-negative result means a delay in crucial treatment, which can decrease the likelihood of survival for some. 

The addition of an AI model than can detect breast cancer risks can close that gap and provide women with more resources to protect their health. However, how will AI fit in the world of mammography?  

What is the Deep Learning AI Model? 

Deep learning is basically a process that analyzes big data similar to how the human brain learns and interprets knowledge. The deep learning model is designed to find signals in a mammogram that might be linked to increased cancer risk. Based on the results, women can be sorted into one of three pathways. They can be sorted into the low risk of breast cancer group, elevated screening-detected risk group, or elevated internal invasive cancer group. As of now, women are advised to get an annual mammogram in order to detect any signs of breast cancer, but with the AI, women would be monitored based on their risk. 

Women with low risk can get scans once every three years instead of once a year while women with higher risk can increase their scanning frequency in order to monitor their breast cancer risk. This can increase breast cancer prevention efforts and provide additional security for women. 

Insights on the Deep Learning Model 

Deep learning may provide additional insight on the risk of breast cancer for people, but what does that mean for the future of mammography? How will the deep learning model impact radiologists?

As of now, deep learning still needs more research before it can be implemented into routine use. This particular study had a small sample of 6,369 women. Additionally, while the AI outperformed in assessing clinical risk factors, it underperformed in interval cancer risk factors. The researchers plan to replicate the study with a group of underrepresented populations.  

The Predicted Impact of AI on Breast Cancer Detection 

While AI is still in the early phases, in the future, it may play a big role in breast cancer risk detection. Due to the benefits of big data, the deep learning model can work without the biological constraints that humans have, making them more efficient and faster. Mammography has experienced several setbacks during to Covid-19, such as staffing shortages.

While more research has to be conducted to figure out the role of AI in regards to mammography, using deep learning and mammography technologists together to get a second look can provide more insights and increase breast cancer prevention. In a world overrun with technology, having humans and AI work together to better society is the best possible outcome. 

A Look Towards the Future 

Regardless of how the deep learning model research progresses, healthcare workers will continue to do their best to provide for patients and detect any signs of breast cancer. In the future, if the deep learning model and mammography technologists can work together, breast cancer prevention will become easier, and we can hopefully see a bigger decrease in fatal breast cancer diagnoses.  

The Advanced Health Education Center provides insights on the industry, and we also provide continuing education for practicing technologists. If you are a mammography technologist in need of CE credits, take a look at our upcoming courses. If you have any questions, please contact us via email office@aheconline.com or by calling us at 1 (800) 239-1361. 

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.