Experts developed a deep learning model that can estimate breast density

Example of the four types of breast tissue density. The density of fibroglandular tissue inside the breast impacts the ability to easily see cancers. Cancers are very easy to spot in fatty breasts, but are almost impossible to find in extremely dense breasts. These examples show craniocaudal mammogram findings characterized as almost entirely fatty (far left), scattered areas of fibroglandular density (second from left), heterogeneously dense (second from right), and extremely dense (far right). RSNA

Example of the four types of breast tissue density. The density of fibroglandular tissue inside the breast impacts the ability to easily see cancers. On X-ray mammography, cancer and dense breast tissue both appear as white and can hide smaller cancers on 2D mammography. Dense breasts are also a risk factor for cancer. Cancers are very easy to spot in fatty breasts, but are almost impossible to find in extremely dense breasts. These examples show craniocaudal mammogram findings characterized as almost entirely fatty (far left), scattered areas of fibroglandular density (second from left), heterogeneously dense (second from right), and extremely dense (far right). Read more. Image courtesy of RSNA

When tested, the model achieved a performance comparable to that of human experts.

HeartFlow raises $215M to keep up with growing demand

An example of the FFR-CT technology from Heartflow. The vendor's AI can take a patient's cardiac CT scan and non-invasively assess FFR hemodynamic flow for all the coronary arteries to determine if blockages are significant enough to require revascularization. Photo by Dave Fornell

An example of the FFR-CT technology from Heartflow. The vendor's AI can take a patient's cardiac CT scan and non-invasively assess FFR hemodynamic flow for all the coronary arteries to determine if blockages are significant enough to require revascularization.  

The company is still riding the momentum of its technology being included in the 2021 American College of Cardiology/American Heart Association chest pain guidelines.

What is the ROI on AI adoption in radiology?

Brain imaging artificial intelligence is a primary area of concentration for AI because oif the critical nature of fast detection and treatment for patients. This is an example of the AI applications displayed by third-party advanced visualization vendor TeraRecon at RSNA 2022.

Brain imaging artificial intelligence is a primary area of concentration for AI because oif the critical nature of fast detection and treatment for patients. This is an example of the AI applications displayed by third-party advanced visualization vendor TeraRecon at RSNA 2022.

Radiology makes up the vast majority of FDA-cleared AI algorithms, but with minimal or no reimbursement, hospital administrators may ask whether AI’s value justifies its expense.