AI-powered robot scans hearts in groundbreaking new study with Milton Keynes University Hospital

Researchers at Cranfield University have been conducting a study in partnership with Milton Keynes University Hospital using an AI-controlled robotic system to deliver echocardiograms. The technology could eventually bring reduced waiting times and greater accessibility to complex scans of this kind.

The system uses a robotic arm with an imaging device on the end, paired with a laptop running AI software that evaluates the quality of the images in real-time.

The imaging probe is initially placed on the skin by a human and then starts a spiral scanning pattern and autonomously moves the position and orientation of the probe according to feedback received during the scan. The pictures are sent to a computer where AI assesses them in real-time to make determinations about both the quality of the imaging and starting point for further scanning.

Once the best position is located, the orientation of the probe is adjusted automatically, informed by AI again, to achieve the optimal quality of data acquisition.

In testing, the system scored 80% accuracy in achieving the A4Ch view - a standard ultrasound picture of the heart that shows all four chambers. This shows proof of concept for the system, and future refinements will help to improve that score.

Echocardiography is one of the most critical tools for diagnosing cardiovascular diseases, but the National Health Service is currently struggling with a lack of trained sonographers. According to the Society of Radiographers, the vacancy rate for sonographers in the NHS went from 6.7%-13.4% between 2019-2023. This is due to increased demand, difficulty of training and retaining staff and factors like repetitive strain injury affecting staff.

Yifan Zhao, Professor of Data Science at Cranfield University, said "This is exactly the sort of AI-driven process that has the potential to positively impact our lives. Because it removes the requirement for staff with specialist skills to conduct the scan, it opens up the possibility of tests in a community health setting in the future, rather than people having to wait for hospital appointments - taking some of the pressure off the NHS. While we are a long way off robots conducting these tests completely on their own, the technology does open the door to much more flexible and accessible testing in the future."

Dr Gilbert Tang, Senior Lecturer in Robotics at Cranfield University, added, "This technology has a really promising accuracy level already, and this will only improve as it is developed further. Using robotics for scans could improve the consistency of diagnosis as well, and potentially improve patient outcomes."

Professor Attila Kardos, Consultant Cardiologist at Milton Keynes University Hospital said, "Using AI to guide autonomous echocardiographic image acquisition is an exceptionally challenging area, but one with enormous potential to deliver rapid diagnostics for at-risk individuals in the community and streamline workflow in the emergency departments and other secondary care settings. This research aligns well with the NHS England 10-year plan, aiming to make diagnostic tests widely accessible and reduce waiting times across the healthcare system."