Researchers led by Concordia University have developed an AI-driven robotic system capable of performing cardiac ultrasounds autonomously. The technology could play a significant role in making diagnostics more accessible, particularly in regions where specialist healthcare professionals are in short supply.
Cardiac ultrasound, or echocardiography, normally requires the expertise of experienced sonographers who position the probe precisely to obtain the correct images. This process not only relies on skill but can also be physically demanding. The new solution automates this process using artificial intelligence and robotics.
AI replaces manual expertise
The system, published in IEEE Transactions on Medical Robotics and Bionics, consists of a robotic arm holding an ultrasound probe and controlled by an AI algorithm. This algorithm has been trained to independently find the correct imaging angles required for a clinical diagnosis. In doing so, the system replaces a large part of the manual expertise normally required for ultrasound.
The AI uses deep reinforcement learning, a technique in which the model learns by receiving feedback on its performance. The system is rewarded when it gets closer to the desired image and adjusts its movements accordingly. As a result, it learns step by step how to position the probe and how much pressure is needed to obtain clear images.
According to the researchers, this approach can not only standardise the quality of scans but also reduce the workload for healthcare professionals.
Training in a virtual environment
A striking aspect of the research is the way in which the AI was trained. Instead of relying solely on real-world data, which is often difficult and time-consuming, the researchers developed an advanced simulation environment based on generative AI.
This virtual environment produces synthetic ultrasound images that closely resemble real medical images. This allowed the AI model to be trained safely and efficiently before being applied to physical equipment. This shortens development time and reduces the need for extensive data collection in clinical settings.
In tests using a robotic setup and a training model of the heart, the system succeeded in producing standard images more quickly and accurately than human operators working remotely. These results proved consistent across multiple test sessions. The researchers emphasise that further validation with real patients is necessary before the technology can be widely applied. Nevertheless, they see clear opportunities for the use of autonomous ultrasound in clinical practice.
Growing role for AI and robotics
This development highlights the growing role of AI and robotics in medical imaging. Last year, Philips introduced a portable ultrasound system that uses AI to make cardiac diagnostics faster and more accessible. The system is suitable for bedside use and can be deployed in various healthcare settings.
Thanks to AI-supported features, such as Auto Measure, scan time can be reduced by up to 50 per cent without compromising accuracy. This leads to faster and more consistent measurements and better healthcare outcomes. The technology is particularly valuable in acute situations and for vulnerable patients, such as newborns and the elderly.
By automating diagnostic processes, healthcare systems can be organised more efficiently and specialist care becomes more accessible to a larger group of patients. In the long term, this technology could contribute to a new standard in cardiac diagnostics, whereby AI-driven systems support doctors or even partially replace them in routine examinations. It thus represents an important step towards scalable, data-driven and future-proof healthcare.
