What does edge computing mean for the use of AI in medical diagnosis and treatment?

Tim Jensen: Edge computing is the notion of putting scalable processing power in the equipment being used in the field, or in the diagnosis room, where low-latency, real-time computing is required or can provide a real benefit. Edge AI allows immediate analysis of sensor data and on-the-spot AI diagnosis where, previously, data or images would have to be sent to a remote data centre, incurring delays in the diagnosis chain.

For treatment, high-power edge computing allows live-rendered 3D images of things like live brain imagery for brain surgery. AI can be used to monitor vital signs and anticipate problems or irregularities and indicate the cause. Medical applications are increasingly only limited by the imagination of the designer.

How can high-resolution sensors and displays impact the use of AI in medical devices?

High-resolution scanners mean more data, which takes more bandwidth and remote storage to be processed. On-site processing and analysis means more data can be processed, faster. So higher-resolution sensors require higher processing power to handle, but modern edge processors are designed for this challenge. Intel’s OpenVINO is designed to make it easy to dedicate processing power to image processing at the edge, and delivers specific benefits for these sorts of applications.

High-resolution touchscreen displays mean that live, interactive imaging is possible. Instead of just basic controls, images can now be reviewed in real-time on a good-quality screen, instead of needing to be printed or exported elsewhere.

What are the most pragmatic opportunities for the use of AI for medical technology?

You can save a lot of time and resources by training AI to analyse data in bulk to recognise and identify irregularities like tumours, for example. AI can do bulk initial diagnosis, to then be verified by a medical professional.

For research purposes, AI can be used to identify patterns in large data sets, thus helping to assess effective treatments. In epidemics, it can be used to identify the source and help to prevent the spread of infection.

On the other side, AI can be used to monitor and optimise treatment dosage, monitor patients in palliative care to make sure they’re receiving treatment at the correct intervals and detect changes in life signs to adjust care levels accordingly. For example, it could be incorporated into blood sugar monitors to optimise sugar intake, or regulate physical therapy in bedridden inpatients.

What factors are limiting its uptake?

The initial investment required to modernise equipment is an issue, but would deliver huge savings in cost over time due to faster diagnosis, increased efficiency and reduced errors.

Similarly, people still want to be diagnosed and cared for by a real person, and patients and medical professionals are reluctant to entrust a computer (with or without AI) to handle medical data and treatment. Data security is also regularly perceived as a problem. In actuality, modern computing and AI is less likely to make mistakes than humans because it can handle larger amounts of data, it doesn’t get tired and it can learn faster. Also, digital documentation is often much more secure than printed records if it is set up correctly.

What are your predictions for the development and use of AI in medicine over the next few years, and what is Avnet Integrated’s role in driving it?

Patients already expect faster service as a result of consumer-level digitisation, so the healthcare industry will be driven to modernise. In turn, perceptions of machine-based diagnostics will improve as the benefits are realised.

As a result, more people will be able to get medical attention more quickly, even as AI-powered diagnostics and treatment will reduce errors and costs. The same number of medical professionals will be able to give a larger number of patients more-effective treatment, making good-quality medical care more accessible in previously underserved areas.

Furthermore, as diagnostics and treatment become more compact and mobile, physical hospitals will be under less pressure to handle large numbers as patients will be able to be diagnosed and treated in the field.

Given Avnet’s experience providing edge-based computing power, graphics processing, medical-grade touchscreen interfaces, and secure data processing and storage to medical businesses, we’re able to support this change by reducing the risk and complexity of implementing integrated technology.