Reaping the benefits of AI innovations – from ingestible capsules to home monitoring

Three new AI tools in development demonstrate the potential that the technology offers in both health and social care

17th June 2024 about a 4 minute read
“Our study shows how machine learning tools can uncover genetic insights that traditional methods might miss when comparing cases and controls. This could lead to new ways to identify biological mechanisms of heart disease or gene targets for treatment.” Dr Ron Do, professor at the Icahn School of Medicine, Mount Sinai

Artificial intelligence (AI) tools are helping with diagnosis and treatment of health problems in a variety of innovative ways.

In one of the most exciting developments, researchers have created ingestible capsules that not only detect stomach gasses but provide real-time location tracking.

Developed by Yasser Khan, an assistant professor of electrical and computer engineering at the University of Southern California (USC), the capsules can detect gasses associated with gastritis and gastric cancers. Khan’s team have created a wearable coil that generates a magnetic field on a t-shirt. This field, coupled with a trained neural network, allows them to locate the capsule within the body.

The capsules are equipped, not just with electronics for tracking location, but with an optical sensing membrane that is selective to gasses. This membrane is made up of materials whose electrons change their behaviour within the presence of ammonia gas. Ammonia is a component of H pylori, the gut bacteria that could be a sign of peptic ulcer, gastric cancer, or irritable bowel syndrome.

The device is still in relatively early stages of development, however. The next step is to test the wearables on swine models. Khan believes that as well as using it to detect gastric illnesses, it has potential to monitor brain health.

Understanding the genetic basis of heart disease

Another AI tool is being used to identify rare coding variants in 17 genes that shed light on the molecular basis of coronary artery disease (CAD).

The investigators used an in silico, or computer-derived, score for coronary artery disease (ISCAD) to represent CAD. This score incorporates hundreds of different clinical features from the electronic health record, including vital signs, laboratory test results, medications, symptoms, and diagnoses. To build the score, the team trained machine learning models on the electronic health records of 604,914 individuals across the UK Biobank, All of Us Research Program, and BioMe Biobank.

The score was then tested for association with rare and ultra-rare genetic variants.

Dr Ron Do, study author and a professor at the Icahn School of Medicine, Mount Sinai, said: “Our study shows how machine learning tools can uncover genetic insights that traditional methods might miss when comparing cases and controls. This could lead to new ways to identify biological mechanisms of heart disease or gene targets for treatment.”

Helping family members keep an eye on vulnerable relatives

Finally, an AI tool is helping to alert family and care providers of potential medical emergencies among elderly or vulnerable people. The tool, created by a consortium of partners including CENSIS (Scotland’s innovation centre for sensing, imaging, and IoT technologies), the University of Edinburgh, Mydex CIC, Carebuilder and Blackwood Homes and Care, has been piloted across 19 households in Scotland.

The device is linked wirelessly to a smart meter or conventional electric meter and then disaggregates the data to identify certain high-power electrical items within the home, such as kettles, microwaves, washing machines and electric showers. Using machine learning, it can tag each item and determine when it has been turned on and off, and, most importantly, spot any anomalies.

For example, if a person usually wakes up and boils a kettle to make tea by 8am, the monitoring device will identify this as normal behaviour. However, if the kettle has not been turned on by 9am, the individual will receive an automated text message. If there is no response, an alert will be sent to their nominated contacts – a family member, carer, neighbour, or a response service – who will be notified to check on them.

Stephen Milne, director of strategic projects at CENSIS, said: “This project is all about repurposing energy data to help inform social care and supporting healthy aging. The system learns the typical activity of the individual living in the household and then spots any erratic behaviour, helping to identify when they may have issues. These could be one-off events, like a fall, and with further research, the system may be able to track changes over a longer time period that may indicate gradual, and more difficult to spot health issues, such as the onset of a condition such as dementia.”

Milne noted that although other technologies can monitor activity, this was the first full-service deployment implemented through passively monitoring a property’s smart meter system. “The device can also pick out each item being monitored, making it much more likely to spot any anomalies, and is barely noticeable for the householder,” he said.

FCC Insight

These three AI tools, all at different stages of development, show the wide variety of ways in which AI can help either improve our understanding of illness, improve diagnosis or support those most in need of help. This is a technology that has tremendous potential, and although we can’t be sure all these ideas will come to fruition, it is clear that there is a real opportunity here to create huge improvements in our understanding and management of illness.