Future diagnosticians may be able to use their smartphones, an AI app, and sounds, to detect dreaded diseases, even in remote areas of the globe. Google and other organizations are developing AI systems that can detect the sounds of disease and offer accurate diagnostics from a distance. The term for the technique is bioacoustics and once perfected, the system will be able to screen people in secluded areas for severe diseases like tuberculosis.
According to Bloomberg, nearly 4,500 people die daily and 30,000 fall ill from tuberculosis. Google engineers created a foundation AI model with 300 million pieces of audio that includes coughs, sniffles, sneezes and breathing. The data feeding Google’s HeAR (Health Acoustic Representations) includes 100 million cough sounds that can help detect TB.
The AI tool can be used on a smartphone and taken anywhere in the world to help detect disease so that patients can receive the care and treatment needed. Google collaborated with Salcit Technologies in India to develop an AI tool called Swaasa, the Sanskrit word for breath. Swaasa is being used to help screen people in remote areas through a mobile app that allows a 10-second cough sample upload and test for diseases with 94% accuracy, says Salcit’s co-founder Manmohan Jain. The tests costs around $2.40 which is a lot cheaper than a $35 spirometry test in an India-based clinic.
Other potential uses for bioacoustics include a test to detect early breast cancer, being developed at the Chang Gung Memorial Hospital in Taiwan. Montreal-based Ubenwa has created a foundation model for infant cries that can interpret their needs and health. According to the McGill University News and Events, the team has developed a cost-effective and non-invasive tool that allows clinicians to flag the risk of newborn asphyxia which, if detected early, is highly treatable. The AI tool may save lives and improve outcomes for millions of newborns each year, according to the researchers.
Other AI tools may be able to detect autism based on the sounds made by infants, says Bloomberg.
While Google’s bioacoustic-based AI model is not close to entering the market, according to Salcit, voice and sound are the new frontiers in medicine.