App may soon diagnose cancer, Parkinson’s by the sound of your voice

If you want a diagnosis without the hassle of a visit to the doctor’s office, there may soon be an app for that.

Scientists are working on artificial intelligence that will analyze speech and breathing patterns and vibrations to pick up on changes indicating illness, according to University of South Florida’s Health Voice Center.

While it’s already known that conditions like Parkinson’s disease or a stroke affects speech, and lung diseases affect breathing, researchers are hoping that a computer program will be able to diagnose a wide range of health conditions — including cancer and depression — with just the sound of your voice.

Dr. Yael Bensoussan, director of USF’s Health Voice Center, believes that “voice has the potential to be a biomarker for several health conditions.”

Experts from 12 research institutions will collect voices of people with health conditions in five areas: neurological disorders (Alzheimer’s; Parkinson’s; ALS, also known as Lou Gehrig’s disease), voice disorders (laryngeal cancer), mood disorders (depression), respiratory disorders (lung diseases) and pediatric disorders (autism and speech delays).

They hope to gather a data set of 30,000 voice samples over four years, as well as other factors such as clinical data and genetics. As more people enter their voices into the app and the database grows, the accuracy of the algorithm is expected to improve. 

Using the app, a patient would speak into their phone and record their voice, then the app would cross-check the vocals with its database, according to Fox 13 in Tampa.

The researchers believe that vocal patterns can offer insight into a person’s underlying health conditions. For example, a person who slurs their words may have suffered a stroke, or someone with low and slow speech patterns may have Parkinson’s disease.

Previous studies have been attempted, but Bensoussan is the first to lead a study with a large data set while keeping data privacy at the forefront.

“Creating an effective framework that incorporates huge data sets using the best of today’s technology in a collaborative manner will revolutionize the way that voice is used as a tool for helping clinicians diagnose diseases and disorders,” she told USF Health.

If the technology works and an app could be created, it would be revolutionary for health care. An app would allow for health-care equity and access in hard-to-reach and rural areas, giving people in those areas access to licensed medical providers. It would also be a game changer for telehealth practices.

“Our future findings could lead to a revolution in health care where continuous voice monitoring could alert physicians earlier than currently possible to certain conditions, such as infections or neurological diseases,” Dr. Olivier Elemento, a co-investigator and director of the Englander Institute for Precision Medicine, said.

Although the app will be available for use by both professionals and nonprofessionals, Bensoussan told NPR’s “Morning Edition” that it’s not supposed to replace a regular medical screening, and that it will only flag signs of disease.

The research project is funded by the National Institutes of Health, as part of the $130 million Bridge to Artificial Intelligence program — an initiative to speed up common and widespread use of AI in medicine.

The goal is to develop an app that can be used in the medical field, without having to go to the doctor.