The promise of artificial intelligence (AI) in healthcare is rapidly taking shape. Pharmaceutical and medical device companies are tapping into the ability of machine learning to make their products more efficient and effective for patients. Even FDA is becoming more comfortable with AI and has recently given approval to several technologies that have embraced machine learning.
ThinkGenetic, a digital health startup, has leveraged the power of AI to help patients understand themselves and the likelihood if they have a genetic disease. The company’s CEO and co-founder David Jacob will be at MD&M East, moderating the panel: How Artificial Intelligence is Moving the Needle in Medtech.
“The panel will be looking for ways to use artificial intelligence to make their device more knowledgeable,” Jacob told MD+DI. “And one of the underlying things will be how do we shorten the diagnosis odyssey - whether that’s with machine learning or with a wearable device.”
ThinkGenetic originated from IBM Watson’s commitment to invest about $1 billion in healthcare back in 2014.
“When [IBM Watson] said they were going to invest a billion dollars in healthcare- we thought, oh that’s the space we’re in… we need to take a look at this,” Jacob said. “Being a consulting partner, we got to have one of the first commercial agreements with IBM to use Watson. We got to use [Watson] for free for two years to build our application.”
ThinkGenetic locates individuals with diagnosed and undiagnosed genetic diseases using AI solutions. Jacob said he likes to think of the company’s solution as a “GPS for genetics.”
“One of the problems the pharmaceutical industry has is that they develop these drug treatments or orphan drugs with genetic conditions and then they have problems finding the patients because they are so rare,” Jacob said. “What ThinkGenetic does is finds these patients on the internet when they’re out searching for answers and walks them through the process of learning more about themselves.”
He added, “[Our application] will walk you through the process of figuring out if it’s genetic, then what it might be. This is all artificial intelligence behind the scenes saying you’ve told me these symptoms – let me ask you so more questions. What we’re trying to do is say if it’s genetic, then let’s talk about what it could be and then what you need to talk to your doctor about. We actually have genetic counsellors on staff who will direct you to the next steps to find out what’s going on.”
AI has been coupled with a wide array of technologies – but mostly diagnostic applications. Last month, South San Francisco, CA-based Freenome revealed it had initiated a clinical study to help bring an AI-Genomics blood test for colorectal cancer to the U.S. market. The test would be able to learn from its mistakes and make colorectal screening more accurate. During an interview with MD+DI, the company’s CEO said AI was a natural step in the evolution of diagnostics that should have taken place “years” ago.
AI has also been paired with applications that can monitor a woman’s health. Startup firm, Ava has developed a bracelet that uses AI to help monitor and track a woman’s cycle. The firm recently raised $30 million in a series B round to help its technologies’ applications.
But perhaps the company with the loftiest ambitions for AI is Beta Bionic. The Boston-based firm is developing a bionic pancreas called the ILet, that could monitor and control a patient’s blood sugar automatically. The company recently received FDA approval to begin recruitment for the home-use studies testing an insulin-only configuration of the device.
Jacob said that the number of applications for AI will continue to grow and that people could expect the scope of machine learning in healthcare to change over the next 10 years.
“In healthcare, I see AI making us more proactive rather than reactive,” Jacob said. “We’re going to see things coming before they actually come. We’re going to have all these devices – your Fitbit, your pacemaker, and your Apple Watch. All these devices are going to be sending data that can be useful to the healthcare system. If the algorithms are written correctly then we can basically see problems [ahead of time]."