8 AI Diagnostics and Imaging Startups for Digital Health
Investing in emerging technologies can be extremely risky. It can also be extremely rewarding – and not just for your bank account. Technologies like artificial intelligence have the potential to change the world in many different ways. One of the industries where AI is already making real advances is healthcare, such as the ability to design and validate drug candidates to treat disease in less than two months. That has attracted the attention of plenty of deep-pocketed investors into AI healthcare startups, which have made more deals than any other AI industry since 2014, according to research firm CB Insights, with more than 80 AI diagnostics and medical imaging companies leading the way across 150 deals and counting.
We’ve spent quite a bit of time talking about the future of AI medical imaging, featuring a couple of lists of AI medical imaging startups here and here. At one point, we expected IBM to dominate the AI diagnostics and medical imaging industry with its Watson Health Imaging platform, but that has yet to happen. Meanwhile, the AI diagnostics landscape has expanded well beyond medical imaging to areas like breath diagnostics and chatbots, while also shrinking hardware down to a handheld ultrasound device, for example, or creating diagnostic software that operates using a smartphone camera. In the second quarter of this year alone, AI healthcare startups raised a record $864 million, CB Insights reported.
In this article, we want to look at some of the AI diagnostics startups that are leading the way in the digital health revolution, based on research from CB Insights and elsewhere. We have a ton of companies to cover, so let’s jump right in with an AI healthcare unicorn called HeartFlow.
AI Diagnostics for Heart Health
Founded in 2007 as Cardiovascular Simulation, HeartFlow changed its name in 2009 and has raised $476.6 million since then from notable investors like BlueCross BlueShield, with a valuation of $1.5 billion. HeartFlow uses deep learning and some pretty smart humans to build personalized, 3D models of a patient’s heart and coronary arteries based on imagery from a computed tomography angiography. The company’s computer algorithms then go to work to solve millions of complex equations to simulate and assess coronary blood flow. So far, the HeartFlow Analysis has been used to diagnose 40,000 patients.
More recently (as in last month), HeartFlow won approval from the Federal Drug Administration (FDA) for a new product, HeartFlow Planner, a non-invasive, real-time virtual modeling tool that helps determine the best therapy based on the HeartFlow Analysis. A recent study found that HeartFlow Planner led to a change in treatment strategy in 45% of patients with coronary disease. The company is now moving forward with getting clearance in the UK.
AI Diagnostics for Eye Health
Founded in 2010, IDx out of Iowa has raised $52.1 million, including a $32.6 million Series A about a year ago. IDx is focused on developing autonomous algorithms that can detect eye disease in, well, a blink of an eye. First up: An FDA-approved algorithm that analyzes images of the eye from a special ophthalmological device called a fundus camera to detect diabetic retinopathy, a condition caused by diabetes that damages the blood vessels of the light-sensitive tissue of the retina. A 2017 pilot test of 900 patients throughout the United States showed that the machine was spot-on about 90% of the time. The IDx-DR system, as it’s called, reportedly costs about $30 to $50 versus a $200-plus screening by a retinopathy specialist – and returns results in about a minute.
The company is also developing AI diagnostics algorithms to detect glaucoma and age-related macular degeneration, perhaps one day reducing the need for vision enhancement devices for the blind.
AI in Blood Testing Diagnostics
The epic fraud – alleged, of course – committed by blood-testing startup Theranos and its infamous founder, Elizabeth Holmes, has led to books and documentaries, with a feature film in the works. You would think that all of that bad blood would have turned off investors from pouring more money into other blood-testing startups. But Israeli startup Sight Diagnostics, founded in 2010, raised $27.8 million in February, bringing total funding to $52.8 million. The company has developed a blood-testing platform that combines point-of-care hardware with a computer vision algorithm that analyzes a blood cell’s shape, size, and more than a dozen other characteristics, returning a complete blood count in minutes from just a couple of drops of blood. The company’s first product, Parasight, was an AI diagnostic blood test for malaria with 99% accuracy.
Sight Diagnostics certainly has the bona fides, with connections to both Mobileye, an Israeli firm now owned by Intel that uses computer vision to develop collision-avoidance systems for vehicles, and geneticist George Church, who has founded a number of genomics startups.
Mobile AI Diagnostics for At-Home Urinalysis
Another Israeli startup, Healthy.io, has created an at-home urinalysis platform that uses a smartphone and computer vision to analyze a dipstick to detect infections, chronic illnesses, and pregnancy-related complications. Founded in 2013, Healthy.io has raised $90 million, including a $60 million Series C last month, with Samsung being a repeat investor. One clinical-grade test helps people with diabetes or high blood pressure to test themselves at home for kidney disease, achieving 99% reliability. The technology has been approved for use in both the United States and Europe. The company says its computer vision algorithms and unique calibration method make accurate testing as easy as taking a selfie. They haven’t seen our selfies.
Early Detection Using Electronic Health Records
Let’s stay with the Startup Nation for our next startup, Medial EarlySign, a 10-year-old company that has raised $50 million so far for its AI-powered platform that leverages big data to detect diseases earlier. The company is developing different algorithms to detect subtle, early signs of disease in high-risk patients based on ordinary electronic health records and existing lab results. The company is targeting a range of conditions, including lower GI disorders, prediabetic progression to diabetes, chronic kidney disease, and coronary artery disease.
Medial EarlySign’s latest product offering is a suite of what it calls AlgoMarkers for diabetes, including one that identifies prediabetic patients with the highest risk of progressing to diabetes within a one-year period and type 2 diabetic patients at high risk for developing chronic kidney disease within three years. In both cases, its algorithms outperformed more traditional diagnostic methods.
AI Voice Diagnostics for Neurology
We’ve written quite a bit about voice recognition technology that uses AI for transcription or even fraud protection, not to mention the increasingly crowded space of AI voice assistants. Now comes along Aural Analytics, an Arizona startup founded in 2015 that has raised $4.7 million in disclosed funding, including a $4.3 million Seed round in September, along with a grant from the National Science Foundation. The company has built what it calls a speech analytics engine that uses AI algorithms to detect changes in speech to monitor different kinds of neurological diseases.
Just this month, Aural Analytics announced the results of a clinical trial that found the company’s platform could identify subtle changes in patients with Lou Gehrig’s disease who were treated with reldesemtiv, an investigational drug candidate that has shown potential to treat debilitating diseases and conditions associated with muscular weakness and/or muscle fatigue. The AI analyzes vocal biomarkers – including phonation, velopharyngeal (soft palate) control, articulation, and speech patterns – to detect and measure speech changes in patients treated with reldesemtiv.
AI Diagnostics for Ultrasound
We’ve written quite a bit about a company called Butterfly Network since 2016 that has taken in hundreds of millions of dollars to develop a handheld ultrasound device. The instrument retails for about $2,000 – an incredible bargain considering the cost of a high-end ultrasound machine can push into six-figure territory. San Francisco-based Caption Health, founded in 2013, has raised a reported $18 million in disclosed funding, including from top VC firm Khosla Ventures, for its AI software that helps non-experts acquire and interpret ultrasound imagery. The company is helmed by Andy Page, a former president of 23andme and chief financial officer of Livongo (LVGO), which went public earlier this year and uses AI to help manage patient care through big data.
A clinical study that tested how well eight registered nurses with no prior cardiac ultrasound experience could perform the diagnostic tests found that the platform was used successfully between 92.5% and 98.8% of the time, depending on the measurement. The baseline was 80%. The FDA has already granted Breakthrough Device Designation to Caption, but final approval is still pending.
AI Diagnostics from a Chatbot
Just about everyone has Googled his or her symptoms at one point, going down a rabbit hole of paranoia and hypochondria. New York-based K Health, founded in 2016, has built an app to give patients a reliable resource of health information, raising $56 million in the process. (While not insignificant, that’s pocket change compared to the more than $650 million raised by Babylon Health, which picked up $550 million from Saudi Arabia and a few other investors in August.) The AI chatbot from K Health has been trained on more than two billion anonymous medical records, enabling it to find patterns in the data provided by consumers, with personalized diagnoses and recommendations based on how each case compares to similar ones. The app even shows users how doctors approached similar conditions, including medication, tests, and other treatments.
Update 02/27/20: K Health has raised $48 million in Series C funding to scale its model and move primary care to mobile devices in an effort to improve access to health care on a global scale. This brings the company’s total funding to $97.3 million to date.
We want to stress that this list is only representative and not comprehensive, as there are dozens of AI healthcare startups now operating in the diagnostics space. CB Insights recently highlighted some of the best of the best in its 150 digital health startups list, many of which we’ve covered or will in the near future:
Another thing to note is that many of these companies are focused on publishing their data, and several have or are in the process of receiving FDA approval or certification in the European Union for their technology. That’s an important benchmark in terms of discerning companies producing real results from those building AI hype platforms. In fact, CB Insights believes AI diagnostics and imaging is one of the more mature AI technologies today:
The next article in our series on AI healthcare startups will focus on AI diagnostics for cancer.