9 Artificial Intelligence Startups in Medical Imaging
You don’t have to be a gambler to appreciate the complexities of the card game Texas Hold ‘Em. It involves a strategy that needs to evolve based on the players around the table, it takes a certain amount of intuition, and it doesn’t require the player to win every hand. Just a few days ago, an artificial intelligence (AI) algorithm named Libratus beat four professional poker players at a no-limit Texas Hold ‘Em tournament played out over 20 days.
If you have even the slightest understanding of how to write code, you would realize that it is impossible to actually code a software program to do that with such “imperfect information”. The AI algorithm did exceptionally well and was utilizing strategies that humans had never used before. Professional poker players are in no danger of losing their jobs, but the incredible capabilities of what AI is mastering these days should make everyone wonder just how safe their jobs actually are.
Let’s take the $30 billion medical imaging market. It’s no secret that AI is now performing certain medical imaging tasks better than human doctors. Pundits say “well, people will always trust a human doctor over an AI” and the answer we’d have to that is “not if the AI is going to give a more accurate answer“. It’s only a matter of time before every X-ray machine is connected to the cloud and one human doctor per hospital puts his hand on your shoulder when he reads you the output from the AI algorithm. Kind of like this:
Here are 10 startups applying artificial intelligence to medical imaging. (Yes, the title says 9 startups but when you do these types of lists you’ll inevitably overlook a company and then have to go back and add them after the fact).
Founded in 2011, Butterfly Network has taken in $100 million in funding from just a few investors, one of whom is Jonathan Rothberg. Mr. Rothberg is somewhat of a serial entrepreneur having founded Raindance Technologies, a company we profiled before, along with Ian Torrent which he sold for a cool $795 million. We previously highlighted Butterfly Network as one of the top-5 artificial intelligence companies to watch in healthcare. The Company is working on reinventing the ultrasound machine by squeezing all of its components onto a single silicon chip. An ultrasound that could sit on your smartphone could completely disrupt the entire market not to mention provide medical imaging for the 60% of people in the world that don’t have access to it.
Founded in 2014, Israeli startup Zebra Medical Vision has taken in $20 million funding so far to develop a cloud based radiology offering. Zebra made the CB Insights AI 100 list with their applicaton of machine learning that identifies abnormalities in CT scans. Not so long ago we were teaching you how machine learning can recognize cats in pictures and here it is already doing things that would take years and years of medical training for humans to perform. Last year Zebra Medical Vision announced the availability of algorithms that automatically detect low bone mineral density, breast cancer, fatty liver, coronary artery calcium, emphysema, and more. Pretty much anyone can upload a CT scan and get an answer within an hour. It’s no surprise that they’ve partnered with AI hardware provider Nvidia. If you’re a radiologist presently enjoying your $250K salary, don’t worry too much because Zebra says there’s a shortage of radiologists at the moment. Looks like you’re in no danger of
losing your job being freed up to do more value added activities.
Founded in September of 2011, California-based startup Arterys has taken in $13.72 million funding so far to develop an “automated, intelligent imaging analytics cloud platform to revolutionize medical imaging“. The Company recently made news with their medical imaging platform receiving the first FDA approval for a deep learning application to be used in a clinical setting. (Note that we said “deep learning” here. The first application of AI actually goes to a company called RADLogics). Arterys helps doctors diagnose heart problems in just 15 seconds while it takes a human around 30 minutes. FDA approval means that the results are at least as accurate as a human. Arterys has been working with GE Healthcare to take MRI images and use cloud-based GPUs (like those from Nvidia) to produce 3D animations of the heart which it then analyzes for defects. Can we have one less news headline about politics and instead highlight this freaking amazing accomplishment from Arterys?
Founded in 2014, California-based startup Enlitic has taken in $12 million funding so far to develop a deep learning algorithm that can increase the accuracy of a radiologist’s interpretation by 50-70% and at a speed 50,000 times faster. Last February we wrote an entire article dedicated to Enlitic and their business model which is to take a cut off the cost savings that firms realize by adopting their solution. Our back of the napkin math shows that radiologist salaries in the U.S. alone total over $9 billion in annual spend. They’re starting out in Australia first so U.S. based radiologists don’t have to cancel their country club memberships just yet.
Founded in 2015, New York-based startup Imagen Technologies took in $5.6 million in seed funding last year so far to develop AI capable of detecting pathologies and early disease identification within medical images. They want you to “imagine a world where all patients are diagnosed instantly by leading experts“. They appear to be operating in stealth mode as there just isn’t any information on them in the public domain.
Founded in 2010, Milipitas California startup RADLogics has taken in $5 million in funding to develop their cloud-based imaging platform that is guarded by US Patent 8,953,858 along with other pending patents. RADLogics was the first company to receive FDA approval (on 4/13/2012) for a machine learning application to be used in a clinical setting. As of today (2/18/17) they may be the only startup with an FDA-cleared product deployed commercially in multiple clinical sites in the U.S., which is using advanced machine learning image analysis technologies. You can check out their interface at the top of this article where we show a screenshot from their platform.
Founded in 2013, California-based startup Bay Labs has taken in $2.5 million in funding so far to develop technologies that apply deep learning to ultrasounds and alleviate the leading cause of death–cardiovascular disease. In addition to seeing what your baby looks like, ultrasounds can also be used to detect heart disease. Bay Labs was over in Africa trying out their tech and helping the less privileged enjoy better medical treatment. According to Import AI, the prototype device they brought over looks like an ultrasound but has an embedded GPU so the AI solution is integrated with the hardware. Bay Labs came out of stealth mode in January of 2016 and they’re backed by Khosla Ventures.
Founded in 2014, California startup Mindshare Medical has taken in $2 million funding so far to develop “a breakthrough in evidence-based clinical decision support technology utilizing personalized diagnostics complete with effective follow up procedures and treatment plans”. That’s about all this stealthy startup has to say on their website, but our understanding is that their focus is on reducing costs by tracing the effectiveness of treatments back to the original AI driven diagnosis in order to add an additional layer of insight.
Founded in 2011, Florida-based startup Entopsis has taken in $1.73 million funding so far to develop cheap platforms aimed at increasing regular health monitoring. Specimens like urine or blood samples are taken and processed into a digital image that is uploaded to the cloud as seen below:
Founded in 2014, Maryland-based startup Proscia has taken in $1 million funding so far to develop “the #1 pathology cloud platform” that uses image analysis and big data analysis to provide organizations a platform to work with whole slide images. Pathology is the study and diagnosis of illness through the microscopic analysis of samples from bodily fluids, tissues samples, etc. You can immediately start to see the benefits of a platform that can make all the images anonymous and then begin analyzing them using artificial intelligence. Now pathologists can receive “suggestions” as to what sort of information can be gleaned from the images they upload.
Here we have 10 startups, most of which are trying to do essentially the same thing. Will we see a future where all these startups are acquired by hardware manufacturers and this just becomes a “feature” of the machines that do the actual imaging? We’ve had some discussions before with our readers on this topic, and we’d love to hear your comments below on where this whole “AI and medical imaging” thing is going.
If you pay more than $4.95 a trade, you're paying too much. Ally Invest is one of the lowest-fee brokers around so you spend less money on transaction fees and more on stocks. With more than 30 trades a quarter it drops even lower to $3.95 a trade. Open an account and begin trading today.