COVID-19 Special Feature – Healthcare Performance Insider
How Imbio is using AI to close the gap of missed diagnosis
Mike Hostetler lost his father when he was 13 to pancreatic cancer. Admittedly curious as to what happened and the cause of his father’s disease, he set his sights on medical school. During his early studies, a retired surgeon gave him some advice that changed his course: “Stay out of medicine. You can do more good by going into business.” So Mike got a business degree and went into the medical device space, having successfully launched over 50 products in his 25 year career, inherently focused on working with products that can ultimately change the course of experiences like he had as a kid. Today this work is in artificial intelligence, with Imbio, as the Director of Sales and Marketing.
Who is Imbio?
Imbio uses artificial intelligence to look at different patterns or anomalies in medical images, primarily focusing on CT scans of the chest. Their software essentially trains computers to recognize diseases that humans have previously determined to be present in patients with that specific disease, using thousands of patient scans, to help radiologists find diseases that are difficult to see with the human eye. They personalize each image for that patient and that treating physician, transforming standard images into rich visual maps with quantitative reports. This level of detailed data on the type and extent of present abnormalities helps physicians develop a treatment plan specifically for that patient.
If we have AI, do we need Radiologists?
Yes! Mike is adamant to drive this point home, noting that when artificial intelligence first hit the scene at an RSNA conference there was a lot of hype around it, including fear and doubt about whether it would replace radiologists. The reality is, it’s just another tool in the physician toolbox.
The radiologist is the mechanic. We’re just giving them a new wrench – a really nice wrench. In Imbio’s case, the wrench looks at chest CTs for chronic and some acute diseases. Over the last couple of years there’s been this sigh of relief because there is a radiologist shortage globally. AI is helping them become much more efficient. Some studies have shown efficiency to be more than 30%. Now, it’s up to the radiologist to choose which algorithms and AI technologies they want to deploy to keep the momentum going.
Leaning in with industry partners
While consumer products are adopted in minutes (iPhone 12 anyone? Oh sorry, it’s been out for 2 weeks and they’re sold out everywhere), healthcare has significant constraints. It’s heavily regulated because we’re working with people on top of minuscule profit margins.
Mike notes that the primary roadblock in many countries is not the radiologist but the IT and security systems within the hospital. They have their own data centers; new systems and technology like Imbio mean there will be some upfront cost as in people resources – neither of which is readily available anywhere in healthcare today.
A single integration can take over a year to install simply because of the proverbial red tape.
How many patients are missed during that time?
Mike points to IT as the main roadblock in integrating with hospitals and health systems. Highlighting an example that has currently been delayed 9 months from doing a 90-day trial just because the IT systems can be so slow. With a perpetual focus on solutions, Mike highlights the value of channel partners in cutting through delays. Working with an external partner that has already cut their teeth in the space and made it through the red tape and security gauntlet, allows higher levels of imaging to be used in a matter of weeks versus years.
The partnership model with companies like 3DR Labs has allowed Imbio to expand fast, with the potential to deploy AI technology in 70 countries and thousands of providers. And we’re just getting started.
Now that we’ve got these channel partners, we’re freed up to do the things that we want to do – build algorithms. We’ve got six or seven algorithms in development. We’re going to stick to our lung and cardiothoracic space, because there’s a lot of disease there that needs artificial intelligence to help. And we owe all that to our channel partners. No question.
Closing the gap
There are big health systems across the globe that admittedly miss patients with disease. In radiology, it can be from the sheer number of chest CTs on a daily basis. They can’t all be treated and there is no historical software failsafe. Patients unfortunately slip through the cracks. Sadly, when patients slip through the cracks, they go home, they can be okay for a year or two, then come back with exacerbations. It costs the health system more money – on top of potential fines from CMS. AI solutions, like Imbio, can help screen and catch patients who can benefit from early stage disease treatment.
In October 2020, MIT released survey results from 900 health professionals showing that 64% have either deployed or evaluated medical imaging AI. That means hundreds of hospitals across the US have deployed some form of AI in medical imaging, for use in everything from triage cases (like intracranial hemorrhages and pulmonary embolisms) to broader chronic diseases (like COPD and emphysema).
AI in the hands of radiologists and treating physicians together is truly personalized medicine. The increased collaboration and use of quantitative data I’ve seen firsthand has been just fantastic.
Imbio and other AI companies are influencing that, and it’s just fun to watch.
Imbio is improving healthcare by increasing the number of patients that can be treated and removing an additional burden of cost in missed diagnosis. That’s a pretty profound impact. It looks like that surgeon was right Mike. You’ve done good!