I had a discussion with a chief information officer the other day asking me if I felt artificial intelligence (AI) was real. My answer was, "Define real."
The PACSman, Mike Cannavo.Does AI have benefits? Absolutely. Is it going to show a solid return on investment in this lifetime? In terms of patient care, yes. But financially, probably not in the very near future. Thankfully, I have invested my entire $572 million retirement fund in other areas, so investing in AI isn't a consideration at least at this time.
I read a report put out by Signify Research that says imaging AI startups have raised nearly $3.5 billion in venture capital (VC) funding since 2015. That's "B" as in billions. That's a whole lotta dough.
Last year alone, the industry saw $815 million in venture capital funds raised, with an average investment of $33 million. Just a few weeks back, a company got a $100 million infusion of cash, bringing its valuation to the billion-dollar level as well.
While over 200 companies made 290 investment deals, the 25 highest-funded companies were responsible for 80% of all capital raised. Interestingly, most of it went to Chinese firms, which raised $260 million versus $150 million for all the U.S., South Korean, and Australian companies combined.
The Signify report says although the funding total in 2021 of $815 million seems big for a category that's figuring out its path towards full adoption, the amount venture capital firms are investing in other areas of healthcare makes it seem pretty reasonable.
Pretty reasonable? Seriously? Well, when you consider seven digital health startups had funding rounds with a total value of $267 million in just one week, that's reasonable to some, but certainly not to me.
The Signify report points out that imaging AI funding remains strong despite a list of headwinds (COVID, regulatory hurdles, lack of reimbursement), while showing more signs of AI market maturation (larger funding rounds to fewer players) and suggesting that market consolidation is on the way.
Ya think? I have been saying for nearly four years that the AI market cannot sustain 200+ vendors, with more entering the marketplace every day. Do I get a commission on their report? I love the ending, "... the next steps in AI's evolution won't be as straightforward as some might think." It can't get any more convoluted that it already is.
Is AI a wise investment?
So is AI the next place to invest all my money? Using AI technology is a plus if you can justify the investment one way or the other. That said, AI's limitations are starting to show.
Just a few weeks ago, the U.S. Food and Drug Administration (FDA) issued this warning on computer-aided triage and notification (CADt) software for patients suspected of having large-vessel occlusion (LVO): "... LVO CADt devices are intended for prioritization and triage only ... These devices should not be relied on when making any diagnostic decisions."
Let's see -- an end user getting paid over $1,000 under the new technology add-on payment (NTAP) program to use software that shouldn't be relied on for any diagnostic decisions ... makes sense to me.
I had another discussion with a company that shamelessly ghosted me after I took exception to its claim that it was the greatest thing since sliced bread. The organization said it had over 1,000 users, had 20 original equipment manufacturer (OEM) agreements, and made over $35 million last year just from software sales. What more could one ask for?
My answer: "How about not losing $10 million in the process as well?"
This same vendor felt its AI solution was the be-all and end-all, and that every PACS/enterprise imaging systems (EIS) vendor would want to allow them to install its AI solution on its PACS/EIS -- for free of course.
The company seemed to take offense when I said that there needs to be something in it for the vendor to offset sales, marketing, product development, integration, and other costs. Typically, a margin from 20% to 30% covers this, but with all the competition in the market it may even take more.
In this vendor's case there were numerous other vendors offering similar products in the same area, and sadly it couldn't differentiate what made its software better than others. In the end, it was just bigger and sang Queen's "We are the Champions" louder, that's all.
In some cases, you will see AI companies breaking even, and a few even making a few bucks, but the vast majority are hemorrhaging cash big time -- that's not including factoring in the startup capital that it has taken to get where they are now.
Truthfully, it's hard to tell where any company is financially because only publicly traded companies release their financial results, and what they share is very scary sometimes as well. Even those who have major cash infusions from VCs and are on their third and fourth round of investments and have a scary pro forma. Why? Because a lot of the investments being made now are in large part to protect the investments that have been made in the past.
That said, there are a few select companies that are making some money -- maybe a few handfuls, tops -- but they aren't setting the world on fire any more than most PACS/EIS companies are. Some AI companies are also making money from programs that help underwrite and jump-start the use of the technology.
But alas, I digress. It's like me getting 50 cents for every "A" I brought home on my report card and 25 cents for every B ... yet seeing it all disappear when I got "unsatisfactory" for conduct. Imagine that. AI vendors also don't lose TV privileges when they get Cs, Ds or, God forbid, an F ... and often still get rewarded to boot.
Is AI technology filled with potential? Absolutely! Is the technology promoted properly? In most cases no.
Most of the AI marketing collateral I have seen is filled with fluff and very little substance. Now, few would question that AI can and does have a positive impact on patient care, and in many cases can even help improve the workflow of radiologists. But in the end, it still comes down to how we justify the cost of the technology.
Will the PACS/EIS you have support AI? That is the number one question. If the AI you want to use (not just what the vendor offers) hasn't been validated, then you have to take a Rube Goldberg approach to using it, launching it separately.
Each time you use AI, you also hope you match up with the right patient until the system can be validated and interfaced. Where you buy it from matters too. Do you buy it directly from the AI vendor or from the PACS provider? If so, who does the interface, provides support, updates, upgrades, and more?
Many of the AI offerings you see today will become integrated into imaging modalities and offered as a value-add by the modality manufacturers. The cost to go it alone with a direct sales effort is expensive, and even establishing OEM relationships is simply too much for the vast majority of vendors.
Some will make it into PACS/EIS as well, but PACS/EIS vendors can afford to be very selective in whom they choose as a partner and who they don't. If we're lucky, one out of four AI vendors will make it (if that) and I'm not saying which ones, either.
Let's just hope that the interest in "Survivor-AI" rates as high as "Survivor-The Australian Outback" did, with over 45 million viewers back in 2001, and doesn't end up not like "Survivor 42" which this year hasn't even garnered 5 million viewers.
Stay tuned.
Michael J. Cannavo is known industry-wide as the PACSman. After several decades as an independent PACS consultant, he worked as both a strategic accounts manager and solutions architect with two major PACS vendors. He has now made it back safely from the dark side and is sharing his observations.
His healthcare consulting services for end users include PACS optimization services, system upgrade and proposal reviews, contract reviews, and other areas. The PACSman is also working with imaging and IT vendors developing market-focused messaging as well as sales training programs. He can be reached at [email protected] or by phone at 407-359-0191.
The comments and observations expressed are those of the author and do not necessarily reflect the opinions of AuntMinnie.com.
