The 2025 PACSMan Awards -- The last hurrah

The year was 1983. Three fresh-faced 20-something guys from Orlando were in Chicago, ready to take RSNA '83 by storm. The equipment and our luggage were loaded in the rental van, and we were off to find our downtown hotel. Somewhere along the way, we made a wrong turn, so we decided to stop and ask for directions.

“There’s a bar! Someone there will know!” Scott said. We parked the van and entered the bar. The scene was like one out of an old western where the piano player stopped playing, everyone got quiet, and all eyes were on us.

“Um … Scott, I think we chose the wrong bar,” I said.

“Relax. Cabrini-Green sounds like a nice Italian neighborhood.”

Yeah, right. “Um, Scott, these are either very dark-skinned Sicilians or we’re all gonna die…” I said softly. Now, in 1983, race relations were far from stellar, and here come three white boys entering a bar located in the middle of the worst housing project in the nation … Not good. Not good at all. Fast forward about a minute and a very nice older white-haired African-American gentleman gave Scott the directions we needed with the admonition, “Now you boys better hop in that van and get going while you still can.”

We headed on down the road to find our luxury digs at the Days Inn. And that is how my first RSNA started.

Michael J. Cannavo, aka the PACSMan.

The DigiPAC 8 failed to take the show by storm, even though our revolutionary camera-on-a-stick teleradiology system worked perfectly. The analog Pioneer laser disk cost about 4x that of film, but it was oh so sexy to look at and well ... We did get good exposure, although kicking Raytel and Dataspan’s butt never happened.

Shortly thereafter, the $50,000 in total that was budgeted for hardware, software development, my salary, travel expenses, and the like ran out, so I was on to greener pastures working for AT&T Bell Labs and then Philips. Both were equally clueless companies back then, but at least they had significantly deeper pockets. By 1987, I'd had enough of the corporate world. I launched my independent PACS consulting career as the PACSMan with an article in Diagnostic Imaging magazine.

With the exception of two very short forays into the corporate world again over the next 35+ years -- both mistakes -- I have been on my own, working with clients looking to evaluate and purchase PACS. For nearly the past decade, this has also included AI. It’s been a blast, but now it’s time to slow things down a bit.

This year marks the 31st anniversary of my 39th birthday, and my knees simply can’t handle all the walking the way they used to. My patience is also a lot thinner than it used to be, and, if you couldn’t tell, my cynicism is much more prevalent, too. Sadly, there is still so much fodder for the PACSMan Awards that I hate to consider not coming to Chicago anymore. So even though this is intended as my swan song, there may be an encore.

Companies are letting seasoned marketing pros go in place of puppies who think they know it all, and it shows. PR agencies are also still providing 35-word intros in press releases because they are paid by the word. The list of frustrations is endless, but it does make for very good award material.

Show summary

It wouldn’t be a PACSMan Awards without a summary of the show. The number of teleradiology companies here is surpassed only by the number of companies trying to recruit radiologists for positions with them. It’s a good time to be a rad, especially since so many positions can be done remotely.

With PACS, I expected to hear competing music from booths all promoting unified systems. Bob Marley over here singing “One Love” (pass the ganja please), “We Are the World” from USA for Africa at another (Michael Jackson, rest in peace), “I’d Like to Teach the World to Sing” from the New Seekers (the Coca Cola song), “One” by Three Dog Night at another. Everywhere, it was all about unified systems.

We spent years developing industry-standard interfaces like DICOM, HL-7, and this year they seem to have been displaced in favor of a complete system from a single company. You may have a great subsystem here, an acceptable subsystem there, and a mediocre subsystem for all the rest. But, hey, it’s unified. Many companies had AI-based reporting, so that was a plus. There were exceptions, but they were few and far between.

Clouds were everywhere, with few talking about on-premise systems except in teleradiology applications. So many of the majors are still years behind some of the newer independents who have cloud-native solutions versus cloud-enabled, but they should be much closer by this time next year. The challenge of native-cloud systems is that few have been clinically implemented for more than six months, if even that. Most of the sales announced are just now coming live, so we have little clinical experience with people actually using the cloud. Almost all is anecdotal.

Again, there are exceptions, notably those systems implemented by a few of the larger cloud-based vendors. That said, the 80% to 85% of facilities looking at cloud now are still in the evaluation phase. Radiologist-owned imaging centers are the largest potential implementers of cloud solutions to date.

While mergers and acquisitions are primarily seen within the AI domain, GE HealthCare’s planned purchase of Intelerad for $2.3 billion raises more than a few eyebrows. This would be GEHC’s fifth (or maybe sixth) PACS and/or teleradiology acquisition, with twice as many vendors having worn a GEHC hat if you count partnerships. This is by far the largest investment made by GEHC or any vendor, for that matter. It also has many speculating which PACS vendor will be scooped up next and by whom.

AI

AI is a conundrum. Do you believe what you read about and hear but don’t see, or do you take the leap of faith that AI is growing, albeit at a much slower rate than most of the 600+ vendors offering over 1,000 algorithms still hope for? Not in academics but in the real-world. Reimbursement and showing a return on investment (ROI) remain AI’s Achilles heel, although several independent insurers are starting to pay to use the technology.

Confidence in the technology remains a close second among AI challenges. AI-based reporting is also growing in leaps and bounds and is included in many of the cloud-native PACS solutions, as well as being implemented as a standalone solution for existing PACS.

Some of the newer algorithms developed, like abdomen/pelvis CT triage, should be easily adopted since that is the highest volume CT category with over 20 million exams annually. Many of the mega radiology groups are also offering AI as a complement to their mammographic interpretations as well although at an additional cost (typically $40).

Reviews on the success of this are mixed as the general public, especially Medicare patients, still don’t feel the need to spend the additional cost. Therein lies the rub with AI. Patients will accept the use of AI as a complement to the radiologist’s interpretation, but few are willing to pay extra for it. It needs to be bundled in the charges or reimbursement, whether the procedure is billed to Medicare, Medicaid, or private insurers.

Soon, facilities won’t be able to charge extra for AI. This eliminates the haves getting the benefits of AI technology that the have-nots do not. Either everyone gets the benefit of AI or no one gets it. Selective use will create the basis for a class-action lawsuit, and then everyone loses.

There are two truths associated with this comment. First, widespread (universal) use of algorithms will drop the reimbursement to the vendor to very low rates, typically under $20 per study and often under $10. The high reimbursements being seen today will be reduced significantly, offset instead by volume. This will vary based on the algorithm and the number of times it is used in a particular study but generally will be the rule. Second, all the studies being published talking about how AI reduces transcription time by X% will be noticed by the U.S. Centers for Medicare and Medicaid Services (CMS). This will no doubt lead to Medicare and Medicaid reimbursement for a study reduced by X% as well.

Getting real

Let’s also be real. As good as AI can be, do you really think a radiologist is going to say, “I missed something, but thankfully AI caught it” in their report? Even with the AI report included as an addendum, the bottom line is what the radiologist attests to in his or her report, no matter how they reached that conclusion. And if AI and the radiologists disagree? The radiologist needs a CYA statement saying AI’s findings (or things they didn’t find) that X cannot be excluded and XYZ should be considered.

Now, what I have stated in the above few paragraphs is sure to be controversial, but if you wait two to three years, when someone with Most Divine after their name says the exact same thing, “their” comments suddenly become wisdom. That has been the way it has been my entire career -- but as long as it finally gets said, it’s all good.

That said, I was also almost brought to tears by the number of people who seem to have recognized that honesty has a place in the industry all these years, and they appreciate my honesty. If nothing else, I have always been honest, although sometimes brutally so. I never thought it mattered to anyone or that they noticed. Apparently, they did, and it does. Thank you from the bottom of my heart.

Before I hand out the 2025 PACSMan Awards, it is important to note that these are all done in fun with no malice, ill will, or anything else negative ever intended. I have a unique (some may say weird) insight into how companies market themselves, but that’s what makes these awards so special. There could be 50 this year -- easily -- but I am pushing the limits of what we have space for with what I have listed here. So with that in mind, here are the 2025 “winners”:

The “One Hit Wonder” Award

He was 25, she was 17. Every time I hear this song, I think of this company...“Ooh, my little pretty one, pretty one ... when you gonna give me some time ….” The company will no doubt have a lot better success than the band did.