The PACSman, Mike Cannavo.
More often than not though, especially with AI roundtables, I feel like I’m following a decorated car with a bobblehead in the back window, everyone nodding up and down in agreement with what is said by one presenter or another when very little substance is usually discussed. Just once, I wish I could see that people have different opinions.
I’ve also never been able to figure out how some of these AI clinical studies get published. I just read a study of 2,500 participants in which almost 600 cases did not mention the nodules in the original report. It would be a revelation if the 24% not found were a concern. Of that initial 24%, only one in five (about 120 cases) were radiologist-confirmed nodules, and of those less than 20 were considered potentially malignant – and ultimately only two probably malignant nodules.
The burning question here is: does the 0.08% improvement in results really justify the cost of AI technology in this use case? This question is particularly crucial because no one knows whether findings identified as potentially or probably malignant have even been confirmed until a biopsy and pathology report is completed.
It can be argued that saving a single person makes the difference for people whose discovery would have been missed without the technology, but is it worth the time and cost? When you factor in the cost of reviewing the 24% of the more than 2,500 studies in which the nodules were not mentioned in the original report that the AI would have found (nearly 600), then discover that three out of four of those identified by the AI were false positives…well….what’s the cost of that too? After all, the last time I checked, AI was being touted as a tool to save interpretation time for radiologists, not add to it.
I like to see positive stories about imaging technology. I was thrilled when I read a story that began, “Up to 60% of radiologists intend to adopt artificial intelligence tools in clinical practice in the near future.” As the article progressed, it said, “…the opinions of those who will inevitably be most affected by its use – radiologists – still remain relatively elusive.”
Now, “elusive” is usually a code word for “uncertain,” implying that radiologists are unlikely to use the technology. But what about that 60% figure? It turns out that the study they used surveyed 66 radiologists. Now, according to the US Bureau of Labor Statistics, there were nearly 30,000 practicing radiologists in the US in 2021. How can you extrapolate a 60% adoption rate from a sample that equals 0.22% of the total population? It just defies logic.
So where is AI imaging technology most likely to be adopted? The answer is simple — where there is the most immediate need. There is a shortage of radiologists worldwide, although the shortage is not as dire as many would like. Europe has 13 radiologists per 100,000 people while the UK has only 8.5 per 100,000. Malaysia has 30 radiologists per million or 3 radiologists per 100,000.
It is not only the population density that makes the difference, but also the number of studies commissioned. This is where the United States leads the pack in one area and lags behind in another. With 11 radiologists per 100,000, the United States is doing well. But France and Germany, for example, have more radiologists per capita. Additionally, the more specialized modalities used in the United States have longer – and in some cases much longer – reading times.
Medicare population growth has outpaced the diagnostic radiology (DR) workforce by approximately 5% from 2012 to 2019. Interestingly, the number of diagnostic radiology trainees entering the workforce has only increased by only 2.5%, compared to a 34% increase in the number of adults. over 65 years old. This is the age group in which most radiology studies are commissioned. To make matters worse, 40% of radiologists currently practicing are expected to reach retirement age within the next decade.
So what is going to be accepted first? In the United States, growth will continue to be slow until usage reimbursement is in effect. In other markets, tuberculosis (TB) screening, COVID-19 screening and other areas will make the adoption of AI crucial, especially where resources are limited.
Remote digital x-ray units on the trucks can drive to where the patient needs to produce the x-ray, and then the AI can produce a real-time reading before the patient leaves. A new AI model used 165,000 chest X-rays from 22,000 people in 10 countries and tested it against chest X-rays from 1,236 patients from four countries, 17% of whom had active TB. Compared to radiologists, the AI system actually detected TB better with greater sensitivity and specificity, reducing the cost of TB detection by 40% to 80% per patient.
This does not mean that AI is better than radiologists. It’s just that in this selective situation, the AI worked well for the application being used, especially in developing countries.
AI also has incredible potential to identify the most dangerous potential mutations linked to COVID-19, so researchers can get a head start on developing protective vaccines. A Swiss team generated a collection of one million lab-created mutated spike protein variants, then trained machine learning algorithms to flag potential harmful variants that might arise in the future. It is hoped that this knowledge can help create next-generation vaccines and treatments.
This is another area where AI plays a role in diagnostic imaging, but not in the more “traditional” sense of imaging data processing. This is one of the challenges of AI in healthcare: where is it used and how?
There are literally dozens of AI applications in healthcare. AI can address everything from improving robotic surgery, to connecting and taming millions of data points, to improving the patient experience. That’s why a report says the AI market is expected to triple by 2030 to over $200 billion.
Interestingly, most prognosticators have predicted only $500 million in sales for the AI medical imaging market in 2022 and just over $1.2 billion by 2025. That number might seem like a lot , but when you divide it by 200+ vendors with maybe a dozen companies (if that) currently making money instead of hemorrhaging it…you get the conundrum here.
Where AI is going and how and when it gets there remain question marks, along with most new technologies. Above all, we need to be honest with ourselves about the answers to these questions and not just nod our heads in agreement with everyone else, hoping whoever nods first is right.
Michael J. Cannavo is known throughout the industry as the PACSman. After decades as an independent PACS consultant, he has worked as both a strategic account manager and solutions architect with two leading PACS vendors. He has now returned safely to the dark side and is sharing his observations.
Its healthcare consulting services for end users include PACS optimization services, system upgrade and proposal reviews, contract reviews and other areas. The PACSman also works with imaging and IT vendors to develop market-driven messaging and sales training programs. He can be reached at firstname.lastname@example.org 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.
Copyright © 2022 AuntMinnie.com
#PACSman #Pontificates #Webinars #Bobbleheads