Can AI Help to Save the Practice of Radiology for the Future?

Dec. 5, 2017
At the RSNA Conference last week, the breadth and intensity of the discussions around artificial intelligence revealed the universal concern over how radiology can prove itself in the new healthcare

In what was perhaps one of the most memorable openings in literature in English, Charles Dickens began his immortal A Tale of Two Cities with this: “It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness, it was the epoch of belief, it was the epoch of incredulity, it was the season of light, it was the season of darkness, it was the spring of hope, it was the winter of despair, we had everything before us, we had nothing before us, we were all going direct o heaven, we were all going direct the other way—in short, the period was so far like the present period, that some of its noisiest authorities insisted on its being received, for good or for evil, in the superlative degree of comparison only.” And yes, that was one long, run-on sentence….!

And yes, participating in RSNA 2017, this year’s edition of the annual RSNA Conference (sponsored by the Oak Brook, Ill.-based Radiological Society of North America), did bring to mind Dickens’ astonishing opening to his great 1859 novel.

And though I saw no one at RSNA 2017 who reminded me at all of Sydney Carton, Lucie Manette, Charles Darnay, or Madame Defarge, I did actually think a bit about France in 1775 (on the eve of the French Revolution). Here’s the thing: the practice of radiology, as we’ve all known it, is moving into uncharted territory now, as the financial, operational, and medical practice model on which it’s been based, is shifting under the feet of today’s radiologists. With both Medicare and private-insurer payment under accelerating threat (let’s face it, diagnostic imaging procedures are an easy target for reimbursement deficit-hawk types), and with the demands for speed of turnaround for interpretive reports also accelerating, there are literally not enough hours in the day for practicing radiologists to make up growing income shortfalls from ongoing reductions in payment from all sources.

And really—let’s be real—none of us are going to be seeing radiologists standing on street corners attempting to sell $1 homeless-resident newspapers anytime soon. In 2016, the median compensation for non-interventional radiologists in the U.S. was $503,255, according to the American Medical Group Association (AMGA), as reported by RSNA; that was up from $490,399 in 2015.

Indeed, RSNA’s Richard Dargan, in his story, quoted Howard Forman, M.D., a professor of radiology, public health, economics and management at Yale University, as stating that “The take-home message here is that we are faster and better at reading studies, and we’ve improved the way we deliver images and the way we process reports and communicate results. There’s no reason to think that this will change anytime soon,” Dr. Forman said.

But, be that as it may, radiologists do feel themselves under threat, as fee-for-service medicine gradually begins to collapse, and radiologists feel more and more need to prove their value in the new, value-based, healthcare. That challenge was the subject of so many conversations at RSNA this year; and every single radiologist I spoke to agreed that radiologists need to begin to seriously leverage artificial intelligence/machine learning/deep learning technologies and strategies in order to demonstrate their value in the emerging healthcare.

Multiple uses envisioned for AI in radiology

In that context, on Friday, Dec. 1, Mia DeFino published an interesting article in Diagnostic Imaging, entitled “Learning from Deep Learning in Radiology,” which highlighted the AI/deep learning emphasis at this year’s RSNA. As DeFino pointed out, “There are many opportunities to use AI and deep learning in medical imaging: image quality control, imaging triage, efficient image creation, computer-aided detection, computer aided-classification, and automatic report drafting.” And she reported on a presentation by Lucio Prevedello, M.D., of The Ohio State University, who told his audience that “[D]eep learning can help in other ways aside from helping label images, such as improving process efficiency when dealing with many high priority cases. For example,” DeFino noted of Dr. Prevedello, “[H]is lab has been able to filter incoming images based on priority using deep learning. The algorithm looks at the images to identify brain hemorrhage or stroke, if the computer detects one of the flagged factors, the patient will move up on the priority list to have their images analyzed first. If the algorithm does not detect any critical factors, the patient’s case falls towards the bottom of the priority list.”

That is just one of numerous potential ways in which radiologists might leverage AI capabilities. Citing another initiative, this one involving Curtis Langlotz, M.D., Ph.D. of Stanford University, DeFino wrote, “For example, at Stanford, Langlotz described a deep learning algorithm that can improve MRI image quality and suggested a future where the MRI machine can notify the technologist that the images are too fuzzy to be read accurately. Through this type of approach, it is possible to improve MRI image quality and have the patient spend less time in the machine.”

Of course, Langlotz said in his presentation, “There is a hype cycle for emerging technologies—we are at the peak for inflated expectations about deep learning and machine learning, the trough of disillusionment is two to five years away. Some have predicted that radiologists will be replaced by robots, but because of the nature of a radiologists’ work, it is unlikely that a computer will be able to fully develop the complex analytic and reasoning skills required to completely replace human radiologists,” as it will still require human judgment to discern the meaning behind automated processes and make real clinical decisions. Still, Langlotz had said in his presentation, “Up to 10 percent of patient deaths are related to some type of diagnostic error and 4 percent of radiology interpretations contain clinically significant errors,” with deep learning potentially significantly improving error rates and patient safety incidents.

Looking—gingerly—towards the future

“I think we’re in the eye of the storm now in terms of the hype cycle,” Rasu Shrestha, M.D., the chief innovation officer at the UPMC health system in Pittsburgh, told me, on the exhibit floor last week. “You could see the storm brewing; and we in the imaging industry had a lot to do with that, because it’s a technology whose time has come,” at a moment when the volume of data and information involved in radiological practice is exploding, and the need for speed is accelerating as never before. And, despite the level of hype right now, Dr. Shrestha underscored his view that radiologists will perforce need to leverage AI tools and strategies simply to stay productively and effective in practice.

Jonathan Messinger, M.D., a practicing neuroradiologist at the six-hospital Baptist Health South Florida integrated health system, put it even more bluntly. “We need to find a way to prove our value in the arena of care; otherwise, we’re going to get passed by,” Dr. Messinger told me, adding that, “You’ve seen everyone starting to freak out over the idea of artificial intelligence potentially replacing human beings, but that’s not going to happen at all. It will create augmented intelligence. And in fact, the radiologists who don’t use AI will be left out,” as the push towards value accelerates.

Meanwhile, James Whitfill, M.D., the chief medical officer at Innovation Care Partners in Phoenix (formerly Scottsdale Health Partners), a physician-led clinical integration network, told me that “Machine learning could provide us with a once-in-a-generation type of breakthrough that could dramatically increase our efficiency in new ways, as physicians, including as radiologists, but also across all the specialties. The reality,” he added, “is that physician offices and hospitals lack the resources to change a lot of what we do; but if machine learning ended up living up to its hype, it might truly provide a breakthrough in improved efficiency and effectiveness that we haven’t seen happen yet.”

Of course, none of this is going to happen automatically—or even with exceptional speed. “Clearly, machine learning is everywhere this year”—but, “Not much of it is for real yet,” imaging and imaging informatics consultant Joe Marion told me on the exhibit floor last week. “Everybody wants to jump on the bandwagon.” Still he added, development is accelerating apace. In that context, he said, “[T]here are numerous options that vendors could take here. They could develop their own applications, do third-party applications, or work with academic researchers to develop solutions. So there's a mix of approaches.”

In any case, all those I spoke with at RSNA are agreed on one thing: artificial intelligence/machine learning/deep learning is a phenomenon whose time has come, in the radiological world.

Sponsored Recommendations

How to Build Trust in AI: The Data Leaders’ Playbook

This eBook strives to provide data leaders like you with a comprehensive understanding of the urgent need to deliver high-quality data to your business. It also reviews key strategies...

Quantifying the Value of a 360-Degree view of Healthcare Consumers

To create consistency in how consumers are viewed and treated no matter where they transact, healthcare organizations must have a 360° view based on a trusted consumer profile...

Elevating Clinical Performance and Financial Outcomes with Virtual Care Management

Transform healthcare delivery with Virtual Care Management (VCM) solutions, enabling proactive, continuous patient engagement to close care gaps, improve outcomes, and boost operational...

Examining AI Adoption + ROI in Healthcare Payments

Maximize healthcare payments with AI - today + tomorrow