What Radiology Tells Us

About AI in Medicine

No other medical specialty has adopted AI greater or faster than radiology. According to a survey of 1,000 physicians working at Brigham and Women's Hospital in Boston, Massachusetts in 2001, 64% of surveyed radiologists reported using AI ⁽¹⁾.

Number of physicians reporting they use artificial intelligence in their medical practice (n=1000) ⁽¹⁾

That survey is certainly not meant to represent the state of the industry, but given Brigham is known for innovation it might be a hint of the future. Consider that the number of AI tools cleared by the FDA for clinical use right now in the U.S. are dominated by radiology applications ⁽²⁾. One begins to see and recognize that this specialty probably has already learned some important lessons.

Distribution of FDA cleared artificial intelligenceproducts in the U.S. FDA⁽²⁾

I attended the 2022 annual meeting of the Radiology Society of North America (RSNA). Over four days, nearly 35,000 attendees presented the latest information on all kinds of topics. My impression was AI was prominently discussed and presented just about everywhere. Here are my key take-aways and what each might indicate for the broader industry.

Physicians are enthusiastic about the POTENTIAL for AI in medicine, however they remain concerned about safety and quality. Most of the AI tools are trained on large data sets originating from patient populations in the northeastern and west coast of the US. Thus the amount of bias inherent in the AI's on the market are unknown, and thus represent a new kind of risk.

• There is a need to validate and measure the impact of local population data sets before implementing AI. However, most healthcare delivery organizations (HDOs) do not have specialists like data scientists who can shepherd that kind of study. The broader industry has to acknowledge this gap, with AI vendors including such validation as part of their service offering.
• Today, AI tools are marketed by statistical factors such as listing their specificity and sensitivity. Clinicians aren't too interested in theory, what they want to see is clinical studies that demonstrate results.
• Radiologists are acknowledging their own bias to trust AI more than their own training. This can lead to errors and safety issues since AI results are not perfect. Even in radiology it is still early days regarding understanding biases and limitations, both on the AI side and the human using the AI. The industry appears to be ahead of the clinicians here on acceptance.

Using AI tools for diagnostic assistance is the most common use case in radiology today, such as through computer vision recognizing a lesion on in a large series of CT images (e.g. a tumor). Helping the radiologist be more productive and possibly lowering stress is a good thing, but is not seen as driving a lot of value.

• More value will be realized when AI is helping with treatment decisions, which is based on a complex set of inputs like patient history, genetic makeup, and many more. In the case of oncology, it gets more complicated by knowing what type of cancer, where it is located, and its rate of growth for example. Some vendors and thought leaders are already speaking about how AI will help enable a new era of decision intelligence ⁽³⁾.
• AI tools have great potential to relieve the burden and toil in everyday work streams, which would generate a lot of value given the current workforce crisis at HDOs. Some vendors are already marketing AI built into their imaging machines, so that a less-skilled technician can perform advanced imaging studies. This same concept shoe applied across the healthcare enterprise.

Radiology vendors seem to be coalescing around three business models for AI.

1. The first was discussed already, which is selling an AI with a specific algorithm used assisting clinicians make diagnosis or interpret information, with the future being decision intelligence. Example vendors in this space are AIDOC and Lunit.
2. The second model is around platforms that are used to build, test, host, manage, and govern AI usage across the enterprise. Nuance and Rad.ai are examples of this approach. The value here is all about making AI's manageable in an era expected to be full of them, so that an HDO can select the best of breed clinical AI and manage them centrally.
3. The third business model is around applying AI to make workflows more efficient, meaning focused on the operations of the radiology department. These vendors promise higher asset utilization, higher image quality, and better patient experience. Example vendors in this space are Viz.ai, V7Labs and LeanTaaS. Philips is a large vendor who has AI capabilities across all three spaces.

I expect to see other AI tools targeted at other medical specialties evolve similar to radiology, and hopefully listen to this group's experience. In fact, the RSNA and American College of Radiology have several opportunities to get involved in standard setting. I encourage readers to seek these groups out to learn more.

— Seth Feder | seth@ontargetadvisors.tech | OnTarget Advisors LLC (C) 2023

---------------------------

References:

1. The Use of Artificial Intelligence in Medicine: A Survey of Physicians. Authors: James G. Anderson, MD, MS; David A. Asch, MD, MS; and Atul Gawande, MD, Published: JAMA Internal Medicine, March 1, 2022
2. FDA Announces More Than 500 Market-Cleared Artificial Intelligence (AI) Medical Algorithms Available in the United States, Published: February 24, 2023, FDA
3. Decision Intelligence in Healthcare: A Primer. Authors: David A. Asch, MD, MS; Atul Gawande, MD; and James G. Anderson, MD, MS, Published: JAMA Internal Medicine, March 1, 2022 — Seth Feder | seth@ontargetadvisors.tech | OnTarget Advisors LLC (C) 2023

Cover image source: Pixabay