New computing technologies and the vast amounts of medical data generated now make machine learning (ML) and artificial intelligence (AI) algorithms feasible in ways that have been unachievable in the past. Lately, new AI algorithms are announced almost daily, which may lead us to conclude that access to the enormous data sets powering these innovations is no longer an issue in healthcare. Unfortunately, diverse data sets for medical algorithm training are still hard for developers to find.

Empowering radiologists to develop AI algorithms at their own institutions, using their own patient data, to meet their own clinical needs, could potentially accelerate the development of AI for diagnostic radiology.Currently, developments in AI for the radiological sciences are being driven by exciting research happening primarily at institutions with extensive informatics and data science resources — and primarily using single institution patient data. While some of these algorithms are achieving performance levels similar to the performance of diagnostic radiologists, it is not clear if and how these tools will be generalized to widespread clinical use across multiple institutions. Without that transition, AI will fail to have broad impact on patient care.

Despite the prowess of newer deep learning methods for recognizing patterns in images and other complex forms of data, the same fundamental barriers from the past two decades still stand in the way. Here’s a take on four major barriers that are slowing the integration of AI into daily clinical radiology practice and how they are slowing radiological advances.

As a radiologist, it is difficult to escape being bombarded with discussions of AI, machine learning, big data, and how the practice of the radiology profession is going to change. Balancing the potential of AI with the current capabilities and practical considerations for implementing successful projects can be a struggle.

We at the ACR DSI have begun collaborating on the next steps in creating an AI ecosystem, including building out the Assess-AI process and registry

Radiomics is an exciting new field of radiology. It uses AI to detect granular patterns in the pixels of medical images to predict molecular and genetic phenotypes of tumors — thus enabling precision medicine.

Working to solve problems that don't exist is a waste of everyone's time. Preventing that fruitless pursuit is a key goal of the ACR Data Science Institute (DSI) Data Science Subspecialty panels.

To date, no other national medical specialty organization has developed AI use cases for developers so they can create AI algorithms that are relevant and valuable to healthcare professionals. The 50 use cases in the newly released ACR DSI TOUCH-AI Directory are building a framework to facilitate the development and implementation of AI applications that are poised to help radiology professionals in disease detection, characterization, and treatment.

To be usable, massive datasets must be trimmed and categorized — largely by computer algorithms. That’s where ethical sharing of patient data comes in. Proper oversight of those data management algorithms is critical to ensure protection of sensitive patient data, especially when it comes to health systems that are sharing their information with AI and big data companies.

The fundamental question with any new imaging-related technology is whether and how it can add value — to the patient, the health system, or the radiologist. Here are three applications where AI methods can have a positive impact in different ways: 1) pre-screening of screening mammograms, 2) assessing bone age, and 3) rapid identification of life-threatening conditions.

When Wilhelm Roentgen first used his machine to extract new information about what was happening inside a patient in 1895, he defined the original cyborg radiologist. This human-machine combination has evolved steadily ever since, with new machines building on the old to increase the information we obtain.

Dr. Abadi enters the dimly lit reading room for the first time since her medical student rotations in radiology. After a grueling intern year in surgery, she appreciates the relative calm of the reading room and looks forward to beginning her radiology residency. As she logs into her workstation for the first time, she turns to an upper-level resident and asks, “So who’s the best radiologist in the department?”

It's said that "data is the new oil." Radiologists should agree — our bread and butter is interpreting complex visual data. Enter artificial intelligence (AI), with its predictions of doom for human radiologists.

Hype aside, AI and machine learning analytics are advanced techniques used in data science, a hybrid field of computer science and statistics. These are not the statistics we remember from medical school.

It’s a classic medical school scenario: an attending physician on rounds in the medical ward leads a group of trainees as they stand in a circle outside of a patient’s room. One of the trainees presents the case, and then the attending fires off a series of questions for anyone in the group to answer.

A key ingredient in moving Artificial Intelligence (AI) algorithms into routine clinical practice will be ensuring our healthcare system supports fair compensation for their development. However, figuring out how that will happen may not be as simple as it might seem.

Computers are now very good at extracting meaningful information from pixel data in images. Applications of AI, such as automatic detection and recognition of faces in photos, have already become familiar in consumer products. The use of AI in medicine is taking off, and medical imaging is one of the fastest-growing application areas.

I feel a vibration on my wrist that gently wakes me from my sleep. I roll over, touch the device on my wrist, and lay there for about five more minutes. Then I feel another vibration. I check the time on my phone. It’s 5:30 — time to get up.

My morning ritual starts in earnest: I weigh myself on my digital scale, which shows that my weight is the same as the previous day and automatically syncs with my wearable fitness tracker. As I get dressed, I ask Alexa, a voice-controlled speaker and personal assistant, to play Lose Control by Missy Elliott.

As we begin the new year, the ACR Data Science Institute (DSI) is celebrating its nine-month anniversary. While it might seem odd to count anniversaries in months, the speed of the advances in data science makes just a few months seem like years. At the 2017 RSNA meeting in November, the enthusiasm for learning more about artificial intelligence (AI) in radiological care was palpable, and the ACR DSI was well represented in the RSNA Machine Learning Showcase to share our work with RSNA attendees and industry leaders.

Artificial intelligence (AI) has its own language. So does economics. I feel pretty confident with econ-speak and can throw out phrases like intensity of work or acronyms like CPT, HOPPS, and QPP in rapid fire. But I am not as well-versed in the language of AI — a point that became apparent during a recent “Economics of AI” talk I gave to a room full of radiology AI experts.

Today, plenty of hope, fear, and hype surround the use of artificial intelligence (AI) in radiology. With media attention and many startups focused on using AI to identify findings within medical images, it’s easy for us in the radiology profession to have "tunnel vision" about AI in our field. However, it’s important to widen our aperture to see the many other ways AI can benefit medical imaging.

If you’ve ever used the search program on your smartphone to find all your photos of a special person or your pet, you are already using artificial intelligence (AI) for image interpretation

While it looks like magic, AI programs that differentiate photos of cats from dogs, and even your own pet from others, are usually based on conventional mathematical optimization.

So you want to learn about Artificial Intelligence (AI) but don’t know where to start? A Google search for “resources for artificial intelligence” yields more than 25 million results, so it’s understandable why you might feel overwhelmed. Adding “in radiology” prunes the list down to an only slightly less cumbersome 459,000 results. What follows is a brief collection of resources culled from these nearly half-million hits — a starting point to help you learn the basics of AI in medical imaging.

To explain how I view artificial intelligence (AI) in medicine, I like to use car repair as an analogy. As car repair is about vehicle maintenance, medicine is, in essence, about the maintenance and repair of people.

For the past five years, ACR’s Imaging 3.0 initiative has been the toolkit that we, as radiologists, have used to enhance our value beyond our interpretations to be better stewards of imaging appropriateness, provide a more patient-centered focus within our practices, and prepare for the evolution of healthcare financing and federal payment policy. Imaging 3.0 continues to challenge us to take ownership of our patients’ entire imaging experience, and now, we have a new tool in the kit that will help us further increase the value we bring to patient care at each step in the imaging value chain: artificial intelligence (AI).