Thank you very much. It really is a pleasure to be here. The idea behind this talk is just to give an overview for what we do in radiology and give some background for those of you that are new to radiology. So we'll begin with a question. Here's a very simple question. Who is this a picture of? And again, I'm going to give you four choices. Is it A, Wilhelm Roentgen, the discoverer of X-rays? Is it B, a picture of a Brooklyn hipster? Is it a dude from ZZ Top, if you actually remember who ZZ Top was, and I certainly do? Or is it a picture of Mark Watson from the 1980s? OK, just let's start with this. Show of hands, who thinks it's Wilhelm Roentgen, the discoverer of X-rays? OK, good. We've got consensus. So we all agree it is Wilhelm Roentgen. And I'm not tricking you. It is Roentgen. But here, can I convince you that if you give Roentgen a small haircut, trim the beard a little, and pop a nose ring in him, you could drop that dude off in Brooklyn tonight at any hipster bar, and he would fit in perfectly. I literally googled Brooklyn hipster. That's the picture that showed up when I googled Brooklyn hipster. It looks like Roentgen. The guy was ahead of his time in so many different ways. But again, it actually was Roentgen. And for those of you that might be younger that might not know ZZ Top, by the way, you should know ZZ Top. They're fantastic. But the two front guys have these really long beards. This is a picture of ZZ Top. I actually don't have a picture of Mark Watson from the 80s, but we can all imagine it'd be a good picture to have. So a little bit about Roentgen. So Roentgen was a mid- to late-career physicist in Germany in the late 1800s. And he was well-respected, not famous. He hadn't made any big discoveries, but people thought of him as a good scientist. He was very meticulous. And he was experimenting with something called a Crookes tube, and you can see it here, a Crookes tube. It was like the thing to do if you were a physicist in the late 1800s. It had been discovered about 20 years earlier. It's a glass cylinder that's partially evacuated. They've taken all the oxygen out. And what they noticed was, if you ran a really high voltage through it, through these two metal areas, that there was this weird, eerie green-yellow glow on one end. And there was one side called a cathode side and one side called an anode side. And I mention that because there was a shadow, and that shadow was being cast on the anode side. So it was like there was an invisible light or ray coming out of the cathode side, striking the anode, and then going to the back wall and glowing. And they didn't know what that ray was, so they called it a cathodic ray, because they don't know it's coming out of the cathode. It's a cathodic ray. And everybody wanted to know, what's the deal with this cathodic ray? What's this eerie glow? And so they were all studying the cathodic ray. So was Roentgen. So Roentgen's in his lab one day with this Crookes tube, trying to understand the cathodic ray. He turns the Crookes tube on, and he happened to notice something in the corner. He made two incredible eureka moments. This was the first one, his first eureka moment. He noticed that the board in the corner had some phosphorus on it, and when the Crookes tube was on, the phosphorus glowed. That was eureka moment one. What he recognized instantly, to his credit, was there was something come out of the Crookes tube, not the cathodic ray, something else, that was striking the phosphorus and making it glow. He turned the Crookes tube off, it stopped glowing. He turned the Crookes tube on, it started glowing. There's some ray coming out of the Crookes tube that's not the cathodic ray. It's hitting the phosphorus and making it glow. He didn't know what that ray was, so he called it an X-ray, X for unknown, and that's how we get the term X-ray today. X-ray, X for unknown. So he knew there was an X-ray coming out, but he didn't know anything about it. So he was like, let's experiment with this. So the first thing he did was he took a cardboard box, painted black, thick cardboard, put it over the Crookes tube, turned the tube on, it glowed. OK, it goes through cardboard. Then he took a wooden thing, it goes through wood. He's experimenting to see what stops the ray. And then metal, OK, metal blocks it. And as he was moving stuff around, he thought he caught a glimpse of the bones of his hand. And he had his second Eureka moment. So he did what he thought was the best thing to do at the time. Hey, honey, come here, I need you. What is it, Wilhelm? Can you just come here for a minute? Can you just please come here? So he called over his wife. Honey, put your hand right here. Why? Just put your hand right here. Is it going to help? Is it going to hurt? I don't think so. So he had his wife put her hand down. And he took the first radiograph of a human. And you can see it right here. By the way, I have no idea why. She has a German accent and he doesn't. And if that's even a German accent, I have no idea. But that's the most famous radiograph ever taken. It's the first one ever taken. It's of his wife's hand. You can clearly see the bones of her hand, as well as her wedding ring. When she saw it, she exclaimed, I have seen my death. You can imagine how shocking it was. With this image, Rankine has now founded an entire new specialty of medicine, radiology. This was in February of 1895. Rankine would go on, by the way, to win the first Nobel Prize in physics. He would change the course of history. In his day, he was a household name. This is February of 1895. That was November of 1895. Now it's February of 1896, a few months later. It's a cold day in New England. People are ice skating on the frozen Connecticut River, when a 14-year-old boy, the McCarthy boy, falls and slips, hits his wrist. They take him to the nearby hospital at Dartmouth. Turns out the doctor taking care of him at Dartmouth, this Dr. Frost, his brother was one of the physicists at Dartmouth and had read about Rankine's discovery of X-rays. And they'd been playing around with it, because look, they got a lot of Crookes tubes. Let's see if we can do this. So like, have the boy come to the physics lab. Let's see if we can take a picture of the bones of his wrist and see if we can see them look for a fracture. It's a great idea. Let's do that. This is the picture they took. They knew it was the first radiograph of a person being taken at Dartmouth. I don't think they knew this was the first one being taken in North America. The first radiograph of a human taken in North America was of a child. And I say that with pride, because I'm a pediatric radiologist. It kind of started with us. So in this picture, which is a real picture, because they knew it was momentous, they took a photograph of it, you can see Professor Frost seated. You can see the McCarthy boy with his hand out. Then you can see Dr. Frost standing. And then Dr. Frost's wife, Nurse Frost, seated. It was a Frost affair, the whole thing. This is the image they took. And when they saw it, they were like, oh my gosh, it's amazing, it's so clear. He broke his ulna, the little bone here on your pinky side in your forearm. He broke his ulna. I'm a pediatric radiologist. I specialize in reading x-rays of bones of children. I've spent the last 20 years focusing on this. I have no idea what they're talking about. And I say it because it's so blurry. It took 20 minutes. The kid had to sit there for 20 minutes. Do you think he held still? I can't hold still for 20 minutes. He tried his best, he did decent, but it's too blurry. There's too much motion. But nonetheless, for them, this was a milestone. It was shocking how great it was. For contrast, this is a modern day x-ray of a child's wrist. We take it in a fraction of a second, not 20 minutes. A fraction of the radiation dose. And if you look really closely, you can see my son's fracture, which doesn't look that bad, although he tells me it hurt really bad. But the Pokemon cards made it feel better. But with that image that Renkin took, again, changed the course of history. Almost immediately, radiology became a part of modern health care delivery. It became important that every hospital had a radiology department. And we were often located in the basement. And radiologists and physicians used to come down to radiology to see the x-rays on their patient. And we took them on film back then. So you actually had to be in the hospital, because we actually were developing film and then hanging film up. And the radiologists would look at the x-rays. And the doctors on the hospital floor would come down to radiology to speak with the radiologists and talk about what was going on with their patients. That's how we got the moniker, the doctor's doctor. Because we weren't directly taking care of the patients. We were taking care of the patients, but we were doing so in partnership with our referring physicians, with the clinicians on the floor. We were their helper. We were the doctor's doctor. Now, a lot has changed since those days when we were located in the basement and people had to come down. We don't read things on film anymore. Everything's digital. Because it's digital, they don't have to come down to radiology. They can see it on the floor. We're not in the basement anyway, but it doesn't matter. They can just read it on the floor. They're not going to come down enough. But what hasn't changed is that we are still the doctor's doctor. We are still those that they come to for information on their patients that are the specialists in the interpretation of these various imaging modalities. Now, I was asked to talk about how you become a radiologist, how I got here. So we go to medical school. So this is me on my graduation day from medical school. So let's just take a moment and appreciate the hope in my eyes, the lack of wrinkles, the two-toned hair. Again, it was the late 90s. That was a thing. So after medical school, you are a doctor. Graduation medical school, you're a doctor. If you don't get picked up by NSYNC or Backstreet Boys, which I clearly was shooting for in this picture, you go on to be an intern. You work as a first-year resident as an intern in the hospital. And then you go specialize. You do your residency. And I did a four-year residency in radiology. So you spent four years just specializing in becoming a radiologist. And at the end of that, you can go off and practice radiology, or you can choose to specialize further, which is what I did and what's called a fellowship. A fellowship is an extra year or two where you specialize in one area of radiology. So to become a pediatric radiologist, I went to Boston and studied for an extra year at Boston Children's just doing pediatric radiology. And at the end of all of that, medical school, intern of your residency, fellowship, you come out an attending radiologist. A lot more wrinkles. I still have two-tone hair. But let's all agree the tones are different. So this is a chest X-ray. It's the most common image that we take in radiology. And it's still basically what Renkin discovered. The idea is you shoot X-rays at a body. And the X-rays are really high energy. And they go through things that are low density, like the air in the lungs. They look black. And they get stopped a lot by things that are dense, like the bones. And so you can expose film based on that. And that's essentially what we're doing. Has the physics principles changed? No. We've refined the process. It's all digital now. It takes a fraction of a second. The radiation dose is a lot lower. But the basic idea of that is kind of what Renkin discovered. But we can do a lot more now. Things that Renkin never imagined, we do now on a daily basis. But the X-ray is still, in many ways, our foundational element. This is a schematic of a current X-ray tube. It's just a fancy crook's tube. That's all it is, a fancy crook's tube. Turns out if you ram really high energy electrons, which, by the way, that's what the cathodic ray was. They didn't know what electrons were yet. The cathodic ray, it's an electron beam. You smack electrons really hard into something, they spit out X-rays. That's exactly what was happening. So an X-ray tube, a fancy crook's tube. So the most common thing we do right now is just plain films, plain X-rays. That's the most common thing we do. But we can use X-rays in other ways, mammography, fluoroscopy, and CT. So let's begin with this, a question about ionizing radiation. Because X-rays are a form of ionizing radiation, radiation that can actually damage your body. So here's the question. A large dose of gamma rays, which is another form of ionizing radiation, would lead to which one of the following? Is it A, sickness and death? B, incredible strength and green color? C, the ability to leap tall buildings in a single bound? Or D, the ability to do whatever a spider can? Who thinks it's A, sickness and death? Who thinks it's B, incredible strength and green color? Some votes for B. Good. Anybody think C or D? We're good? It's not? OK, no. So let's be really clear. It's A. If you get a big dose of gamma rays, you will get sick and you might die. And I say this because this is how Dr. Bruce Banner, a.k.a. the Incredible Hulk, got his powers. And you might be thinking, well, wouldn't I just get that if I got a big dose of gamma rays? No, you wouldn't. And I know that because some students at Texas A&M actually did the study. I can't love this anymore. So this is what the students actually did. What would happen to Dr. Bruce Banner, who, by the way, is fictitious, but let's let that go for a moment, if Dr. Bruce Banner actually got a big dose of gamma rays? And what they determined was he would not have turned into the Incredible Hulk. In all likelihood, he would have died. And again, I love this so much. They quoted Tony Stark, Iron Man. Now, a brief aside, as somebody who's been going to the RSA Annual Meeting for many, many years, and I love looking at the posters, I think this is an opportunity for us as a society. We don't quote Avengers enough in our posters. And I think this is a good example of what we could do better. That much gamma exposure should have killed him. Amen, Tony Stark. Amen. So if you're thinking about you want to be Incredible Hulk, trust the students from Texas A&M. Don't do it. Gamma rays would be bad. But we can use x-rays in a lot of other ways that Renkin didn't yet know about. One of them is fluoroscopy. So fluoroscopy is real-time x-rays, which is really important if you want to look at things that are moving in the body. And you can watch them real-time. But what moves in the body? Blood flow does. So what if we could inject something into the blood vessel that shows up on x-ray, and I could watch that, and I can see it flow through the blood vessels? Well, that's an angiogram. That's fantastic. So what they've done here is they've put a small catheter, a tube, a really small tube, into the aorta. It's the big blood vessel here in your chest and abdomen. And they've injected material that shows up on x-ray. So you can see the aorta and all those tree-like branches coming out of it. That's an angiogram. But what else moves? Peristalsis, your intestines. So what if I gave somebody to drink that shows up on x-ray, or I did an enema? Well, that's a contrast enema, an upper GI or lower GI. In this case, it's a lower GI. It's a baby who they thought might have a colonic obstruction. So I did an enema with this contrast that shows up on x-ray. And I can see their colon really well, and it flows through the colon. And I can watch it real time using x-rays. Mammography uses x-rays, uses delicate technique to create exquisite images of breast tissue, looking for very subtle findings of breast cancer. But again, based in the use of x-rays. CAT scans uses x-rays, but in a whole new way. So it uses x-rays to create slices of the body. So you lay on this table, and you go through this tube, and it shoots x-rays at your body. And then it uses a really powerful computer to then reconstruct all this data that it gathered from doing that. And it creates slices. So here's the way to think of a CT slice. You can see a slice of the abdomen right here. Imagine taking a one-inch thick slab of my body, just taking an inch right through here, and taking that slab, and then going and taking an x-ray of that. That's what a CAT scan does. Takes slices of the body. You don't have to actually slice into the person, which actually turns out really nice. And then what you can do is you can put all those slices together, and then you can tell the computer to do things like take away everything but the bones. And that's the other picture you see. We've created a skeleton. It's stacked up all the slices, took away everything but the bones, and I'm left with a skeleton that I can rotate in three dimensions and look at it in different ways. If you ever want to go down the rabbit hole, by the way, you may notice I've got a picture of the beetles there. Not right now, not at this moment, but if you ever want to go down the rabbit hole and look into this. Did the beetles have a fundamental or not fundamental role in the invention of the CAT scan machine? By the way, CT and CAT scan are synonyms. Did the beetles have a role or not? Feel free to go down that rabbit hole when you have time. But not all the things that we use use ionizing radiation. Some modalities don't, including ultrasounds and MRI. So ultrasound, you're all probably familiar with ultrasound, probably one of the obstetric uses, right? You put the jelly on the belly, you see the baby. Ultrasound uses sound waves that are outside of the audible range of humans. And it's not ionizing radiation. And it gives you real-time images. We use it a lot, particularly in pediatric radiology. Because again, there's no ionizing radiation and I can look real-time. So it's a great way to look at things like the kidneys, the gallbladder, the liver. It's a fantastic use. So we use it all the time, particularly in children. And then MRI, no ionizing radiation. MRI uses powerful magnets and radio waves to take images of the body. And it again gives exquisite, detailed images of the body, particularly good for the soft tissues. So many of you, I'm sure, have had MRIs of things like your back or your joints, like your knee, your shoulder because it really lets you see the articular cartilage, the ligaments. It lets you see all that that you're not gonna see on things like an X-ray. Now, people sometimes get confused between CT and MRI. They're like, they kind of look alike. Are they the same? Turns out, no. Here's how I explain the difference. The CT is like the donut and the MRI is the tube. So on the CT unit, you lay on the table, you pass through the tube. It takes a few seconds. It's literally like, hold your breath, and breathe. Okay, you're done. How was it? You got radiated, but you didn't know it, but you didn't feel anything. MRI, you go into the tube, you lay in the tube for like, I don't know, 20, 30 minutes. It's loud, it can be claustrophobic, gets beautiful pictures, but no radiation. So the physics of them are completely different and they give complementary, different but complementary pictures. Now, we even talked about nuclear medicine, so I'll briefly mention that because it's another area of imaging. Nuclear medicine uses radioactive material to image the body. And I'll give you two quick examples. One is a tagged red cell study. You take somebody's red blood cells and you attach a radioactive material, an isotope to it, and then you inject it back in the body. And it goes where all the red blood cells go. But if they're bleeding somewhere, it will collect where the bleeding is happening. And then you can image them and you can be like, aha, they're bleeding right there because I can see where all the radioactivity is. That's a red tag red cell study. And then there's PET scans. PET scans, you inject a radioisotope into the body that goes to areas of high metabolic activity. Well, what has high metabolic activity? Turns out the cancer does. So this is a way to look for cancer. So you see that splotch in the middle of the chest that's lighting up really bright? That's the cancer. And it's a way to look for other sites of cancer. So you may very well know the cancer is located here in the chest, but I wanna know is it anywhere else. But you have to also know where the tracer normally goes. So you see how bright the brain is and the head is? That's normal. You gotta know where it normally goes. You can see it being excreted by the kidneys and being stored in the urinary bladder. And that's where the radiologist comes in, knowing where it should be and where it shouldn't be. Now, everything we've talked about so far is in what I would call diagnostic, is known as diagnostic radiology. We in radiology make diagnoses. Does the patient have cancer? Has the cancer spread? We make a diagnosis. There's another branch of radiology called interventional radiology. Interventional radiologists use these medical imaging technologies to do minimally invasive procedures. Think of them almost like surgeons. You can see here what looks like a surgeon, but they're in the angiography suite. It's a fancy fluoroscopy room. They've got fancy fluoroscopy cameras using real-time X-rays. And you can see that image. What they've done is they've taken a small catheter, a tube, and they've threaded it into the artery, feeding the kidney, and injected contrast that shows up on X-ray. And you can see the tree-like branches of the artery of the kidney. And real-time in the exam room, you can watch the blood flow into the kidney and through the arteries, and then watch it flow out through the veins. And while you're there, you can make diagnoses. And because you're right there with a catheter, you can actually intervene and do stuff. So let's talk about a day in the life of both a diagnostic radiologist and an interventional radiologist. So let's start with diagnostic. What do diagnostic radiologists do on a daily basis? So here's the lifecycle of an imaging exam. It begins and ends with the patient and the physician taking care of them. So you'll start with a clinical question. For example, a patient twists their ankle, and the emergency room doctor wants to know, did they break their ankle? OK, so that's our clinical question. Did they break their ankle? So they do a request for a radiology exam. The patient goes down to radiology. A technologist takes the X-rays of the ankle and then loads them into a computer system, where they are then read by a radiologist. And you can see here Dr. Bala, who is a radiologist at WashU in St. Louis and a board member of the RSNA, reviewing the exam. And then we dictate a report. And you can see here the resident dictating with Dr. Bala. We dictate a report that then gets sent to the referring physician. It used to actually be a literal letter. Now it's obviously electronically done, and they get it immediately. But we send an answer to them. So it begins with a question. Is the ankle broken? And it ends with an answer. No, it's not. It's swollen. It's not broken. And it all circles around what's best for the patient. Now where do we practice? Now I showed you, we used to have to practice in the hospital, right? Because we were developing film. There were literal films, like develop the film, hang up the film, look at the film. It was all in the hospital, and it all had to be there. But it's all digital now. So it doesn't matter if you're in the room down the hall from where the exam was taken, or across the country. Because I can transmit those digital images anywhere at the speed of light. So I don't care where you're sitting. So a lot of it's still practiced in the hospital. But you don't have to be in the hospital anymore. In fact, a lot of our imaging gets done in imaging centers, like in strut malls anywhere. I'm sure some of you had your imaging done outside the hospital, probably in an imaging center somewhere. And by the way, there may or may not have been a radiologist sitting in the imaging center, because we can sit anywhere. And in fact, many of us actually work from home. You can work wherever you want. You can work in a home office. You can work in another office. You can work wherever you want. But that's one of the things people like about radiology is, honestly, you can work anywhere. Now, for myself, on a typical day, I actually go in. I work on site at the University of Chicago Children's Hospital. And on a typical day, like tomorrow, I will be there reading plain films. I'll read a lot of plain films. I'll read a lot of ultrasounds. And I'll read a sprinkling of CT, MR, fluoroscopy. Not so much the nuclear medicine stuff. We've got people who do that. But I'll read different body parts, too. It won't all be chest. It'll be chest, and abdomen, and arms, and legs. It'll be a huge variety of things. And because, remember I mentioned we're the doctor's doctor, I'll be speaking to a whole spectrum of doctors tomorrow. I may very well be on the phone with the emergency department talking about a patient of theirs who got a CAT scan for appendicitis, but I have a note to please call back the pediatric surgeon, or to call back the cardiologist or the orthopedic surgeon, because they want to talk to a different patient. And when I say that radiology is the nexus, that's what I mean, that we're sort of that connector, that we have to be able to speak the language of the cardiologist, of the orthopedic surgeon, of all these different areas in medicine. The radiologist has to be able to have the communication with all of them, because we look at all of their imaging examinations. Interventional radiology. They're often doing procedures. So you can see them here. It looks like a surgery room, but it's, again, an angiography room. And in this case, what they've done is they've threaded a catheter, a small tube, into the blood vessels of the brain. And they're injecting material that's just one x-ray into the blood vessels of the brain. And that's what you're seeing on the screen. And they can actually see the vessels in the brain, and they can actually do things that they need to fix something of the blood vessels of the brain. So here's an actual case study. So here's a real case that I took from one of my interventional colleagues. It's a 64-year-old who had intermittent lower GI bleeding. So they were bleeding from below. And they knew they were probably bleeding from their colon. And it was coming and going, but they weren't sure what was happening. So the first thing they wanted to do was do a colonoscopy, because if they can see the bleeding right there from the colon, then maybe they can fix it. Well, colonoscopy, as you know, you've got to prep them. It's a hard thing to do. It's not as easy. So that failed. They couldn't do a colonoscopy. So remember I mentioned a tagged red blood cell scan? That's what they did. They did a tagged red cell scan. They took the patient's red cells, put a radioisotope in it, injected it back. OK, they're bleeding. They were able to see that blood collecting in the right lower quadrant. They knew they were actively bleeding. What are we going to do about it? So they called interventional radiology. So this is the fluoroscopy image. And what you can see is they've taken a catheter, a small tube, and they've injected it. They've gone in through the right groin, and they've threaded it up into the aorta, that big blood vessel. And they're then selected, gotten into the blood vessel that feeds the intestines. And they're injecting contrast, which shows up dark here, into the blood vessels, which feeds the intestines. And if you look really carefully on your left side of the screen, you can actually see a little blush of contrast, a little blush of contrast. That's active bleeding. That's the contrast leaking out of the blood vessel actively. So we can now say, OK, yes, the patient's bleeding. They're bleeding actively right now as we speak. And I can tell you exactly what vessel it's coming through. Well, that's pretty cool. Let's go one step further. What they did was, next, they took a small catheter, which is well-named a microcatheter, and they threaded it out there further. And they got it into that blood vessel that was leaking. And they injected contrast again. And now you can really clearly see where there's leaking. Now you can see that amorphous blush. That's active bleeding. That's the contrast actively leaking out of the vessels. And so what can we do about that? Well, what they did was, they put these thin metal coils through that microcatheter that goes into the lumen in the inside of the blood vessel. And it clots off the blood vessel. So the bleeding stops. It's like building a dam. They put the coils in. Blood flow stops. The bleeding stops. So what they did was, they diagnosed the bleeding. And then they treated the bleeding. With no surgery, all they did was go through a little nick in the groin, thread a catheter up there. And they were able to diagnose the patient and treat the patient and stop the bleeding. It's incredible what interventional radiologists can do. So let's pivot and talk about AI for a minute. Because I hear it's kind of a thing. Has anybody here heard about maybe AI and radiology being a thing? Oh my gosh. So let's talk about AI and radiology for a moment. So this is Geoffrey Hinton, who is considered, and this is the term that gets used, one of the godfathers of AI. If you Google him, that's the term you're going to see. Godfather of AI. And he said this in 2016. I think if you work as a radiologist, you're like the coyote that's already over the edge of the cliff, but hasn't yet looked down. By the way, he's also dating himself. And I get it too, because it's a reference to the Bugs Bunny cartoons. And I don't know if kids know Bugs Bunny, but I get the reference. People should stop training radiologists now. It's just completely obvious, within five years, deep learning is going to do better than radiologists. He said that in 2016. Now, I'm willing to believe that Professor Hinton knows a lot about AI technology. He literally is one of the godfathers of AI technology. But I think this is a good example of him not maybe fully understanding radiology and what's going on. Because not only was he wrong, and through his credit, he has walked this back, but he was profoundly wrong. So this was in 2016. Thank goodness we did not listen to him and stop training radiologists. So I looked at the job board last week for radiology. It's almost 2,000 jobs. Now, admittedly, this is not all radiologists. It includes things like medical physicists and other things, too. But it's mostly radiologists. Now, a couple comments. One, we train. We put out about 1,000 radiologists. A little over 1,000 radiologists per year come out of training. Not all the jobs that are available are posted. I know that because we don't post all the jobs on this site. So the site is not complete, which means that if nobody died, if nobody retired, if nobody left the practice of radiology, it would take more than a full cycle of trainees coming out just to fill the demand we have right now. And that demand is growing up, is increasing. And by the way, people are leaving. Things happen. People retire. So it's not static, and we don't have enough physicians. So thank goodness we didn't listen to him and stop training. In fact, I had the honor of representing RSNA at the Korean Congress of Radiology this past fall. And in talks given by the president of the Korean Society and the Royal College of Radiology in the UK, the lack of radiologists was a main theme. This was highlighted by the president of KSR in her address. And then the head of the Royal College said in the UK, they have a 30% shortage of radiologists, and it's getting worse. So the problems we're having right now in the US with the shortage of radiologists, they're not unique to us. So Hinton wasn't just wrong. He was profoundly wrong. But let's talk about AI and radiology, because it is having an impact. And people usually say, is going to a future tense? I want to show you some things that are being done right now, today, in practice using AI. This is probably not what Hinton meant. Hinton was thinking about AI reading the exams. This is an example that my practice uses. Today, using AI technology, it's helping us. It's helping us with our radiology report. So this is a CAT scan of a patient. And see that big white thing in the middle? That's the aorta I keep talking about. That's the big blood vessel. And if it looks too big to you, yeah, it is. It should be two centimeters or less, and it's a lot bigger than that. It's 4.8 centimeters. That's an aneurysm. The risk of an aneurysm, as you would imagine, is that if it ruptures, that's really bad. It's got a very high mortality rate. So that's a big deal. So when you're dictating it, you want to make sure the patient gets the right follow-up. So let's say I'm dictating this, and I go, there's a 4.8-centimeter abdominal aortic aneurysm. We saw that screen pop up. That's a software that we developed with natural injury processing that uses AI. And it's reading along with the radiologist. And it knows all the different ways to say aneurysm. I could have just said, the abdominal aorta measures 4.8 centimeters. I don't have to say aneurysm. It knows it's an aneurysm. And the screen pops up and goes, hey, Dr. Heller, that's an aneurysm. Do you want to put into your report that's going to be sent to the referring physician what that referring physician should do with that patient as a next step? Well, that would be really helpful, yeah. Click of a button. It inserts it in there. There is evidence-based guidance on what should happen to a patient that has an abdominal aortic aneurysm this size. I don't have to memorize it. I don't have to know that there is one. The computer does all that for me. And then it inserts it with the click of a button into the report. So I'm telling the referring physician what the evidence-based guidance is for a patient with a AAA, an abdominal aortic aneurysm, that's this size. That's a meaningful way to help patients using AI technology. It's probably not what Hinton meant then. Hinton was thinking about stuff like this, AI visualization. So here's another patient. Came in with left-sided chest pain, no history of trauma. Their concern was a heart attack or a pulmonary embolism. A pulmonary embolism occurs when you've got a blood clot in the veins of your leg that breaks off, then travels through the veins in your abdomen and your chest, and then gets caught in the blood vessels of your lungs, and it's an emergency. How do you diagnose it? CAT scan. So the patient got a CAT scan. Good news, didn't have a pulmonary embolism. But there was an AI algorithm running that looks for rib fractures. And the AI goes, hey, look over here. There's a rib fracture. Well, that's important. Why? Because now we can explain the patient's chest pain. So you go to the patient. Hey, you said you had no history of trauma. Yeah, that's right. Does it hurt here? Ow! Yeah, it does. Are you sure nothing happened? You know what? Now that you mention it. And we don't have to work them up for all this other stuff. We can say, OK, we can explain your chest pain. We know why your chest is hurting. You fell, and you've got a rib fracture. It's a great way to help patients. Here's another one. This is a patient who came into an outpatient imaging center for staging of their endometrial cancer. And there was an AI algorithm running the background that looks for pulmonary embolism, that blood clot in the lungs. And it found it. The white is the contrast in the blood vessels. And you see those dark areas within the white? That's the actual thrombus. It's the actual clot of blood that broke off from the legs. And you can see it sitting there. Would the radiologist have missed this? Unlikely. This is pretty big pulmonary embolism. This isn't going to get missed. But this wasn't being done in the emergency department. This was being done at an outpatient imaging center. This is an emergency. So this patient would go, get their exam, and then go off and run errands, go home, whatever they're going to go do. But the exam might not get read for hours or days, maybe even longer. Remember, there's a workforce shortage. So it's going to get read. But it might not get read immediately. With the AI algorithm running, the AI looks at it immediately and goes, hey, look, over here, I think there's pulmonary embolism. Flags it for the radiologist. The radiologist looks at it and goes, that's pulmonary embolism. Don't let the patient leave. Keep them in the imaging center. They're going to send them straight to the emergency department. This is lifesaving. It's not that the radiologist was going to miss this. But we won't see it fast enough. This is a way to actually save lives. But it's not perfect. This is an example. The AI goes, hey, look over here. There's pulmonary embolism. And the radiologist goes and looks and goes, yeah, no, I see what you're saying. But you're wrong. You're getting confused by a branch vessel. It looks like a PE, but it's not. AI is not perfect. We need to have a radiologist. I'll show you another reason why we need a radiologist. Imagine you've got an AI algorithm that's really good for shoulder fractures and dislocations, a common cause of shoulder pain. People come into the emergency department all the time, shoulder pain. It'd be great to know if they had a fracture or a dislocation. So you run the AI algorithm. And you look at it. It says negative. No fracture, no dislocation. And you can imagine a chance where they didn't have the radiologist right there. And they think, you know, it's fine. I looked at it. I don't see a fracture. The computer doesn't see one. And you imagine somebody looking at it and being like, OK, and then going to talk to the patient and being like, well, good news, bad news. Good news, you don't have a fracture. The bad news, you've got some arthritis here. I'm going to give you some Advil. But if I'm being honest, look, you're not 18 anymore. It's probably going to come back. The AI algorithm is not wrong. There's not a fracture. There's not a dislocation. And that person who looked at it isn't wrong. There's not a fracture. There's not a dislocation. But they missed the lung cancer. That's what's causing the patient's pain. Their shoulder pain is caused by the lung cancer. And the AI was not looking for that. It was looking for fracture or dislocation, not lung cancer. So the headline was the lung cancer. And that's what it missed. Give you another example. An AI algorithm looking for fractures. It puts arrows on it. Over here, look, there's a fracture. You're absolutely right. There's a fracture. That's not the headline. The headline is this fracture is caused by cancer. That's the headline, the cancer. And that's what the AI might not pick up. That's why you want to make sure you have radiologists in the loop. So you sometimes see headlines like this from this past spring, will AI replace doctors who read x-rays or just make them better than ever? The best answer to this question I've ever heard came from Dr. Kurt Langlotz, the current president of RS&A, a radiologist and a world's authority on AI. AI won't replace radiologists. But radiologists who use AI will replace radiologists who don't. Now, I give talks to medical students. And I talk to them about why I chose radiology and why I think it's so exciting. And my theme for them is, whatever you're interested in as a medical student, there's a home for you in radiology. And I'll show you what I mean. My first question for them, how important is it to you to have direct patient interaction? Do you need to have that direct patient interaction? And if they're like, yeah, that's very important to me. Can I get it in radiology? The answer is yes. In areas like pediatric radiology, women's imaging, interventional radiology, they're very interactive directly with patients. But what if they're like, absolutely not. I don't ever want to see a patient. I want to help them. I'll become a doctor. I just don't actually want to see them or interact with them. You can work from home and literally never see one of the patients you're helping. You can have it either way. If you tell me you want to work with children, pediatric radiology, you're interested in neurology stuff and you're thinking about neurology or neurosurgery, great specialties. There's also neuroradiology. I want to do procedures. I'm thinking about surgery. Interventional radiology. And you can mix and match. I like procedures, but I like working with kids too. Great. Pediatric interventional radiology. You can mix and match. So I've mentioned that I'm a pediatric radiologist, as is my father. You can see him here. Dad had his career at Vanderbilt in academics. And I went into private practice. You can practice in lots of different ways. You can practice in the big university settings, like my dad did, but you can also just go work at a suburban hospital, or you can work for an imaging center. You can be somebody who does all clinical, somebody who does all research, or you can do part clinical, part research. There really is that home for everybody within radiology. Now, a few words about radiology societies. I think we're all familiar with this society. So the RSNA focuses on research, science, education. It's really the home for science and education for radiology. We have other big societies too. For example, the American College of Radiology, which really focuses on policy and practice. And the Renkin-Ray Society, a very old society, again, which does a great job in education. And those are sort of broad societies. But then we each have our sub-specialty societies. An example, I'm a pediatric radiologist. So I belong to the Society for Pediatric Radiology. IR has theirs, and Neuro, and so on. And what I encourage my radiology colleagues to do is to belong to more than one society, that RSNA is your home for science and education. ACR is your home for health policy. And then I encourage people to belong to their sub-specialty. So I belong to RSNA, ACR, SPR, and I'm active in all the societies. I probably need to mention the RSNA annual meeting, because it's a big deal. I understand that not all of you have been to the RSNA annual meeting. It's tough for me to emphasize how big a deal it is in our community. It's such a big deal in radiology that it's not referred to as the RSNA annual meeting. It's literally referred to as, am I going to see you at RSNA? We all know that RSNA does a million things besides an annual meeting. But in slang, when radiologists are speaking to one another, the annual meeting is such a big deal, you can just say, am I going to see you at RSNA? Or, am I going to see you in Chicago? Because the meeting always takes place around Thanksgiving in Chicago. I don't have to say RSNA. I can just say, I'll see you in Chicago, right? And we all know what we mean, because tens of thousands of people in radiology every year from around the world descend upon McCormick Place in Chicago every year. It's that big a deal. And I always tell radiologists, at some point in your career, you have to go to the annual meeting. It's that big a deal. Now, here's my trivia question for you. Who has the longest continuous streak of on-site attendance of RSNA annual meetings? The answer is me. And here's why. There are those that are senior to me that have been going to meetings longer. But in 2020, the whole meeting went virtual. Which, by the way, was a huge credit to anyone who was there, like it's incredible to move the whole meeting and go virtual, that was fantastic. I can't believe we were able to do that. One person went in person and showed up in McCormick. That's right. I brought a folding chair and a card table and my laptop and freaked out the security. But I set up shop outside McCormick Place to prove that I went on site. So I now have the record of the continuous on-site attendance at RSNA annual meetings. So who is this guy? Wilhelm Roentgen, the discoverer of x-rays. He founded an entire specialty of medicine, and I'm a proud descendant of that. The things that we can do now, he would have never imagined. This is a CT scan where I've taken away everything but the blood vessels and the bones. This would have amazed him, the things that we can do with his discovery, with what he founded. So this is Wilhelm Roentgen. It's not a Brooklyn hipster. It's not a dude from ZZ Top. It's not even Mark Watson from the 1980s. Yes, I do have the picture. I do have the picture. But because of Mark Watson and the RSNA, because of the work that you all do, radiology has never been stronger, more important for patient care. So on behalf of the global community of radiologists, I want to say thank you to all of you for the tremendous work you do. Thank you.

radiology

X-rays

Wilhelm Roentgen

AI in radiology

radiologist training

medical imaging

radiology societies

RSNA