Is Artificial Intelligence Actually Taking Over Radiology? let’s find out

Is Artificial Intelligence Actually Taking Over Radiology?

You have to know what a radiologist actually does before you can say that radiology is going to be taken over by some artificial intelligence, robot or computer. So I wanted to go over a few things that radiologists actually do and what makes us important.

First and foremost our main job is to read studies, we read CTS, MRIs, chest x-rays etc etc. Our second job and arguably our most important job is to answer phone calls throughout the day from other clinicians and providers, and give them answers to questions they may have about their patients. We oftentimes will walk a provider through a study over the phone or in person and this actually helps providers tremendously, because it kind of gives them the lay of the land of what’s going on with their patient and it can also help them decide certain management approaches to treating their patient and also just as importantly it can affect surgical management and surgical planning prior to a surgery.

Another thing that really I’ll just do is we customize studies for individual patients in individual scenarios. For example some studies may need contrast, some studies may not need contrast, some patients may have an allergy of contrast, in addition to that a lot of times some of these studies can be very detailed and specifically designed for a certain patient. For instance I’ll give you an example of a cardiac MRI which is completely designed around a certain patient, it has to do with like contrast timing and heart rate control etc.

Another responsibility of a radiologist is to attend multidisciplinary conferences, so an MDC conference or a multidisciplinary conference is where a ton of different commissions get together and discuss one patient. So you’ll have an oncologist, an internist, a hepatologist, an interventional radiologist, a radiologist, an oncological surgeon, they all come together and talk about a certain patient and try to come up with a treatment plan specifically designed for that patient. so in this scenario, we will discuss the imaging as it pertains to the patient and offer our guidance as well.

Another thing we do that I know you guys have heard me say a million does that we do a ton of procedures throughout the day, so at any given point when we’re in the middle of our studies we will be taken out of our chair and we will go do a biopsy or a small procedure. now I’m doing interventional radiology so a majority of my time will be spent doing procedures however almost all radiologists do procedures in the middle of their day. and just out of curiosity which one of those affirmation themes do you think would be replaced by a computer? I don’t think any of those things to be replaced by a computer.

Also we need to keep in mind that AI has been incorporated in radiology for the better part of a decade, a lot of people don’t realize this we have been using a program called CAD, C ad and it stands for computer aided detection and what that does is it helps radiologists identify even the smallest of breast cancers or calcifications that could lead to breast cancers. Not all mammographer Zoar radiologists actually use it the reason being because it’s not perfect, you can use it as a supplement but it would never replace a radiologist or a mammographer reading a mammogram and I’m willing to bet you wouldn’t want it to replace a radiologist reading a mammogram either.

I hope that makes sense too there are also other models or artificial intelligent models that will help us identify pulmonary nodules on chest x-rays and acute hemorrhage on a CT head, now what these programs do just that they aren’t entirely accurate all the time, and sometimes they can even be too accurate and what I mean by that is sometimes for instance on a CT head without contrast, a computer programmer will recognize a miniscule acute hemorrhagic stroke or a possible hemorrhagic stroke on a CT head, which to a normal radiologist we would not see because it’s so small which has its benefits because it will be able to identify things that we miss, however a lot of times those tiny hemorrhagic fossa aren’t clinically relevant and even if we find them you don’t do anything for them anyways and since they may be of little clinical significance or clinical relevance it may lead to more harm than being able to identify them, and what I mean by that is we end up having to follow them and follow them and re image and re image and I think that may lead to increased healthcare costs in the long run, so there are pluses and minuses to being able to identify everything.

So what do I think about AI? that’s the age-old question you guys keep asking me that all the time so what do I think of AI. I think our field is becoming more and more demanding, our scans are increasing exponentially over the past few years our stand time has decreased, so we can crank out more scans faster and we’re reaching the point where we almost can’t keep up, as a radiologist you can only read so fast and still be accurate.

So I think AI has the potential to facilitate our reading and has the ability to make us more proficient, which like every radiologist in every physician efficiency is key, in addition to that I think it would allow radiologists to focus on more complex studies, like difficult MRIs or difficult rectal cancer staging etc. rather than sifting through a list of mundane ICU chest radiographs which take a ton of time and they aren’t that valuable.


so in conclusion please do not let AI steer you. I think as radiologists, we should embrace it, because it can only help us. It is here to stay it will only get better and it is here to help us, not replace us radiologists are way more valuable than you may think.

Share this post ...

Leave a Reply

Your email address will not be published. Required fields are marked *