This series highlights the insights and learnings from medical residents engaged in the use of artificial intelligence for their medical practice
AI continues to make inroads into medicine, providing solutions that support medical practitioners in managing complex workloads and busy daily lives. For Dr. Eric Gray, a resident at UCSD San Diego, artificial intelligence is a useful tool that helps to cement decisions and offer insights that only serve to enhance the work that he does.
AIDOC: Can you tell us a bit more about yourself?
Dr. Eric Gray: I am from the Central Coast of California and have managed to work my way down the coastline to UC San Diego, which has an excellent reputation for being at the forefront of clinical innovation and research. I am in my second year of Radiology residency and I have another two years of residency left, with my fellowship following after that. I have not decided on my future fellowship plans, but I am currently leaning towards neuro or body as both offer a wide breadth of pathology.
AIDOC: Did you use AI or solutions like Aidoc before you started work here?
Dr. Eric Gray: I had never used an AI solution before, but I had been interested in the radiology-specific application of machine learning. Once I had access to the Aidoc solution at UCSD, however, I found it to be a very intuitive program that you can pretty much learn how to use right away. The Aidoc application stays in the background with very inconspicuous notifications and it was very easy to integrate into my normal workflow. When I come in on a shift, I’ll take a look at the notifications – as there may be a number of cases to dictate – and I’ll first look at the Aidoc flagged cases to do a quick triage of which cases to dictate first. Then as the day goes on, I will check when a notification goes off. It only takes a second and you can quickly tell if that case requires prompt attention.
AIDOC: In what ways do you feel Aidoc supports your work as a radiologist?
Dr. Eric Gray: Currently, I feel that Aidoc functions as a useful triage system. Every day is usually busy, and there are always distractions whether it be phone calls, consulting physicians, or just other cases that you are working on. For us residents, this is especially true at night when the workload is higher and we are fielding calls and reading studies from multiple locations. Having a system that can efficiently notify you of an emergent finding that requires your immediate attention is incredibly helpful.
It is also reassuring that the Aidoc algorithm is reviewing every study and that both you and Aidoc agree that a particular emergent finding is not present. The more I’ve been able to use Aidoc over the past several months, the more confident I’ve become using it. It’s exciting to see the new algorithms being introduced. We currently have the intracranial hemorrhage solution active at the moment and the pulmonary embolism solution will hopefully soon be deployed. I’m hoping we will continue to add the additional Aidoc solutions as they become available.
AIDOC: How do you believe Aidoc will make an impact over the next ten years?
Dr. Eric Gray: I think Aidoc will eventually be so powerful that it will be able to create a preliminary report that includes all of the pertinent findings. Aidoc may become like a super resident that will be able to quickly and reliably evaluate every study. The radiologist will then make edits to this preliminary report and provide additional clinical context as needed. In the future, I think Aidoc will allow you to get through studies faster and be confident that you did not miss an important diagnosis. I think Aidoc will be especially valuable for junior radiologists, but in the end it will be a valuable asset for every radiologist.
AIDOC: Have you used Aidoc for peer review or training?
Dr. Eric Gray: I have thought about it and I am interested in the idea of having Aidoc retroactively evaluate cases to see if a particular diagnosis was ever missed. We haven’t yet established how this systematic peer review process would work, but I think this will be something we will implement in the future. I also think Aidoc could be valuable for training as it could compile examples of important diagnoses that trainees could then review.
AIDOC: Can you share your first-hand experience of working with Aidoc?
Dr. Eric Gray: As I mentioned earlier, Aidoc is a very intuitive program that you can pretty much learn how to use right away. Most radiologists will find it easy to use within a day, and the fact that it runs seamlessly in the background is great. It doesn’t cause unnecessary distractions. It’s like any simple app – like a sports score widget that you’re occasionally checking!
We had a case the other day that really showed the value of this solution and the potential of where it is going. A patient came in with a headache, they suffered from migraines but this headache was more severe than normal, and there had been no history of trauma. Aidoc picked up a little hyperdense focus that was in a location for a potential bleed, but it didn’t fit with the clinical scenario because there had been no trauma. There was definitely something there, though, so the patient was sent for a MRI and it turned out to be an arteriovenous malformation in the brain. The patient ended up having several. This really showcased how these programs function effectively as a triaging system, but also it is easy to see how these programs will continue to become more powerful and will eventually be able to provide a differential diagnosis. This was an interesting case and demonstrated how Aidoc could help expedite patient care.