What does the radiologist of the future look like?
I believe the radiologist of the future will work with machines. But how does one get to be that radiologist?
In the era of social media, simply hitting the submit button on a tweet or a post can make it go viral or trend in a matter of seconds. Recently, I came across a tweet about a new course in AI and deep learning from the Massachusetts Institute of Technology. Embedded in the tweet was an image of a chest X-ray on the left and an activation map on the right.
Alexandre Cadrin-Chenevert, MD, a radiologist and computer engineer from Quebec, Canada, immediately pointed out a flaw in the activation map — that the pneumothorax heat maps were overlapping both diaphragms instead of the lung apices, which could be a result of overfitting the model.
Five years ago, when I graduated with a master’s degree in health informatics, I knew nothing of deep learning. Now, during my radiology residency, I am learning new technologies and techniques that are poised to disrupt the current workflow of radiology as we know it. I cannot think of a more exciting time to be practicing radiology. The future belongs to the Cadrins of today, those who understand how the machine works, what its abilities and limitations are, and how it will help us take care of our patients and keep us true to the calling of medicine.
Judy W. Gichoya, MBChB, MS, is an informatician and radiology resident at the Indiana University School of Medicine.