If you want to understand why robots are taking over radiology, you should first understand the concept of “Deep Learning.”
So what is it?
Deep learning in radiology is a stepping stone towards Artificial Intelligence (AI) — it is a branch of machine learning based on a set of algorithms that attempt to replicate high-level abstractions of data.
How does it help radiology?
Deep Learning has found a stronghold in the radiology department. It is used in the interpretation of medical images to help countries like Scotland that are suffering from a shortage of radiologists. At a time when imaging orders are often increasing faster than radiologists are entering the field, deep learning is the answer. The market is currently standing around $40 million but will rise to $300 million by 2021, and while that may seem quite a ways out, 2020 is less than three years away.
When will we all be using it?
Deep learning is still in its early stages in medical imaging. There aren’t many products on the market and extensive research still needs refinement to discover the nuances for protocols, demographics, and the like. Since it is still in the early stages, skepticism is high among both radiologists and investors.
Many radiologists still have the memory of the underwhelming nature of the earlier-generation computer-aided radiology. This early technology relied heavily on feature engineering, and memories of it are a driving force behind today’s skepticism surrounding deep learning’s impact on radiology.
Before you get too concerned about robots taking over the radiology world, however, remember this. When deep learning is coupled with a radiologist’s experience, it can make a great addition to value-based service. Deep learning could potentially help cut down on re-admissions, repeat visits, or scans.
I will leave you with the knowledge that we don’t need to worry about deep learning becoming Skynet… yet.
This post was written by Mike Schwartz, a technical writer with Novarad.