Elad Walach: Using AI in medical imaging is transforming patient care
How a devotion to accuracy helped Aidoc create a foundation model from the ground up
What kind of impact does aiOS have on patient care?
Imaging backlogs across the U.S. mean it can take weeks to get results. If something serious is hiding in the data, you shouldn’t have to wait. For example, around 2% of cancer patients have pulmonary embolisms that need to be flagged immediately. In rural emergency rooms, it can take two hours to get scan results. If you have a brain bleed, two hours is too long. With strokes, every 60 seconds of delay can cost a patient nearly two million brain cells.
Instead of reading scans in the order they arrive, Aidoc flags critical findings — brain bleeds, pulmonary emboli, air in the chest and abdomen — almost instantly. AI isn’t making the diagnosis; it’s giving the doctors the information they need to look at a case, confirm what’s happening, and act quickly. It’s a whole new way of working. You’re no longer waiting on an overstretched radiologist to get to your scan.
What’s surprised you most about Aidoc’s journey?
How bad healthcare’s last mile problem is — in other words, the gap between jaw-dropping demos and real-world deployment. It’s 10 times worse than most industries. Interoperability is difficult; safety constraints are very demanding. So we’ve had to think carefully about how to implement AI in ways that clinicians actually want to use. That’s why our platform is vendor-neutral, customizable to our customers’ existing IT stacks, and designed to run in any clinical setting while orchestrating numerous AI applications at once.
The mistake many innovators make is assuming they’re bringing fully-formed solutions into the world. We can bring a lot of the technology, but need leaders and evangelists inside health systems — whether that’s a CEO or innovative radiologist — who genuinely want to integrate these tools into their workflows and transform how care is delivered. If you don’t have partners to work with on implementation, you’re just going to get fancy technology on a broken workflow.
Today, our platform is used by cardiologists, neurosurgeons, and ER physicians, not just radiologists. Radiology is still in the center, but this is a health system transformation.
.jpg)