The Reading Room with Diagnostic Imaging
Sep 8, 2026 · 16 min · 5 segments
In the first of a two-part podcast episode, Lauren Nicola, MD, Christoph Wald, MD, PhD, MBA, and Peter Shen discuss current challenges with reimbursement for AI in radiology.
Lauren NicolaGuest
Christoph WaldGuest
Jeff HallHost
Peter ShenGuest
So I think probably for the purpose of this conversation, it helps to take one step back and explain the different ways that Medicare pays for things.

There's another one called the hospital outpatient prospective payment system, which is how Medicare pays hospitals and facilities.

So you could have codes that get paid in that system that don't get paid with physician fee schedule category one CPT codes.

And then there's a third called the inpatient prospective payments system, which is how Medicare pays for inpatients.


And there are examples in each one of those that CMS pays for different types of healthcare services.

Particularly in the physician fee schedule, there are very few category one CPT codes for AI technology.

But when we talk specifically about the position fee schedule, which is how doctors get paid for the work, the position work they do, and they get paid for the technical component in their offices.

I think the main reason for that is probably that a lot of the AI algorithms we have now are ways that we can detect findings or triage findings so that it duplicates the payments that already exist in the CPT system.

So if you have an algorithm that pays for finding, or I'm sorry, that finds lung nodules, well, CMS already pays us to find lung nodules in the code that already exists for a chest CT, right? So they're not going to make a new CPT code that's going to give us more money for using a software that's going to help us find the lung nodules.

Even if it does it faster or better, they generally do not like duplication of work.

So specifically for those detection and triage algorithms, the CPT system in general doesn't work well for those types of things, for that duplication of work.

So I think probably for the purpose of this conversation, it helps to take one step back and explain the different ways that Medicare pays for things.

There's another one called the hospital outpatient prospective payment system, which is how Medicare pays hospitals and facilities.

So you could have codes that get paid in that system that don't get paid with physician fee schedule category one CPT codes.

And then there's a third called the inpatient prospective payments system, which is how Medicare pays for inpatients.


And there are examples in each one of those that CMS pays for different types of healthcare services.

Particularly in the physician fee schedule, there are very few category one CPT codes for AI technology.

But when we talk specifically about the position fee schedule, which is how doctors get paid for the work, the position work they do, and they get paid for the technical component in their offices.

I think the main reason for that is probably that a lot of the AI algorithms we have now are ways that we can detect findings or triage findings so that it duplicates the payments that already exist in the CPT system.

So if you have an algorithm that pays for finding, or I'm sorry, that finds lung nodules, well, CMS already pays us to find lung nodules in the code that already exists for a chest CT, right? So they're not going to make a new CPT code that's going to give us more money for using a software that's going to help us find the lung nodules.

Even if it does it faster or better, they generally do not like duplication of work.

So specifically for those detection and triage algorithms, the CPT system in general doesn't work well for those types of things, for that duplication of work.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.