_{Published October 25, 2021}

Back in February, American Roentgen Ray Society (ARRS) representatives made a call. On the other end was the American Medical Association (AMA). They weren’t alone. Representatives from three more of this country’s biggest imaging stakeholders were also on the line.

This meeting was the first of three virtual convenings for the AMA’s Current Procedure Terminology (CPT) Editorial Panel, the assembly ensuring “that CPT codes reflect the latest medical care available to patients.” Top line of the agenda for ARRS and the American College of Radiology (ACR), Radiological Society of North America (RSNA), and Association of University Radiologists (AUR) was updating AMA CPT category codes for the 2023 cycle.

It was hardly a modest proposal. ARRS et al. were lobbying AMA to approve not one, but two new CPT codes specifically for medical imaging. One of the code requests, for quantitative ultrasound tissue characterization, seemed innocuous enough, especially compared to the other glaring one. That second request, for incidental vertebral fracture detection, had two small, historic words preceding it: “automated analysis.”

Apropos of North America’s first radiological society, indeed, ARRS was seeking radiology’s first-ever CPT code for artificial intelligence (AI).

Artificial life comes at you fast, too. Should AMA’s CPT panel ultimately approve these two proposals (in the middle of a pandemic, no less), the concomitant codes would be released first-thing on July 1. Their live, in-effect date would start January 1, 2022, thus allowing clinics to start the billing.

Regarding reimbursement, the prevailing ideology still stood. If an AI code were ever granted AMA’s blessing, it would then need to stand before their Specialty Society RVS Update Committee (RUC), that group “dedicated to describing the resources required to provide physician services which the Centers for Medicare & Medicaid Services (CMS) considers in developing Relative Value Units (RVUs).” Whereas the RUC in Chicago is free to advise, of course, it’s the CMS in Washington that would have the final financial consent. Even ACR’s own journal was asking: “Who Pays and How?” Literally.

By now, you know what happened on the first of July. 55 years after publishing its inaugural CPT code, AMA had heard the plea of ARRS, ACR, RSNA, and AUR, found it good, and the specialty’s first Category III CPT codes for AI were born. Three Category III codes, in fact:

• 0689T—Quantitative ultrasound tissue characterization (nonelastographic), including interpretation and report, obtained without diagnostic ultrasound examination of the same anatomy (e.g., organ, gland, tissue, target structure) 
• 0690T—Quantitative ultrasound tissue characterization (nonelastographic), including interpretation and report, obtained with diagnostic ultrasound examination of the same anatomy (e.g., organ, gland, tissue, target structure) (List separately in addition to code for primary procedure)
• 0691T—Automated analysis of an existing computed tomography study for vertebral fracture(s), including assessment of bone density when performed, data preparation, interpretation, and report

Naturally, 0691T hogged the headlines, but imaging professionals paying closer attention were quick to point out the utility that this pair of CPT codes promised for quantitative ultrasound tissue characterization. Able to be used together with or separately from existing ultrasound examination codes (0690T in conjunction; 0689T on its own), they represented a giant leap forward for mainstay ultrasound examinations (e.g., breast and thyroid). Perhaps most importantly, as the Imaging Wire duly noted, they also bode well for ultrasound AI decision supporters, like Koios Medical.

Meanwhile, many radiologists could see that 0691T for “automated analysis” of existing chest CTs to detect vertebral fractures could pave all the way for AI-based population health osteoporosis programs.

“I believe that this type of ‘population’ health application will lead to requests for many additional CPT codes such as coronary artery calcification on conventional thoracic CT, abdominal aneurysm detection on thoracic, abdominal and pelvic CT scans, evaluation of cardiac chamber enlargement, [chronic obstructive pulmonary disease (COPD)], renal calculi, and many others,” Eliot Siegel of the University of Maryland and Baltimore VA Center told Aunt Minnie.

This single CPT code represented quite a milestone for Israeli AI developer Zebra Medical Vision, which garnered FDA 510(k) clearance back in May 2020 for its software that can detect incidental vertebral compression fractures on chest CT scans.

Right now, Zebra is the only company with an FDA-cleared product for population-scale detection. However, on August 10, 2021, Nanox—Israel’s publicly traded x-ray disruptor—signed an agreement to purchase Zebra for up to $200 million in Nanox shares, half based on future milestones. Hours later, after posting a net loss of $13.6 million (compared to a net loss of $6.4 million for the second quarter last year), Nanox’s founder, Ran Poliakine, announced he will relinquish his role as CEO in January 2022. This Zebra-to-Nanox story, the most momentous acquisition in AI imaging’s brief history, is still developing.

It’s important to point out that the three CPT codes above are not Category I CPT codes. Category III CPT codes are wholly provisional, intended to aid data collection for developing technologies, incipient procedures, and service paradigms. Their endgame is to accumulate enough clinical documentation for the FDA to validate and clear for general application. Often dubbed experimental or, at best, tentative by most major insurers, reliable reimbursement remains unlikely until AMA sees fit to assign a permanent Category I CPT code.

Despite that lack of instant reimbursement, it’s still worth submitting claims for CPT III codes; the data from their utilization become primary sources of support for the eventual creation of a CPT I code. Likewise, many advocates advise referencing similar Category I procedures when reporting Category III CPT codes.

Most experts agree that a standard CPT III code should lead to a permanent CPT I code within five years. So, come 2026, radiologists should expect to see the very first Category I CPT code regarding the automated analysis of an existing CT study for a vertebral fracture—brought to you, in part, by your medical imaging society, ARRS.