The Economics and Strategic Deployment of AI in Radiology

By Chad Hudnall, senior writer, ACR Press

The uncertainty around AI and reimbursement is more relevant now than ever. There is a need to balance the interplay between maximizing revenue and minimizing risk in coding, billing and collections for both the technical and professional components of AI tools.

To that end, ACR is launching the Artificial Intelligence Economics Committee to help consolidate current resources and develop financial strategies for the development and deployment of radiologic AI tools. The committee will specifically address Medicare, Medicaid and private insurance coverage and reimbursement for AI.

“There is a large unmet need to not only consolidate long-standing AI infrastructure but also to include economic expertise,” says Frank J. Rybicki, MD, PhD, FACR, chair of the ACR AI Economics Committee. “We are at the precipice of AI tools and AI-related tools being considered by CMS and third-party payers for reimbursement because so many AI applications are radiology-based. That means radiology needs to be at the forefront of that discussion and leading that economic charge.”

Codes and Agency

FDA regulates some — but not all — AI-enabled products used in healthcare, and the agency plays an important role in ensuring the safe and effective use of medical devices (including AI tools and software) under its jurisdiction.

There is a gap to fill in terms of CPT® codes for imaging services connected to AI technology. FDA-approved Class I codes are necessary for appropriate reimbursement — and currently there are many Category III CPT codes for AI. As spelled out by the AMA, technologies and services with CPT Category III codes could transition to reimbursable CPT Category I codes if they satisfy certain criteria within five years.

While many emerging AI technologies and procedures are initially assigned Category III codes to track their use and facilitate data collection, they can eventually transition to Category I codes as they become more established and widely adopted.

Billing and Data

“It is all about appropriate compensation for physicians’ work and appropriate technical fees for services that are going to be provided,” Rybicki says. “The goal of this is to evaluate what we can and cannot bill for and then figure out how we’re going to bill for it. We will do this with some urgency, as ACR members are constantly bombarded by vendors trying to sell them AI.”

The new AI Economics Committee will have 12 members with opportunities for guests to participate and offer their expertise. All committee members will be experts in the various components of medical economics, including:

• Medicare and Medicaid payment schedules (MPFS and HOPPS)
• The AMA/Specialty Society RVS Update Committee
• CPT code development and implementation
• Informatics and the work of the ACR Data Science Institute® (DSI)
• Computer science and convolutional neural networks (a cornerstone of deep learning)

“ACR is at the forefront of navigating an intricate and rapidly evolving imaging AI landscape,” says Gregory N. Nicola, MD, FACR, chair of the ACR Commission on Economics. “The AI Economics Committee will enable us to consolidate resources across the College — bringing together a diverse range of experts to accelerate this important work.”

 

 

“We plan to have a large database of every AI product, but it’s unclear at this point how we will share the information from that database,” Rybicki says. Within the committee, there may be a smaller group to monitor FDA decisions around new AI software. That group would populate the database to inform the committee’s work.

The committee’s plan will encompass the following responsibilities and tasks:

• Provide ACR leadership with evaluations of products cleared by the FDA under the AI medical device clearance pathway.
• Monitor, cull and prioritize the literature that supports and debunks AI applications in radiology, with an emphasis on findings that support CPT codes and reimbursement benchmarks.
• Advise other ACR committees and members on future projects and literature (computer science, clinical trial, white papers, etc.) that best support new billing codes and optimize reimbursement for existing codes.
• Engage and interact with patient advocacy groups that may facilitate the committee’s goals.
• Review new CPT applications and meet with vendors who approach ACR about AI products.

The committee will also work closely with the DSI, which convenes imaging AI users and thought leaders through its annual DSI Summit and year-round events. “In addition, our ARCH-AI and ASSESS-AI programs help practices safely implement and monitor imaging AI,” says Christoph Wald, MD, PhD, MBA, FACR, vice chair of the ACR Board of Chancellors and chair of the ACR Commission on Informatics. “We are eager to inform AI economic strategy with real-world insights from previous landmark ACR initiatives.”

The new economics-based committee will draw on human resources from multiple commission stakeholders within the College, and the committee will report at each of the three yearly CPT meetings, Rybicki says: “This committee will be a sounding board for the truth.”