CPT 2026 introduced the first set of dedicated codes for services involving augmented intelligence, including codes for AI-assisted coronary plaque assessment and perivascular fat analysis derived from cardiac CT angiography. According to the AMA’s CPT 2026 release, the new codes are part of a broader effort to create payment infrastructure for AI-augmented diagnostic services that have been delivered as bundled or non-billable adjuncts. For cardiology and radiology practices using these technologies, the 2026 update creates a billable category for work that was previously unreimbursed, but only for practices that update their workflows.
What the new codes cover
The CPT 2026 AI-augmented codes are specific to defined services: AI-assisted analysis of cardiac CT angiography for coronary plaque burden, plaque morphology, and perivascular fat attenuation. They do not apply to general imaging interpretation or to AI-assisted detection in non-cardiac imaging (which has its own separate code paths or remains bundled). Practices using AI tools that are not specifically referenced in the CPT descriptor cannot bill the new codes; they continue to bill the underlying imaging code without an AI add-on. Reading the full descriptor before billing is essential.
FDA clearance and the technology requirement
The CPT descriptors generally require that the AI tool used is FDA-cleared for the specific clinical use, that the analysis is performed using the cleared software, and that the output is reviewed and interpreted by a qualified physician. The most common audit failure on new technology codes (across many specialties, not just AI) is billing for services performed with non-cleared or off-label software. Practices should keep a current record of the FDA clearance number for each AI tool in use, the version of the software, and the dates of clearance and any updates. That documentation lives with the imaging report.
The interpretation requirement
AI output is not the billable service. The billable service is the physician interpretation of the AI output, integrated with the rest of the imaging study. The audit-defensible report includes a specific section that addresses the AI-derived findings (plaque burden by territory, calcified versus non-calcified composition, perivascular fat metrics where applicable), the physician’s assessment of the AI output’s clinical accuracy, and any clinical recommendations that flow from it. A report that simply attaches AI output without physician synthesis is at risk on review.
Payer coverage in 2026
Payer coverage for the new AI codes is mixed in 2026, as is typical for any first-year code. Medicare Administrative Contractors are issuing or revising LCDs throughout the year, with most expecting full coverage decisions by Q3 or Q4 2026. Commercial payers are setting their own policies, and several have announced they will not cover AI-augmented codes pending more outcomes data. Practices should not assume coverage; they should run a payer-specific coverage check at the time of the first billed service, document the coverage decision, and update fee schedules as policies change.
Bundling rules and modifier 26 versus TC
Like other imaging codes, the AI-augmented codes have professional and technical components. Modifier 26 indicates the professional (interpretation) component, and modifier TC indicates the technical (acquisition and AI processing) component. Hospital-based radiology and cardiology practices typically bill the professional component while the facility bills the technical component. Independent imaging centers may bill both. Bundling rules with the underlying imaging study (cardiac CT angiography, CPT 75571 through 75574) need to be checked against the National Correct Coding Initiative edits to avoid bundled-service denials.
Documentation that holds up at first audit
First-year codes attract review. The audit-defensible record for an AI-augmented service includes the imaging study itself, the AI software output (preserved in the imaging archive, not just summarized in the report), the physician interpretation report addressing the AI output specifically, the FDA clearance reference for the AI tool used, and the order or referral that justified the imaging. Practices that build this packet at the time of service spend zero time reconstructing it under audit later.
How MHB helps cardiology and radiology practices
For practices integrating AI-augmented services into their imaging workflow, our team supports specialty medical coding for cardiology and radiology, with FDA-cleared technology tracking, modifier 26/TC discipline, NCCI edit compliance, and payer-by-payer coverage monitoring. The work runs in the practice’s existing PACS and billing systems.
The bottom line
The 2026 AI-augmented CPT codes are the start of a long sequence. More codes will follow as AI tools clear FDA review and as the AMA Editorial Panel keeps pace. Practices that build clean documentation discipline now (FDA clearance tracking, physician interpretation, payer coverage monitoring) will scale with the codes. Practices that bill first and document later will have to defend years of claims when the audits arrive.
