For radiology AI, August 13, 2026 was a date worth circling. That is when CMS published the FY2027 Inpatient Prospective Payment System final rule and, buried in hundreds of pages of hospital payment policy, confirmed what the industry had been waiting years to see: a New Technology Add-on Payment for an AI-powered radiology triage tool.
The winner is Aidoc, whose CARE Multi-Triage CT Body (commercially known as BriefCase-Triage) will receive a maximum Medicare add-on payment of $137.53 per eligible inpatient case, effective October 1, 2026. But the number, while historic, is almost secondary to the strategic context in which it arrived. Because at the same moment CMS handed Aidoc its prize, it quietly shut the door on the shortcut most competitors have been counting on.
The alternative NTAP pathway, the fast lane that allowed FDA Breakthrough Device designation to substitute for hard clinical evidence of substantial improvement, is gone. Starting with FY2028 applications, every radiology AI company that wants Medicare add-on payment will need to do it the hard way.
That changes everything.
What Just Happened: The Numbers
The $137.53 figure is the ceiling, not the floor. Under standard NTAP mechanics, CMS pays up to 65 percent of the amount by which a qualifying case's actual costs exceed the standard DRG payment. So the real payment any individual hospital receives depends on case-level cost accounting, how well Aidoc's solution is coded on the claim, and whether the hospital's costs genuinely exceed the base DRG.
Still, even a partial payment is transformative for hospital budgets. Aidoc's eligibility runs for three years, which is the standard NTAP window. Hospitals that deploy CARE Multi-Triage before October 1, 2026 are immediately eligible to bill for it on qualifying inpatient cases. For a mid-sized academic medical center running 5,000 or more inpatient chest and abdomen CTs per year, the financial math is worth doing carefully.
The approval appears in the FY2027 IPPS final rule alongside a broader payment landscape that includes a 2.3 percent overall payment update for hospitals, a $2.1 billion aggregate increase in inpatient payments, and $779 million specifically allocated to new technologies. Nineteen new NTAP approvals begin in FY2027, and 41 previously approved technologies continue. Aidoc's is the one that will define the radiology AI conversation.
How NTAP Works: A Primer for Radiology AI Leaders
NTAP was created in 2001 to solve a specific problem: when CMS bundles hospital payments by diagnosis-related group, hospitals have no financial incentive to use genuinely new, expensive technologies, because the incremental cost is absorbed in the fixed DRG rate. NTAP provides a temporary supplement to bridge that gap.
To qualify through the traditional route, a technology must meet three tests. First, it must be new, meaning it is not substantially similar to technologies already factored into DRG rates. Second, the case costs associated with using it must be high enough that the standard DRG payment is inadequate, typically defined by a cost-threshold formula. Third, and most consequentially, the technology must demonstrate substantial clinical improvement over existing options.
CMS has interpreted "substantial clinical improvement" broadly but not casually. Evidence that has historically satisfied the standard includes reduced mortality, fewer hospitalizations or readmissions, shorter inpatient length of stay, or meaningfully faster time-to-treatment for acute conditions. The bar is real, and the evidence requirements are specified in the application package.
Because the evidence bar was genuinely high, Congress created the alternative pathway in 2020, allowing technologies with FDA Breakthrough Device Designation to bypass the substantial clinical improvement test. The logic was that the FDA had already certified clinical promise, so CMS would defer to that finding.
The result was a surge in applications. In FY2020, before the pathway existed, CMS received 18 NTAP applications total. For FY2027, the agency received 47, with 32 arriving through the alternative pathway and only 15 through the traditional route. CMS looked at those numbers and concluded the alternative pathway had become a workaround, not a supplement.
The Technology: What CARE Multi-Triage Actually Does
Aidoc's CARE Multi-Triage CT Body is not a single-condition AI tool. It is a foundation model, a single underlying architecture that simultaneously triages 14 acute conditions across chest, abdomen, and pelvis CT scans, including both contrast and non-contrast studies.
The 14 indications include findings like liver injury, spleen injury, appendicitis, and a range of other time-sensitive abdominal and thoracic pathologies. Aidoc received FDA Breakthrough Device Designation for the platform in September 2025 and FDA clearance in January 2026.
The FDA-reviewed pivotal study is what made the clinical evidence argument possible. Across the 11 newly cleared indications, the tool achieved a mean sensitivity of 97 percent and mean specificity of 98 percent. In certain settings, sensitivity reached 98.5 percent and specificity reached 99.7 percent. Critically, Aidoc's internal data suggests the multi-condition foundation model approach achieves roughly an order-of-magnitude reduction in false alerts compared to best-in-class single-condition tools, which has historically been one of the most damaging adoption barriers for radiology AI.
The technology is delivered through Aidoc's aiOS enterprise platform, which handles data normalization, continuous performance monitoring, and governance across health system deployments. That infrastructure layer matters for NTAP purposes: CMS will be watching whether real-world performance during the three-year window matches the pivotal study. If it does not, the renewal conversation in FY2030 will be difficult.
The Bigger Story: The Alternative Pathway is Gone
Here is what Aidoc's NTAP win cannot obscure: the company is the last major beneficiary of a regulatory shortcut that no longer exists.
CMS finalized the repeal of the alternative NTAP pathway alongside Aidoc's approval. Starting with applications submitted for FY2028 consideration, every technology, regardless of FDA Breakthrough Device Designation or any other expedited FDA status, must demonstrate substantial clinical improvement to be eligible for Medicare add-on payment. There are no exceptions, no grandfather provisions for new applicants, and no alternative routes.
This is not a hypothetical risk. It is a policy fact with a defined implementation date. Companies that had been planning FY2028 or FY2029 NTAP applications on the assumption that a Breakthrough Device designation would carry them past the evidence bar need to fundamentally revise those plans right now.
CMS was transparent about its reasoning. The agency looked at 47 applications for FY2027, saw that 32 of them leaned on the alternative pathway rather than substantial clinical evidence, and concluded that the pathway was being used to subsidize technologies whose clinical value was unproven. The core bargain of NTAP, temporary taxpayer-funded support for genuinely superior technologies, was being diluted.
The FDA and CMS did offer something in exchange. In April 2026, the two agencies jointly announced the RAPID pathway (Regulatory Alignment for Predictable and Immediate Device), which is intended to create a more coordinated coverage review for breakthrough devices. But RAPID does not eliminate the evidence requirement. It changes the timing and coordination of the review, not the standards. Clinical evidence is still required.
For radiology AI specifically, this lands in a complicated moment. The FDA has cleared more than 1,000 AI applications for clinical use as of early 2025, with the large majority concentrated in radiology. Very few of those products have been subjected to the kind of prospective, controlled clinical studies that CMS's substantial clinical improvement standard demands. The industry built an enormous regulatory footprint on the back of retrospective validations, reader studies, and regulatory pathway optimization. That foundation is not what CMS is looking for.
Three Questions Every Radiology AI Leader Needs to Answer Before FY2028
The Aidoc decision clarifies the competitive landscape in ways that matter for every organization building, buying, or deploying radiology AI. Here are the three questions that should be on every strategic agenda right now.
Question 1: Can your technology demonstrate substantial clinical improvement, and do you have the evidence to prove it?
"Substantial clinical improvement" is not a marketing claim. CMS evaluates it against a defined set of criteria: reduced mortality, fewer hospitalizations, shorter length of stay, faster time-to-treatment, or other measurable clinical outcome improvements over existing alternatives. A technology that helps radiologists read faster is not the same thing as a technology that demonstrably reduces adverse outcomes.
If your product pipeline does not yet have prospective clinical data showing one of these outcome improvements, the window to design and execute that study is narrowing. FY2028 NTAP applications are submitted roughly 18 months before the fiscal year begins, meaning the relevant application cycle opens in early 2026 for October 2027 effective dates. If you are not already designing or running the study, you may be too late for FY2028 and working against the clock for FY2029.
Aidoc had pivotal study data with mean sensitivity of 97 percent and mean specificity of 98 percent across 11 indications, plus a clinical argument about time-to-treatment for acute findings. That combination, plus Breakthrough Device Designation, satisfied both the traditional NTAP standard and gave CMS the evidence base it needed. Companies planning future applications need a similarly concrete evidentiary package.
Question 2: Does your DRG cost structure make NTAP worth pursuing?
NTAP's 65-percent-of-excess-cost formula means the reimbursement amount is highly sensitive to your product's actual cost structure and how cases are coded. A tool that adds meaningful clinical value but carries a low per-case cost may not generate enough excess cost to make the NTAP calculation compelling, either for CMS at the application stage or for hospitals at the billing stage.
The $137.53 maximum for Aidoc's tool implies a cost structure that, in at least some cases, produces enough excess to trigger meaningful payment. Before you build an NTAP strategy, model the DRG cost math with your actual deployment costs, and model it across the full range of DRGs your product touches. A radiology AI tool used across a broad set of CT studies will have different cost-threshold dynamics than a single-condition tool used in a narrow subset of cases.
This analysis is also important for hospital purchasing conversations. Health systems evaluating AI tools increasingly understand that NTAP eligibility changes the ROI equation, but only if the hospital can actually capture the payment, which requires accurate coding and documentation of the technology's use on qualifying cases. The operational infrastructure to capture NTAP payment is not trivial, and hospitals that have not previously built it will need implementation support.
Question 3: What is your FY2028 strategy, given that the alternative pathway no longer exists?
This is the most urgent question for companies that had been counting on Breakthrough Device Designation to simplify the NTAP pathway. The answer requires an honest assessment of where your clinical evidence currently stands and how long it will take to generate what you actually need.
Three scenarios are possible. First, if you already have prospective clinical data meeting CMS's substantial clinical improvement criteria, you should be actively preparing a traditional-route NTAP application for FY2028 or FY2029. The application requirements are defined, the timeline is predictable, and the pathway now has a successful precedent. Second, if you have suggestive data but not a completed pivotal study, you need to prioritize study design and execution immediately. FY2029 may be your realistic target, which means applications submitted in late 2027. Third, if you do not have a clear path to the clinical evidence, NTAP may not be the right strategy, and you should evaluate whether the new RAPID pathway changes the calculus for your situation.
The companies that come out of this transition period in the strongest position will be the ones that treated clinical evidence as a core product development investment, not an afterthought, while their competitors were optimizing for regulatory pathway efficiency.
What This Means for Health Systems Buying AI Today
For hospital and health system leaders evaluating radiology AI investments, the Aidoc NTAP decision creates a practical filter. A product with NTAP eligibility is a product that has satisfied CMS's evidence standard, which is a meaningful quality signal independent of the payment. It also changes the budget conversation: a $137.53 per-case add-on payment, even if only partially realized in practice, converts what was an operational cost center into a partial revenue offset.
That does not mean every hospital should rush to deploy Aidoc's product specifically. It means every hospital should understand which AI tools in their current or prospective portfolio are on a credible path to NTAP eligibility, and structure vendor evaluations accordingly. As the FY2028 and FY2029 cycles play out, more products will either succeed or fail to clear the evidence bar, and the market will develop a clearer picture of which technologies have durable clinical value versus which ones were primarily regulatory positioning.
The three-year NTAP window also creates a practical timeline for health systems. The first wave of institutions to deploy CARE Multi-Triage before October 2026 will receive reimbursement support through approximately fiscal year 2029. Institutions that delay will eventually be paying full cost, which changes the implementation urgency calculation.
The Broader Picture: Medicare's Signal to the Market
CMS is not a passive observer of the AI economy in healthcare. The decision to simultaneously approve Aidoc's NTAP and eliminate the alternative pathway is a deliberate policy signal: Medicare will support AI technologies that demonstrate clinical value, and it will not subsidize technologies that have not.
That signal has downstream effects well beyond NTAP. Commercial insurers watch Medicare coverage decisions carefully. Hospitals evaluating capital expenditures weigh Medicare reimbursement status. Investors assessing radiology AI companies will increasingly distinguish between companies with a credible path to CMS reimbursement and companies that were optimizing for FDA clearance volumes. The evidentiary bar CMS has now reset will, over time, reshape investment allocation, clinical research priorities, and product development strategies across the entire sector.
For a sector that has sometimes prioritized regulatory momentum over clinical outcomes research, that is a significant adjustment. The companies that have been building evidence alongside their regulatory programs are about to find out that they were right to do so. The companies that treated clinical evidence as a downstream problem will be facing a harder conversation with their investors and their hospital customers than they had anticipated.
Aidoc's win is genuinely historic. The first NTAP for a radiology AI foundation model, with Medicare dollars flowing to hospitals for AI-assisted CT triage starting October 1, is a milestone the industry has been working toward for years. But the more consequential development is not the $137.53. It is the policy framework in which that number exists, and what it demands of everyone who wants to follow.
The alternative pathway is closed. The evidence bar is set. The clock is running.
Oatmeal Health covers the intersection of healthcare data, AI, and policy for health system and life sciences leaders. Subscribe for weekly analysis.