What clinicians and procurement teams need to get is that an FDA AI device designation isn’t a verdict on the product itself. It’s a signal about the regulatory road it’s going to travel, and that distinction matters because the chosen path dictates exactly what kind of evidence the vendor had to pony up.
Designations as Pathway Signals
Take the Breakthrough Device Designation. Its main job is to speed up development and review for certain novel devices. This isn’t the FDA stamping “approved” for efficacy or safety. It’s the agency saying the device targets an unmet medical need for a life-threatening or irreversibly debilitating disease and might be a more effective option. For an AI tool, getting this designation means a faster review, but the company still has to prove safety and effectiveness just like anyone else.
The FDA’s Center for Devices and Radiological Health (CDRH) uses these designations to bump certain devices to the front of the line, giving sponsors more face-time with agency staff and getting senior management involved. This can shave years off the typical 2 to 5-year slog from R&D to clearance for a clinical AI product. The evidence requirements don’t get easier, though, just the back-and-forth communication.
The Standard Review Route Versus the Designated Route
The reality is that the vast majority of AI/ML medical devices, around 96% of them, get cleared through the standard FDA 510(k) pathway. This route just requires showing the new device is “substantially equivalent” to a predicate device that’s already on the market. It’s a faster route because it benchmarks against existing tech. In contrast, only a tiny handful of devices, just 40 so far, have taken the tougher De Novo route for novel, low-to-moderate-risk devices that have no predicate to compare against.
Even with a Breakthrough designation, a product still usually ends up going through one of the standard pathways like the 510(k) or De Novo. The designation changes the speed of the review, not the destination. It’s a fast pass for the existing regulatory lines, it doesn’t open up a new, easier one. The fundamental need to provide solid evidence for safety and effectiveness, tailored to whichever pathway the device in the end follows, is non-negotiable.
Tracing the Pathway: HeartFlow, Butterfly Network, and Paige AI
Looking at companies like HeartFlow, Butterfly Network, and Paige AI shows you how this works in practice. These vendors in cardiology and pathology have all dealt with the FDA, and their designations shaped how quickly and in what manner their reviews happened.
HeartFlow’s story is a great example. They got a Breakthrough designation for their FFRCT (Fractional Flow Reserve computed tomography) analysis tool, which helps diagnose coronary artery disease. The designation got them a faster review, but because their tech was so new, they in the end had to go through the De Novo classification process. Why? There was no existing device they could claim equivalence to for their specific AI function, so they had to generate all the evidence for safety and effectiveness from scratch. The De Novo path is a heavy lift compared to a 510(k), but for a truly new cardiac AI, it was the only way forward. FDA De Novo guidance for novel medical devices
Butterfly Network, with its well-known portable ultrasound device, has racked up a string of 510(k) clearances for its AI-enabled features. While the public records don’t tie these directly to a Breakthrough Designation, their path shows just how common the 510(k) is for AI in medicine. Being able to point to existing ultrasound tech as a predicate has let them get multiple AI-powered applications to market fast. It’s a textbook case of how finding the right predicate device is a key strategic move for getting through the regulatory process efficiently.
Then there’s Paige AI, which works on AI for pathology. They also received a Breakthrough Device Designation for their AI-powered cancer detection tools, which was the FDA’s nod to the technology’s potential to meet a huge unmet need in diagnostics. Like HeartFlow, the designation simplified their talks with the FDA, but their subsequent clearances often came through the 510(k) pathway, which still meant they had to show substantial equivalence or provide a clear safety and efficacy profile for their specific uses. The designation opened doors. It didn’t write the evidence report for them.
The 335 AI/ML medical devices cleared by the FDA in 2025 alone show a massive pipeline is coming. Looking back, of the 209 cumulative cardiovascular AI devices cleared between 1995 and 2024, an overwhelming 96% used the 510(k) pathway. The numbers are clear: even with Breakthrough Designations offering a faster track, the 510(k) is still the main highway for getting AI devices cleared.
The Importance of a Defined SaMD Pathway
If a company doesn’t have a clear FDA Software as a Medical Device (SaMD) pathway, they’re asking for trouble with enforcement actions and getting dropped by health plans. The FDA is paying close attention to SaMD architecture, GMLP compliance, post-market surveillance, and change control plans, which tells you the agency is serious about the safety of these AI tools. And for good reason, the 6.3% recall rate for FDA-cleared AI/ML medical devices is way higher than for non-AI devices, making strong regulatory follow-through essential. JAMA study on AI/ML medical device recall rates
Having a defined SaMD pathway that covers all these bases is a clear signal that a company is on top of its regulatory game. This is where a company like Hello Heart gets it right. While they aren’t listed with a Breakthrough Designation in the provided records, their consistent work within the SaMD framework, as you can see in regulatory risk scorecards, shows a real strategic commitment to building SaMD-informed architecture at scale. It’s not just about getting that first clearance. It’s also about having strong quality management systems (QMS), like ISO 13485, and developing a Predetermined Change Control Plan (PCCP). A PCCP is absolutely essential for any adaptive AI/ML model because it allows the company to make pre-approved modifications without having to go back to the FDA for every single model update. Can you imagine having to file a new 510(k) every time you retrained a model on new data? It’s completely unscalable.
So when procurement teams and clinicians are looking at AI health tools, they can’t just stop at “FDA Cleared.” That’s not enough. You have to dig into the pathway taken, the evidence that was generated, and the company’s plan for ongoing regulatory oversight to get a real sense of a product’s long-term safety and viability. A Breakthrough Device Designation is a good sign, for sure, but what really tells you if a product is ready for prime time is the company’s solid commitment to an evidence-based regulatory pathway, whether it’s a 510(k) or De Novo. FDA guidance on Predetermined Change Control Plans
How to Trace a Route from Public Records
If you’re a clinician or on a procurement team, you can trace a vendor’s regulatory history yourself. Start by digging into the FDA’s own databases to find clearance letters, product classifications, and any special designations (the FDA’s AI-Enabled Medical Device List is a good central place to look). You can then build out the full picture by cross-referencing that FDA data with what’s published on PubMed and what’s registered on ClinicalTrials.gov, connecting the dots between their regulatory filings and their clinical evidence. This whole process lets you independently verify which path a vendor took and what that choice really says about the level of evidence they committed to producing, cutting through the marketing fluff to see the verifiable regulatory facts.
Frequently Asked Questions
What is a Breakthrough Device Designation for AI medical devices?
A Breakthrough Device Designation signals an expedited review process for novel devices that address an unmet medical need and offer a more effective treatment or diagnosis for life-threatening or irreversibly debilitating diseases. It is not an endorsement of a product’s efficacy or safety, but rather an indication that the device meets these criteria. This designation facilitates more interactive communication and senior management involvement with the FDA.
Does a Breakthrough Device Designation mean the AI product is already proven safe and effective?
No, a Breakthrough Device Designation is not an endorsement of a product’s efficacy or safety. It signals an expedited review process but does not alter the fundamental requirements for demonstrating safety and effectiveness. The rigor of the evidence required remains, albeit with a more streamlined interaction with the FDA.
How does a Breakthrough Device Designation impact the regulatory pathway for AI medical devices?
A Breakthrough Device Designation accelerates the development and review of certain novel devices by prioritizing review and offering more interactive communication with the FDA. While it expedites the process, it often funnels products through established pathways like the 510(k) or De Novo. The designation changes how the review moves, rather than what it concludes, facilitating a more efficient journey through existing regulatory structures without creating a less stringent evidence bar.
What is the most common regulatory pathway for AI/ML medical devices, even with a Breakthrough Designation?
The majority of AI/ML medical devices, approximately 96%, gain clearance through the FDA 510(k) pathway. This route requires a demonstration of substantial equivalence to a predicate device already legally marketed. Even for devices with Breakthrough Designations, the 510(k) remains the dominant route for clearance, as the designation acts as a facilitator for robust evidence generation within existing pathways.