Media coverage and marketing materials routinely compress seven distinct regulatory and evaluative concepts into a single word, "approved," and the compression matters, because each concept represents a different amount and type of evidence, and confusing them is how a research-stage signal ends up described with the confidence of a proven treatment. This article is the second evergreen reference this cluster is built around, defining each term precisely so every product review can link here rather than re-explaining regulatory status from scratch.
The FDA pathways
FDA 510(k) clearance: the most common pathway for medical devices, including most AI-enabled ones, requiring the manufacturer to show substantial equivalence to an existing legally marketed device. It is a real regulatory hurdle and not a rubber stamp, but it is a different and generally lower evidentiary bar than full premarket approval. FDA De Novo authorisation: used for novel, lower-to-moderate-risk devices without an existing equivalent to compare against, effectively creating a new device category, a pathway several genuinely novel autonomous AI systems in this cluster have used precisely because nothing quite like them existed before. Premarket approval, PMA: the most rigorous FDA pathway, generally reserved for higher-risk devices, requiring the strongest evidence of safety and effectiveness among the three. Breakthrough Device designation: an accelerated-review status granted to promising technologies addressing unmet needs, valuable and worth noting, but critically not the same as clearance or approval; a designation means the FDA has agreed to prioritise and expedite review, not that review has concluded or that the device may yet be marketed, a distinction this cluster's Aidoc coverage specifically flags because the two get conflated in reporting.
The UK and European pathways
CE marking: the European conformity mark, indicating a device meets EU regulatory requirements for its stated intended use and risk classification, with higher classes requiring more evidence. UKCA marking: the UK's post-Brexit equivalent, following a broadly similar risk-based framework. MHRA registration: registering a device with the UK's Medicines and Healthcare products Regulatory Agency is a regulatory requirement for placing it on the UK market, and it should never be represented as MHRA approval, accreditation or endorsement of clinical performance, a distinction the regulator itself is explicit about and that this cluster treats as a standing caution across every UK-facing product review. NICE evaluation: the National Institute for Health and Care Excellence's assessment of whether a technology should be recommended for use, at what evidence standard, and under what conditions. And NICE Early Value Assessment specifically: a mechanism allowing promising technologies to be used in the NHS while further evidence is generated under defined data-collection conditions, materially different from NICE's mature routine-use guidance and worth its own full explanation, which the companion article on Early Value Assessment provides in depth.
Why this hierarchy matters practically
The consequence of collapsing these seven concepts into one word is specific and predictable: a device with Breakthrough Device designation, a signal that regulators consider it promising and worth expedited review, gets described in coverage as though it were already cleared for use, when clearance may still be pending. A device registered with the MHRA gets described as MHRA-approved, when registration is a market-access requirement rather than an endorsement of clinical performance. A NICE Early Value Assessment recommendation, use while evidence is generated under specified conditions, gets reported as unconditional NICE endorsement, when the entire mechanism exists because the evidence is not yet considered complete. Each of these compressions moves a technology up the evidence ladder in the reader's mind without any corresponding movement in the underlying evidence, which is precisely the gap this whole cluster exists to close.
The FDA's own caveat, worth stating explicitly
The FDA maintains a public list of AI-enabled medical devices, a genuinely useful resource, and the agency itself states that the list is not comprehensive. Inclusion on it indicates that a device completed an applicable premarket process for its stated indication; it does not mean the FDA has proved improved patient outcomes in every clinical setting the device might be used in, a distinction worth holding onto specifically because the list's existence tempts exactly the opposite reading, that presence on an official government list functions as a broader outcomes guarantee than any individual clearance actually provides.
The standard regulatory-status box
Every product review in this cluster carries the same fields, deliberately, so status is comparable across companies rather than described in whatever language each company's own marketing prefers: product name, manufacturer, territory, pathway, exact intended use, intended user, intended population, input, output, and the date the status was last checked, because regulatory positions change and a status confirmed today may not hold in six months. Reading any diagnostic-AI claim without these fields filled in is reading a claim with its most important context missing.
Frequently asked questions
Does Breakthrough Device designation mean a device is already usable clinically?
Not necessarily: designation signals expedited FDA review priority for a promising technology, and the device may still require full clearance or approval before it can be legally marketed for its intended use, a status worth checking specifically rather than assuming from the designation alone.
Is MHRA registration the same as clinical validation?
No: registration is a market-access administrative requirement, and the MHRA itself distinguishes it from any assessment of clinical performance, which is why a registered device still needs the same clinical-evidence scrutiny as any other.
What should change if a device moves from NICE Early Value Assessment to routine-use guidance?
The evidence base: routine-use guidance follows completed data collection meeting NICE's evaluation standard, a materially stronger position than conditional use while evidence accumulates, and any article referencing a technology's NICE status should state which of the two applies and when it was last checked.