If you are building software that draws boxes around nerves and blood vessels on a live ultrasound screen, FDA just told you exactly what evidence it wants before you sell it — and, almost as important, how long the rulebook can lag behind the product that made it necessary. On June 1, 2026, FDA published a final order creating 21 CFR 868.1980, a new Class II regulation for what it calls the "real-time ultrasound anatomy visualization and labeling device for ultrasound guided regional anesthesia." The device that triggered it, Intelligent Ultrasound Limited's ScanNav Anatomy Peripheral Nerve Block, was actually granted De Novo status back on October 18, 2022. That is three years and seven and a half months between the day FDA decided this technology could be safely marketed and the day the agency finished writing the generic rule other companies can now build against.
That gap is the lesson. If you are developing a similar product, or you already hold a De Novo grant and assumed the regulatory picture was settled the day you got your decision letter, it is not. The codified regulation — the thing that creates a real product code, a real predicate, and a real 510(k) pathway for competitors — can trail the actual market clearance by years. Plan your competitive strategy, your patent timing, and your investor conversations around that reality, not around the press release date.
What FDA Actually Did on June 1, 2026
The Federal Register notice (Docket No. FDA-2026-N-5722, 91 FR 32338) is a classification order, not a new law and not a recall. FDA used its authority under section 513(f)(2) of the Federal Food, Drug, and Cosmetic Act — the De Novo pathway — to formally place this device type into Class II with special controls, and to codify that decision as 21 CFR 868.1980 under the anesthesiology devices regulations in Part 868. The order took effect on the date of publication, June 1, 2026.
In plain terms: FDA has now written down, in regulation, what this category of device is, what class of risk it belongs to, and what four categories of evidence a company has to produce to get one cleared. Before this order, the only company with a clear path to market this exact device type was Intelligent Ultrasound Limited, working from its own De Novo decision summary. After this order, any manufacturer can reference 21 CFR 868.1980 and pursue a 510(k) using the newly classified device type as a predicate, provided they meet the special controls.
I want to be direct about what this device actually is, because "real-time ultrasound anatomy visualization and labeling" sounds abstract until you picture it. It is software — frequently AI-driven image analysis — that sits on top of a live ultrasound feed during a nerve block procedure and highlights or labels anatomical landmarks (nerves, vessels, fascial planes) in real time, so the clinician performing the block has an overlay to help confirm what they are looking at. FDA's own intended use language describes it as providing "real-time interpretation and enhanced visualization of live ultrasound images by highlighting anatomical landmarks in preparation for performing regional anesthesia." It does not replace the clinician's judgment. It is not a diagnostic device. It is a visualization aid layered onto an existing ultrasound-guided regional anesthesia workflow.
The Device Behind the Rule, and Why the De Novo Pathway Matters Here
The De Novo pathway exists for exactly this situation: a device with no valid predicate, that FDA nonetheless believes is low-to-moderate risk and doesn't need the full weight of Class III premarket approval. Congress created the pathway through the 1997 FDA Modernization Act and streamlined it in 2012 through the FDA Safety and Innovation Act, allowing a manufacturer to request classification directly instead of first filing a doomed 510(k) and waiting for a not-substantially-equivalent decision. Intelligent Ultrasound Limited used that pathway for ScanNav Anatomy Peripheral Nerve Block, and FDA granted it in October 2022.
Here is the part that trips people up: a De Novo grant already creates a Class I or Class II classification for that specific device the moment it's granted. The manufacturer that filed the De Novo can market immediately. What takes years longer is the Federal Register rulemaking that turns that individual grant into a generic, codified regulation in the Code of Federal Regulations — the step that opens the door for every other manufacturer in the space. Anyone who read the 2022 De Novo grant and assumed the regulatory landscape was locked in for the category, not just for that one product, was working from incomplete information for almost four years.
I think this is worth saying plainly: FDA's own device classification data shows that Class II devices make up roughly 47 percent of all regulated device types, which tells you the agency treats this risk tier as its default working category for moderate-risk technology — not an exception, not a special favor to one manufacturer. This device landing in Class II with special controls, rather than Class III, is consistent with how FDA has treated ultrasound-adjacent visualization software generally.
The Four Special Controls, and What They Actually Require
Special controls are FDA's substitute for the heavier premarket approval process — a defined, device-specific set of evidence requirements that, if met, gives reasonable assurance of safety and effectiveness without a full PMA. For 21 CFR 868.1980, FDA specified four categories.
Clinical performance testing for anatomical landmark accuracy. The manufacturer has to demonstrate, under conditions that represent anticipated use, that the device correctly identifies and locates the anatomical structures it claims to label — nerves, vessels, fascial planes — across the patient population and body sites the labeling supports. This is not a bench test. FDA wants performance data that reflects real anatomical variability, not a curated dataset of clean images.
Human factors testing on instructions-for-use alone. The device has to be shown to work correctly when a clinician uses it based only on the labeling, with no additional training beyond what's provided in the instructions. This control exists because overlay and labeling tools live or die on whether a busy anesthesiologist in a procedure room reads the screen the way the manufacturer intended, not the way a company trainer explained it in a conference room.
Software verification, validation, and hazard analysis, including ultrasound device compatibility. Because this software runs on top of another manufacturer's ultrasound hardware and image output, FDA wants documented compatibility testing across the ultrasound systems the labeling supports, plus the standard software lifecycle documentation. If you are the software manufacturer, your hazard analysis has to account for image quality variation coming from a machine you don't control.
Labeling that discloses training, clinical data, and the limits of the device. The labeling has to describe recommended user training, summarize the clinical data behind the claims, and — this is the one I'd flag hardest for anyone drafting labeling copy — include an explicit warning against over-reliance on the device's output. FDA is telling manufacturers, in the regulation itself, to warn clinicians not to trust the overlay more than their own clinical judgment. That is a meaningful signal about how FDA views the risk profile of AI-assisted visualization tools generally: the failure mode it's most worried about is not the software being wrong, it's the clinician stopping the independent verification that ultrasound guidance was supposed to preserve in the first place.
Before and After: What Changed for Manufacturers
| Before June 1, 2026 (De Novo grant only) | After June 1, 2026 (21 CFR 868.1980 codified) | |
|---|---|---|
| Regulatory status of the device type | Established only for the original De Novo grantee | Generic Class II device type, codified in the CFR |
| Pathway for follow-on manufacturers | None — a new De Novo request required for each entrant | 510(k), citing 21 CFR 868.1980 as predicate |
| Evidence standard | Whatever FDA required in the individual De Novo review | Fixed set of four special controls, publicly known in advance |
| Predicate available for competitors | No | Yes |
| Applicable FDA authority | Section 513(f)(2), FD&C Act (De Novo) | Section 513(f)(2) De Novo + Section 510(k) for subsequent entrants |
| Public visibility of requirements | Buried in a De Novo decision summary | Codified regulatory text, searchable in 21 CFR Part 868 |
The practical effect: competitors who were waiting on the sidelines because a second De Novo request for a nearly identical device type is expensive and slow now have a much cheaper, much faster 510(k) pathway available. That is very likely to accelerate competitive entry into AI-assisted regional anesthesia visualization over the next 18 to 24 months.
Effective Dates and Deadlines to Track
The final order took effect June 1, 2026 — that is the date the classification became codified law, not a proposal or draft. There is no delayed compliance date or transition period specified in the order; the regulation applies to submissions from the moment it published. If you already have a related product in development, the practical deadlines you need to track are internal, not statutory: when your clinical performance data collection needs to be locked to support a 510(k) built on this predicate, when your human factors study needs to be designed and run, and when your labeling draft needs legal and regulatory review against the four special controls. None of those have an FDA-imposed clock. All of them have a competitive one, now that the pathway is open to everyone.
If you are a hospital or ambulatory surgery center evaluating one of these products for purchase, the relevant date is simpler: any device cleared after June 1, 2026 under this classification has been reviewed against a known, public standard. A device on the market before that date — including the original ScanNav clearance — was reviewed under the De Novo process directly and should have equivalent, if not identical, supporting evidence. I would ask any vendor for the human factors study summary specifically, since that is the control most directly tied to whether your staff will actually use the overlay correctly in a real procedure room.
Practical Compliance Guidance
For manufacturers building a competing or adjacent product, start by mapping your existing clinical and usability data against the four special controls before you draft a 510(k). Landmark accuracy data collected for internal development purposes rarely satisfies FDA's expectation of performance testing under anticipated use conditions across a representative patient population — you may need a dedicated study, not a retrospective analysis of your training dataset. Budget for a human factors study that tests naive users against your actual instructions for use, not a study that trains participants beforehand. And write your labeling's over-reliance warning now, early in development, because it shapes what claims you can responsibly make about the device's accuracy elsewhere in your labeling.
For manufacturers who already hold or are pursuing a De Novo grant in an adjacent category, do not assume your grant date is the end of the regulatory story for your device type. Track whether FDA has published, or announced intent to publish, a codifying final order — that is the event that opens your category to competitors, and it can trail your own clearance by years, as this one did. If you were counting on a multi-year window of exclusivity based purely on the complexity of a repeat De Novo filing, that window may be shorter than you think once FDA finishes its rulemaking.
For hospital quality and risk management teams, treat AI-assisted anatomy labeling devices the same way you would treat any new clinical decision-support software entering your OR or block room: confirm the vendor's clearance basis (De Novo direct grant versus 510(k) under 21 CFR 868.1980), request the human factors summary, and build the over-reliance warning into your own staff training rather than assuming the device's labeling alone will change clinician behavior. Labeling warns; training changes practice.
Why This Reaches Beyond Regional Anesthesia
This order is a small regulation about a narrow device type, but it is a useful data point for anyone tracking how FDA handles AI-enabled software more broadly. FDA's own running list of AI/ML-enabled medical device authorizations has grown past 1,000 entries in recent years, dominated by radiology and cardiology, and this order shows the same review logic FDA has applied there — clinical performance testing, human factors testing, software validation, and an explicit anti-over-reliance labeling requirement — being exported into anesthesiology. Ultrasound guidance itself has been the accepted standard of care for peripheral nerve blocks for well over a decade precisely because it lets the clinician see what a landmark or nerve-stimulator technique can only estimate; layering AI-assisted labeling on top of that guidance is a logical next step, and FDA's special controls make clear it intends to regulate the layering, not just the base ultrasound hardware.
For readers building broader regulatory strategy across device categories, our breakdown of the FDA De Novo pathway walks through when a De Novo filing beats a doomed 510(k), and our guide to software validation for AI-enabled medical devices covers what auditors actually check when IEC 62304 and FDA's software documentation expectations intersect with a Class II special-controls device like this one.
FAQ
What is the new regulation number, and where does it live? 21 CFR 868.1980, codified under Part 868 (Anesthesiology Devices), effective June 1, 2026.
What device caused FDA to create this classification? Intelligent Ultrasound Limited's ScanNav Anatomy Peripheral Nerve Block, granted De Novo status on October 18, 2022. The final order published in 2026 codifies the generic device type based on that grant.
Does this rule apply to every ultrasound machine used in regional anesthesia? No. It applies specifically to software that visualizes and labels anatomy in real time on top of an existing ultrasound feed for regional anesthesia guidance — not to the base ultrasound imaging hardware itself.
If I already market a similar device, do I need to resubmit anything? Not automatically. If your device was cleared through its own De Novo grant before this order, your clearance stands. The new codified regulation mainly changes the pathway available to new entrants, who can now file a 510(k) referencing 21 CFR 868.1980 instead of pursuing a separate De Novo request.
What are the four special controls in plain language? Prove the labeling is accurate against real anatomy, prove clinicians can use it correctly from the instructions alone, validate the software and its compatibility with the ultrasound systems it runs on, and label it with training guidance, clinical data, and a clear warning against over-relying on the device instead of clinical judgment.
Is there a compliance deadline I'm at risk of missing? There is no delayed compliance date — the order was effective on its publication date, June 1, 2026. The deadlines that matter now are the ones you set for your own submission timeline, since the competitive pathway is open to any manufacturer prepared to meet the four special controls.
Last updated: 2026-07-29
Jared Clark
Principal Consultant, Certify Consulting
Jared Clark is the founder of Certify Consulting, helping organizations achieve and maintain compliance with international standards and regulatory requirements.