Stryker's Vision Pro Surgery App Spent 318 Days at FDA and Came Back With a Device Class of Its Own

Stryker announced on September 1 that a surgeon at Duke Health had completed the first hip arthroscopy performed using SportSuite Vision, the company's software for the Apple Vision Pro. That is the readable half of the story, and it is the half every outlet covering the news this week reported. The consequential half sits in two database records at the Food and Drug Administration, where De Novo request DEN250040 was granted on July 17 and, in the granting, brought a device classification into existence that had not existed before.

What the software is authorized to do is narrower than the headlines suggest, and the wording is worth reading rather than paraphrasing. Stryker's release states that SportSuite Vision Software "is indicated for the intraoperative display of arthroscopic video and medical imaging during femoroacetabular impingement and labral repair hip arthroscopy procedures, and the display of the same information as presented by the HipCheck software and HipMap FAI Analysis." Two procedures on one joint. Not orthopedics, not surgery, not spatial computing in the operating room, but hip arthroscopy for impingement and labral repair, with the arthroscope feed and the planning output placed in the surgeon's field of view rather than on a monitor across the room.

The classification is the asset, not the app

De Novo is the route a manufacturer takes when its device is low or moderate risk but has no legally marketed predicate to point at. There is nothing to be substantially equivalent to, so the applicant asks the agency to create a classification and then becomes the first device inside it. Stryker asked, and the agency wrote one. The De Novo record lists regulation number 888.1150, product code SIT, and the Orthopedic advisory committee.

The classification record for that product code is the more revealing of the two pages. FDA's entry for SIT puts the class in Device Class 2 under the Office of Orthopedic Devices, and defines it this way: "This device consists of augmented reality software on a head-mounted display that displays arthroscopic or endoscopic video and medical imaging during orthopedic surgery. The device is intended to intraoperatively augment visualization with contextually relevant information and data, including that received from other devices." The page's technical method field describes the user viewing "through the augmented reality video see-through head-mounted display."

Read those sentences for two things. The first is what they do not contain, which is any mention of Apple or of the Vision Pro. The class is written around a head-mounted display, generically, so the regulation Stryker paid for does not attach to Apple's hardware in a way a competitor would have to work around. The second is scope. The class covers arthroscopic and endoscopic video during orthopedic surgery, which is a considerably wider field than the two hip procedures SportSuite Vision itself is authorized for. Stryker's product sits inside a room much larger than the product.

The field that matters most to everyone else in this market is the submission type, and FDA lists it as 510(k). The first device through a new De Novo class becomes the predicate for that class, so the next company building an intraoperative arthroscopy display for a headset is not facing the road Stryker just walked. It files a 510(k) demonstrating substantial equivalence. That is not a free pass and should not be read as one, since the agency can still ask for clinical data and equivalence has to be shown rather than asserted. But a predicate and a product code are the difference between a defined path and an open question, and until July there was no path at all.

One wrinkle for anyone who goes looking for the rule itself. The section number is live in FDA's device databases but is not yet codified in the Code of Federal Regulations, where the Part 888 edition current to September 2 runs from 888.1100 for arthroscopes straight to 888.1240 for dynamometers. Codification follows a De Novo grant rather than accompanying it, so this is ordinary sequencing and not a defect. It does mean a compliance officer reading the CFR today will not find the class their vendor is citing.

318 days, and no change control plan

The De Novo record dates the request. Stryker Endoscopy, the San Jose entity that filed it, submitted on September 2, 2025, and the decision came on July 17, 2026, which is 318 days later. For a first-in-class request establishing a new regulation, that is not a slow review, and it is a useful figure for anyone modeling a comparable submission. It also shows how far ahead of the announcement the regulatory work sits. The case at Duke Health made the news 46 days after the authorization that permitted it.

Another field on that record carries a cost that will not surface for a while. The entry includes a line reading "Predetermined Change Control Plan Authorized", and its value is No. Such a plan is the mechanism that lets a manufacturer specify in advance which modifications it may make without returning to the agency, which is the tool most useful to software expected to change. Stryker does not have one, so a modification that could significantly affect safety or effectiveness follows the ordinary route. That matters more than it sounds, because this class is defined around a headset somebody else builds and discontinues. VR.org raised the same dependency in August, covering MediView's XR90, whose 510(k) clearance was granted on a HoloLens 2 that Microsoft stopped producing in 2024 and will support only through the end of 2027. Stryker is better placed there, since Apple has announced no end of support for anything. The dependency has not gone away. It has moved to a vendor with a longer runway.

What a purchasing committee actually has here

The release carries two quotes, and they are its least informative part. Matt Moreau, vice president and general manager of Stryker Sports Medicine, said the product "marks a new chapter in digital surgery, giving surgeons a more intuitive and ergonomic way to access critical information when and where they need it." Chad Mather III, M.D., M.B.A., the orthopedic surgeon at Duke Health who performed the case, said that "using spatial computing enabled me to customize the placement of key clinical information to fit my workflow and access it within the sterile field." Both are the sentences a first case produces. Neither is evidence.

The evidence that exists is thin, and it is worth being honest about how thin. One completed case, announced by the manufacturer. Stryker did not date the procedure itself, and the release carries no pricing and no rollout timing. Against that, the strongest published data on a headset in an operating room remains the UC San Diego trial this site covered in August, which measured 32 tear duct operations, found the headset group roughly 19 percent faster with identical outcomes, and which its own authors called preliminary. Different specialty, different workflow, no FDA-authorized product involved.

So the honest summary for a hospital evaluating this is that the clinical case is still one case and the regulatory case is now settled. Those are not the same thing, and only the second one changed in July. When SightMD disclosed its Vision Pro cataract workflow back in May, the constraint on adoption was never whether the headset worked. It was that a procurement officer had no authorized product to buy, no product code to reference, and no classification to write a policy against. All three now exist, and the second company that wants them will find the paperwork considerably shorter than 318 days.