On June 4, 2026, the FDA cleared DEEPVESSEL Plaque under 510(k) number K260497. The applicant is Keya Medical Technology, and the openFDA record files the product under code QIH and regulation 892.2050, the agency's bucket for automated radiological image-processing software, classed as a Class II device in the radiology specialty. The device name in the classification is plain to the point of understatement: "Automated Radiological Image Processing Software." What that flat label conceals is a strategically loaded event. A Chinese cardiac-AI company has just earned a regulated foothold in the single most valuable medical-imaging market on the planet.
Plaque analysis is not a peripheral corner of imaging. Coronary artery disease remains the leading cause of death in the developed world, and the clinical question that drives an enormous volume of testing is deceptively simple: how much plaque has built up in a patient's arteries, what is it made of, and how dangerous is it? Answering that from a CT angiogram by hand is slow, subjective and operator-dependent. Software that automates the measurement and characterization of plaque turns a specialist-bound, variable process into something fast, reproducible and scalable. That is precisely the kind of bottleneck that AI monetizes, and it is why every serious cardiac-imaging player is building or buying into this category.
Why a 510(k) is the entry ticket that matters
For a company headquartered outside the United States, the 510(k) clearance is not a formality; it is the business model's foundation. Without it there is no legal U.S. market, no path to reimbursement, and no credibility with the American academic medical centers and integrated health systems that set purchasing standards the rest of the world watches. With it, the door opens to a market that pays more per study, adopts new diagnostic technology faster, and writes the reimbursement codes that eventually propagate globally. The U.S. is where cardiac-AI revenue is richest, and K260497 is Keya's admission ticket.
The Class II 510(k) route the company used is the same pragmatic lane that has made the AI-imaging market explode. By demonstrating substantial equivalence to a predicate device rather than running a years-long premarket-approval gauntlet, a developer can reach the market with manageable capital and a defined timeline. That accessibility is a double edge. It lowers the barrier for Keya, but it also lowers it for every competitor, which is why the durable advantage in this space comes not from any single clearance but from clinical validation, workflow integration, and a portfolio broad enough to become a standard rather than a feature.
The competitive landscape Keya is walking into
The U.S. cardiac-AI imaging market is not empty. It already has a marquee incumbent in coronary CT analysis that built its position on outcome studies, payer coverage and a recurring per-analysis service model, and the large imaging OEMs are all assembling cardiac-AI offerings of their own. Keya enters as a challenger with a differentiated plaque-characterization angle and, presumably, a cost structure shaped by a large domestic Chinese installed base on which to refine its algorithms. The strategic logic of expanding from a vast home market into the U.S. is familiar from consumer electronics: build scale and data volume at home, then attack the high-margin Western market once the technology is mature and a regulatory clearance is in hand.
The contrarian read worth holding onto is that a clearance is the start of the hard part, not the end. The graveyard of medical AI is full of cleared products that never achieved adoption because they lacked reimbursement, failed to integrate into the radiology reading workflow, or could not produce the outcome evidence that risk-averse health systems demand before they will pay. Keya's clearance proves the technology is regulator-acceptable. It does not prove that an American cardiology department will buy it, that a payer will reimburse it, or that it can dislodge an entrenched competitor. Those are commercial battles that a 510(k) does not win.
What the clearance unlocks
What K260497 does unlock is optionality, and for a company at Keya's stage optionality is valuable. The clearance lets the company pursue U.S. pilots at academic centers, generate the American clinical-validation data that anchors reimbursement applications, and pitch partnerships to the imaging OEMs and health systems that would never engage with an uncleared foreign product. It also de-risks the company in the eyes of investors and potential acquirers; a U.S.-cleared cardiac-AI asset is a far more financeable and far more acquirable thing than a promising algorithm with only home-market approval. In an environment where the imaging incumbents are buying their AI catalogs rather than building them all in-house, a clean U.S. clearance is exactly the kind of asset that attracts a strategic bid.
There is also a broader signal in this single record. The flow of cleared AI imaging devices is increasingly international, with developers from outside the United States establishing American footholds via the 510(k) pathway. That globalization of the cleared-AI pipeline intensifies competition for the incumbents and accelerates the commoditization of individual capabilities like plaque analysis. The winners over the next several years will be the firms that turn a cleared algorithm into a reimbursed, workflow-embedded standard of care. The clearance is the price of admission to that race.
The bottom line
DEEPVESSEL Plaque's clearance under K260497 is a small line in the FDA's daily output and a meaningful strategic event for Keya Medical. It converts a coronary-AI capability honed at home into a regulated American product, opens the door to the world's richest cardiac-imaging market, and gives the company the optionality to pursue partnerships, reimbursement and, plausibly, an exit. The technology is now regulator-acceptable. Whether it becomes a business depends on the unglamorous work, reimbursement, integration and outcome evidence, that no clearance can shortcut. For now, a challenger has its ticket, and the incumbents have a new name to watch.