Federal Circuit Finds Dental AI Patent Claims Abstract Under Section 101

The U.S. Court of Appeals for the Federal Circuit has upheld the invalidation of five patent claims covering Dental Monitoring’s artificial-intelligence-based dental image analysis technology, concluding that the claims were directed to an abstract idea under Section 101 of the U.S. Patent Act, 35 U.S.C. §101.

The July 7, 2026 decision in Dental Monitoring SAS v. Align Technology, Inc., No. 24-2270, concerned two patents covering the use of deep learning to analyze images of teeth and orthodontic aligners. The Federal Circuit affirmed the Northern District of California’s determination that the challenged claims were patent-ineligible. The opinion is nonprecedential.

The decision is significant beyond the dental industry. It provides another example of how U.S. courts are treating patent claims that apply established artificial intelligence or machine-learning techniques to a particular industry without claiming a sufficiently specific technological improvement.

What happened?

Dental Monitoring SAS sued Align Technology, Inc. in the Northern District of California in November 2022, alleging infringement of patents covering dental image analysis technology. The litigation involved Align Technology’s Invisalign ecosystem and remote-monitoring capabilities.

The patents at issue in the July 2026 Section 101 appeal were:

Patent	Broad subject matter	Claims at issue
U.S. Patent No. 11,049,248	Deep-learning analysis of orthodontic aligner fit	Claims 1 and 14
U.S. Patent No. 10,755,409	Deep-learning analysis and acquisition of dental-arch images	Claims 1, 7 and 12

The Federal Circuit affirmed the invalidity determination for all five challenged claims.

But the important point is what the court did not say.

It did not hold that artificial-intelligence patents are inherently unpatentable.

Instead, the court examined the specific claims and concluded that the claimed use of deep learning did not provide enough of a technological improvement to make those claims patent-eligible.

The technology behind the patents

The claims were more specific than simply saying “use artificial intelligence in dentistry.”

For example, claim 14 of U.S. Patent No. 11,049,248 described a learning system using a database containing more than 1,000 dental-arch images.

The system would use those images to train a deep-learning device and then analyze a new image to determine information concerning the relationship between a tooth and an orthodontic aligner.

The patent also described using a cellphone to acquire the dental image.

The other patent, U.S. Patent No. 10,755,409, similarly involved acquiring an image of a dental arch, analyzing that image with a deep-learning device, determining an image-attribute value, comparing that value with a setpoint, and communicating information that could guide the acquisition of another image.

These limitations made the claims look technologically specific. The Federal Circuit nevertheless found that they did not cross the Section 101 threshold.

Why did the Federal Circuit find the claims abstract?

The court applied the two-step patent-eligibility framework associated with the U.S. Supreme Court’s decisions in Alice Corp. v. CLS Bank International and Mayo Collaborative Services v. Prometheus Laboratories, Inc.

The first question is whether the claims are directed to a patent-ineligible concept, such as an abstract idea.

The second question is whether the claim contains an inventive concept sufficient to transform that abstract idea into a patent-eligible application.

For Dental Monitoring SAS, the Federal Circuit concluded that the claims were directed to abstract ideas and that the additional limitations did not supply the necessary inventive concept.

The court’s reasoning focused heavily on the deep-learning device.

The Federal Circuit found that the claimed deep-learning technology was essentially generic. The patent specifications themselves contemplated the use of neural networks developed by companies such as Google or Microsoft rather than describing a new neural-network architecture or a specific improvement to the operation of deep learning.

That became central to the court’s analysis.

The question was therefore not simply whether Dental Monitoring had created a useful system. The question was whether the patent claims captured a patent-eligible technological improvement.

The Federal Circuit concluded that they did not.

The 1,000-image detail is important

One of the more interesting parts of the decision is the treatment of the more-than-1,000-image training dataset.

At first glance, specifying a particular training dataset could appear to make an artificial-intelligence invention more concrete.

The Federal Circuit was not persuaded.

The court explained that training a machine-learning system using a particular set of data was part of the ordinary process of applying machine learning. In the court’s analysis, the claims did not explain a technological improvement arising from the particular dataset itself.

That creates an important distinction for artificial-intelligence patent drafting:

A patent claim can specify the data being used without necessarily claiming a new technological way of processing that data.

The latter is much more important to a Section 101 analysis.

A familiar Federal Circuit pattern: Recentive Analytics

The Dental Monitoring decision follows the Federal Circuit’s April 18, 2025 precedential decision involving Recentive Analytics, Inc. and Fox Corporation.

In Recentive Analytics, Inc. v. Fox Corporation, the Federal Circuit considered four patents covering the use of machine learning to generate television network maps and optimize schedules for live events.

The court held those claims patent-ineligible under Section 101.

The significance of Recentive Analytics, Inc. is that the company is not a technology or a type of artificial intelligence. It is simply the name of the plaintiff in that earlier patent case.

The 2025 case was important because it addressed the patent eligibility of machine-learning inventions in a direct way.

The Federal Circuit found that the patents essentially applied generic machine-learning techniques to new environments without claiming a sufficient technological improvement.

The court also made clear that machine learning itself is not excluded from patent protection. The problem was the way the particular inventions had been claimed.

That reasoning is relevant to Dental Monitoring.

The two cases involved completely different industries:

Federal Circuit case	Artificial-intelligence application	Result
Recentive Analytics, Inc. v. Fox Corporation, 2025	Machine learning for television scheduling and network maps	Patent claims found ineligible
Dental Monitoring SAS v. Align Technology, Inc., 2026	Deep learning for dental image analysis	Challenged claims found ineligible

The common thread is not the industry. It is the court’s focus on whether the claimed artificial intelligence implementation provides a technological improvement, rather than merely applying an existing machine-learning approach to a new field.

But Dental Monitoring is not just a patent concept

There is an important real-world dimension to this case.

The U.S. Food and Drug Administration granted DentalMonitoring a De Novo classification on May 17, 2024 under De Novo request DEN230035.

The device classification name is “dental image analyzer.”

The U.S. Food and Drug Administration classified DentalMonitoring as a Class II medical device.

This is important context, although it does not determine patent eligibility.

Regulatory approval and patent eligibility answer completely different questions.

The U.S. Food and Drug Administration is concerned with whether a medical device meets applicable regulatory requirements for its intended use.

The Federal Circuit was asking whether particular patent claims satisfy the legal requirements for patentable subject matter.

So a technology can be real, useful, clinically deployed and subject to medical-device regulation, while particular claims covering that technology can still fail Section 101.

The artificial intelligence has also been studied clinically

DentalMonitoring is not simply a theoretical system described in a patent.

A 2026 study published in Scientific Reports evaluated DentalMonitoring’s artificial intelligence for detecting orthodontic aligner-tracking issues using data from 1,014 patients in the United States.

The study examined three types of orthodontic conditions and reported the following performance:

Condition assessed	Sensitivity	Specificity
Bracket debonding	98.4%	99.6%
Open self-ligating clips	91.1%	88.3%
Tie loss	93.3%	96.5%

These figures demonstrate that the system has measurable performance in a clinical research setting.

But they do not change the Section 101 analysis.

Clinical accuracy is not the same legal test as patent eligibility.

The Federal Circuit was not deciding whether DentalMonitoring’s artificial intelligence works. It was deciding whether the particular claims before it constituted patent-eligible subject matter.

That distinction is critical.

The broader Dental Monitoring patent battle continues

The July 7 decision is also only one part of a much larger patent dispute between Dental Monitoring SAS and Align Technology, Inc.

The parties have been involved in multiple Federal Circuit proceedings involving overlapping patents and different patent-law issues.

This is important because the July 7 ruling should not be described as the Federal Circuit invalidating “all of Dental Monitoring’s patents.”

It did nothing of the sort.

The July 7 decision concerned five specific claims from two patents under Section 101.

Other Federal Circuit proceedings involving Dental Monitoring SAS and Align Technology, Inc. have addressed obviousness under Section 103, which is a different patentability requirement.

That distinction matters.

A patent claim can be challenged under several different provisions of U.S. patent law, including:

Section 101: whether the subject matter is eligible for patent protection;
Section 102: whether the invention is new;
Section 103: whether the invention would have been obvious;
Section 112: whether the patent adequately describes and enables the invention, among other requirements.

The Dental Monitoring litigation demonstrates how the same technology can face several different patentability questions.

What this means for artificial-intelligence patents

The practical lesson is not that companies should stop patenting artificial intelligence.

It is that claim drafting matters enormously.

A claim that effectively says:

“Use a deep-learning model to analyze dental images and produce a result”

may face a very different Section 101 analysis from a claim that identifies a specific technical mechanism that improves the operation of the artificial-intelligence system or the underlying computer technology.

The U.S. Patent and Trademark Office has similarly emphasized the importance of identifying technological improvements when evaluating computer-implemented and artificial-intelligence-related inventions.

The Federal Circuit’s 2025 decision involving Recentive Analytics, Inc. provides the strongest recent precedential warning: applying established machine-learning techniques to a new environment, without more, may not be enough.

The 2026 Dental Monitoring decision reinforces that point in a medical-technology setting.

The takeaway

The most important message from the case is simple:

Artificial intelligence is not the problem. A lack of claimed technological improvement can be.

Dental Monitoring’s technology uses deep learning to process real dental images. Its product has received a Class II medical-device classification from the U.S. Food and Drug Administration, and peer-reviewed research has evaluated its performance in more than 1,000 patients.

Yet those facts did not establish patent eligibility for the specific claims before the Federal Circuit.

For companies developing artificial-intelligence technologies, particularly in healthcare, the case highlights a critical patent strategy question:

What is technically new about the artificial-intelligence system, and does the patent claim actually capture that technical improvement?

That may increasingly matter as artificial intelligence moves from general-purpose software into medical imaging, diagnostics, drug discovery, robotics and other technology-intensive fields.

The July 7 ruling does not close the door on artificial-intelligence patents. It makes the door narrower for claims that rely primarily on applying generic machine-learning technology to a new field of use.

And for AI patent applicants, that distinction could be decisive.

SOURCES:

Author

Surbhi Ashri

Federal Circuit Finds Dental AI Patent Claims Abstract Under Section 101