Patent Trends for AI in MedTech
As part of our spotlight on AI in MedTech, we had a look at the data surrounding the patents being filed in this space. That data reveals a number of interesting stories around the patents being filed and for this report, we are going to focus on two of those stories.
For a discussion on how we have classified a patent or patent application as falling within the AI in MedTech space, please see the section on how the data was obtained.
General overview
Before we delve into the details of the filings in this space, it is worth taking a look at the general filing trends over the recent years. Figure 1 shows two plots, the first of which represents the number of published European patent applications (grouped by patent family, i.e., multiple publications in the same family are not counted separately) each year. The other plot represents the number of those published European applications that have been granted (again, grouped by family). Figure 2 shows the same plots as Figure 1, but including publications and granted patents in the United States of America, China, the Republic of Korea and Japan. The plots in Figures 1 and 2 are for publications which have both an A61 and a G06N classification code, which is the combination generally assigned to inventions involving AI in MedTech.
As shown in Figures 1 and 2, the AI in MedTech space is a relatively young sector that is currently undergoing a period of rapid growth. The data for Figures 1 and 2 was collected in July 2024. To provide some context, AI & MedTech (A61 and G06N) publications made up about 2.5% of the total AI (G06N) publications, again highlighting that this is a relatively young sector with plenty of room for innovation and growth.
As expected, the number of grants lags behind the number of publications, particularly in Europe (see Figure 1) where the route to grant can take longer than other jurisdictions. Looking to Figure 2, it can be seen that trends in the grant numbers roughly follows that of the publication numbers.
The apparent dip in grant numbers in 2022 and 2023 should not be alarming to prospective applicants in this space. A similar dip was seen in the numbers of applications granted by the EPO for applications from all countries across all technology fields and recent statistics point to a recovery in the positive trend of grants from the EPO.
Small applicants
The development and sale of physical medical devices tends to have a relatively high barrier to entry. This can be due to the company, and the medical device being developed, needing to reach certain regulatory requirements, as well as the often significant upfront costs of developing, testing and making the products themselves.
The high barrier to entry can present a challenge to SMEs or start-ups trying to break into this space. For more insights on the patent trends for medical devices, we have previously published a Medical Devices report, which can be found here.
The same barriers to entry do not necessarily apply to software and AI being developed for use in the MedTech field. Often, the largest barrier to entry is obtaining the data, often patient data, needed to train an algorithm.
It is all very well hypothesising about the barriers to entry in the field of MedTech/AI, but what does the data show?
Figure 3 shows a pie chart illustrating the distribution of applicants in the MedTech/AI sector based on the number of published European applications belonging to each applicant.
Jurisdictional differences
When it comes to protecting AI-based inventions, or any software based invention, there can be a large variance in the approaches taken by different national patent offices. Unsurprisingly, the same is true for inventions within the AI in MedTech space.
Figure 7 shows a waterfall chart illustrating the distribution of national/regional phase entries stemming from PCT applications (between 2003 – 2024).
Figure 7: Distribution of national/regional phases for PCT applications
It is clear to see that the United States of America is by far the most popular jurisdiction to enter into for applications in the AI in MedTech space, followed at some distance by Europe, with roughly half the number of regional phase entries compared to the US.
“When it comes to protecting AI inventions in the medtech space, it’s not surprising to see that the EPO is a popular jurisdiction. Here in Europe, case law relating to these type of inventions is stable and well tested, meaning that we are able to offer good predictions as to what may or may not be patent eligible”
Emma Graham
Partner, Patent Attorney
How was the data obtained?
The data was obtained by performing a classification search using international patent classification (IPC) codes. The IPC codes of interest in this report are G06N - Computing arrangements based on specific computational models, which is the code assigned to most AI-based inventions and A61 – Medical or veterinary science, which is the code related to medical science. For the purposes of this report, a patent or a patent application was considered to be relevant to AI in MedTech if it received both an A61 and a G06N classification code.
"The field of medical technology is a well established one, but the recent rapid development in the capabilities of AI has opened an exciting new avenue for the development of life-saving technologies."
LET'S TALK
Callum Anderson - Associate, Patent Attorney
callum.anderson@mewburn.com
website profile
© 2025 Mewburn Ellis
