Developing An Effective Patenting Policy To Protect AI-Enabled Medical Devices
By Ryan N. Phelan, Marshall Gerstein & Borun LLP
Artificial Intelligence (AI) is revolutionizing various sectors, and the medical device industry is no exception. AI's role in software-based medical devices, ranging from diagnostics to treatment plans, is rapidly growing. As these innovations emerge, protecting intellectual property (IP), particularly through patents becomes critical. A well-defined AI policy is essential to safeguard these medical devices from infringement while encouraging further innovation.
This article explores how to develop an effective patenting policy for protecting AI-powered medical devices, focusing on the FDA's evolving stance, patent eligibility, and patent marking requirements.
Patenting Software-Based Medical Devices
FDA's Recognition of Software-Based Medical Devices
The U.S. FDA has increasingly acknowledged the significance of software-based medical devices, including those incorporating AI. As highlighted in the FDA’s Digital Health Innovation Action Plan (DHIAP), AI-driven health technologies are shaping the future of medical care, from mobile medical apps to AI-driven clinical decision support tools. The agency anticipates continued innovation in this space and recognizes the need for clear regulatory definitions and protections for AI-based medical devices.
The FDA classifies software-based medical devices into two main categories:
- Software in a Medical Device (SiMD): Software embedded within a medical device.
- Software as a Medical Device (SaMD): Software intended for medical purposes but functioning independently of any hardware device.
The FDA has categorized AI-enabled medical devices as types of SiMD innovations or SaMD device innovations. For example, the FDA keeps a list of AI-based medical devices that have met premarket requirements; there are approximately 950 such AI-based medical devices to date. See Artificial Intelligence and Machine Learning (AI/ML)-Enabled Medical Devices. These innovations are categorized as SaMD or SiMD type devices.
These categories also correlate with the patentability of AI-based medical devices. SiMD innovations, which incorporate physical components, are generally more likely to meet patent eligibility standards, whereas SaMD innovations, which rely on data processing, often face challenges under patent law.
Patent Eligibility of AI-Based Medical Devices
To protect AI-based medical devices through patents, companies must navigate complex legal frameworks that distinguish between eligible and ineligible inventions. The U.S. Patent and Trademark Office (USPTO) and courts have developed specific criteria for patent eligibility, particularly for software-driven technologies like AI.
SiMD innovations — due to their connection to physical devices — are typically seen as patent-eligible. The combination of software and hardware elements forms a more tangible innovation, which courts and the USPTO have repeatedly deemed patentable.
SaMD innovations, however, often face challenges with patent eligibility. These innovations frequently involve abstract processes, data processing, or medical algorithms, which courts have found insufficient to meet patent eligibility requirements. A key factor in securing patent protection for SaMD is demonstrating how the software provides a technological improvement, either by enhancing the performance of a computing device or improving the medical field.
Crafting AI Patent Claims
For companies developing AI-based medical devices, drafting patent claims requires strategic precision. An effective AI policy must ensure that patent claims emphasize the device's structural components or functional steps. In cases where a device lacks physical components, patent claims should focus on how the AI-based innovation improves an underlying computing device or enhances a medical procedure.
When drafting claims for medical devices involving AI, linking AI-driven processes to a specific machine or computing device is crucial. Including these elements helps demonstrate that the invention operates on a "particular machine" rather than a general-purpose computer, thereby reinforcing its patent eligibility.
USPTO 2024 Guidance on AI Patent Eligibility
The USPTO's 2024 Guidance Update on Patent Subject Matter Eligibility offers a more detailed framework for patenting AI-related inventions, including AI-based medical devices. Effective July 17, 2024, this update aims to clarify examination procedures for AI-related patent applications under 35 U.S.C. § 101.
The 2024 AI guidance includes several hypothetical examples (claims 47-49) to illustrate how AI-related patent claims will be analyzed for subject matter eligibility. These examples emphasize that non-technical claims — those that merely describe an AI model without explaining how it functions or improves a technology — are likely to face Section 101 rejections.
For medical device manufacturers, this means that an AI policy should incorporate the latest USPTO guidance. Specifically, AI patent claims should include detailed explanations of how the AI improves the functioning of a medical device or enhances a medical procedure. For example, explaining how an AI algorithm refines diagnostic accuracy or optimizes treatment plans can strengthen the case for patent eligibility.
Patent Marking For AI-Based Medical Devices
Importance of Patent Marking
Once a patent is granted, medical device manufacturers must ensure their products are appropriately marked. Patent marking provides constructive notice to potential infringers and allows patent holders to claim damages from the time of marking. Under U.S. law, patent holders can mark products with the patent number or provide virtual marking, typically a URL that links to patent information. The Federal Circuit’s ruling in Maxwell v. J. Baker Inc. emphasizes that consistent and accurate marking is necessary to inform the public about a product’s patent status.
Challenges of Marking AI-Based Medical Devices
Marking AI-based medical devices can be challenging, particularly for software-driven innovations. The nature of the device — whether it is external, implantable, cloud-based, or third-party hosted — determines the most appropriate marking approach. Some common challenges include:
- External Devices: If AI software is embedded in an external device, marking the wrong component can hinder the recovery of damages. The McKesson Automation Inc. v. Swisslog Holding AG case illustrates this issue. Here, marking a different product than the one practicing the patent invalidated claims for damages.
- Implantable Devices: Marking implantable medical devices is particularly difficult due to size and visibility constraints. In these cases, marking the packaging or accompanying documentation may suffice. The Global Traffic Technologies LLC v. Morgan case underscores the importance of evaluating the “character of the article” (e.g., where and how a medical device can physically be marked) when determining the appropriate marking method.
- Cloud-Based Medical Devices: Cloud-based AI software presents unique challenges for patent marking since there is no physical product to mark. Virtual marking is an effective solution, such as providing a URL linking to patent information. The case of Lexos Media IP LLC v. Jos. A. Bank Clothiers Inc. demonstrates that virtual marking (i.e., providing a link to a website that includes relevant patents per 35 U.S.C. § 287(a)) can be sufficient for software-driven products.
- Third-Party Devices: When AI software is installed on third-party devices, patentees should require third-party manufacturers to mark the hardware. If the software is installed via the cloud, marking within the software or its documentation is essential. In McKesson Automation Inc. v. Swisslog Italia S.P.A., the court clearly emphasized the need to mark hardware configured by the patented software.
Accordingly, an AI IP policy should include where and how to mark a specific medical device product.
Conclusion
As AI-based medical devices continue to transform the healthcare industry, it is crucial for companies to develop comprehensive AI policies that address patent protection. These policies should focus on drafting strong patent claims highlighting AI's technological improvements in medical devices. Companies must also stay current with evolving USPTO guidelines on AI patent eligibility and implement effective patent marking strategies to maximize the protection of their innovations.
By carefully considering these factors, medical device manufacturers can protect their AI-driven technologies from infringement while fostering further innovation in this rapidly growing field.
About The Author:
Ryan N. Phelan is a partner and patent lawyer with the Chicago-based intellectual property boutique law firm Marshall, Gerstein & Borun LLP. He may be reached at rphelan@marshallip.com.
DISCLAIMER: The information contained in this article is for informational purposes only and is not legal advice or a substitute for obtaining legal advice from an attorney. Views expressed are those of the author and are not to be attributed to Marshall, Gerstein & Borun LLP or any of its former, present, or future clients