Covering AI and Machine Learning Features in Medical Therapeutics and Diagnostics
Learn how pending patents can strengthen your IP portfolio, attract investors, and boost your company’s valuation. Build credibility and protect your innovations with Lin IP, LLC.
Artificial Intelligence (AI) and machine learning are transforming medical technology—from traditional mechanical devices to diagnostic imaging and pharmaceuticals. Innovators everywhere want to integrate AI features into their products and protect these devices with intellectual property. Here are strategies for crafting digital therapeutic patent applications involving AI and machine learning features.
Identifying New and Useful Features
The first step is to identify new and useful features that may contribute to revenue streams while monitoring public disclosures and competitor products. These features may be designed around, blocked, or further improved.
Next, assess whether these features primarily relate to:
Software generating therapeutic content, or
Therapeutic use of the generated content (e.g., a method of treatment).
This distinction may determine whether the patent application is classified in different technology areas at the USPTO. Applications classified as therapeutics may have a greater chance of overcoming patent eligibility issues. Supporting data showing the effectiveness of applying digital therapeutics is helpful for both eligibility and enablement. Once assessed, applications should be filed quickly around these improvements.
Drafting Patent Applications for AI and Machine Learning
When drafting applications, emphasize technical improvements in the following areas:
Input Data – How the data used to create the model is structured.
Model Training – How the model is trained, or whether it is applied to a new technology area.
Output Evaluation – How output data indicates patient characteristics or sets therapeutic dosage levels.
User Interface – How results are presented to users (visual, audio, or multi-modal).
Each improvement should be described in detail, and, when possible, supported with experimental data and charts.
Overcoming Patent Eligibility Rejections
If eligibility rejections arise, arguments can be made that:
These AI-driven features cannot be performed by a human, and
The improvements are “significantly more” than abstract concepts.
With a well-crafted application supported by evidence, applicants can obtain strong patent coverage for AI and machine learning-based digital therapeutics.
Case Examples
Vanda v. West – Therapeutics involving treatment with specific dosages based on obtained data were ruled patent eligible.
Exergen v. Kaz – Digital diagnostics using an unconventional method to detect core body temperature via a temporal artery radiation detector were ruled patent eligible.
Coincidentally, I examined and granted the patent at issue in this case while serving as a USPTO patent examiner.
View the patent
Conclusion
AI and machine learning are reshaping healthcare. Protecting these innovations requires a thoughtful IP strategy that highlights technical improvements, provides supporting data, and anticipates eligibility challenges. With careful planning and drafting, companies can secure robust patent protection for digital therapeutics and diagnostics.
For guidance on drafting AI-related medical patent applications, overcoming USPTO eligibility issues, or building a strong IP portfolio in therapeutics and diagnostics, contact Lin IP, LLC at office@liniplaw.com. Our experienced attorneys are ready to help safeguard your innovations.
