- in China
- with readers working within the Advertising & Public Relations industries
The rapid advancement of artificial intelligence technology is reshaping the technical boundaries within the medical and healthcare techniques. From early machine learning algorithms to AI large models, healthcare is evolving toward greater precision, intelligence, and convenience. Today, AI is applied across multiple medical departments. AI-powered image interpretation can automatically identify lesions, achieving over 95% accuracy in common disease detection. AI surgical assistance systems enable intraoperative navigation and real-time analysis, providing surgeons with procedure recommendations based on current surgical conditions to enhance success rates and efficiency.
These innovative AI applications often focus on intelligent diagnosis and precision treatment. However, diagnostic and therapeutic methods are not patentable under China's Patent Law. Therefore, how to protect AI+healthcare innovations warrants careful consideration.
According to Article 25 of the Patent Law, "methods for diagnosing and treating diseases" are not eligible for patent protection. In examination practice, drugs used for disease treatment can typically employ Swiss-type claims to circumvent patentability issues. Medical devices for diagnostic or therapeutic purposes may generally obtain patent protection based on innovative mechanical structures or operational procedures. However, technical solutions directly implemented on human or animal bodies and aimed at obtaining disease diagnosis conclusions or achieving therapeutic effects are ineligible for patent protection.
In AI-assisted diagnosis, AI-powered medical image interpretation represents the most mature technical approach. Its conventional workflow involves: annotating medical image data—such as lesion locations and characteristics—from historical medical images. This processed data then serves as training data for neural network algorithms. Ultimately interpretation tasks like lesion detection and organ segmentation are enabled. The core of medical image interpretation lies in achieving AI-powered image recognition, whose fundamental principles align with AI image recognition in other technical fields. The distinction is that medical image interpretation ultimately outputs suggested diagnostic conclusions—specifically, whether a tissue structure exhibits abnormalities and whether the severity of such abnormalities constitutes a medical indicator.
In this process, AI system typically processes medical images as input and generates discriminative recommendations for specific regions within the images. This means that if an AI-powered medical image interpretation solution focuses solely on enhancing image recognition accuracy—such as refining neural network models to improve detection precision—it does not involve diagnosing the progression of the disease itself. Such a technical solution is patentable.
Conversely, if the AI-powered medical image interpretation solution focuses on providing physicians with reference diagnostic results—such as translating medical diagnostic consensus into computer-implementable protocols—it directly addresses the disease diagnosis method itself. This does not solve a technical problem but rather involves natural laws, falling outside the scope of patentability.
Compared to AI medical image interpretation, AI-assisted treatment system often proves more contentious. AI-assisted treatment system can be applied in medical devices such as surgical robots, minimally invasive equipment, and brain-computer interfaces. During the operation of these surgical devices, AI system relies on various sensors to collect human body data. Through neural network algorithms and similar methods, it identifies key data and patient status, subsequently providing physicians with reference information such as preoperative path planning, surgical operation guidance, tissue state recognition and treatment suggestion. Throughout this process, both the sensor data relied upon by the AI system and the final reference information provided originate from or are applied to the human body. Consequently, the overall technical solutions for AI-assisted treatment system are generally difficult to obtain patent protection.
Upon detailed examination of the formation process of AI-assisted treatment systems, it becomes evident that they resolve multiple technical challenges, such as neural signal processing and navigation algorithms for surgical robots. The solutions to these technical problems are not directly related to the treatment plan itself, as their final output remains technical information rather than therapeutic steps. Accordingly, an AI-assisted treatment system can be broken down into multiple distinct technical solutions to achieve patent protection.
In summary, whether for AI-assisted diagnosis or AI-assisted treatment, when seeking patent protection, the focus should be on identifying the technical information within the solution to form a technical solution under the framework of patent law. Then can a patent protection be obtained.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
