Great news for the Digital Health sector and developers of AI-supported diagnostic systems! In a recent landmark decision (X ZB 5/25), the German Federal Court of Justice (BGH) clarified that computer-implemented methods for analyzing medical image data often do not fall under the patentability exclusion for diagnostic methods.

⚖️ The Case: Automated Detection of Findings

The proceedings concerned a system for the automated determination of examination results in image sequences (e.g., angiographies). The method uses graph theory to track and evaluate “finding candidates” (such as vascular constrictions) across consecutive frames.

The Federal Patent Court had initially rejected the application, arguing that the method ultimately serves to diagnose pathological conditions and relies on data collected from the body, thus making it an unpatentable diagnostic method.

🔍 The Ruling: The “All-or-Nothing” Principle

The BGH has now overturned this decision, confirming a narrow interpretation of the exclusion criteria under § 2a (1) No. 2 of the German Patent Act (PatG).

Key takeaways from the decision include:

• The Completeness Doctrine: A patent exclusion only applies if all essential steps of a diagnostic method (data collection, comparison with norm values, identification of a deviation, and the final medical interpretation) are performed on the human or animal body.
• Physical Presence Required: A procedural step is only considered to be performed “on the body” if it involves an interaction that necessarily requires the presence of the patient.
• Data Processing is Patentable: Purely technical steps that operate on previously collected data (such as image sequences) do not take place “on the body”—even if they prepare or support a later diagnosis.

With this ruling, the BGH aligns itself with the restrictive line of the EPO’s Enlarged Board of Appeal (G 1/04), ensuring legal certainty in harmony with European practice.

💡Practical Tips for Patent Strategy

What does this mean for your patent applications?

1. Optimize Claim Drafting: Ensure that the independent claims do not include steps that strictly require the patient’s presence (e.g., “capturing an X-ray image”). Instead, define the process as a computer-implemented method that operates on provided data.
2. Focus on Technical Analysis: Emphasize the algorithmic evaluation, such as the generation of graphs, nodes, and “communities” of image areas as seen in this case. This highlights the technical character of the invention rather than the medical diagnosis itself.
3. Be Careful with Purpose Limitation: While the description should highlight medical utility (e.g., automated stenosis detection), the claim must not be so narrow that it only covers the final clinical decision-making phase (“deductive medical decision phase”).
4. Leverage Cloud Solutions: The BGH noted that execution via a cloud infrastructure can be an indicator against the method being performed “on the body”, as the location of data collection and the location of evaluation are separated.

Conclusion: The path for protecting innovative software solutions in radiology, cardiology, and beyond is clearer than ever. By drafting claims strategically, innovators can successfully navigate the diagnostic method exclusion.