In T 0655/24, Technical Board of Appeal 3.3.04 of 20 January 2026 considered when post-published evidence may be used to support an improved technical effect for inventive step. The decision is relevant because the Board expressly addressed the application of G 2/21 where the application as filed discloses a technical effect, but the proprietor relies on a later-demonstrated improvement of that effect.
Background
The case concerned European patent No. 2 943 507, relating to proteins having an Fc region, like antibodies. Claim 1 of the main request defined specific mutations in the human IgG1 heavy chain of the proteins. The claim also required that the Fc region does not bind to any Fcγ receptor and that plasma clearance deviates from the corresponding wild-type protein by no more than 10%.
For inventive step, the Board considered the FE variant disclosed in D10 and D21 to be the closest starting point. The claimed subject-matter differed by one additional modification (D265A), resulting in the novel FEA variant. The proprietor relied on reduced CD69 expression as evidence of reduced T-cell activation and thus reduced effector function compared with the FE or FES variants. It also referred to post-published documents D25 and D26.
Key findings of the decision
The Board first addressed whether the post-published data could be taken into account. Referring to G 2/21, it held that such evidence can only be considered where the technical effect it is intended to show is already credible from the application as filed. Importantly, the Board did not accept that an improvement of an effect is automatically encompassed by the original technical teaching merely because the underlying effect itself was disclosed.
The Board recognised that T 1989/19, T 2716/19 and T 840/22 had taken a different approach. It disagreed with an approach under which an improvement may be treated as implicitly derivable once the general effect is derivable. In the Board’s view, this could allow inventive step to be established solely by post-published evidence where the application disclosed only an effect already known from the prior art or common general knowledge. The Board considered that result to be in tension with G 2/21 and with the principle that the invention must have been made by the effective date.
Turning to the application as filed, the Board found that Figure 3A and Example 3 directly compared the FE and FES variants with the claimed FEA variant. The differences over a broad concentration range made it credible that decreased CD69 release, reflecting reduced effector function, was achieved by the FEA variant compared with FE or FES. Example 11 and Figure 13A did not raise serious doubts, because they lacked a direct comparison with FE or FES and a proper negative control.
The Board therefore formulated the objective technical problem as providing a protein having an improved Fc region that reduces T-cell activation as measured by CD69 expression while retaining plasma clearance comparable to the corresponding wild-type Fc region. The post-published data in D25 and D26 were not strictly necessary, but supported the effect already credibly disclosed in the application as filed.
Analysis and implications
The decision draws a clear distinction between reliance on a disclosed technical effect and reliance on an improved version of that effect. For applicants and proprietors, it underlines that post-published evidence may corroborate, but not create, the relevant improved effect. Where inventive step depends on an improvement over the closest prior art, the application as filed should contain a basis from which that improvement can be derived, preferably through a meaningful comparison.
On obviousness, the Board found that neither D10 nor D21 gave the skilled person a reasonable expectation that combining the FE mutations with D265A would provide the claimed functional profile. The same conclusion was drawn even under the less ambitious problem of providing an alternative Fc-region protein due to the . The claimed subject-matter therefore involved an inventive step, and the case was remitted with an order to maintain the patent on the basis of the main request.
Conclusion
T 0655/24 confirms that an improved technical effect must be supported by the application as filed before post-published evidence can be used to substantiate it. The decision is also notable for its explicit disagreement with earlier Board decisions that treated an improvement as implicitly derivable from a disclosed effect.
