Antibodies have become one of the most successful classes of therapeutic agents, with hundreds of approved products across oncology, immunology, infectious disease, and other indications. For applicants seeking patent protection in Europe, however, the EPO's approach to inventive step can present a significant hurdle.
A further antibody against a known target is often regarded as obvious unless it can be shown to possess an unexpected technical property or advantage. Unlike the more common approach to small-molecule inventions, the EPO generally does not consider structural (sequence) differences to contribute significantly to the inventive step of an antibody.
This approach, which is set out in Section G-II, 6.2 of the EPO Guidelines for Examination, reflects the EPO’s view that antibody generation is a relatively mature field and that obtaining antibodies binding to a known target is routine. As a result, obtaining patent protection may prove challenging even for a specific lead clinical antibody. Furthermore, applicants are frequently required to define the claimed antibody by detailed sequence features, often extending beyond the complementarity-determining regions (CDRs) to include framework-region residues.
Advances in antibody engineering have considerably expanded the therapeutic applications of antibody-derived binding domains beyond standalone antibody products. Antibody-derived binding domains are now incorporated into a variety of more complex therapeutic modalities, including multispecific antibodies, antibody-drug conjugates (ADCs), and chimeric antigen receptor (CAR) therapies. Although successful commercial products already exist within each of these classes, the corresponding case law on inventive step remains relatively limited.
The question therefore arises whether these modalities will be assessed by analogy with conventional antibody inventions or under a less restrictive approach. Based on recent Board of Appeal decisions, this article provides suggestions for avoiding the restrictive inventive step framework traditionally applied to antibody inventions when patenting next-generation antibody-derived therapeutics.
The choice of closest prior art
The EPO assesses inventive step using the problem-solution approach, the first step of which is to identify the closest prior art. A common examination approach is to start from an antibody directed to the same target as the closest prior art, and to treat the overall architecture of a CAR, ADC, or multispecific antibody as part of the common general knowledge that can be implemented using routine methods.
However, a monoclonal antibody directed to the same target is not necessarily the most appropriate closest prior art for a claim relating to a CAR, ADC, or multispecific antibody. These modalities represent distinct therapeutic architectures involving complex additional design considerations and are often more difficult to develop compared with traditional monoclonal antibodies. Rather than focusing solely on the binding domain considered in isolation from the other components, this entire context should play a role in the selection of a suitable closest prior art document.
It is interesting to note that the Unified Patent Court (UPC) has developed an approach to the assessment of inventive step that has been described as more ‘holistic’. This was recently articulated by the Court of Appeal in Amgen v Sanofi and Meril v Edwards (both issued on November 25 2025). Unlike the EPO’s problem-solution approach, which starts from the closest prior art and derives an objective technical problem from specific distinguishing features, the UPC first determines the objective technical problem and only then considers the starting point.
The Court of Appeal emphasised that the objective problem should not be defined by dissecting individual distinguishing features in isolation but by assessing the claimed invention as a whole in its technical context. This may reduce the risk that inventive-step analysis becomes focused on an individual binding domain while losing sight of the claimed molecule as a whole.
Similar considerations in fact exist at the EPO. Indeed, one of the central considerations in determining the closest prior art at the EPO is that it must be directed to a similar purpose or effect as the invention. As set out in the Case Law of the Boards of Appeal (CLBA), the aim is that the assessment should start from a situation as close as possible to that encountered by the inventor, avoiding ex post facto considerations, and taking into account the real-world circumstances (see CLBA, 11th edition, Section I.D.3.4). These considerations lean closely towards those expressed by the UPC.
A recent Board of Appeal decision dealing with inventive step of a CAR illustrates that a related CAR can be the closest prior art, even when it carries an unrelated type of binding domain and when binding domains as claimed were already described in the art. In T 203/22, an anti-B7-H6 CAR was claimed without any sequence limitation. B7-H6 is a tumour-associated ligand for NKp30. Even though anti-B7-H6 antibodies were known in the prior art, the board assessed the claimed invention starting from NKp30 CAR-T cells, which use B7-H6’s natural receptor instead of an antibody-derived binding domain as a targeting moiety. The anti-B7-H6 antibodies were treated as secondary references rather than the starting point.
Interestingly, the patent in T 203/22 also claimed anti-B7-H6 binding domains in the context of an anti-B7-H6/anti-CD3 bispecific T-cell engager (BiTE). Inventive step of these claims was assessed separately, and for those an anti-B7-H6 antibody was taken as the closest prior art. The claims nevertheless survived. Although the BiTE format was commonly known, the board found that the person skilled in the art would not have considered approaches based on a different mode of action, such as recruitment of T cells, and therefore would not have arrived at the BiTE format.
T 203/22 illustrates that next-generation antibody-derived therapeutics need not always be assessed starting from a known antibody against the same target. Depending on the invention, a related CAR, ADC, or multispecific construct may provide a more realistic starting point. Even where the closest prior art is a known antibody, inventive step may still reside in its incorporation into a more complex therapeutic format, in particular when a different mode of action is involved.
Considering the properties of the molecule as a whole
The inventive contribution of antibody-derived therapeutics often resides in the structure and function of the molecule as a whole. When inventive step can be shown to not solely depend on the advantageous properties of the binding domain, this may have the additional benefit that there is no need to introduce CDR or entire variable region sequences into the claim.
Under the EPO’s established ‘could-would’ approach, the relevant question is whether the prior art would have motivated the skilled person to arrive at the claimed construct with a reasonable expectation of achieving the relied-upon technical effect. Thus, even where CAR, ADC, or multispecific formats form part of the common general knowledge, inventive step may still be present if the prior art provides no reason to expect the resulting functional properties.
T 203/22 is a useful confirmation of this established case law in the specific context of CAR constructs. The technical contribution resided in the specific CAR design and functionality, rather than merely in identifying another binding domain against a known antigen. The board acknowledged that anti-B7-H6 antibodies could be assumed to be suitable for constructing a CAR, and that the structure of CARs was common general knowledge. Nevertheless, it held that the skilled person would not have expected the improved properties observed with anti-B7-H6 CAR T cells compared with NKp30 CAR-T cells, and the claims were found inventive.
Also of interest is T 957/18 concerning a bispecific anti-CD20/anti-CD47 antibody. Although the prior art disclosed combination therapy with separate anti-CD20 and anti-CD47 antibodies, the board did not accept the Examining Division’s position that the skilled person would generally prefer a single medicament over administration of two separate antibodies. Instead, inventive step was assessed at the level of the overall molecule. The board found that the prior art did not suggest that the bispecific format would reduce toxicity through decreased binding to healthy CD47-expressing cells. The claims were therefore held inventive. Thus, like T 203/22, T 957/18 is an example of the boards recognising inventive step based on the functional properties of the architecture as a whole.
In T 1019/22, involving bispecific nanobody technology, the board considered the claimed bispecific construct as an integrated molecule and assessed whether the prior art provided motivation to arrive at the claimed combination. Inventiveness did not turn exclusively on the properties of the individual binding domains, and there was no requirement to incorporate CDR or variable region sequences.
Together, T 957/18, T 203/22, and T 1019/22 suggest that boards are willing to focus on the technical contribution of the molecule as a whole rather than reducing the analysis to individual binding domains.
Key takeaways
Antibody-derived therapeutics are expected to grow further in both clinical and commercial importance, meaning that the Boards of Appeal will increasingly be required to address inventive-step questions relating to CARs, ADCs, multispecific antibodies, and other next-generation formats.
Although the case law remains limited, decisions such as T 957/18, T 203/22, and T 1019/22 suggest that these constructs need not necessarily be analysed through the same lens as conventional antibodies directed to known targets. Meanwhile, patent litigation concerning complex antibody-derived therapeutics has also reached the UPC, with an anti-BCMA CAR-T dispute pending before the Brussels Local Division.
As more cases involving advanced biologics come before the EPO and UPC, the coming years should provide insight into whether, and to what extent, these technologies will be accommodated within or distinguished from the EPO’s established inventive-step framework for conventional antibodies.
