T 1418/22: Acalabrutinib provides guidance on establishing inventiveness of polymorphs at the EPO

The EPO Board of Appeal’s decision in T 1418/22 gives useful guidance on the inventiveness of polymorph claims in Europe. Starting from a known amorphous form, a specific crystalline form was found inventive considering: (1) data in the application demonstrated the “technical difficulty” of crystallising this polymorph; and (2) the crystalline form exhibited unexpected non-hygroscopicity. This follows a common reasoning followed at the EPO that a combination of beneficial properties is non-obvious as discussed in our recent article at point 1.

The patent relates to acalabrutinib (Calquence) licensed for the treatment of non-Hodgkin lymphoma. Amorphous acalabrutinib was known in the prior art, with the claims of the patent relating to polymorphic Form I.

When considering the objective problem to be solved, the Board confirmed that, in line with the European Pharmacopoeia classification, the data in the application demonstrated that the claimed Form I was non-hygroscopic (rather than “less hygroscopic,” as previously argued by the opponent-appellant).  The Board therefore formulated the problem to be the provision of a more stable and non-hygroscopic form of acalabrutinib.

The opponent-appellant cited D6 and D9 to assert that it was common general knowledge that crystalline forms tend to absorb less water and are more stable than amorphous forms, and that when starting from an amorphous form the skilled person would routinely screen for polymorphs.

The Board acknowledged that it would be advisable for the skilled person to screen for polymorphs early in drug development and that they would be familiar with routine crystallisation methods for screening polymorphs from a variety of solvents under different conditions.  However, the Board considered that the mere fact that prior art suggests screening for polymorphs to isolate the crystalline form with the most desirable properties is not, by itself, sufficient to render a specific polymorphic form with a particular desired property obvious.

While common general knowledge may suggest that crystalline forms are typically less hygroscopic, the Board emphasized that no indication in the cited art pointed to the existence of polymorphs with no hygroscopicity, as classified by the European Pharmacopoeia.

Furthermore, the Board placed significant weight on the data in the application showing that only 4 out of 75 screened solvents yielded Form I, and that anti-solvent experiments did not lead to Form I either.  This demonstrated the technical difficulty of crystallising Form I.

Practice points

The Board discussed that the relative difficulty in finding the claimed polymorph contributed to its inventiveness, being formed in a relatively low number of tested conditions. This highlights that there can be a tension between inventive step and sufficiency of disclosure to be considered when filing European polymorph applications. A polymorph which is hard to find may be less likely to be considered obvious, but reproducing its synthesis based on the limited textual disclosure of a patent application may be challenging. Detailed disclosure of the conditions used to screen for polymorphs (beyond just the solvent, such as concentration, temperature, etc.) is therefore always advisable.

This decision also highlights that there can be considerable importance to how technical effects are characterised by the parties, in this case the relatively subtle distinction between reduced hygroscopicity and the polymorph being non-hygroscopic was significant. Whilst reduced hygroscopicity may have been obvious in light of the common general knowledge, the lack of hygroscopicity of the polymorph was a surprising technical effect.

The case also shows one way that a patentee may stave off attacks based on post-filed data relating to polymorphs. In a late-filed document submitted by the opponent-appellant, samples of starting material were characterised by XRPD spectra as being amorphous. The proprietor argued that admission of this document into the proceedings would increase the complexity of the case because it needed to be discussed whether the amorphous nature could be inferred from the data or whether it might include some crystalline material. The Board didn’t admit the document, stating that discussion of this point would raise complex issues and be detrimental to procedural economy. Given the inherent challenge of fully characterising a material as wholly amorphous, this could be a useful argument for patentees to have to hand generally.

For any questions relating to the above, please contact the authors, Harry Gregson or Robert Scanes.