UPC Weekly - Same parties, same prior art, same patent – different claim interpretations in the UPC and UK lead to different outcomes

2025 Week 47

We don’t often get to see the same patent being assessed in an EPO opposition, in the UK Patents Court and also at the Unified Patent Court. So we’ve lucked out with Advanced Cell Diagnostics v. Molecular Instruments, a decision on infringement and validity from The Hague Local Division of the UPC, coming after EPO opposition proceedings and recent UK litigation.

The two patents in issue in the case are the parent EP 1910572 B1 and divisional EP 2500439 B2. The invention is defined by method claims in the parent, while the divisional uses kit claims.

Before diving into the detail, the table below shows the different outcomes at the EPO, UK and UPC:

Technical background

The invention is a clever way of increasing the sensitivity of RNA detection in situ (in whole cells or tissue samples). By coupling a labelled probe to the target RNA via a pair of “capture probes”, instead of one, unwanted non-specific binding is greatly reduced. The labelled probe binds weakly to a single capture probe, but very strongly when both capture probes bind the target alongside each other. “Noise” is reduced because labels bound via single random, non-specific interactions are minimised.

The case turned on the interpretation of a single feature of the independent claims: The capture probes are defined as having sections that are complementary (and therefore bind) to non-overlapping regions of the labelled probe.

All the capture probes exemplified in the patent have sections that are complementary (and therefore bind) to adjacent regions of the labelled probe that are distinct and do not overlap with each other.

In contrast, the alleged infringing product from Molecular Instruments (MI) contains capture probes that are complementary (and therefore bind) to regions of the labelled probe that contain a small degree of overlap. In other words, there is a small region of the labelled probe that can be bound by one end of one of MI’s capture probes and by one end of the other MI capture probe. MI had found that this configuration enhances the subsequent signal generating steps.

The EPO opposition

An opposition was filed against the divisional (but not the parent). The opponent was Affymetrix, rather than MI. But the opposition proceedings fizzled out without a hearing. Affymetrix seemed to lose interest and the patent was maintained in an amended form, everything being decided in writing at the beginning of 2017.

The EPO opposition division considered various prior art attacks, including documents “Kern” (D5) and “Collins” (D6) which feature heavily in the later UK and UPC proceedings. For the EPO, the key point about Collins was that its relevant disclosure was presented as speculation amongst many other possibilities. The fact that Collins had been published about 8 years before the priority date of the patent was also taken as an indication of inventive step – no-one had in fact made the combination of Collins and Kern to provide a public disclosure falling inside the claims before the patent was filed.

The UK decision

In the UK Patents Court proceedings, the judge decided that a small degree of overlap did not alter the fact that the two capture probes nevertheless also bind regions of the labelled probe that do not overlap. After considering expert and experimental evidence, the judge decided that, although the skilled person would consider “completely non-overlapping” regions to be the simplest design, nothing in the patent would make them think some overlap should be avoided:

… the simplest approach would be to have no overlap (of this kind) at all, but that so long as there was enough non-overlapping sequence on each side of the label probe beyond the overlap to allow two stable duplexes to form, the invention would work and the key benefit would be achieved. (paragraph 273 of the reasons.)

The judge decided that the experimental evidence showed that there was some improvement for MI’s overlapping configuration compared to non-overlapping, at least at some concentrations. But this did not change his view that the “non-overlapping” feature should be given a purposive interpretation tied to the intended function of the feature.

Even though the court found infringement based on a normal interpretation of the claims, the decision briefly went on to consider infringement by equivalence. The UK doctrine of equivalents (DoE) is well known and sets out a series of three questions in the UK Supreme Court in Actavis v. Eli Lilly (pemetrexed litigation).

Having found infringement on the basis of normal interpretation, it’s unsurprising that the UK judge decided that there would also be infringement under DoE. Even though it was accepted that the alleged infringement provided an improvement, the judge considered that the core of the invention was not strict non-overlapping, but the formation of two stable duplexes. So the assessment got green lights on each of the three UK DoE questions.

Validity went the other way. This did not turn on the same point of claim interpretation. The judge considered that the skilled person would have a reasonable expectation of success when implementing the teaching of Collins combined with Kern.

The UPC decision

The UPC decision was reached on the basis of the same prior art and the same agreed product description for the alleged infringement. The decision was also based on the same agreed common general knowledge as in the UK case.

For the question of infringement, the key point in the UPC proceedings was the same issue of claim interpretation. But the UPC decided that “non-overlapping” should be given a strict meaning:

The patentee chose to use the restrictive term non-overlapping which has a specific meaning to the skilled person. Two sequences are either non-overlapping or they are overlapping, … (paragraph 38 of the reasons)

The UPC decision accepts that this is a literal interpretation of the claim and looks to the disclosure of the patents to confirm this.

This literal interpretation is confirmed when considering the teaching of the patent specifications. The function of the non-overlapping regions is to create a stable duplex at the core of the matter. There is no single example or passage in the description that allows for partial overlapping sequences for the binding regions… . Also, the figures of the patent are silent on this. To the contrary, in all figures there is no overlap whatsoever. (paragraph 38 of the reasons)

The court therefore considered infringement by equivalence. We need to bear in mind here that The Hague LD is the only division of the UPC that has laid its cards on the table with a proposed UPC DoE, based on the Dutch national approach with a sequence of four questions, which was set out also in the corresponding pemetrexed case. We reviewed this in UPC Weekly 2025 w24 (Where have we got to with the UPC doctrine of equivalents?).

In this decision, the LD does not explicitly step through their four DoE questions, but it seems clear that they consider at least the first two. While accepting that the first question (technical equivalence) might be satisfied, the court queries whether it would be fair (second question) to extend the scope of protection to include some overlap. This is on the basis that, although inventive, the patents are not ground-breaking. More substantively, the court also considered that the overlap is not a mere workaround of the claim, but provides an accepted function of its own. So there was no infringement by equivalence.

On validity, the key point was again inventive step. The parties aimed to use the “next step” approach to inventive step, but the court noted in practice that their arguments were closer to the problem solution approach. Having said that, the court noted that the same outcome would result for either approach.

The UPC decision seems to take a markedly different assessment of the disclosure of Collins compared with the UK case, and how Collins would be read with Kern. The result was that the LD decided that at least two modifications would be needed to Collins to reach the claimed invention, and that there was no motivation for the skilled person to make these changes. The was also passing reference to the time that had elapsed between Collins and the priority date, as a secondary indicator of inventive step, as in the EPO opposition.

The outcome was therefore that the patents are valid at the UPC, but not infringed.

Lessons learned?

The two significant differences between the UPC and the UK Patents Court were on claim interpretation and the assessment of the disclosure of the prior art. These are not easily ascribed to differences in legal approaches to these topics. Both courts are in principle applying the same law based on the European Patent Convention. Probably there were differences in expert evidence in the two cases, although the UPC decision does not rely heavily on this for its assessment of what the skilled person would understand from the wording of the claims.

The overall lesson is a tricky one for parties looking for legal certainty in European patent litigation. Patent disputes can be complex and it’s difficult to predict the outcome of finely-balanced cases. Reasonable judges can come to different conclusions on the basis of the same overall facts.

This blog was co-authored by Matthew Naylor and Eliot Ward.