Personalised medicine and patent protection: The key issues so far

It is critical that the emerging field of personalised medicine receives appropriate patent protection, explain Joel Coles and Theo Cooper of Powell Gilbert.

Personalised medicine (PM) or ‘precision medicine’ is a major growth area within the life sciences sector, with an increasing influence on how treatments are being developed and delivered.

Indeed, looking for genetic markers to predict disease occurrence (for example, BRCA1 mutations indicative of increased risk of breast cancer), or sensitivity to a particular therapeutic (for example, HER2 overexpression indicative of benefiting from trastuzamab treatment) is already common in clinical practice.

Similarly, chimaeric antigen receptor (CAR) T-cell therapies such as Kymriah and Yescarta, which involve reprogramming a patient’s own cells to fight disease, have been approved for use in treating certain cancers.

Last year, the UK’s Medicines and Healthcare products Regulatory Agency (MHRA) became the first drug safety regulator to pilot its own genetic ‘biobank’ with a view to informing approaches to PM, and investment from both the public and private sectors continues to grow in PM applications across a broad range of fields.

In order to maximise growth in this relatively new area, it is critical that innovation in the field is afforded appropriate patent protection to ensure that businesses are incentivised to invest in research and development—-as we have previously seen in the fields of biotechnology and medical devices.

What is personalised medicine?

PM broadly refers to the delivery of therapeutics according to an individual’s response/risk profile, for example by reference to genetic or molecular/cellular biomarkers.

By identifying a subpopulation of patients most suitable for a therapeutic, PM can offer a more precise means of delivering treatment than the traditional ‘one-size-fits all’ model of pairing patients with the indication on a therapeutic drug’s label. Some modern applications of PM go further—targeting a patient’s own cells and affecting genetics to obtain therapeutic results. PM inventions therefore often consist of a combination of diagnostic and treatment elements.

Several issues arise when patenting PM inventions:

• Commonly, PM applications do not present a traditional target for patent protection, such as a compound claim. A particular challenge concerning claims for PM applications is that the individual features (ie, the diagnostic and treatment elements) will often already be known in the prior art—the invention comes from the combination of these features.
• The fact that PM inventions often involve a combination of diagnostic and treatment elements creates enforcement questions concerning territoriality and identifying infringers, which will depend upon the scope of the patent claims in issue and the specific factual circumstances.
• PM inventions may include features which are excluded from patentability under the European Patent Convention (EPC), such as software elements and methods of treatment.
• Given the importance of machine-learning algorithms and data-driven approaches to PM development, issues of rights ownership may arise.The validity and infringement of PM inventions remain relatively untested in national courts and the Unified Patent Court, creating uncertainty over how patentees will fare in asserting such patents, and whether adequate relief is available if they are successful.

Current approaches to these issues and possible future challenges are explored further below.

Patient subcategories

Many PM inventions seek to target a newly-identified subpopulation of patients most suitable for an existing therapeutic, meaning that neither the therapeutic itself nor the labelled indication are the invention.

Moreover, patients receiving treatment according to those inventions may have received the same therapeutic absent invention, which creates prima facie concerns regarding novelty and non-arbitrariness of identifying the population for treatment.

This issue was considered by the European Patent Office's (EPO) Board of Appeal (TBA) in 2019, in decision T0694/16.  That particular case  looked at the composition for use in preventing or treating dementia in a person having “characteristics” of a patient with early-stage dementia.

Those ‘characteristics’ were the presence of certain markers in the cerebrospinal fluid (CSF).  In overturning the Opposition Division’s finding that the claim was not novel, the TBA held that: “the issue of whether patients displaying the markers of claim 1 were present among a population of previously treated patients and were already ‘inevitably’ or ‘inherently’ treated is irrelevant for assessing novelty in the present case.

“The only thing which counts is that D1 and D3 do not disclose a method whereby a patient or a group of patients displaying the relevant CSF markers but not affected by dementia was purposively and selectively targeted for carrying out the preventive treatment defined in claim 1.”

Therefore, it would appear that a claim relating to an existing drug for an existing indication—conditional upon detecting a newly-established biomarker—is an acceptable form in which to protect such PM inventions according to the EPO. Nonetheless, the validity of claims in this form is yet to be tested in many European national courts, including the UK.

Innovators in PM may also seek to claim aspects of the software, algorithm or other computerised process used to identify patients or in the process of delivering a therapeutic. While such features can successfully be incorporated into claims to a PM invention, additional issues surrounding excluded subject matter under the EPC, establishing infringement and IP ownership may arise, which are explored further below.

Gene therapies

As described above, another area of PM involves reprogramming a patient’s cells to fight a disease, as in the case of CAR T-cell therapies. In such a scenario, and given that the ‘product’ is effectively the patient’s own cells, claims are unlikely to target the product’s entire structure or sequence (as in the case of small molecule pharmaceuticals or monoclonal antibodies).

Instead, developers of these therapies may claim protection by referring to the modifications that are made to the cells, or to the nucleic acids encoding the receptors that are inserted for recognising the target molecule.

However, where a claim potentially covers a wide range of therapies or target sequences, patentees should be wary of claiming a broader range than the patent enables, given the Supreme Court’s finding in Regeneron v Kymab  (2020) UKSC 27 that a patent must be enabled across the whole scope of a ‘relevant range’ (being one significantly affecting the value or utility of the product in achieving its purpose).

In addition, aspects of the manufacturing process, methods of treatment, synergistic combinations with other therapeutics and dosing regimens represent potentially patentable aspects of such therapies. Similar considerations are likely to apply to new therapies developed using CRISPR/Cas9 technology, which enables targeted gene-editing within a patient’s cells. The first therapy using this Nobel Prize-winning tool, Casgevy, received MHRA approval late last year.

Litigation relating to CAR T-cell therapies has been reported in the US in recent years (eg, Juno Therapeutics et al v Kite Pharma). Similarly, the ownership of valuable patents in the CRISPR space is the subject of long-running and well-publicised litigation.

This recent activity highlights the importance of adequate patent protection for technologies at the cutting edge of PM, and could signal a growing potential for patent disputes in this area.

Data-driven approaches

Artificial intelligence (AI) is a tool with a growing influence on how PM is being developed.  Machine-learning algorithms have the potential to assist in identifying and validating the biomarkers that PM inventions are rooted in, for example by detecting patterns in the genomes of patients and their responses to therapeutics based on clinical trial data.  Inevitably, businesses investing in this area will expect protection in the tools they have developed, but the patenting of machine-learning algorithms is a difficult and fast-evolving area.

New UK Intellectual Property Office guidance for patenting AI inventions was published in September 2022, which has since been suspended following the High Court’s ruling in Emotional Perception AI v Comptroller-General [2023] EWHC 2948  last November.

Issues of IP ownership may also arise when machine-learning is involved in the research and development process. This has been the subject of much discussion in the aftermath of the Supreme Court’s recent decision in Thaler v Comptroller-General (2023) UKSC 49, which held that an AI could not be named as the inventor of a patent.

Given the considerable uncertainty in this area, use of externally-developed or performed AI tools should be approached with caution, and the provenance of any data used to train machine-learning algorithms (even those developed in-house) should be carefully considered to avoid ownership disputes.

Excluded subject matter

Patentees seeking to protect a diagnostic feature in their PM inventions should be wary of the EPC’s exclusion from patentability of “methods for treatment of the human or animal body by surgery or therapy and diagnostic methods practised on the human or animal body” (Art.53(c) EPC). However, products for use in methods of diagnosis/treatment are specifically carved out of the exclusion, making use-limited product claims an attractive option. Additionally, according to EPO case law the exclusion applies only to those methods practised directly on the body, i.e. requiring its presence, and not to methods including an in vitro step (such as testing a biological sample).

Similarly, for claims including a computing or software element, the EPC’s exclusion from patentability of “programs for computers” may apply (Art.52(2)(c) EPC). In order to avoid this exclusion, which is applied narrowly by the EPO, patentees seeking to protect a software/computing feature of their PM invention might consider formulating their claim as a computer-implemented method.

Establishing infringement

The combination of diagnostic and treatment features in many PM inventions raises several possible issues for patentees seeking to assert their patents against infringers. First, the diagnostic element may require different expertise to the treatment provider.  Many modern therapeutics are now accompanied by companion diagnostics devices, which provide information on safety or efficacy, before or during treatment.

However, performance of the diagnostic step may be outsourced, and may even be arranged directly by the treating clinician.

A competing business may therefore seek to escape infringement on the basis that they are not performing the invention. Moreover, questions of territoriality arise: does infringement take place where the patient is treated, or where detection of the biomarker (and therefore arguably the assessment of a patient’s suitability for treatment) takes place, if they are different? The answer in each case will turn on the specific facts of the scenario in issue, in the context of the inventive core of the relevant patent and the proper construction of its claims.

Relief available

In the event of infringement being proven, questions of the appropriateness, or otherwise, of the available relief arise.  Issues such as the availability of springboard relief may need to be considered and whether, in view of the public interest, damages may be a more appropriate alternative to the grant of an injunction in certain scenarios.

In the case of infringing second medical use patents, ‘skinny labelling’ to carve out an infringing indication from the label of an otherwise non-infringing product may be sufficient to avoid infringement, according to obiter comments by the Supreme Court in Warner-Lambert v Generics [2018] UKSC 56. Given that a PM invention may involve treating a new sub-population of existing patient groups, with the same product and for the same indication, the discussion of how to avoid infringing use of a non-infringing product outside the newly-identified patient subpopulation may be reopened. This remains an area of relative uncertainty and could add an additional layer of complexity to what is already a challenging area for the pharmaceutical industry to navigate.

Conclusion

At present, our patent system appears to be coping with the arrival of PM and the challenges it presents. Pharmaceutical and biotechnology businesses innovating in the field are finding ways to protect their PM inventions within the existing framework of patentability.  However, further challenges lie ahead as more patents are filed in the field and the solidity of those rights are tested at the national level.  As the area continues to mature, we could see further developments to current approaches to patentability and enforcement, as well as consideration of the most appropriate forms of relief available in cases relating to PM inventions.

Joel Coles is a partner at Powell Gilbert. He can be contacted at: joel.coles@powellgilbert.com

Theo Cooper is an associate at Powell Gilbert.  He can be contacted at theo.cooper@powellgilbert.com