Moderna’s melanoma result highlights a widening gap between how oncology creates clinical value and how medicines are owned, evidenced and paid for.

A cancer vaccine – and a joint effect

There is something inherently arresting about the idea of vaccinating against cancer. On 19 August, Merck (MSD) and Moderna announced positive topline results from the Phase III INTerpath-001 trial of intismeran autogene plus pembrolizumab in patients with completely resected high-risk melanoma. The combination significantly improved both recurrence-free survival and distant metastasis-free survival compared with pembrolizumab alone. Detailed Phase III results have yet to be presented and overall survival follow-up continues, but this is the first positive Phase III readout for an individualised neoantigen therapy and an important milestone for mRNA-based cancer treatment.

The setting matters. These patients have undergone complete surgical resection; the aim is to prevent recurrence by training the immune system to recognise neoantigens derived from the unique mutational profile of an individual patient’s tumour. Pembrolizumab complements this by blocking PD-1-mediated immune suppression.

These are not even radically different therapeutic modalities. Both intervene in the antitumour immune response: one seeks to prime and direct it, the other to release a brake upon it. Yet the clinically relevant effect emerges from their interaction.

That raises a question extending well beyond cancer vaccines:

What happens when the value of cancer treatment increasingly resides not in an individual asset, but in what several assets achieve together?

When value is created between components

Combination treatment is hardly new to oncology. Nor is the thoughtful design of regimens according to mechanism, sequence, disease stage or patient characteristics. Presenting either as a new discovery would rightly irritate oncologists.

What is changing is the range and specificity of the components now available. Targeted small molecules, checkpoint inhibitors, ADCs, multispecific antibodies, radioligands, cell therapies and individualised neoantigen approaches allow us to intervene in cancer biology through an increasingly diverse set of mechanisms and therapeutic principles. At the same time, tumour heterogeneity, phenotypic plasticity and acquired resistance remind us that the biological system we are treating is itself dynamic.

The important development, therefore, is not simply that we can combine more treatments. It is that more clinical value may be generated through interactions between increasingly specific components.

This is where the idea of a therapeutic architecture becomes useful.

I do not mean another name for a regimen. An architecture includes the components, but also the interfaces between them and the context in which they operate: sequence, timing, duration, biomarker gating, disease stage, patient selection and therapeutic backbone. In complex systems, important properties often arise at the interfaces rather than residing completely within any single component. Oncology increasingly seems to present a similar problem:

We own the components. But who owns the interface?

The asset remains essential – but it may no longer be sufficient

Pharmaceutical companies are understandably built around assets. Assets sit in pipelines, carry IP, receive investment, have development plans and forecasts, and ultimately generate revenues. HTA and pricing processes likewise require identifiable interventions against which incremental benefit and cost can be assessed. None of that is going away. But the asset may no longer be a sufficient unit through which to understand value.

If Drug A generates a modest benefit alone and Drug B does the same, but their interaction produces something clinically much more important, part of the value resides neither neatly in A nor in B. It resides in what happens between them. One component may serve as the backbone, another may unlock activity, prevent escape or make a response more durable. Neither its scientific contribution nor its economic value can necessarily be understood in isolation.

This creates a structural mismatch: clinical value can be jointly produced by assets that remain separately owned, developed, priced and commercially accountable. That is where the issue becomes more than an interesting feature of cancer biology.

Who created the benefit?

Regulators are already confronting the attribution question. In July 2025, the FDA issued draft guidance specifically addressing the “contribution of effect” of individual drugs within novel oncology combinations. The agency’s concern is straightforward: showing that a regimen works does not always establish what each component contributed to the observed benefit.

For Evidence Architecture, that distinction is consequential. A development programme needs to make an asset’s role within the wider therapeutic architecture interpretable – including its incremental contribution, the patients in whom that contribution occurs, the relevance of sequence and duration, the backbone against which it acts and the counterfactual against which benefit will eventually be judged. Evidence sufficient to demonstrate that a regimen is clinically effective may therefore not be sufficient downstream to attribute and value that effect. And attribution becomes harder still once the question moves from clinical benefit to economic value.

Jointly created value meets separately priced assets

The most striking illustration comes from health economics: when a new treatment is added to an already expensive on-patent backbone, a clinically effective combination can under certain conditions fail a conventional cost-effectiveness threshold even if the new add-on is priced at zero. This is not merely a theoretical curiosity. The phenomenon has been explored extensively in the combination-therapy literature and illustrated in previous NICE assessments.

The underlying logic is revealing. If improved outcomes also prolong the period during which the patient receives an expensive backbone, the additional cost of that backbone may consume the economic headroom created by the clinical benefit. There may then be little or no value left, within the conventional assessment framework, to reward the medicine that helped create the additional benefit. This is more than a high-price problem. It exposes a mismatch between jointly created clinical value and separately priced components.

Recent work on value attribution for oncology combinations continues to wrestle with precisely this issue. Various frameworks attempt to divide the value of a combination among its constituent medicines, yet all require assumptions about benefits that may be incompletely observed and about how jointly produced value should be allocated. Science is becoming increasingly sophisticated at constructing therapeutic regimens, while our assessment systems remain much better at valuing medicines one at a time.

You do not own your future counterfactual

For pharmaceutical strategy, this creates a less obvious risk: an asset enters development within a therapeutic environment that will continue to change while it progresses through the clinic. The standard of care against which a pivotal programme is designed today may no longer be the relevant comparator when the product reaches HTA several years later. A promising combination partner may itself move into a different position in the treatment pathway. A new backbone may emerge. Biomarkers may redefine the relevant population. Treatment may migrate from metastatic disease into earlier settings. In other words, you may own your molecule, but you do not own its future counterfactual.

That matters because decisions taken today can determine what remains possible later. A biologically attractive partner may create a difficult future cost structure. A trial optimised for regulatory efficiency may leave the incremental contribution of its components poorly characterised. A sequencing decision can alter the comparator against which benefit is ultimately assessed. None of these decisions needs to carry a “Market Access” label when it is taken. Yet each can expand or constrain the future access proposition.

Access optionality is therefore being created – or quietly lost – years before launch planning begins.

For companies developing oncology assets, this suggests a different set of strategic questions. Where, exactly, is the future value expected to arise – within the asset itself, within the combination, or at the interface between components? Which parts of that future value proposition does the company actually control, and which depend on a partner, a backbone, a comparator or a treatment pathway that may change before launch? And does the evidence programme preserve enough optionality to explain and attribute value if the therapeutic context evolves?

The answers will look very different for a company that controls several components of an oncology portfolio than for a biotech whose single asset depends on somebody else’s backbone. But in both cases, these are no longer questions that can safely be deferred until the clinical programme is largely fixed. That is one of the lenses I will take with me to ESMO in Madrid this October. Beyond the inevitable flood of new molecules, hazard ratios and late-breaking headlines, I will be interested in where clinical value is actually emerging: within individual assets, between them, or increasingly at the interfaces that connect them.

The Moderna-Merck melanoma result may ultimately prove important because individualised cancer vaccination works. But it may also illustrate a broader challenge.

Modern oncology is becoming increasingly good at creating value jointly. Our evidence, ownership, HTA and pricing systems still largely need to decide who that value belongs to.

You can read the article and connect with Matthias on LinkedIn here.