We explore how the first JCA provides insights into the longstanding methodological challenge of evidence generation and comparative assessment for precision medicine.
The first JCA: a new perspective on familiar challenges
The publication of the first European Joint Clinical Assessment (JCA) marks the transition of the EU Health Technology Assessment Regulation (HTAR) from methodological guidance to real-world implementation. The first assessment evaluated tovorafenib, a targeted therapy for a rare paediatric central nervous system (CNS) tumour harbouring specific gene alterations. While the medicine itself is indicated for a rare tumour, the methodological questions raised by the assessment extend beyond orphan medicines and reflect broader trends in precision oncology.
Over the past decade, oncology has progressively shifted towards biomarker-defined patient populations. This evolution has transformed how clinical trials are designed and how evidence is generated. This is because:
- Eligible patient populations are often small, making sufficiently powered randomised controlled trials (RCTs) difficult to conduct.
- Treatment pathways are heterogeneous, and the absence of a well-established standard of care can complicate the choice of comparator.
- Evidence is increasingly reliant on single-arm, non-randomised studies and indirect comparisons, resulting in greater uncertainty over relatively short periods of follow-up.
These challenges are not unique to the JCA. Precision oncology has long required healthcare systems to adapt their infrastructure, evidence-generation methods, and reimbursement policies to biomarker-defined care. Previous work has highlighted the need for innovative trial designs, greater use of real-world evidence, and more coordinated approaches to value assessment—without lowering evidentiary standards. Yet, HTA systems address these challenges inconsistently, which can produce divergent assessment and reimbursement outcomes.
These tensions are especially visible in tumour-agnostic therapies: treatments defined by a shared molecular target across multiple tumour types, rather than by tumour location alone. They therefore provide a particularly revealing test of whether assessment frameworks can accommodate biomarker-defined populations while still making coherent decisions across indications. Our global landscape analysis showed how differences in HTA approaches can translate into variation in access to these therapies.
Tovorafenib is not itself tumour-agnostic, but its use in a rare, molecularly defined cancer raises several of the same evidentiary and assessment challenges. This first JCA therefore provides an informative early indication of how these longstanding challenges may be tackled within the harmonised European assessment framework. While our previous insight focused on the architecture and governance of the JCA, this insight reflects on what its first report may tell us about the future assessment of biomarker-defined therapies.
When one tumour becomes many assessment populations
One observation from the JCA report is that precision medicine, particularly in oncology, has fragmented what was once considered a single disease entity into multiple clinically distinct assessment populations. This shift is increasingly reflected in HTA.
In this case, although tovorafenib is indicated for paediatric low-grade glioma (pLGG), the JCA report distinguished between separate populations according to molecular subtype, each requiring its own PICOs and comparative evidence base.
This stratification has a clear clinical rationale and it is not new to HTA. Patients with different gene alterations may have markedly different prognoses, with reported differences in both progression-free and overall survival. As a consequence, the JCA recognised that different molecular subgroups require separate comparative assessments. As precision oncology continues to evolve, future JCAs are therefore likely to assess an increasing number of biomarker-defined populations.
From a clinical perspective, this enables treatments to be more closely matched to the characteristics of individual patients, facilitating more personalised care. From an HTA perspective, however, it expands the number of assessment populations requiring separate comparative evidence, setting the stage for the evidence-generation challenges explored below.
The risk is that greater precision becomes equated with assessing each increasingly narrow population entirely in isolation. An analogy is a city evaluating an adaptive traffic-control system junction by junction. Conditions at each intersection matter, but the relevant decision is also whether—and under what conditions—the system improves traffic management across the city. Likewise, precision oncology depends on a connected system of diagnostics, data infrastructure, evidence generation, and targeted treatment—not merely on the performance of a medicine within each narrowly defined subgroup. Assessment frameworks must recognise clinically meaningful differences between populations without losing sight of the recurring system-level decision: whether to adopt, restrict, delay or conditionally reimburse a technology as evidence develops across patients, tumour types and time. Ever narrower assessments should not become a dead-end-street for precision medicine as a whole.
When evidence becomes a fraction of a fraction
Increasing biological precision has a second consequence: it reduces the proportion of available evidence that can be considered directly applicable to an individual assessment.
The principal evidence supporting tovorafenib came from FIREFLY-1, a phase II single-arm study evaluating patients with relapsed or refractory pLGG. Although the trial comprised three study arms, only the tumour-specific populations from Arms 1 and 2 were considered relevant to the assessment scope. The applicable evidence was then narrowed further during the assessment.
For the only population and PICO that could be fully assessed (Population 2, PICO 5), the assessors concluded that fewer than 20% of patients enrolled across Arms 1 and 2 were sufficiently representative of the predefined assessment population. Consequently, much of the available clinical evidence was considered outside the scope of the comparative assessment.
This illustrates an increasingly important trade-off in precision medicine. As assessment populations become more narrowly defined, the evidence becomes better aligned with the specific decision problem facing that particular HTA. However, this greater precision comes at the expense of the amount of evidence available for assessment. In effect, the assessable evidence becomes a fraction of an already small study population, potentially increasing uncertainty when interpreting the comparative findings.
This does not mean that greater uncertainty or lower evidentiary standards should be accepted as inevitable consequences of precision medicine. Rather, it strengthens the case for evidence-generation approaches designed around rare, biomarker-defined populations—including basket, umbrella and adaptive trials, multinational collaboration, and the structured use of real-world evidence. The question for JCA is therefore not only how narrowly available evidence should be filtered, but how assessment requirements can encourage evidence to be generated and updated more efficiently.
This has implications for how member states may implement JCA outputs. Some agencies may choose to align reimbursement decisions closely with the narrowly defined populations assessed and accepted within the JCA, whereas others may place greater emphasis on the applicability of the evidence to their local clinical context. These different approaches could lead to variation in patient access for an already small, rare population. Such uncertainty around the size of the reimbursable population may ultimately reduce incentives to launch and invest in precision medicines across Europe.
Precision requires precision: biomarker-specific evidence only works if both sides of the comparison are equally precise
The challenge is not limited to ensuring that the intervention evidence reflects the intended population. Comparator evidence should also represent the same biomarker-specific population if meaningful comparisons are to be made.
To estimate comparative effectiveness, the developer relied on an unanchored matching-adjusted indirect comparison (MAIC). The JCA assessors evaluated the appropriateness and implementation of this analysis, but its interpretability was constrained by limitations in the underlying comparator evidence. It included a small proportion of patients with high-grade glioma rather than paediatric low-grade glioma, and differences in tumour characteristics could not be adjusted for because individual patient data were not available. In addition, only one objective response rate (ORR) analysis was considered acceptable because the alternative used different response criteria.
These findings highlight that precision medicine requires precision throughout the evidence base. It is not sufficient for the intervention to target a highly specific patient population; the comparator population and outcome definitions must also be equally specific and directly comparable. As indications become increasingly biomarker-defined, assembling evidence that meets this level of alignment is likely to become progressively more difficult.
The challenge may amplify for tumour-agnostic therapies, which are treatments intended for patients who share a molecular alteration across multiple tumour types. Although this broadens the eligible patient population, it also means that a single JCA may need to evaluate evidence generated across multiple tumour types, each with different comparators, treatment pathways, and standards of care, but often with a small or immature evidence base. Meeting the evidence requirements for multiple PICOs across multiple tumour-types within a single JCA may therefore be complex, despite the common biological target.
Looking ahead: What future JCAs may reveal
The first JCA was the first attempt to apply the new EU HTA framework to a new medicine. It provides informative early insights into how relative effectiveness of precision oncology can be assessed under the JCA framework.
As future JCAs will evaluate both biomarker-defined and tumour-agnostic therapies, several methodological questions are likely to become more prominent.
- Should PICO definition evolve as indications become increasingly biomarker-defined or span multiple tumour types?
- How should evidence applicability be judged when narrowly defined assessment populations leave only a small proportion of the available evidence directly relevant?
- How can robust comparative evidence be expected when suitable comparator populations are difficult to identify?
- How should the interpretation of JCA findings reflect the diagnostic pathways and data infrastructure needed to identify and treat the populations being assessed?
- How can methodological guidance support a learning approach to precision medicine—using accumulating clinical, genomic, and real-world data to progressively reduce uncertainty, while identifying and managing the uncertainty that cannot realistically be resolved?
The first JCA does not answer these questions, nor should it. Rather, it provides the first opportunity to observe how longstanding methodological challenges associated with precision oncology are approached within a European assessment framework. As precision medicine continues to evolve, future JCAs and member states appraisals will offer further insights into how comparative assessment frameworks deal with increasingly diverse indications, while balancing evidential rigour with timely patient access.
