Tarlatamab and lurbinectedin: What the first HAS assessments show about the transition from JCA to France

Tarlatamab and lurbinectedin are among the first medicines for which the transition from the European Joint Clinical Assessment (JCA) to assessment by the French Haute Autorité de Santé (HAS) can be observed in practice. The two cases show that clinical evidence generated for the JCA can inform the French assessment, while the Transparency Committee (Commission de la Transparence, CT) continues to apply its own criteria when assessing clinical benefit (SMR), clinical added value (ASMR) and the target population.

For tarlatamab, the key issue is the identification of the PICO questions relevant to France and the robustness of indirect comparisons. For lurbinectedin, the case shows that a single JCA PICO and the availability of direct randomised evidence do not necessarily align the assessments: the choice of analysis, methodology and clinical relevance of the results remain subject to national appraisal.

These first French assessments therefore illustrate a central interface created by the EU HTA Regulation: the JCA establishes a common clinical evidence base at European level, but whether that evidence supports SMR and ASMR remains a question for the French assessment.

JCA changes the available evidence base without replacing HAS assessment principles

The JCA establishes a common clinical assessment at European level, while the Transparency Committee continues to apply the French principles for assessing medicines.

The JCA assesses relative clinical effectiveness and safety against the PICO questions defined in the European assessment scope. In France, the assessment serves a different purpose. The CT assesses, in particular, the Service médical rendu (SMR) and Amélioration du service médical rendu (ASMR) and estimates the population cible.

SMR assesses the clinical benefit of a medicine in relation to reimbursement by the French solidarity-based healthcare system. ASMR assesses the therapeutic improvement provided over existing alternatives. This relative assessment considers, among other factors, the comparateur cliniquement pertinent, the quality of the evidence, the magnitude and clinical relevance of the additional effect, quality of life and medical need.

JCA therefore changes the evidence available for the French assessment, but it does not translate a European finding directly into an SMR or ASMR level.

The JCA dossier becomes an evidence source for the French HAS assessment

The JCA dossier can provide a substantial part of the clinical evidence used in France without replacing the national dossier or the assessment by the Transparency Committee.

France contributes to the European assessment scope by submitting its national PICO requirements. Clinical evidence already structured in the JCA dossier can subsequently be used for the French assessment.

This interaction also changes national dossier preparation. Clinical results already documented at European level do not need to be reproduced in the same way when they can be reused. The French dossier must instead build the argumentation required for the national assessment. This includes, in particular:

• the disease and healthcare context in France;
• medical need;
• comparators relevant to France;
• the target population;
• the proposed SMR and ASMR;
• potential gaps between the available European evidence and French assessment requirements.

The main synergy therefore lies in the reuse of clinical evidence that has already been structured, rather than in the adoption of a European assessment conclusion.

Tarlatamab: Seven JCA PICOs meet a more focused French assessment question

The tarlatamab case first illustrates a scoping challenge: a JCA covering several PICOs must be mapped to the population and therapeutic strategy relevant in France.

The JCA of tarlatamab in adults with extensive-stage small-cell lung cancer (ES-SCLC) includes seven PICO questions. The populations differ, among other factors, according to the timing of disease progression after first-line treatment and the treatment-free interval. Comparators also differ across PICOs.

The evidence base is correspondingly heterogeneous. It includes the randomised DeLLphi-304 trial as well as several indirect comparisons: a network meta-analysis, an unanchored MAIC and an anchored indirect comparison.

For the French early access assessment, the relevant question was more focused. The favourable opinion of the Transparency Committee concerned patients who were not eligible for retreatment with platinum-based chemotherapy.

Tarlatamab therefore illustrates a first practical consequence of European PICO scoping: not every PICO in a JCA necessarily has the same relevance to the French assessment.

Where the population targeted in France corresponds clearly to a JCA PICO, the European clinical evidence may be more directly reusable. Where the French assessment question spans several PICOs, the relevant evidence is more fragmented and requires greater national interpretation.

Tarlatamab shows the limits of indirect evidence for the French assessment

The inclusion of an indirect comparison in the European evidence package does not mean that it will be sufficiently robust to support the French assessment.

For tarlatamab, different indirect comparison methods were used across the JCA PICOs. Some analyses, however, had methodological limitations, including concerns about the homogeneity assumption for the network meta-analysis and the risk of confounding in the unanchored MAIC. These indirect comparisons therefore did not constitute the decisive evidence in the French early access assessment.

The situation differed for the direct comparative evidence from DeLLphi-304. The randomised trial used a prespecified analysis and a multiplicity-control strategy. For overall survival, a hazard ratio of 0.6 with p < 0.0001 was reported. Results for dyspnoea and cough were also assessed within the hierarchical testing strategy for quality of life.

The transition from JCA to the French assessment therefore involves more than mapping available analyses to a PICO relevant to France. It also requires determining whether each analysis is methodologically robust enough to support the clinical value of the medicine in the population concerned.

Tarlatamab therefore highlights an evidence-selection challenge: within a European evidence package covering several PICOs and comparative methods, which evidence is sufficiently relevant and robust for the French assessment?

Lurbinectedin: One PICO and a randomised trial do not automatically lead to the same assessment

The lurbinectedin case illustrates another dimension of the transition from JCA to HAS: even when population, comparator and comparative evidence are closely aligned, the national appraisal remains independent.

The evidence structure is less fragmented than for tarlatamab. The JCA of lurbinectedin in combination with atezolizumab includes one population and one PICO. The comparator is atezolizumab monotherapy. Comparative evidence comes from the randomised, open-label phase III IMForte trial. The case therefore involves neither multiple PICOs nor the same reliance on indirect comparisons as tarlatamab.

Nevertheless, the Transparency Committee issued an unfavourable opinion on the early access request and HAS refused the authorisation. The difference did not arise from the need to identify another study or comparator, but from the appraisal of the available direct comparative evidence.

Lurbinectedin therefore shows that close structural alignment between the European PICO and the French assessment question does not automatically lead to the same interpretation of the evidence.

For lurbinectedin, data cut-off, endpoints and methodology shape the French appraisal

The French assessment of lurbinectedin shows that a statistical result must be interpreted in light of its clinical relevance and the study methodology.

The IMForte trial assessed, among other outcomes, progression-free survival (PFS) and overall survival (OS). At an initial analysis after approximately 15 months of follow-up, the hazard ratio was 0.54 for PFS with p < 0.0001. For OS, the hazard ratio was 0.73 with p = 0.0174. At a later analysis after approximately 21 months, the hazard ratios were 0.56 for PFS and 0.81 for OS; the nominal p-value for OS was 0.0605.

In the French assessment, the PFS result was interpreted in the context of the mortality data. Methodological issues were also considered, including a late protocol amendment affecting the statistical method.

The difference between the two cases can therefore be summarised simply:

Tarlatamab illustrates an evidence-selection challenge. Lurbinectedin illustrates an evidence-interpretation challenge.

For tarlatamab, the task is to identify which evidence from several PICOs and sources can support the French question. For lurbinectedin, a direct randomised evidence base is available, but its clinical and methodological implications must still be appraised according to CT principles.

Tarlatamab and lurbinectedin illustrate two routes from JCA to HAS

The two early French cases show that national use of JCA evidence depends both on the structure of the European assessment scope and on the nature of the available clinical evidence.

These two configurations create different risks for evidence-generation strategy.

Where the JCA contains numerous PICOs, the populations and comparators likely to be relevant in France need to be anticipated. Otherwise, a comprehensive European evidence package may still leave uncertainty around the question that ultimately matters for SMR or ASMR.

Where the European PICO and the French question are closely aligned, the risk shifts towards the quality and interpretation of the evidence: the selected analysis, statistical plan, effect size, endpoints and clinical relevance remain decisive.

The German tovorafenib case confirms the general principle within a different national framework

Germany also does not automatically adopt evidence from the JCA into its national assessment. For tovorafenib, an unanchored MAIC considered in the JCA was not used by the G-BA for the German benefit assessment, while PFS and tumour response were reassessed against German requirements for patient relevance and endpoint operationalisation. The comparison with France remains limited, however, because the systems answer different questions: the CT assesses SMR and ASMR in relation to the comparateur cliniquement pertinent, whereas AMNOG focuses on the national assessment of added benefit and, where applicable, the zweckmäßige Vergleichstherapie defined within the German framework.

The comparateur cliniquement pertinent and a JCA PICO comparator serve different functions

A comparator included in a European PICO is not automatically the comparator that determines the French assessment.

For the Transparency Committee, the comparateur cliniquement pertinent (CCP) is an intervention positioned at the same level of the therapeutic strategy as the new medicine and intended for the same patients. Comparison with the CCP is particularly important for the assessment of ASMR.

European scoping follows a different logic: the JCA assessment scope brings together the PICO requirements submitted by Member States. A JCA may therefore include several populations and comparators even though only part of that scope is decisive for France.

Tarlatamab illustrates this distinction directly. Seven European PICOs do not translate into seven French assessment questions of equal importance.

For evidence-generation strategy, the objective is therefore not merely to cover comparators that may appear in the JCA. It is also necessary to determine whether the clinical evidence supports a robust comparison with the CCP relevant to the French therapeutic strategy.

Clinical relevance remains a French assessment question after JCA

A relative treatment effect described in the JCA must still be appraised in France in terms of the quality of the evidence, magnitude of effect and clinical relevance.

ASMR is not based solely on the existence of a difference between treatment groups. The CT considers, among other factors, the quality of the evidence, the magnitude and clinical relevance of the additional effect, quality of life, tolerability and medical need.

The lurbinectedin case illustrates this step particularly clearly. A PFS difference was observed in the randomised trial, but the French assessment also considered OS, deaths, the data cut-off and methodological aspects of the study.

The national question is therefore not only: what results are available in the JCA? It is also:

Which results can support SMR and ASMR in the French assessment context?

This distinction structures the transition from the European dossier to the French dossier de transparence.

JCA evidence must be translated into a coherent French argument for SMR and ASMR

JCA provides a common clinical evidence base, but the French assessment requires a coherent argument linking the European evidence to SMR, ASMR and the target population.

The transition can be summarised as follows:

1. European JCA: European PICOs and common clinical evidence;
2. French assessment: relevant PICO(s) and comparateur cliniquement pertinent;
3. appraisal of evidence quality, magnitude of effect and clinical relevance;
4. SMR + ASMR + population cible.

The added value of the national work therefore does not lie in exhaustively repeating all clinical evidence already presented at European level. It lies in placing that evidence within the assessment framework of the Transparency Committee.

Tarlatamab and lurbinectedin represent two different starting points. For tarlatamab, the nationally relevant evidence must be extracted from a complex European scope. For lurbinectedin, the challenge lies more in the French appraisal of a relatively clear direct comparative evidence base.

JCA-ready does not automatically mean HAS-ready

An evidence-generation strategy designed to address JCA does not by itself guarantee an evidence package suited to the French assessment.

JCA and HAS do not, however, create two completely separate evidence environments. Structured reuse of European clinical evidence can reduce duplication and facilitate national preparation.

The strategic task is therefore to identify early which parts of the European evidence will be directly usable in France and which elements will require specific national analysis or argumentation. Key considerations include:

• PICOs likely to be relevant in France;
• the comparateur cliniquement pertinent;
• availability of direct comparative evidence;
• methodological robustness of indirect comparisons;
• clinical relevance of endpoints;
• magnitude of effect and quality of life;
• the French target population;
• evidence updates between JCA and the national assessment.

The dossier de transparence can only partly compensate for fundamental gaps in evidence generation. If an important CCP has not been adequately addressed or the endpoints cannot support the French assessment question, the problem arises before the HAS submission itself.

Six questions for anticipating the transition from JCA to the French assessment

The first French experiences highlight six questions that are particularly important for evidence-generation planning:

1. Which JCA PICOs are likely to be genuinely relevant to the French assessment?
2. Does the clinical development programme cover the comparateur cliniquement pertinent expected in France?
3. Which direct comparative evidence can support SMR and ASMR?
4. Are the necessary indirect comparisons methodologically robust and suited to the French assessment question?
5. Are the endpoints, magnitude of effect and quality-of-life data not only available for JCA but also clinically relevant to the CT?
6. Which France-specific elements on the target population, healthcare context and evidence updates need to be prepared in addition to the JCA?

Preparation for the French assessment therefore begins before the JCA is finalised. PICO scenarios, the comparator landscape and national evidence expectations need to be incorporated early into the evidence-generation strategy.