Tovorafenib after the JCA: What the first G-BA benefit assessment reveals about the EU-to-Germany evidence transfer

Germany's Federal Joint Committee (Gemeinsamer Bundesausschuss, G-BA) published its benefit assessment of tovorafenib on 17 August 2026. Tovorafenib has orphan drug status, meaning that an added benefit is considered established by virtue of its marketing authorisation under German law. The current procedure therefore assesses the magnitude of that added benefit. The published benefit assessment is not yet the final G-BA resolution. Its publication initiates the commenting procedure, with the final resolution to follow later in the process.

Nevertheless, the case provides an important first practical test of how the Joint Clinical Assessment, the national AMNOG dossier and the German benefit assessment interact. It shows for the first time which elements of European evidence can be carried into the German procedure and where the G-BA applies its own methodological standards.

The European assessment scope for tovorafenib comprised three populations and eight PICO questions. Population 1 covered the full claimed indication. Population 2 included patients older than one year with a BRAF V600E mutation. Population 3 included patients with a BRAF fusion, rearrangement or V600 mutation other than V600E.

Four PICOs were defined for the full claimed indication. The comparators included individualised treatment, carboplatin plus vincristine and vinblastine. However, no comparative data were available for these four PICOs. Comparative evidence was also unavailable for PICO 6 and PICO 8.

Comparative evidence was submitted particularly for PICO 5: tovorafenib versus dabrafenib plus trametinib in patients older than one year with a BRAF V600E mutation. This comparison was based on an unanchored matching-adjusted indirect comparison, or MAIC.

A MAIC was also submitted for PICO 7, comparing tovorafenib with trametinib in patients with BRAF fusion, rearrangement or a V600 mutation other than V600E. However, this comparison was not included in the JCA report because the available documentation for the comparator study TRAM-01 was insufficient to assess its methods and results adequately.

The JCA therefore already highlighted a central evidence gap: usable comparative results were unavailable for six of the eight PICO questions.

The evidence gaps for tovorafenib arise before the delta dossier

The absence of comparative data for six of the eight PICO questions is not simply an issue of later dossier preparation. The case illustrates how strongly the usability of a JCA depends on the evidence base already available at the time of the European assessment.

Tovorafenib received a conditional marketing authorisation following a clinical development programme in which the pivotal FIREFLY-1 study was a non-randomised, non-comparative Phase II study. The limited sample size, heterogeneous study population, absence of a control arm and still limited long-term data particularly restrict the possibilities for estimating relative treatment effects. The evidence base is being further developed through the ongoing randomised FIREFLY-2 study.

The case therefore illustrates a fundamental challenge for early evidence planning: the evidence available to support marketing authorisation will not necessarily answer all relative PICO questions subsequently defined for the JCA. For manufacturers, this increases the importance of anticipating potential European and national HTA questions during clinical evidence planning.

Individualised treatment creates additional requirements for the comparator strategy

The tovorafenib JCA also provides an important lesson for PICOs in which individualised treatment is specified as the comparator.

For several PICOs, the manufacturer argued that suitable comparative evidence was unavailable because no identified study covered all treatment options included in the requested individualised treatment strategy. The JCA assessors considered this approach inadequate. Studies covering only a relevant selection of the requested treatment options could also have been submitted. Their suitability for a specific population or subpopulation could then have been assessed as part of the JCA.

An individualised treatment comparator therefore does not necessarily mean that a single comparator study must cover every treatment option included in the PICO. For the evidence strategy, the key task is to identify relevant comparator evidence early and as comprehensively as possible, and to assess whether it can inform specific populations or subpopulations.

For future delta dossiers, this broadens Thesis 3: early evidence planning concerns more than the study design of the new medicinal product itself. It also includes the comparator landscape, information retrieval and the question of which external evidence may support subsequent European and national PICO questions.

The tovorafenib dossier provides the first practical example of how European evidence can be reused across the individual AMNOG dossier modules. The pattern is not uniform. While Modules 1 to 3 remain strongly shaped by national requirements, references to the EU dossier are used much more systematically in Module 4.

Module 3 provides a particularly clear example. According to the dossier, relevant passages from the EU dossier were translated into German for the description of the disease and supplemented with additional information. By contrast, epidemiology and the derivation of the German statutory health insurance population required separate national analyses because the EU dossier did not contain the corresponding Germany-specific information.

Module 4 illustrates the other side of the approach. Parts of the study characteristics and methodology are not reproduced in full but incorporated through specific references to the EU dossier. European content is also partly reused for endpoint operationalisation and the presentation of results and then combined with information required for the German assessment.

The first practical case therefore shows that the synergies of a delta dossier are not distributed equally across all AMNOG modules. Reuse is particularly relevant for clinical evidence in Module 4, while Germany-specific content still needs to be developed nationally.

The tovorafenib dossier shows that reuse of the EU dossier does not replace key national workstreams. Germany-specific epidemiology, the derivation of the statutory health insurance target population, annual treatment costs and the national treatment context remain distinct components of the AMNOG dossier.

National supplementation is also required for the clinical evidence. The systematic literature review must be sufficiently up to date at the start of the German procedure. New data cuts can be added. At the same time, endpoints and analyses presented in the dossier still need to be assessed against German requirements for patient relevance, operationalisation and methodological suitability.

For tovorafenib, this division of tasks can be seen particularly clearly:

• Epidemiology and target population: separate derivation for Germany and the statutory health insurance population.
• Costs and treatment context: national presentation of annual treatment costs and expected treatment shares.
• Systematic literature review: update for the national procedure.
• Clinical evidence: selective reuse of study information from the EU dossier.
• Endpoints and analyses: reassessment according to German methodological standards.
• Added-benefit argumentation: national interpretation of the evidence within the AMNOG framework.

The practical value of a delta dossier therefore does not lie in replacing as much national content as possible with references. The key synergy arises where evidence already prepared at European level can be reused in a structured way and then supplemented specifically with the information required for Germany.

The first fundamental difference concerns the assessment question itself. The JCA evaluates relative clinical effectiveness and safety against the comparators defined in the European assessment scope. The German benefit assessment, by contrast, follows the legal and methodological requirements of AMNOG.

Tovorafenib also involves an additional specific feature: the product has orphan drug status. In the current benefit assessment, no appropriate comparator therapy is therefore determined as the basis for a conventional comparative added-benefit assessment. Instead, the G-BA assesses the magnitude of the legally presumed added benefit on the basis of the marketing authorisation studies and other relevant evidence.

The JCA and the G-BA benefit assessment should therefore not be understood as two evaluations of the same question. They partly rely on the same clinical evidence, but assess it within different regulatory and methodological frameworks.

For the delta dossier, this confirms the first thesis: EU HTA does not replace the national benefit assessment in Germany.

The difference becomes particularly clear in the indirect comparison of tovorafenib with dabrafenib plus trametinib.

For PICO 5, the JCA included an unanchored MAIC. The comparison was based on individual patient data from FIREFLY-1 for tovorafenib and aggregated data from the Bouffet et al. study for dabrafenib plus trametinib. The JCA reports relative results for outcomes including progression-free survival, response outcomes and safety.

However, the JCA itself identifies substantial limitations. These include potential residual confounding, differences between the study populations and small effective sample sizes. For some analyses, the effective sample size was in the single digits or low double digits. The assessors therefore state that the resulting estimates may not represent causal treatment effects and may only be transferable to the target population to a limited extent.

For the German benefit assessment, the G-BA goes one step further and does not use the MAIC at all.

The G-BA identifies three key reasons: MAIC analyses based on aggregated study data are generally not considered appropriate in the context of the German benefit assessment; the analysis covers only a small subpopulation of FIREFLY-1; and the submitted effectiveness outcomes, particularly tumour response and PFS, do not represent relevant patient-relevant efficacy endpoints for this assessment from the G-BA's perspective.

The difference is therefore not that the JCA accepts the MAIC without reservation while Germany rejects it. The JCA presents it as relative evidence despite major uncertainty, whereas the G-BA does not consider it usable for the national benefit assessment.

This distinction is important for the delta dossier. An analysis can form part of the JCA without automatically becoming a robust basis for the German procedure.

The European assessment scope for tovorafenib requested outcomes including overall survival, progression-free survival, objective response, health-related quality of life, disease symptoms and safety. For PFS and objective response, the scope explicitly requested analyses according to RAPNO and RANO criteria where corresponding data were available for both intervention and comparator.

The G-BA subsequently reassesses the patient relevance, operationalisation and validity of these endpoints according to national requirements.

The result is striking. Of the endpoints presented in the German dossier, the benefit assessment essentially considers overall survival and adverse events. Tumour response, PFS, long-term tumour control, visual acuity and several patient-reported outcomes are not included in the assessment.

Why does the G-BA not consider tumour response for tovorafenib?

The G-BA does not consider tumour response because the patient relevance of the submitted operationalisation is regarded as unclear.

Radiological imaging findings are not considered directly patient-relevant. At the same time, the study documentation does not make it sufficiently clear whether clinical status and corticosteroid use were incorporated into the overall response assessment according to consistent criteria.

The G-BA also identifies an important methodological issue in the response criteria. In FIREFLY-1, the overall response rate according to RANO-HGG was around 20 percentage points higher than the response rates according to RANO-LGG and RAPNO-LGG. For the paediatric low-grade glioma population assessed here, the G-BA considers RAPNO-LGG more appropriate than RANO-LGG and particularly RANO-HGG.

Why does the G-BA not consider PFS for tovorafenib?

PFS is not considered because the G-BA regards its patient relevance in the submitted operationalisation as unclear.

The RANO and RAPNO criteria used to determine progression can classify an event on the basis of radiological changes. However, the available data do not allow a sufficiently clear distinction between patients whose progression was accompanied by clinically relevant deterioration and those for whom imaging changes alone determined progression. The G-BA therefore considers the overall patient relevance of PFS unclear.

Why are quality-of-life and other patient-reported outcomes not considered?

Patient-reported endpoints covering cognitive function, fatigue, disease symptoms and health-related quality of life were included in the German dossier. The G-BA does not use them because of insufficient questionnaire completion rates. For several instruments, response rates were already below 70 percent at baseline.

Visual acuity is also excluded from the assessment. In contrast to tumour response and PFS, the G-BA explicitly recognises visual acuity as patient-relevant. However, the submitted operationalisation contains uncertainties, including the testing procedures used, standardisation across study centres and the population included in the analysis.

The tovorafenib case therefore strongly confirms Thesis 4 on the delta dossier: endpoints included in the European assessment scope still need to be reassessed nationally for patient relevance, operationalisation and methodological usability.

The G-BA also reassesses analysis populations and statistical methods

The differences between the JCA and the German benefit assessment for tovorafenib are not limited to the selection of patient-relevant endpoints. Analysis populations, data cuts and statistical methods are also independently assessed in the national procedure.

The G-BA benefit assessment identifies limitations in the FIREFLY-1 analysis of overall survival, among other issues. Different analysis sets were used for different endpoints. For the mortality analysis, the G-BA particularly criticises the implementation of the intention-to-treat principle and the resulting uncertainty surrounding the submitted results.

The data cuts are also subject to methodological assessment. The benefit assessment notes that the data cuts were not prespecified. At the same time, the G-BA uses the more recent three-year data cut for the national assessment because it provides the longest available follow-up.

The indirect comparison illustrates the same national reassessment at another level. The unanchored MAIC presented in the JCA is not adopted for the AMNOG procedure simply because it was already part of the European assessment. The G-BA reassesses its methodological suitability for the national benefit assessment and does not use it because of the identified limitations.

For the delta dossier, this means that reuse does not replace national quality assessment. The decisive question is not only whether an endpoint or analysis was included in the JCA, but whether the population, operationalisation, statistical method and presentation of results also meet the requirements of the German benefit assessment.

The European JCA and the German benefit assessment do not rely on the same data cut.

The FIREFLY-1 evidence submitted for the JCA contained the two-year data cut. The JCA report notes that the health technology developer informed the assessors in November 2025 that a three-year data cut was already available, but these clinical data were no longer incorporated into the JCA submission.

The German dossier, by contrast, contained both the two-year and three-year data cuts. For the German benefit assessment, the G-BA uses the three-year data cut from 6 June 2025, because it provides the longest available follow-up.

The first practical case therefore already shows a difference in the evidence base between the JCA and the national assessment.

This confirms Thesis 5 on the delta dossier: evidence updates are not necessarily an exception. They may become a routine part of the transition from JCA to the national procedure.

Tovorafenib also illustrates a second timing issue.

The JCA report was endorsed by the HTA Coordination Group on 30 April 2026. The procedural review by the European Commission is dated 19 May 2026.

The German benefit assessment procedure, however, started on 15 May 2026. According to the G-BA, the JCA report had not yet been published at that point. The G-BA therefore explicitly states that the final JCA report could not be taken into account in the benefit assessment under the applicable German rules.

An important distinction is required here between the JCA dossier and the JCA report. The European JCA dossier was used as one of the sources for the German benefit assessment. The final JCA report itself could not be considered because of its publication timing.

The first practical case therefore demonstrates that European and German procedural timelines will not necessarily be aligned.

Thesis 1: EU HTA does not replace the national benefit assessment

Tovorafenib confirms this thesis. The JCA and the German orphan drug benefit assessment address different assessment questions. The G-BA continues to apply national requirements for patient relevance, methodology and the usability of evidence.

Thesis 2: The delta dossier is about translation, not cross-referencing

The first delta dossier shows for the first time how this translation works in practice. The strongest synergies with the EU dossier can be seen in Module 4. Epidemiology, the target population, costs, the national treatment context and national added-benefit argumentation, by contrast, remain largely national tasks.

Thesis 3: The decisive evidence planning starts before the delta dossier

Comparative data were unavailable for six of the eight JCA PICOs. At the same time, the discussion around individualised treatment shows that comparator strategy and information retrieval need to be considered early. Evidence gaps that exist at the time of the JCA can only be addressed to a limited extent in the subsequent delta dossier.

Thesis 4: Comparators, endpoints and analytical methods remain key risks

PFS and tumour response were relevant within the JCA scope but are not used by the G-BA as patient-relevant efficacy endpoints in the submitted operationalisations. In addition, the G-BA assesses analysis populations, implementation of the intention-to-treat principle and indirect comparisons against national methodological requirements.

Thesis 5: Timing and evidence updates will be decisive

Germany already uses a more recent FIREFLY-1 data cut than the JCA. At the same time, the final JCA report could not be incorporated into the initial German benefit assessment because of its publication timing. The European and national evidence bases can therefore already differ within the same assessment sequence.

Tovorafenib provides the first concrete indication of where a delta dossier can create synergies between JCA and AMNOG and where national work remains necessary. Particularly in Module 4, study information, methods and results already prepared at European level can be reused in a structured way. Across the other modules, many Germany-specific requirements continue to require separate national development.

At the same time, the G-BA benefit assessment shows that reuse should not be equated with automatic acceptance. Endpoints, indirect comparisons, analysis populations and statistical methods continue to be assessed against German methodological standards. New data cuts may further change the national evidence base.

The delta dossier is therefore neither an entirely new dossier nor simply a shortened version of the traditional AMNOG dossier. It combines the structured reuse of European evidence with the continuing independent requirements of the German benefit assessment.

For manufacturers, this shifts part of the critical work upstream. PICO scoping, comparator strategy, HTA-suitable endpoints, evidence updates and the national applicability of analyses should already be considered during evidence planning.

The first tovorafenib case provides an important starting point. Future JCA and AMNOG procedures will show which synergies can be used consistently across different indications and evidence settings.