Beyond the Headlines: Assessing the First JCA Report on Tovorafenib

Following the application of the EU Health Technology Assessment Regulation (HTAR) on January 12, 2025, the first Joint Clinical Assessment (JCA) report has now been published on June 9, 2026. This first published JCA report, developed by NCPE (Ireland) and IQWiG (Germany), describes the assessment of tovorafenib (Ojemda), Ipsen’s innovative treatment for pediatric low-grade gliomas (pLGG). This first report provides insights into both the practical implementation of the Regulation, the Implementing Acts and guidance documents, as well as the assessment of the distinct evidentiary and methodological challenges that commonly arise in orphan indications.

Lessons learned and confirmed

The tovorafenib JCA report has shone a physical light on key procedures previously outlined in the text of EU HTAR and published guidance documents. Many expectations of how the JCA procedure would operate were confirmed, but the evidence assessment also highlighted key ambiguities throughout the published JCA materials.  

Guidance-consistent implementation

The tovorafenib JCA met the procedure and timelines outlined in the Procedural Guidance and were completed as follows:

A broad JCA scope, but limited traceability and transparency

Eight Population, Intervention, Comparator(s), Outcome(s) (PICOs) were requested across three populations in the assessment scope, including:

Population 1: full claimed indication

• PICOs 1 and 2: Two large, individualized baskets
• PICOs 3 and 4: Individual comparisons vs a chemotherapy regimen and an alternative

Population 2: BRAF V600E mutation in patients > 1 year

• PICOs 5 and 6: Subpopulation PICOs with an individual comparison against targeted therapy and the simplified individualized chemotherapy alternative

Population 3: BRAF fusion, rearrangement, or V600 (non-E) mutation

• PICOs 7 and 8: Subpopulation PICOs with a preferred targeted therapy and an individualized alternative 

The “PICO tree” is presented below:

In line with the JCA scope presenting PICOs agnostic to the Member State(s) that requested them, limited information accompanied the assessment scope. As such, the context and origin of each PICO was not specified, save for a brief reference to the European Standard Clinical Practice Recommendations for primary pLGG and for input provided from the carer participant in Appendix A.Notably, treatment pathway diagrams and detailed references to existing guidelines – stalwarts of national health technology assessment (HTA) processes – are not presented in the JCA report, except for a passing reference.¹

In contrast to the health technology developer (HTD) submission, which cited five European guidelines/HTA reports providing treatment recommendations for second-line and later (2L+) relapsed/refractory (R/R) pLGG, the tovorafenib JCA report simply stated “a current European standard for treatment of pediatric LGG patients who have already received one or more prior systemic therapies has not been established. Different chemotherapy regimens are used for the treatment of these patients. In addition, targeted therapies are available for distinct patient groups,“ without supporting references.

Absence of evidence?

No comparative evidence was provided by the HTD for PICOs 1, 2, 3, 4, 6 and 8 due to the lack of comparator data within the specific populations requested by the PICO scope.

The narrow population definition in the PICO request (focusing specifically on the full claimed BRAF‑mutated/BRAF fusion pLGG population) likely enabled the limited evidence package to be submitted. At JCA, the assessment scope defines the scope of information retrieval.² However, this may have had the unintended consequence of excluding evidence from historical or unselected pLGG cohorts that might provide evidence for the BRAF fusion subgroup.³ In practical terms, this can reduce the decision-usefulness of the assessment: national HTA bodies may prefer imperfect but clinically relevant contextual evidence over having no comparative evidence at all.

As a result, there was no assessment whether broader, unanchored comparisons (e.g. positioning FIREFLY-1 outcomes against relevant historical chemotherapy trials or other real-world datasets) could have provided PICO-relevant evidence utilizing broader populations than the full claimed indication, which is particularly relevant if the full claimed indication is tied to molecular profiling. The key question is whether the high specificity of the PICOs ultimately helps local HTAs reach a conclusion, or whether it increases the likelihood that they will request additional analyses that relax the population restriction to make use of the wider literature, and whether such assessment is best placed at the local or European level.

Challenges associated with individualized comparators

The HTD was criticized for how they searched for comparator data for the individualized treatment baskets: only searching for comparator data that contained the exact set and combination of therapies contained within the individualized basket.

“The HTD did not aim to include studies with an individualized treatment comparator which did not consider all of the treatment options. The HTD’s approach to include studies with an individualized treatment comparator, only if the comparator included all treatment options, is not adequate. If comparator data that include only a selection of the requested treatment options had been identified and submitted, the suitability of these data to address the relevant PICOs would then have been a matter of assessment“

A thorough process, a hint of assessor flexibility and a high bar for incompleteness?

The JCA process encompassed systematic and procedural review of all methods of evidence identification and synthesis. Such detailed scrutiny led to some surprising call outs in the JCA report, including a criticism that: “The search dates for studies performed or sponsored by the HTD and for the studies from submission files to EMA are missing.” It is unclear exactly how an HTD can “search” for studies performed by themselves in a systematic manner, and documenting the search date of this is not clearly called out in the JCA materials. This criticism was leveled despite not being included in the requests for additional information to the HTD, which could have provided an adequate response if this had been requested.

Given the requirement of the JCA dossier to be “complete with regard to the available studies and data that could inform the assessment”, it was surprising the HTD was not pressed to re-screen or re-search for articles to demonstrate this completeness. Instead, the assessors used their own additional searches confirming the completeness of the dossier: “In their own searches, the assessors did not identify additional studies… [during] the review of completeness for any of the PICOs.”

The fact that an HTD was able to submit despite a “not adequate” approach to identifying comparator data may suggest a high bar for JCA discontinuation and/or represent an implicit acknowledgment of ambiguity in the original JCA guidance documents, with the Frequently Asked Questions (FAQ) document developing the concept of an individualized treatment comparator only being published following, and perhaps in response to, the tovorafenib submission.⁴

The expected critique of an unanchored MAIC with low ESS

In the face of repeated statements of “no comparator data available,” this first JCA was perhaps as striking for what we have not learned rather than what we have learned about how the JCA will assess evidence.

The critique of the unanchored matching-adjusted indirect comparison (MAIC) with an effective sample size (ESS) of 5.81 was unavoidable.However, it was notable how thorough the assessors were with this, covering each methodological angle outlined in the guidance documents on evidence synthesis. The assessors levied criticism ranging from the use of an unanchored comparison itself to the methods of identifying and adjusting for prognostic factors, differences in study conduct and outcome definition, and the absence of pre-specification and multiplicity correction. The critique of the methodology came from a strongly theoretical standpoint. The assessors highlighted limitations in the approach to covariate selection where problems could have occurred in principle, rather than considering whether any important variables had been excluded in practice. This sets the bar implausibly high since unmeasured confounding can always exist even with the most stringent process to identify variables. The critique of the MAIC itself was motivated primarily by the theoretical properties of the analysis methods and the assumptions required to generate meaningful results. There was little consideration of what else could have been done given the very small sample sizes that are typical of rare conditions such as pLGG.

Due to the paucity of evidence in this submission, the core objective of the JCA – to provide an analysis of scientific uncertainty and strengths and limitations of comparative effectiveness and safety evidence of a medicinal product – was hardly tested. Assessment (and presentation) of more robust evidence packages, including randomized controlled trials (RCTs), anchored indirect treatment comparisons (ITCs) or stronger unanchored ITCs, and the impact on local HTAs, remain the major questions emerging from this first JCA.

With respect to the quantitative bias analysis, the assessors were clear that these are not a “tick-box” add-on for unanchored ITCs. They comprehensively rebutted the HTD’s shifted hypothesis testing because the shift range was justified via minimal clinically important differences (MCIDs) rather than evidence-based bounds on plausible unmeasured confounding. The assessors also rebutted the HTD’s E-values because they were incorrectly calculated and not appropriately interpreted given likely substantial residual confounding.

For single-arm trials, the assessment strictly concerns relative efficacy and safety

No individual assessment or discussion of the results of the FIREFLY-1 single-arm trial were presented, with the assessors simply referring to the HTD’s submission. For many national HTA processes, non-comparative results of single-arm trials, especially in orphan indications, may play a key role in defining the clinical and economic value of a medicinal product. As such, the absence of an assessment of the single-arm trial methods and results leaves this evaluation of non-comparative endpoints down to the individual Member State. 

Implicit value judgments

It was noteworthy that the assessors decided not to present certain results they deemed to be insufficient. For examle, the relative effectiveness results for Independent Review Committee (IRC)-assessed overall response rate (ORR) based on the Response Assessment in Neuro-Oncology criteria for low-grade gliomas (RANO-LGG) are not included in the JCA report “due to clear bias arising from methodological flaws and inappropriate analysis methods used by the HTD”, resulting from clear differences in outcome definitions between FIREFLY-1 and Bouffet 2023.

The conclusion to not present certain results appears to be the strongest, and perhaps categorical, statement of a result’s inappropriateness for decision making. However, it is unclear how these decisions to present or not present results in the JCA report can be made consistently within and between assessments.

For example, while the IRC-assessed ORR based on RANO-LGG was considered so unsuitable that it could not be included in the JCA report, the unanchored MAIC of progression-free survival (PFS) was judged to be of sufficient quality to include in the JCA report. This is despite the assessors identifying major, general uncertainties with the MAIC methods that “threaten[ed] the applicability of the results to the target population.” This is in addition to the HTD providing evidence that the comparator data from Bouffet 2023 provided a clinically implausible PFS for a R/R pLGG population, which provided longer PFS estimates for R/R pLGG than a study reporting data in the first-line pLGG setting. The assessors criticized this reasoning as being based on a naïve ITC that did not adjust for differences between the data sources but did not comment on the lack of face validity when patients receiving treatment at later lines have better outcomes than those treated at earlier lines.

Considerable additional research and clinical interpretation will be required at the national HTA level

Overall, there is still a substantial amount of clinical interpretation and assessment that will be required at the individual Member State level to translate the JCA report on tovorafenib into concrete local reimbursement decisions. Member States will have to re-assess whether the data presented for JCA are complete with regards to their individual needs – noting the potential availability of comparator data from beyond the target population and the lack of evidence for most PICOs. If required, they will have to assess the validity and usability of single-arm non-comparative trial outcomes for their own HTA processes, as well as provide considerable interpretation on the suitability of the unanchored MAIC presented for PICO 5 for their local decision making.   

Notably, the single PICO for which evidence was provided in the tovorafenib JCA report covers only approximately 17% of the indicated population – leaving the rest without evidence. How EU Member States will consider this minority population and identify data or approaches to inform decision making for the remaining 83% of the indicated population remains to be seen.³ Will comparisons against molecularly unselected patients or local data be requested?

Other key learnings:

• Redaction was minimal: The HTD was only allowed to redact details around the databases and patient numbers in their ECA feasibility assessment, as well as remove the statistical analysis code from the appendixes. No clinical trial or comparative effectiveness or safety results were redacted, in line with the guiding principles on data transparency
• There was partial, but not full, transparency: Details of the completeness check were not published, and there is limited information on the PICO scoping procedure (other than the clinician and carer input into this)
• The factual accuracy check was factual: In line with national HTA procedures, the factual accuracy check presented an opportunity for the HTD to correct factual errors in the assessment report and submission dossier. Further requests relating to the substance and presentation of the assessment report were deemed “outside of the scope of the factual accuracy check”
• Despite guidance suggesting a well-executed individual participant data (IPD) – IPD external control arm (ECA) would be preferable to unanchored MAICs in this evidence-sparse setting, the HTD’s feasibility assessment concluded negatively and no ECA was submitted – yet the JCA largely did not engage with or critique this.

Next steps

The tovorafenib JCA report has confirmed the approach as procedural and theoretical, following the letter of guidance on what ought to be done in principle, rather than what may be done in practice. However, this first JCA may not be fully generalizable or representative of future JCAs. It is a succinct dossier for an orphan indication with a limited evidence base and further evolution of the process is expected. According to the June 9, 2026 update, of the 18 initiated, and 16 ongoing, JCAs, three are expected to be published by the end of this summer and a further four by December.⁵ As this first JCA report highlights even more strongly, high-quality preparations are essential.

References

1.         On behalf of the interdisciplinary SLGGWG, Roka K, Scheinemann K, et al. European standard clinical practice recommendations for primary pediatric low-grade gliomas. EJC Paediatric Oncology. 2024; 4.

2.         Directorate-General for Health and Food Safety. Guidance on filling in the joint clinical assessment (JCA) dossier template – Medicinal products. 2024. Available at: https://health.ec.europa.eu/publications/guidance-filling-joint-clinical-assessment-jca-dossier-template-medicinal-products_en. Accessed: 17/06/2026.

3.         Ryall S, Zapotocky M, Fukuoka K, et al. Integrated Molecular and Clinical Analysis of 1,000 Pediatric Low-Grade Gliomas. Cancer Cell. 2020; 37(4):569–83.e5.

4.         Directorate-General for Health and Food Safety. Questions and Answers on general methodological and procedural issues for joint clinical assessments. 2026. Available at: https://health.ec.europa.eu/document/download/b89dc548-10d7-4ef7-9c3b-deea90b639ca_en?filename=hta_20260518_QA_methodological_procedural_questions_HTACG_en.pdf. Accessed: 17/06/2026.

5.         Directorate-General for Health and Food Safety. Updated – List of ongoing joint clinical assessments. 2025. Available at: https://health.ec.europa.eu/latest-updates/updated-list-ongoing-joint-clinical-assessments-2025-09-02_en. Accessed: 17/06/2026.