Two Small-Cell Lung Cancer Drugs, Two JCAs, Two Very Different Evidence Stories

Share:

Focus keyword Joint Clinical Assessment evidence

Joint Clinical Assessment evidence requirements have been tested twice in recent months, through the JCA reports on lurbinectedin (Zepzelca) and tarlatamab (Imdylltra), both indicated for small-cell lung cancer. The two dossiers took very different approaches: lurbinectedin submitted a single well-aligned RCT against one PICO; tarlatamab addressed seven PICOs across four populations using a mix of direct and indirect evidence. Both were picked apart. This analysis draws out what the two JCA reports reveal about the framework in practice, and what the implications are for trial design, evidence strategy, and pre-launch planning under EU HTAR.

Lurbinectedin and tarlatamab were recently assessed under the EU JCA: One brought a single, tightly aligned RCT; the other addressed seven PICOs held together by a mix of direct and indirect evidence. Here is what the two JCA reports reveal and what they mean for evidence strategy.

Lurbinectedin vs tarlatamab at a glance

 Lurbinectedin (Zepzelca)Tarlatamab (Imdylltra)
Assessment scope1 PICO7 PICOs, 4 populations
Pivotal evidenceIMforte, open-label RCTDeLLphi-304, open-label RCT
Indirect evidenceNoneNMA, unanchored MAIC, Bucher ITC
Direction of effectOS attenuated at mature cut, PFS heldConsistently favoured tarlatamab
Statistical controlNominal p-values onlyNominal p-values only
Principal weaknessEffect size and population alignmentCertainty and indirect comparison validity
Likely national consequenceRestricted reimbursement, unfavourable ICERWeaker negotiating position on certainty grounds

What did the JCA say about lurbinectedin’s evidence package?

Lurbinectedin (Zepzelca): A narrow scope, well-aligned evidence package and the lingering limitations

The evidence package

  • Narrow assessment scope: A single PICO.
  • Manufacturer evidence: One open-label RCT (IMforte), direct comparison of lurbinectedin + atezolizumab, vs atezolizumab monotherapy. No ITCs.
  • Evidence closely aligned to assessment scope.

Where the JCA pushed back

  • Efficacy weakened as data matured: At first data cut, both OS and PFS were statistically significantly increased, in favour of lurbinectedin. At the later, more mature data cut, only PFS held, while the OS effect was attenuated, with the HR approaching 1. Additionally, all p-values were nominal only, with no control for multiplicity.
  • Population narrower than label: Study exclusion criteria (ECOG PS ≥ 2, CNS metastases at baseline) meant the trial population did not cover the full claimed indication.
  • Outcome gaps: Some requested outcomes were either not recorded in the study IMforte or not included in the JCA report due to methodological flaws & inappropriate analysis methods. Also, ORR and complete response rate carried limitedrelevance to the full population, being based only on patients with measurable disease at randomisation.
  • Open-label study design underlies the risk of bias, except OS and PFS, where risk of bias is considered low. With PFS, Independent Review Facility (IRF) assessment stopped after the first data cut, and only investigator-assessment was available for the second data cut. However, high IRF-investigator concordance and a short gap between the data cuts kept the overall risk of bias low.
  • PRO reliability is undermined: Systematically shortened observation; asymmetric observations between arms; low, imbalanced questionnaire return rates.
  • Safety reliability is undermined: Systematicallyshortened observation and insufficient prophylaxis in the intervention arm.
  • Some signals work against the value story: Statistically significant disadvantage for lurbinectedin + atezolizumab on certain HRQoL scales and safety outcomes.

How GPI interprets this

  • Value story is thin where payers look closest. No mature, significant OS gain; a nominal-only PFS gain; hints of HRQoL and safety deterioration; all likely pointing to a thin QALY gain. This is where payers will press, even if same evidence landed well with regulators.
  • An unfavourable ICER looks likely, making a price premium harder to defend, on top of the atezolizumab backbone therapy.
  • Reimbursement may be restricted to resemble the trial population, given the trial was narrower than the label.
GPI Horizon

Stress-testing your evidence strategy before the JCA does it for you.

GPI Horizon evaluates pre-launch assets against the evidence domains most scrutinised in EU Joint Clinical Assessments, from trial design quality to indirect comparison methodology and PICO coverage. The time to identify weaknesses is before the trial is designed, not after the assessment.

What did the JCA say about tarlatamab’s multi-population evidence submission?

Tarlatamab (Imdylltra): A wide scope, multi-population, multi-comparator submission, with a mix of direct-indirect evidence

The evidence package

  • Wide assessment scope: 4 distinct populations, 7 PICOs.
  • Manufacturer evidence: Mix of comparative evidence: Direct comparison RCT (DeLLphi-304), NMA, unanchored MAIC, Bucher ITC. However, not all requested outcomes were provided by the manufacturer, due to missing data.
  • When statistically significant differences emerged between tarlatamab and its respective comparator, they consistently favoured tarlatamab, across survival, morbidity, safety and HRQoL.

Where the JCA pushed back:

On the direct RCT (DeLLphi-304)

  • PFS & ORR: Risk of bias considered high. Certainty is limited by solely investigator-based, unblinded assessment in open-label study (RECIST v1.1, no blinded independent central review).
  • OS: Certainty islimited by missing information on subsequent and post-progression therapy.
  • ORR & DOR: Results are potentially biased by unequal proportions of missing assessments between arms.
  • PROs of disease symptoms, HRQoL, health status: Systematically shortened observation; completion rates rapidly declining and uneven between groups.
  • Safety outcomes: Systematically shortened observation, uneven between groups.

On the indirect evidence

  • Unanchored ITCs rely on strong assumptions that are challenging to meet; if violated, effect estimates are likely biased. Inherently greater risk of bias.
  • Residual confounding resultsin a risk of bias of unknown magnitude & direction (confounders were identified but left unadjusted for in the MAIC, or not identified at all).
  • MAIC positivity & overlap assumptions were not assessed, adding uncertainty around robustness & internal validity.
  • No control for multiplicity was applied; nominal-only p values.
  • OS certainty is also limited by missing information on subsequent therapies.
  • PFS is potentially biased by informative censoring and/or differences in censoring mechanisms between studies.
  • NMA certainty is very low: Similarity assumption for relevant studies was severely violated. Additionally, NMA consistency could not be assessed.
  • Assessment intervals differed between studies, meaning progression might have been recognised faster in DeLLphi-304.

On the source studies feeding the indirect comparisons

  • Baize 2020: Inconsistencies in reported data of sources used, flagged errors, unreported derivation of logHR standard errors. For PFS, results were largely descriptive and highly uncertain, with no further information available to interpret comparative effectiveness (e.g., patient characteristics, treatment/observation periods); hence, data should be interpreted with caution and not regarded as reliable.
  • Von Pawel 1999: Transparency issues; extracted data underlying the HR confidence interval were not reported.

How GPI interprets this

  • The efficacy signals are there, but their certainty is discounted.
  • Indirect evidence is more fragile, creating a weaker starting position with national reimbursement authorities.
  • Expect the data quality of source studies behind the indirect comparisons to be scrutinised, too.

What do the two JCAs tell us about EU HTAR evidence strategy?

So, what are GPI’s conclusions across the two JCAs?

1. Evidence strategy starts at trial design

  • The weaknesses JCA can expose are largely locked in years before making the JCA submission, when the trial is designed.
  • Anticipate the assessment scope and PICO matrix. Build your trial(s) and evidence package to cover them, or the gaps will get flagged at European level.
  • It is not just about compiling a complete enough dossier; it is about developing a robust evidence strategy.
  • Identify upfront what historical evidence cannot address and fill the gap with trials and complementary evidence that are robust enough.

2. Methodology is scrutinised, even for RCTs

  • Methods must be scientifically robust and well-justified. Scrutiny is high, even for RCT evidence.
  • Trial blinding and follow-up matter. Unblinded assessments and shortened, uneven observations create uncertainty.
  • Assessors look for the mature data cut. The last pre-specified data cut, with the longest observation, carries the most HTA weight.
  • Fix what can be fixed: for PROs, a long-enough observation period, sufficient completion rates and consistent collection across arms; for PFS, a blinded IRF alongside the unblinded investigator assessment.

3. Breadth versus quality is a real trade-off

  • Both lurbinectedin & tarlatamab submitted results against every PICO (all included in the JCA report), unlike tovorafenib, with its many unaddressed PICOs. However, covering a broad PICO range may force reliance on weaker evidence.
  • Allow enough time to gather RWE and run additional analyses to the necessary quality.

4. Use the engagement tools, and remember that JCA is not the final word

  • Use JSC early to raise issues, close gaps and align on PICOs.
  • Engage local affiliates strategically.
  • JCA harmonises the clinical assessment, not the price. It gives a common starting point, but Member States still layer on their own economic, budget-impact, affordability and country-specific considerations and decide for themselves. The same JCA can produce very different pricing and reimbursement outcomes across markets. The evidence still has to land at country level.

Next up: Watching how each Member State uses the JCA within its own processes and decisions.

About this analysis

Global Pricing Innovations helps pharmaceutical companies anticipate the assessment scope their assets will face and build evidence strategies that hold up under it. GPI Horizon™ provides pre-launch HTA and international reference pricing intelligence. GPI Pulse™ tracks pricing and reimbursement decisions in real time.

If you are designing a trial that will face a Joint Clinical Assessment, the time to stress-test your evidence strategy is now.

Contact: sales@globalpricing.com

sources: Joint Clinical Assessment report on lurbinectedin (Zepzelca); Joint Clinical Assessment report on tarlatamab (Imdylltra)

Acronyms: CNS – central nervous system | DOR – duration of response | ECOG PS – Eastern Cooperative Oncology Group performance status | HR – hazard ratio | HRQoL – health-related quality of life | HTA – health technology assessment | ICER – incremental cost-effectiveness ratio | IRF – Independent Review Facility | ITC – indirect treatment comparison | JCA – Joint Clinical Assessment | JSC – Joint Scientific Consultation | MAIC – matching-adjusted indirect comparison | NMA – network meta-analysis | ORR – objective response rate | OS – overall survival | PFS – progression-free survival | PICO – population, intervention, comparator, outcome | PRO – patient-reported outcome | QALY – quality-adjusted life year | RCT – randomised controlled trial | RECIST – Response Evaluation Criteria in Solid Tumours | RWE – real-world evidence.

FAQs

A: The EU Joint Clinical Assessment (JCA) framework requires manufacturers to submit evidence that addresses every PICO (population, intervention, comparator, outcome) set by Member States. Assessment quality is high regardless of whether evidence comes from a randomised controlled trial or from indirect comparisons. Recent assessments have shown that clean RCT evidence can still be discounted for open-label design, shortened observation, or attenuating survival signals at mature data cuts. Indirect evidence carries additional scrutiny around residual confounding, MAIC overlap assumptions, and source study data quality.

A: Three lessons emerge across the two assessments. First, evidence strategy starts at trial design: most weaknesses the JCA exposed were locked in years earlier, when trials were designed, comparators chosen, and endpoints set. Second, methodology is scrutinised even for RCTs: trial blinding, mature data cuts, observation length, and consistent PRO collection all carry weight. Third, breadth versus quality is a real trade-off: covering a wide PICO range may force reliance on weaker evidence, so allowing enough time for RWE and additional analyses is essential.

A: The Joint Clinical Assessment harmonises the clinical assessment across the EU but does not determine national pricing and reimbursement outcomes. The JCA provides a common starting point on relative clinical effectiveness and safety, but Member States still layer on their own economic evaluations, budget impact considerations, affordability assessments, and country-specific factors. The same JCA can produce very different pricing and reimbursement outcomes across different European markets. The evidence still has to land at country level.

Designing a trial that will face a Joint Clinical Assessment?

GPI helps pharmaceutical teams anticipate the assessment scope their assets will face and build evidence strategies that hold up under it. Whether you are early in trial design or preparing a JCA dossier, we can help you stress-test the approach before it goes to Member States.

Table of Contents