Pemetrexed for the first-line treatment of locally advanced or metastatic non-small cell lung cancer

Article history

The research reported in this article of the journal supplement was commissioned and funded by the HTA programme on behalf of NICE as project number 08/88/01. The assessment report began editorial review in April 2009 and was accepted for publication in May 2009. See the HTA programme website for further project information (www.hta.ac.uk). This summary of the ERG report was compiled after the Appraisal Committee’s review. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the referees for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

none

Copyright statement

© 2010 Queen’s Printer and Controller of HMSO. This journal is a member of and subscribes to the principles of the Committee on Publication Ethics (COPE) (http://www.publicationethics.org/). This journal may be freely reproduced for the purposes of private research and study and may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NETSCC, Health Technology Assessment, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

2010 Queen’s Printer and Controller of HMSO

Description of the underlying health problem

According to Cancer Research UK,3 over 38,000 people were diagnosed with lung cancer in the UK in 2005. Survival from lung cancer is poor with around a quarter of patients (25% men, 26% women) surviving for 1 year after diagnosis, falling to 7% for 5 years after diagnosis, and the disease was responsible for approximately 34,000 deaths in 2006. Reasons for this poor prognosis include the late identification of the disease and low active anticancer treatment rates.

The main subtypes of NSCLC are squamous cell carcinoma, adenocarcinoma and large-cell carcinoma. According to a recent audit of England and Wales,4 33% of patients had squamous NSCLC, 25% had adenocarcinoma, 4% had large-cell carcinoma with the remaining 36% defined as NSCLC ‘not-otherwise specified’ (NSCLC-NOS).

Scope of the evidence review group report

The ERG report presents the results of the assessment of the manufacturer’s (Eli Lilly Ltd) evidence submission regarding the use of pemetrexed with cisplatin compared to platinum-based chemotherapy for the first-line treatment of locally advanced or metastatic NSCLC. The report includes an assessment of both the clinical effectiveness and cost-effectiveness evidence submitted by the manufacturer. The primary clinical outcome measure was overall survival with secondary outcomes of progression-free survival (PFS), response to therapy and tolerability. The cost-effectiveness data were presented as incremental cost-effectiveness ratios (ICERs).

Pemetrexed (Alimta^®) is a multitargeted anticancer antifolate agent that exerts its action by disrupting crucial folate-dependent metabolic processes essential for cell replication. It was approved by the European Commission for the first-line treatment of NSCLC (other than predominantly squamous cell histology) in combination with cisplatin on 8 April 2008. In this group of patients, it is indicated for patients with locally advanced or metastatic NSCLC.

Summary of submitted clinical evidence

The majority of the efficacy evidence described in the manufacturer’s submission is derived from a phase III open label randomised controlled trial (RCT) known as the JMDB trial5 and is presented in Table 1. The trial achieved its primary objective to demonstrate non-inferiority of pemetrexed/cisplatin to gemcitabine/cisplatin for overall survival in all patients with NSCLC [median 10.3 months for both trial arms, adjusted hazard ratio (HR) = 0.94; 95% confidence interval (CI) 0.84 to 1.05]. As pemetrexed is only indicated for the first-line treatment of patients with non-squamous NSCLC, a subgroup analysis of patients with non-squamous NSCLC was presented that reported superiority of pemetrexed/cisplatin on the primary outcome of overall survival compared with gemcitabine/cisplatin (median 11.0 and 10.1 months, respectively, adjusted HR = 0.84; 95% CI 0.74 to 0.96). In the population of patients with non-squamous NSCLC, median PFS was not reported to be statistically superior and, while tumour response rates were reported to be higher for pemetrexed/cisplatin, significance tests were not reported.

The manufacturer also defined a more specific target population of patients with adenocarcinoma or large-cell carcinoma. In this target population, median overall survival was also significantly superior in the pemetrexed/cisplatin group (median 11.8 and 10.4 months, respectively, adjusted HR = 0.81; 95% CI 0.70 to 0.94). It should be noted that defining the target population in clinical practice would require more specific testing than is currently standard practice in the UK (as treatment is currently based on whether patients have squamous or non-squamous NSCLC). A study of preoperative histological classification of lung cancer7 cited by the manufacturer suggests that diagnosing adenocarcinoma may be particularly challenging.

Because no other studies were found comparing pemetrexed/cisplatin with any other relevant comparator, additional efficacy evidence was presented from two phase III RCTs comparing gemcitabine/cisplatin with gemcitabine/carboplatin8 and docetaxel/cisplatin9 (Table 2). The manufacturer’s submission reported from its indirect comparisons’ analysis that median overall survival and PFS and tumour response rates were more favourable for pemetrexed/cisplatin than for any other comparator.

With the exception of nausea, pemetrexed/cisplatin appeared to be more tolerable than gemcitabine/cisplatin in terms of grade 3/4 toxicities. No safety issues related to pemetrexed/cisplatin arose beyond those already previously documented. No significant differences were reported for tolerability regarding the different cisplatin regimens (pemetrexed/cisplatin, gemcitabine/cisplatin and docetaxel/cisplatin). Gemcitabine/carboplatin reported less non-haematologic toxicity in terms of nausea and vomiting, and more haematoxicity in terms of an increased incidence of thrombocytopenia than gemcitabine/cisplatin.

Summary of submitted cost-effectiveness evidence

The manufacturer did not identify any published cost-effectiveness analyses of pemetrexed for the first-line treatment of patients with NSCLC. Therefore economic evidence was derived solely from a de novo economic model developed by the manufacturer.

A Markov model was developed to evaluate the cost-effectiveness of pemetrexed/cisplatin compared with gemcitabine/cisplatin, docetaxel/cisplatin and gemcitabine/carboplatin. The clinical data used in the economic evaluation were primarily generated from the JMDB trial,5 with additional data from the two further trials used in the indirect comparisons analysis. 8,9 Although the economic evaluation was trial-based, the modelling component enabled the extrapolation of health effects beyond the period of 30 months of the trial, adopting a lifetime horizon (taken as 6 years) for the consideration of costs and benefits. The perspective of the model was that of the UK NHS and Personal Social Services.

The ERG identified series problems with this economic model. As a result, three different versions of the model were submitted to NICE and considered by the ERG. The ICERs estimated by this final version of the model ranged from £8056 to £33,065, depending on the comparator, the population and the application of a continuation rule.

Commentary on the robustness of submitted evidence

The JMDB trial was a randomised controlled head-to-head clinical trial that was well-designed, used robust randomisation techniques and was suitably powered to demonstrate the primary non-inferiority objective of the trial for the total population of patients with squamous and non-squamous NSCLC. Subgroup analyses of patients with non-squamous NSCLC and the manufacturer’s own defined target population were conducted. The subgroups appeared to be clinically appropriate and confidence in the robustness of the findings was increased by the fact that these two subgroups were both pre-specified and relatively large in size.

Evidence from the indirect comparisons should be treated with caution as other comparators defined in the original scope and decision problem (vinorelbine and paclitaxel in combination with cisplatin or carboplatin and docetaxel/carboplatin) were excluded from the indirect comparisons analysis. In addition, the statistical approach employed to generate the findings is not considered to be the most optimal, as calculations were based on median survival times and individual trial arm level data from within trials were compared, thus ignoring the benefits of randomisation. Finally, data was only available for all patients with NSCLC (i.e. squamous or non-squamous NSCLC) in all but the JMDB trial. Thus, the HRs for each subgroup in the JMDB trial were used to estimate HRs for subgroups of patients given gemcitabine/carboplatin or docetaxel/cisplatin in the comparator trials. However, it was impossible to confirm from the data reported by the published papers of these trials whether the relative effects found in the JMDB trial would be consistent across subgroups for these patients.

Examination of the final version of the economic model submitted to NICE and considered by the ERG showed that, although minor modifications had been made to correct some of the problems identified by the ERG with earlier versions, the model still failed to adequately address the crucial problems at the heart of the model. These were beyond the remit of the ERG to address, and included:

• The chosen model design was not obviously suitable for modelling the disease and treatments described in the published clinical trial, imposing as it does serious constraints on the possibility of representing the observed patterns of response to treatment and progression of disease.

• The implementation of the model was marked by examples of basic errors with marked consequences.

• There is little evidence of a systematic approach by the manufacturer to identifying and eliminating errors in the development of the model, or of attempting to replicate the prime source of information for the model, i.e. the JMDB trial itself.

• The restriction of comparators to those that are relatively high cost is likely to give a misleading impression of the true cost-effectiveness of pemetrexed regimen. Furthermore, gemcitabine will be off patent in the UK from March 2009 and may soon become available in generic form at a lower price. This was not considered in the manufacturer’s model.

• The methods used for adjusting treatment effects (positive and negative) when a scenario is used with fewer treatment cycles than in the trial evidence are not obviously robust and defensible and may tend to overestimate the outcome benefits to be expected from use of pemetrexed/cisplatin, while underestimating the additional cost.

Thus, the ERG believed that the model requires extensive modification and redesign, subjected to thorough validation against the JMDB trial results. A full quality audit was also required as it is likely that further model inconsistencies may be present that have not yet been identified.

Disclaimers

The views expressed in this publication are those of the authors and not necessarily those of the HTA programme or the Department of Health.