Erlotinib Monotherapy for the Maintenance Treatment ...

Pharmacoeconomics 2011; 29 (12): 1051-1062 1170-7690/11/0012-1051/$49.95/0

REVIEW ARTICLE

ª 2011 Adis Data Information BV. All rights reserved.

Erlotinib Monotherapy for the Maintenance Treatment of Non-Small Cell Lung Cancer after Previous Platinum-Containing Chemotherapy A NICE Single Technology Appraisal Rumona Dickson,1 Adrian Bagust,1 Angela Boland,1 Michaela Blundell,1 Helen Davis,2 Yenal Dundar,1 Juliet Hockenhull,1 Carlos Martin Saborido,1 James Oyee1 and Vidhya Sagar Ramani3 1 Liverpool Reviews and Implementation Group, University of Liverpool, Liverpool, UK 2 North West Medicines Information Centre, Liverpool, UK 3 School of Cancer Studies, University of Liverpool, Liverpool, UK

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. The Decision Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. The Independent Evidence Review Group (ERG) Report. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Clinical Evidence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 Critique of Clinical Evidence and Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Cost-Effectiveness Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Critique of Cost-Effectiveness Evidence and Interpretation. . . . . . . . . . . . . . . . . . . . . . . . 2.3 Conclusions of the ERG Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. National Institute for Health and Clinical Excellence Guidance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Clinical Effectiveness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1 Subgroup Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2 Generalizability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3 Epidermal Growth Factor Receptor Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Cost Effectiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 Erlotinib vs Placebo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 Erlotinib vs Pemetrexed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 End-of-Life Guidance Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

1051 1053 1054 1054 1055 1056 1056 1058 1059 1059 1059 1059 1059 1059 1059 1059 1060 1060

The UK National Institute for Health and Clinical Excellence (NICE) invited the manufacturer of erlotinib (Roche) to submit evidence for the clinical and cost effectiveness of erlotinib as monotherapy for the maintenance treatment of patients with non-small cell lung cancer (NSCLC) and stable disease following previous treatment with four cycles of platinumcontaining therapy. The Liverpool Reviews and Implementation Group

Dickson et al.

1052

(LRiG) at the University of Liverpool was commissioned to act as the Evidence Review Group (ERG) for this appraisal. The ERG reviewed the clinical- and cost-effectiveness evidence in two stages and in accordance with the decision problem defined by NICE. The analysis of the submitted models assessed the appropriateness of the approach taken by the manufacturer in modelling the decision problem. Analysis also included reliability of model implementation and the extent of conformity to published standards and prevailing norms of practice within the health economics modelling community. Particular attention was paid to issues likely to have substantial impact on the base-case cost-effectiveness results. Clinical evidence was derived from a multi-centre, double-blind, randomized, phase III study designed to address the overall population of NSCLC patients. Outcomes included progression-free survival (PFS) and overall survival (OS). The recruited population was mainly from outside of Western Europe and no patients in the pivotal trial had received pemetrexed as a firstline therapy, which is now accepted clinical practice in the UK. The evidence considered in this article includes only the population for whom marketing authorizations has been received – that is, patients with stable disease following first-line therapy. The trial reported a small but statistically significant increase in both PFS and OS in patients with stable disease receiving erlotinib compared with placebo. However, no significant difference was identified in OS when patients with non-squamous disease and stable disease were considered as a subgroup. The economic evidence was focussed on the ERG’s assessment of three economic models that related to patients with stable disease and compared erlotinib with placebo in the squamous and non-squamous populations and erlotinib with pemetrexed in the non-squamous population. The incremental cost-effectiveness ratios (ICERs) reported by the manufacturer were d39 936 per QALY gained (stable disease, all); d35 491 per QALY gained (stable disease, squamous); and d40 020 per QALY gained (stable disease, nonsquamous). In comparison with pemetrexed, in the cases where erlotinib was considered to be superior or equivalent, erlotinib dominated. In the cases where erlotinib was considered to be slightly inferior, then the ICERs ranged between d91 789 and d511 351 per QALY gained; these ICERs appear in the south-west corner of a cost-effectiveness plane, i.e. erlotinib is cheaper but less effective than pemetrexed. The ERG recalculated the base-case cost-effectiveness results in the manufacturer’s submission, considering nine key areas where corrections and/or adjustments were required, related to time horizon, discounting logic, costs of erlotinib and pemetrexed, cost of second-line chemotherapy, unit costs, utility values, PFS and OS. This resulted in ERG-revised ICERs for the stable disease squamous population of d44 812 per QALY gained, in the stable disease non-squamous population of d68 120 per QALY gained, and, when erlotinib was compared with pemetrexed, the result was d84 029 per QALY gained. All values were above NICE’s perceived willingness-to-pay threshold. After the second Appraisal Committee meeting, the Committee did not recommend the use of erlotinib in this patient population.


Pharmacoeconomics 2011; 29 (12)

Erlotinib Maintenance Treatment for Non-Small Cell Lung Cancer

1053

Key points for decision makers

The ‘just in time’system for providing NICE guidance close to the time of marketing authorization means that there will be instances when the indication(s) in the manufacturer’s submission does not match the authorized indication(s), causing delays in decisions related to guidance

Such differences may require the examination of subgroup analysis, for which the trials were not powered

The methods of calculating survival benefit continue to evolve but need to be based on the use of actual trial data in favour of modelling when trial data are available

Uncertainties remain regarding the NICE ‘end-of-life’ criteria

The UK National Institute for Health and Clinical Excellence (NICE) is an independent organization responsible for providing national guidance to the NHS in England and Wales on a range of clinical and public health issues, as well as appraisal of new health technologies. The NICE Single Technology Appraisal (STA) process is specifically designed for the appraisal of a single health technology for a single indication, where most of the relevant evidence lies with one manufacturer or sponsor.[1] Typically, the process is used for new pharmaceutical products close to launch. The evidence for an STA is principally derived from a submission by the manufacturer/sponsor of the technology, which should be based on a specification developed by NICE. The manufacturer’s submission is critiqued by members of the independent Evidence Review Group (ERG) who produce a report to be considered by the NICE Appraisal Committee (AC). The NICE AC considers the submissions from the manufacturer and the ERG alongside testimony from experts, patients and other stakeholders to formulate preliminary guidance. All stakeholders have an opportunity to comment on this preliminary guidance, after which the AC meets again to produce the final guidance (final appraisal determination [FAD]). This article presents a summary of the ERG report for the STA of erlotinib as monotherapy for the maintenance treatment of patients with non-small cell lung cancer (NSCLC) and stable disease following previous treatment with four cycles of platinum-containing therapy. This article is one in a series of STA summaries being published in PharmacoEconomics.[2-10] ª 2011 Adis Data Information BV. All rights reserved.

Full details of all the relevant appraisal documents (including the appraisal scope, ERG report, manufacturer and consultee submissions, Appraisal Consultation Document [ACD], FAD and comments on each of these) can be found on the NICE website.[11] 1. The Decision Problem Approximately 33 500 new cases of lung cancer occur each year in England and Wales.[12] Diagnosis of the disease tends to be late, with at least 40% of patients confirmed as having stage IIIB or IV disease.[13] Survival rates are therefore poor, with less than 30% of patients alive at 1 year and less than 10% at 5 years.[14] Figures from the NHS Information Centre from the National Lung Cancer Audit for 2007[13] indicated that approximately 14% of reported cases were small cell cancer or mesothelioma. Of the remaining ‡29 000 other reported cases, only 10 500 (approximately 55%) were histologically confirmed as NSCLC. Currently, histological confirmation of lung cancer is made in only 72% of patients in the UK, with variation across centres (range 25–85%).[15] In addition, a significant proportion of these cancers are typed as NSCLC without specifically being subclassified (e.g. non-squamous, adenocarcinoma). Recent NICE guidance[16] for the first-line treatment of NSCLC now calls for histological testing, and therefore testing rates across the UK are expected to increase. Current first-line treatments include the use of platinum-based chemotherapy in combination with gemcitabine, paclitaxel or Pharmacoeconomics 2011; 29 (12)

Dickson et al.

1054

docetaxel, and also gefitinib in a more limited population (epidermal growth factor receptor [EGFR] positive) and with cost limitations.[17] Erlotinib currently has limited approval (based on cost) for use in the UK as a second-line treatment for patients with NSCLC who have progressed after initial treatment.[18] This STA was designed to examine the use of erlotinib as maintenance treatment in patients who had not progressed following first-line treatment. In addition, NICE has recently recommended pemetrexed as an option for the maintenance treatment of people with locally advanced or metastatic NSCLC, other than predominately squamous cell histology, if the disease has not progressed immediately following platinum-based chemotherapy in combination with gemcitabine, paclitaxel or docetaxel.[19] It is challenging to assess drugs for use in clinical practice immediately after marketing authorization is awarded. To illustrate, in this STA the manufacturer’s submission and initial ERG report were completed prior to the decision by the EU licensing authority regarding the terms of the marketing authorization; the terms of the authorization were much more limiting than those originally submitted in the manufacturer’s licence application. The result of this was a re-submission of clinical data and three new economic models to NICE by the manufacturer, focussing on the terms of the marketing authorization and following an initial negative opinion after the first AC meeting. This paper summarizes the overall clinical trial data for only the authorized population, which is patients who have stable disease following initial first-line therapy. Stable disease in this case is defined as neither sufficient tumour shrinkage to qualify for partial response nor sufficient increase to qualify for progressive disease, taking as reference the smallest sum lesion diameter since the treatment started.[20] It is noteworthy that, at the time of writing the ERG report, the results of the SATURN (Sequential Tarceva in Unresectable NSCLC) trial,[20] on which the submission was based, had not been published and data were from the clinical study report submitted as a part of the appraisal process. The overall study report is now available in the public domain.[21] ª 2011 Adis Data Information BV. All rights reserved.

The primary clinical outcome in the SATURN trial[20] was progression-free survival (PFS). Secondary outcomes included overall survival (OS), time to treatment progression (TTP) and quality of life. Cost effectiveness was measured in terms of incremental cost per QALY. 2. The Independent Evidence Review Group (ERG) Report The ERG examined and critiqued both the initial and the subsequent submissions from the manufacturer as well as considering the manufacturer’s response to their request for clarification on a number of issues. The ERG report comprised a critical review of the evidence for the clinical and cost effectiveness of the technology and embodied three aims: To assess whether the manufacturer’s submission conformed to the methodological guidelines issued by NICE; To assess whether the manufacturer’s interpretation and analysis of the evidence were appropriate; and To indicate the presence of other sources of evidence or alternative interpretations of the evidence that could help inform NICE guidance. In addition to providing this detailed critique, the ERG modified a number of key assumptions and parameters within the manufacturer’s economic model to examine the impact of such changes. The next section summarizes the submitted evidence and the ERG’s review of that evidence. 2.1 Clinical Evidence

The clinical-effectiveness evidence was derived from a single trial known as the SATURN trial.[20] This was a multi-centre, double-blind, randomized, phase III study designed to evaluate the efficacy of erlotinib or placebo following four cycles of platinum-based chemotherapy in patients with histologically documented, advanced or recurrent (stage IIIB and not amenable for combined modality treatment) or metastatic (stage IV) NSCLC who had not experienced progression or unacceptable toxicity during chemotherapy. This was a large (n = 889), well designed trial that was Pharmacoeconomics 2011; 29 (12)


conducted in over 100 centres across 26 countries. Eligible patients were randomized to receive either erlotinib or placebo using a minimizationallocation technique, an adaptive randomization method using six factors: EGFR protein expression by immunohistochemistry (IHC), stage of disease at start of chemotherapy, Eastern Cooperative Oncology Status (ECOG) status, chemotherapy regimen, smoking status and region. Justification for only one of these factors, EGFR status, is provided in documentation submitted to the US FDA.[22] Although the trial was large, the parameters of the licence awarded by the European Medicines Agency[23] meant that examination of the evidence was limited to patients with stable disease (n = 487). Of this subgroup, 235 patients received placebo and 252 received erlotinib (table I). To allow for comparison with pemetrexed, additional post hoc analyses were carried out in relation to patient response by histology (squamous/nonsquamous). OS and PFS rates for the entire trial population and these three subgroups are presented in table II and demonstrate a small but statistically significant benefit in all analyses except OS in stable disease patients with non-squamous histology. The manufacturer also compared erlotinib with pemetrexed. The manufacturer stated that, due to a lack of publicly available PFS and OS evidence in patients with stable disease and nonsquamous histology from the JMEN[24] trial, evidence synthesis comparing erlotinib and pemetrexed in this population was not possible. Thus, in the submitted base case, the manufacturer assumed pemetrexed and erlotinib were of equivalent efficacy and then performed an exploratory sensitivity analysis using hazard ratios (HRs) for both PFS and OS of 0.9, 1.0 and 1.1 to define scenarios for potential relative efficacy in the comparison of erlotinib with pemetrexed. Table I. SATURN[20] trial population Population groups

Placebo

Erlotinib

Total trial (n)

451

438

Stable disease [n (%)]

235 (52.1)

252 (57.5)

Stable disease (squamous) [n (%)]

93 (20.6)

97 (22.2)

Stable disease (non-squamous) [n (%)]

142 (31.5)

155 (35.4)


1055

Table II. Progression-free survival (PFS) and overall survival (OS) results from the SATURN trial[20] by types of patient populations: erlotinib vs placebo[20] N

PFS [HR (95% CI)]

OS [HR (95% CI)]

Full analysis set

889

12.3 vs 11.1 mo [0.71 (0.62, 0.82)]

12.0 vs 11.0 mo [0.81 (0.70, 0.95)]

All stable disease

487

12.1 vs 11.3 mo [0.68 (0.56, 0.83)]

11.9 vs 9.6 mo [0.72 (0.59, 0.89)]

Stable squamous

190

NR [0.69 (0.51, 0.93)]

11.0 vs 8.3 mo [0.67 (0.48, 0.92)]

Stable nonsquamous

297

NR [0.69 (0.54, 0.87)]

13.7 vs 10.6 mo [0.76 (0.59, 1.00)]

CI = confidence interval; HR = hazard ratio; NR = not reported.

2.1.1 Critique of Clinical Evidence and Interpretation

Although the original SATURN[20] trial was large and well designed, the limitations of the marketing authorization meant that only a portion of the data could be used for analysis. In addition, the ERG identified a number of other limitations related to the clinical trial evidence: 1. The randomization technique used in the trial included stratification by six different factors. Response to treatment (e.g. stable disease) and histological status were not employed as stratification factors. However, the manufacturer’s submission relies exclusively on post hoc analyses of data related to this subgroup of patients. The ERG considers that the results of the post hoc analyses should be considered with caution, as the trial was not designed to perform this type of analysis and did not adjust for multiple testing. 2. Despite efforts to ensure blinding of patients and investigators, the extent to which patients and investigators were truly blind to treatment allocation throughout the trial is uncertain because patients in the erlotinib arm were significantly more likely to develop a rash and experience diarrhoea than patients in the placebo group. 3. The generalizability of the results of the SATURN trial[20] to patients in England and Wales is uncertain for a number of reasons: only seven patients were recruited from the UK, 75% of patients in the trial were recruited from outside of Western Europe, and the trial did not include patients who had received pemetrexed as a first-line treatment, Pharmacoeconomics 2011; 29 (12)

1056

which is becoming the dominant first-line treatment for patients with non-squamous NSCLC in the UK; hence the response of patients to erlotinib after treatment with pemetrexed is unknown. Paclitaxel appears to be used as a first-line treatment for a greater proportion of patients in the trial than might otherwise be the case in clinical practice in England and Wales.[25] The impact of this when generalizing to patients in England and Wales is unknown. A number of patients in the trial received second-line therapies that are not available to patients in clinical practice in England and Wales; this may affect the magnitude of the OS benefit observed in the trial. 2.2 Cost-Effectiveness Evidence

The manufacturer’s submission included a literature review of cost effectiveness and reported that there were no relevant published economic evaluations available for consideration. In the supplementary evidence submission, the manufacturer provided three new economic models adapted from the models that were previously part of the original submission. These models were created to satisfy four cost-utility comparisons: Stable disease: erlotinib versus placebo (model 1); Stable disease with squamous histology: erlotinib versus placebo (model 2); Stable disease with non-squamous histology: erlotinib versus placebo (model 3); Stable disease with non-squamous histology: erlotinib versus pemetrexed (model 3). As a part of this submission, the manufacturer provided incremental cost-effectiveness ratios (ICERs) reported of d39 936 per QALY gained (stable disease, all); d35 491 per QALY gained (stable disease, squamous); and d40 020 per QALY gained (stable disease, non-squamous). The supplementary evidence submission demonstrated that the manufacturer had incorporated the majority of the amendments recommended in the ERG report. However, only four of the amendments were followed directly and appropriately. In addition, to correct an error in the ª 2011 Adis Data Information BV. All rights reserved.

Dickson et al.

original model, a significant and appropriate change was made by the manufacturer to the estimated costs of best supportive care (BSC). The modelling carried out by the manufacturer included calculations to derive the proportion of patients in each health state over time. It was estimated that, compared with placebo, in the stable disease population of patients with squamous disease, the cost per QALY gained by the use of erlotinib was d35 491 while, in stable disease patients with non-squamous disease, this would rise slightly to d40 020. Erlotinib dominated in the scenarios when erlotinib was assumed to be equally or more effective than pemetrexed. In the scenario in which it was assumed that pemetrexed was more effective than erlotinib, the base-case ICERs ranged from d91 000 to >d500 000. However, the manufacturer highlighted that these ICERs were in the south-west quadrant of the costeffectiveness plane, i.e. erlotinib is both less costly and less effective. 2.2.1 Critique of Cost-Effectiveness Evidence and Interpretation

The ERG took the view that the first of the four comparisons (model 1) was redundant and potentially misleading. The separate comparisons relating to the squamous and non-squamous populations encompass the whole of the stable disease population and a combined analysis merely obscures potentially important differences in clinical effectiveness and cost effectiveness, leading to unpredictable results when combined. Therefore, only results generated with models 2 and 3 are discussed. In addition, the ERG recalculated the base-case cost-effectiveness results in the manufacturer’s submission, considering nine key areas where corrections and/or adjustments were required, related to time horizon, discounting logic, costs of erlotinib and pemetrexed, cost of secondline chemotherapy, unit costs, utility values, PFS and OS. The submitted models did not incorporate the ERG’s preferred method of calculating PFS and OS. The ERG noted that the recorded data in the SATURN trial[20] for PFS among the squamous and non-squamous NSCLC population indicate that no patients remained alive without disease Pharmacoeconomics 2011; 29 (12)


progression at the close of the trial (i.e. the PFS data set is complete). In such situations, there is no justification for resorting to projective modelling to establish the mean duration of PFS. The most appropriate and reliable measure may be derived directly from a Kaplan-Meier survival analysis. Squamous Population

In the squamous population, the Kaplan-Meier analysis showed the mean PFS to be 152 (95% confidence intervals [CI]: 121, 182) days for erlotinib maintenance patients, and 111 (95% CI 82, 140) days for the placebo comparator arm, giving a net benefit in favour of erlotinib of 41 (95% CI 11, 70) days. Post-progression survival (PPS) was then calculated for this population of patients. Examination of the cumulative hazard trends for survival following disease progression revealed that, after a brief period (about 10 weeks) of common mortality risk, the two trial arms diverge, each following a steady constant hazard rate. This indicates that the most appropriate parametric function for projective modelling of PPS is an exponential survival curve. However, the recorded data for follow-up beyond disease progression continue for nearly 2 years, so the ERG decided that the most accurate and reliable basis for estimation of the mean PPS was to use the Kaplan-Meier ‘area under curve’ estimates up to a common maturity stage of follow-up (in this case we used Kaplan-Meier survival