Pemetrexed maintenance treatment for non-squamous non-small-cell lung cancer after pemetrexed and cisplatin

Scope of the appraisal

4.2 The committee noted the evidence presented by the company on the use of pemetrexed maintenance treatment for people with locally advanced or metastatic (stage IIIB and IV) non‑squamous NSCLC, with performance status of 0–1, whose disease completely or partially responded or was stable after first‑line treatment with pemetrexed plus cisplatin. The committee was aware that this appraisal was concerned with the extension to the marketing authorisation for pemetrexed maintenance treatment after induction therapy with pemetrexed and cisplatin, and that NICE has already issued guidance on the use of pemetrexed after platinum‑based chemotherapy plus gemcitabine, paclitaxel or docetaxel (pemetrexed for the maintenance treatment of non-small-cell lung cancer).

4.3 The committee considered the decision problem as outlined in the final NICE scope for the appraisal, noting that in the scope, best supportive care (including bisphosphonates and palliative radiotherapy) was identified as the comparator. The committee heard from the clinical experts that standard practice for patients receiving pemetrexed‑containing chemotherapy is observation and further treatment to be considered only at the time of disease relapse. The committee therefore concluded that best supportive care was an appropriate comparator for this appraisal because it was considered equivalent to the current practice of observation after first‑line induction chemotherapy.

Performance status in clinical practice

4.4 The committee discussed performance status in relation to both first‑line chemotherapy for advanced non‑squamous NSCLC and maintenance treatment. It noted that NICE's guideline on lung cancer recommends that:

Clinical effectiveness

4.5 The committee was aware that the only evidence of clinical effectiveness came from 1 randomised clinical trial (PARAMOUNT). It considered that PARAMOUNT was well designed. The committee then discussed the applicability of the PARAMOUNT data to the population of people with NSCLC in England. It heard from the clinical experts that patients in clinical trials are generally younger and fitter than those seen in clinical practice in England. The committee noted that 32% of patients who entered PARAMOUNT had a performance status of 0 at the end of 4 cycles of induction chemotherapy. The committee concluded that patients in PARAMOUNT were generally younger and fitter than those seen in clinical practice.

4.6 The committee discussed the number of pemetrexed maintenance cycles that a patient would be likely to receive, conscious of the evidence review group's (ERG's) comment that the mean number of cycles of treatment with pemetrexed in PARAMOUNT was more than 7 cycles and that 6 cycles might be considered a likely maximum in UK clinical practice. However, the committee heard from clinical experts that patients would have treatment until disease progression or unacceptable toxicity, or patient or physician choice to stop treatment early, rather than with a set number of cycles. Based on the evidence put forward by the clinical experts, the committee concluded that patients would receive pemetrexed maintenance treatment until disease progression or unacceptable toxicity.

4.7 The committee discussed and reviewed the progression‑free survival and overall survival data from PARAMOUNT. The committee concluded that pemetrexed monotherapy for the maintenance treatment of locally advanced or metastatic non‑squamous NSCLC in patients whose disease has not progressed immediately after induction therapy with pemetrexed and cisplatin (and with a performance status of 0–1) provides a statistically significant gain in progression‑free survival and overall survival compared with placebo.

Adverse effects of treatment

4.8 The committee noted the greater rates of grade 3 and 4 adverse reactions associated with pemetrexed maintenance treatment than with placebo, specifically increased hospitalisations, fatigue and blood transfusions. Increased grade 1 and 2 adverse reactions were also noted by the committee, in particular nausea and vomiting, but there was no statistically significant difference in health‑related quality of life between the pemetrexed and placebo arms of PARAMOUNT. The committee concluded that treatment with pemetrexed maintenance therapy in this setting was associated with clinically significant but acceptable adverse reactions.

Cost effectiveness

4.9 The committee considered the assumptions around resource use in the economic model submitted by the company. It first discussed the monitoring requirements for pemetrexed maintenance treatment. The committee was aware that patients in PARAMOUNT had a CT scan every 6 weeks, and heard from clinical experts that a CT scan would be repeated every 2 to 3 months during maintenance treatment in UK clinical practice. The committee noted the company's original assumption that 3% of patients would need additional scans, occurring every 24 weeks. The committee noted that the revised base case and the updated revised base case provided by the company increased the proportion of patients needing additional scans to 100% and increased the frequency of CT scans to once every 12 weeks. Based on the clinical experts' comments, the committee concluded that this was an acceptable assumption.

4.10 The committee discussed the costs of post‑progression chemotherapy in the company's base‑case analysis. It was aware that the company's original base case had assumed that patients on pemetrexed would be 12% less likely to receive additional chemotherapy after progression than patients on placebo. It noted that the trial data did not support this assumption, because a similar proportion of patients in both groups received additional chemotherapy after progression. The committee also considered that the time at which a patient's disease progresses on maintenance pemetrexed treatment would be later than for those who received placebo and that this might therefore have affected the timing and numbers of patients recorded in the trial as having post‑progression chemotherapy. The committee noted that the company had accepted this as an amendment in the revisions to its base case and updated revised base case, assuming equal rates of post‑progression chemotherapy for pemetrexed and placebo. The committee concluded that this was an appropriate amendment.

4.11 The committee welcomed the company's revisions to its base case and updated revised base case but noted that some concerns remained about concomitant medication costs, utility model design and survival projection. The committee discussed the absence of the cost of the concomitant medications that are needed with pemetrexed (vitamin supplements and dexamethasone) in the company's revised and updated revised base‑case analyses. The committee heard from the company that a free 'supplementation pack' that includes vitamins and dexamethasone had been introduced to hospitals in the UK. The committee was aware that the impact on the incremental cost-effectiveness ratio (ICER) of including the concomitant medication costs was small (around £100). The committee concluded that the effect on the ICER of including the concomitant medication costs was not significant, particularly when compared with the other outstanding issues, and so did not need to be considered further.

4.12 The committee considered the method used for estimating utility in the economic model. It heard from the company that its preferred method of estimating utility was the unadjusted mixed model based on the PARAMOUNT EQ‑5D individual patient data because it gave intuitive utility values. The committee agreed that the values were intuitive but remained concerned that including a cycle variable to account for changes in utility with treatment cycles in the adjusted model would cause significant instability. The committee discussed the alternative utility values from the Nafees model, and was aware that these were based on patients with NSCLC receiving second‑line treatment, rather than maintenance treatment. In addition, the committee had reservations about using the Nafees utility values (which were not obtained using EQ‑5D methods) in preference to EQ‑5D data from PARAMOUNT. The committee welcomed the availability of EQ‑5D data from the trial and agreed that they should be used to provide the utility values for the model. However, the committee remained cautious about how the unadjusted regression model had explored the effect of treatment on utility. Furthermore, the committee was aware that to accommodate the impact of a loss in utility (disutility) from an adverse effect of treatment, the company had calculated an average disutility for all pre- and post‑progression health states, and applied these to the pre‑progression health states only. Although the committee accepted that disutility from treatment‑related adverse reactions would only occur during the pre‑progression phase (while a patient is still on treatment), it was not appropriate that the disutility value should be estimated from an average of the on‑treatment and off‑treatment times. The committee concluded that although the unadjusted model had not been optimally executed and disutility had not been correctly estimated, the values were still preferable to those from the Nafees model, which were neither EQ‑5D based nor from the population of interest.

4.13 The committee discussed the source of resource use data within the economic model used to calculate the cost of adverse reactions. It was aware that the economic model allowed 2 methods for calculating resource use, the company's preferred approach, 'JMEN method', and the ERG's preferred approach, 'PARAMOUNT'. The committee understood that the PARAMOUNT method used data directly from PARAMOUNT and was not limited to including only 4 adverse reactions. It agreed with the ERG that the more detailed PARAMOUNT approach was reasonable. The committee concluded that it was more appropriate to use the PARAMOUNT method because this did not limit the adverse reactions included and was more detailed.

4.14 The committee considered the evidence in support of a post‑progression benefit of pemetrexed over placebo. The committee noted the ERG's Kaplan–Meier analysis of post‑progression survival indicated that for the 2 trial arms, survival corresponded very closely. The committee also understood from comments made by the clinical experts at the first committee meeting that a continued benefit of pemetrexed over best supportive care after disease progression is difficult to explain. The committee heard from the company that it was not considered plausible that a patient would receive a benefit from pemetrexed throughout treatment and that the benefit would suddenly stop immediately on discontinuation of treatment. However the committee noted that although it may not be plausible to assume an immediate end to the benefit of treatment on disease progression, this was not the same as assuming a significant benefit of pemetrexed over and above that of placebo. During consultation the company highlighted Stein et al. (2009) and Stein et al. (2011) as evidence of treatment effect reducing tumour growth rates after treatment is stopped (using non‑pemetrexed treatments) in patients with advanced prostate and renal cancer. The committee considered that the findings in these papers did not support an extended benefit of chemotherapy after disease progression, because the only scenario in which post‑treatment benefit was suggested occurred with a vaccine treatment. The committee did not find any reason for basing its decision on anything other than the PARAMOUNT data, which did not show any evidence for a post‑progression benefit of pemetrexed over placebo. The committee concluded that no evidence to support a post‑progression benefit for pemetrexed over placebo had been provided throughout the appraisal.

4.15 The committee further discussed the approaches to survival modelling. It noted that the company had challenged the committee's review of the ERG's modelling approach, suggesting that the ERG should have done statistical tests, such as goodness of fit. The committee suggested to the company that, in the single technology appraisal process, the onus is on the company to provide the evidence, including an economic model. It considered that the ERG's role is to critique the evidence, rather than build a new model. The committee heard that the company agreed that this is the case but still considered that the committee should have commented more specifically on the company's supporting statistical tests, which the company considered to provide justification for the use of the gamma distribution to project overall survival. The committee then heard from the ERG that additional statistical tests were not necessary because, in the ERG's opinion, the PARAMOUNT data were sufficiently mature to allow calculation of the survival advantage of pemetrexed without any extrapolation. This was because the trajectories of the pemetrexed and placebo curves were parallel and could be overlaid once overall survival was less than about 37% by shifting the overall survival curve of the control arm to the right by approximately 200 days. This approach allowed the difference in survival to be calculated from the differences in areas under the curves, and was approximately 106 days (3.49 months). The committee understood from the ERG that this approach removed the need to use a hazard function to model the survival data, and was based on a new exploration of data that the company had previously provided, and that had been available throughout the course of the appraisal. After further discussion, the company agreed that there appeared to be no statistically significant difference in post‑progression survival between the trial groups. The committee concluded that extrapolation of the data was not needed and that its decision on the cost effectiveness should be made based on the actual data.

4.16 Based on its discussions (see sections 4.12–4.15), the committee considered that the most appropriate ICER should be calculated using the revised assumptions about cost and resource use, the unadjusted utility model and the ERG's approach to survival modelling. The result of combining these assumptions was confirmed by the ERG to produce an ICER of approximately £74,500 per quality-adjusted life year (QALY) gained.

End-of-life considerations

4.17 The committee considered supplementary advice from NICE that should be taken into account when appraising treatments that may extend the life of patients with a short life expectancy and that are licensed for indications that affect small numbers of people with incurable illnesses. For this advice to be applied, all the following criteria must be met:

4.18 Noting evidence from the National Lung Cancer Audit (2013) and the survival time of patients on placebo and best supportive care in PARAMOUNT, the committee concluded that the life expectancy of patients with advanced non‑squamous NSCLC is normally less than 24 months. For the criterion about extension to life, the committee noted the results from PARAMOUNT showing that there was a statistically significant increase in median overall survival of 2.85 months for pemetrexed compared with best supportive care. Although this was not greater than 3 months, the committee was aware that all of the modelled estimates provided by the company and the ERG were greater than 3 months. The committee therefore concluded that there was sufficient evidence to indicate that the treatment offers extension to life of at least 3 months.

4.19 The committee considered the patient population for which pemetrexed is licensed, taking into account all the therapeutic indications for pemetrexed identified in the summary of product characteristics. The committee noted that pemetrexed has a UK marketing authorisation for the following indications:

4.20 The committee discussed the small patient population criterion. It heard from NICE that, for treatments for small groups of patients, higher prices, and therefore reduced cost effectiveness, were more likely to be justified given the need to recoup costs of development of the product if the licensed indications only apply to a small potentially eligible patient population. It further heard that the case for reduced cost effectiveness weakens as the potential total population for a product increases. Therefore, taking into account the cumulative population covered by all the indications in the marketing authorisation needs to be considered. The committee understood that the small patient population criterion was intended to recognise the long‑term benefits to the NHS of innovation. The committee was aware that, for this reason, it was appropriate to add the potential populations for all indications covered by the marketing authorisation together rather than consider them based on actual use. As advised by NICE, the committee considered that the calculation of the total population should reflect only the population covered by the licensed indications in the countries where NICE guidance has formal effect (since April 2013, that is England, rather than England and Wales). The committee recognised that in the case of patients having first‑line chemotherapy with pemetrexed plus cisplatin and then continuing on maintenance pemetrexed, this represented additional opportunities for the company to recoup the costs of development for pemetrexed.

4.21 The committee considered the population size for pemetrexed as first‑line therapy. It was aware that the licensed indication is that pemetrexed should be given in combination with cisplatin, and, based on comments from the clinical experts, that only patients with a performance status of 0–1 would be considered for treatment with cisplatin. However, for patients who are fit enough to tolerate this combination, the committee heard from the clinical experts that this would be the first‑line treatment of choice. The committee noted that, according to the National Lung Cancer Audit (2013), the number of patients in England with confirmed NSCLC who have a performance status of 0–1 and who have stage IIIB or IV disease is 6,735. It understood that 68% of these patients would have non‑squamous histology (see NICE's guideline on lung cancer), therefore the potential population eligible for first‑line therapy with pemetrexed would be 4,580. The committee was aware of comments received during consultation suggesting that this figure included a number of people with lung cancer that is epidermal growth factor receptor tyrosine kinase (EGFR-TK) mutation positive and that in clinical practice; these people would not receive pemetrexed. The committee discussed this comment but was of the opinion that it is appropriate to estimate the potential population as defined by the licensed indication. The committee was aware that the decision to estimate the potential population as defined by the licensed indication, rather than actual use, was in line with an appeal panel decision for erlotinib monotherapy for maintenance treatment of non-small-cell lung cancer. The committee concluded that the population to be included in the calculations for first-line treatment was therefore 4,580.

4.22 The committee considered the population size for the maintenance indication for pemetrexed. It was aware that the licence extension meant that pemetrexed would be an option for patients as 'continuation maintenance' (that is, pemetrexed maintenance treatment after induction therapy with pemetrexed plus cisplatin) or 'switch maintenance' (that is, pemetrexed maintenance treatment after induction therapy that does not include pemetrexed). The committee noted that the National Lung Cancer Audit (2013) reported that, of those eligible for first‑line treatment (4,580), 57.2% (2,620) receive first‑line chemotherapy and of these 40% (1,048) receive pemetrexed plus cisplatin. Of these, the company estimated that 58.4% (612) would be eligible for pemetrexed continuation maintenance treatment. The committee accepted this number. For switch maintenance, the committee noted the comments received during consultation, which suggested that pemetrexed would either be used in a first‑line setting or as switch maintenance, but not as both. The committee found the comments received about switch maintenance to be reasonable and therefore decided that it was not appropriate to account for switch maintenance treatment in addition to first‑line treatment. The committee concluded that the population to be included in the calculations for maintenance treatment was therefore 612.

4.23 The committee considered the population size for the second‑line treatment of NSCLC for pemetrexed. It noted that anyone who did not receive pemetrexed as induction or maintenance therapy (that is, those patients with a performance status of 2) would be potentially eligible for second‑line pemetrexed therapy after disease progression. It noted the company's most recent estimate that 429 people in England and Wales receive first‑line chemotherapy that does not include pemetrexed, and that of this group an estimated 20% of people will die before disease progression, leaving a potential second-line treatment population of about 340. The committee agreed that including patients who had died would be perverse and that they should not be included in the total population size. The company then further refined the estimate of 340 by performance status, resulting in an estimate that 180 patients in England and Wales with a performance status of 2 would be eligible to receive pemetrexed in a second‑line setting. The committee did not accept that the population should be reduced according to performance status, preferring instead to estimate the population size based on the licensed indication (which does not restrict treatment by performance status). Aware that the population size should be based on the population in England alone (rather than England and Wales) the committee accepted a further adjustment based on data contained in the National Lung Cancer Audit (2013), reducing the estimated number who would be eligible for pemetrexed as per its licensed indication in a second‑line setting from 340 to 320 people. The committee concluded that the population to be included in the calculations for second-line treatment was therefore 320.

4.24 The committee was aware that pemetrexed also has a marketing authorisation for mesothelioma. The committee noted that, according to the National Lung Cancer Audit (2013), the number of patients with mesothelioma in England and Wales is 1,964. Aware that the population figures should be based on the population in England alone (rather than England and Wales) the committee reduced the number of mesothelioma patients from 1,964 to 1,872 based on data from a Cancer Research UK report from 2010. The committee understood that 88% of these patients would have advanced disease; therefore the potential population with mesothelioma eligible for pemetrexed would be 1,647. The committee was aware of comments received during consultation suggesting that the mesothelioma population eligible for pemetrexed should be limited to those patients with a performance status of 0–1. However, in line with its previous discussions about the licensed population (rather than the eligible population; see sections 4.21 and 4.23), the committee considered it was appropriate to estimate the potential population, as defined by the licensed indication. The committee concluded that the population to be included in the calculations for mesothelioma treatment was therefore 1,647.

4.25 The committee considered the total population size for which pemetrexed has a licence (approximately 7,160). The committee was of the opinion that this figure estimated the maximum population size of patients who could receive pemetrexed for its licensed indications in England. Therefore, the committee was persuaded that the population eligible for pemetrexed would not be higher than this figure. The committee considered the population size in the context of the other end-of-life criteria for this appraisal (see section 4.18). It acknowledged that the benefit of pemetrexed had been demonstrated in all modelled estimates of mean overall survival, and that pemetrexed therefore offered a valuable treatment option for a population of people for whom no other treatment options existed at this maintenance stage. It further considered that the estimate of the population size was very close to 7,000 (see NICE's guide to the methods of technology appraisal). Taking these 2 factors into consideration, the committee concluded that the total patient population should be considered as a small population for the purposes of meeting the criterion for the supplementary advice on end of life. The committee therefore concluded overall that pemetrexed maintenance treatment after induction therapy with pemetrexed and cisplatin could be considered under the supplementary advice to the committee on end‑of‑life treatments.

4.26 The committee noted that even taking into account end‑of‑life considerations, all the estimates of the ICER (including the one the committee felt represented the most plausible ICER, that is, approximately £74,500 per QALY gained) for pemetrexed maintenance treatment after induction therapy with pemetrexed and cisplatin were substantially higher than would normally be considered a cost‑effective use of NHS resources. Therefore the committee concluded that pemetrexed maintenance treatment should not be recommended for treating locally advanced or metastatic non‑squamous NSCLC in people whose disease has not progressed immediately after induction therapy with pemetrexed and cisplatin.

4.27 The committee discussed whether its recommendations for pemetrexed as a maintenance therapy after induction with pemetrexed plus cisplatin were associated with any issues related to equality legislation and the requirement for fairness. The committee was aware that NICE's technology appraisal guidance on pemetrexed recommends it as an option for the maintenance treatment of locally advanced or metastatic NSCLC other than predominantly squamous cell histology if disease has not progressed immediately after platinum‑based chemotherapy plus gemcitabine, paclitaxel or docetaxel. The committee discussed whether its recommendations could be considered unfair given the recommendations in NICE's technology appraisal guidance on pemetrexed, and that the difference between the populations in that appraisal and the current appraisal was in terms of the first‑line treatment received. The committee agreed that first‑line treatment is not linked to the protected characteristics covered in the equality legislation. The committee was aware that it needs to make a decision for each appraisal based on the evidence before it and this is what it has done in this case. The committee agreed that its decision on pemetrexed as a maintenance therapy after induction with pemetrexed plus cisplatin was made because pemetrexed maintenance was not cost effective in this population. Furthermore, even if there was any unfairness, given the high ICER of approximately £74,500 per QALY gained, the committee agreed that the recommendation could be justified and was in line with the committee's role and the application of the cost‑effectiveness criteria, and was a proportionate means of achieving a legitimate aim. The committee had not identified any special factors that would require or justify making a positive recommendation despite the very high ICER.

4.28 The committee discussed an issue raised by the company that a negative recommendation for pemetrexed as a maintenance treatment after induction therapy with pemetrexed and cisplatin would amount to a withdrawal of the treatment when it appears to be working. The committee noted that the first-line and maintenance indications for pemetrexed are separate; that they have been supported by separate trial development programmes and that they are considered to be separate stages of treatment, especially since the first-line indication specifies that pemetrexed is given with cisplatin whereas the maintenance indication specifies that pemetrexed is given as a monotherapy. The committee did not therefore accept the company's assertion that there would be an ethical implication to a decision not to recommend pemetrexed as a maintenance treatment after induction therapy with pemetrexed and cisplatin.

Cancer Drugs Fund reconsideration

4.29 This appraisal was a Cancer Drugs Fund reconsideration of the published NICE technology appraisal guidance on pemetrexed for maintenance treatment following induction therapy with pemetrexed and cisplatin for non-squamous non-small-cell lung cancer. The committee considered the updated company submission and cost–utility analysis that included:

Summary of appraisal committee's key conclusions