Erlotinib monotherapy for maintenance treatment of non-small-cell lung cancer

Clinical effectiveness

4.4 The Committee discussed the evidence on the clinical effectiveness of erlotinib for the maintenance treatment of patients with stable disease. It noted that one randomised trial, the SATURN trial, was presented, which showed that maintenance treatment with erlotinib was associated with a statistically significant improvement in progression-free survival and overall survival compared with placebo for the overall population of patients with stable disease. It noted that the improvement in progression-free survival with erlotinib was similar for patients in the squamous and non-squamous disease subgroups, but a statistically significant improvement in overall survival was only demonstrated in the subgroup of patients with squamous disease (HR 0.67; 95% CI 0.48 to 0.92 compared with HR 0.76; 95% CI 0.59 to 1.00 in the non-squamous disease group). The Committee was aware that the results for patients with stable disease were based on a post hoc subgroup analysis of 55% of the SATURN trial population. Furthermore, the results for the subgroups of patients with squamous and non-squamous disease were also post hoc analyses based on a disaggregation of the stable disease population and there were relatively small numbers of patients in each subgroup (190 and 297 respectively). The Committee was aware that the SATURN trial had not been designed for such analyses. It therefore regarded that the true magnitude of the benefits of erlotinib in these patient populations was uncertain. The Committee considered that the adverse events associated with erlotinib (most commonly, rash and diarrhoea) were well known and noted that most patients in the SATURN trial did not require their treatment to be changed or stopped because of adverse events.

4.5 The Committee discussed whether the results of the SATURN trial could be generalised to UK clinical practice, noting that in the trial there were few UK patients, a high proportion of south-east Asian patients and a high proportion of patients who had never smoked. The Committee was aware that Asian patients are known to respond better to lung cancer treatments than patients of other races and patients who have never smoked respond better than those with a history of smoking. It also noted that in the SATURN trial 30% of patients with stable disease still had an ECOG performance status of 0 after four cycles of chemotherapy and this was likely to reflect a fitter group of patients than the population who would receive maintenance treatment with erlotinib in UK clinical practice. The Committee also observed that 66% of patients in the SATURN trial were under 65 years of age and represented a younger population of patients with non-small-cell lung cancer than would be seen in UK clinical practice. It also noted that a high proportion of patients went on to receive further systemic therapy after the SATURN trial, some of which is not routinely given to patients in UK clinical practice. The Committee therefore concluded that the results from the SATURN trial had limited generalisability to UK clinical practice, therefore adding further uncertainty to the true magnitude of the benefits of erlotinib that would be achieved for UK patients.

4.6 The Committee acknowledged that the manufacturer provided updated analyses for the stable, non-squamous disease group, during consultation on the appraisal consultation documents, which adjusted for some prognostic factors (ECOG status and smoking history), which the manufacturer suggested may have biased results against erlotinib in the SATURN trial. However, the Committee was concerned about the reliability of the data because of the small numbers of patients included in these further subgroup analyses. The Committee heard from the ERG that the differences in ECOG status and smoking history between the erlotinib and best supportive care groups were not statistically significant in the non-squamous disease population and that the differences in these baseline characteristics would not artificially decrease the overall survival estimate for erlotinib. The Committee also acknowledged comments from the ERG that no adjustments for other prognostic factors that could have had an impact on overall survival, such as gender and disease stage, had been made in these analyses by the manufacturer. The Committee further heard from the ERG that in a previous trial, patients with non-squamous disease experienced an additional gain in overall survival when treated with first-line pemetrexed and cisplatin compared to other chemotherapy regimens. In the light of these findings, it is unclear whether erlotinib maintenance treatment will supplement the extended survival advantage seen when patients receive pemetrexed and cisplatin first-line chemotherapy, because no patients received pemetrexed plus cisplatin before the SATURN trial. The Committee therefore concluded that the manufacturer's adjusted estimates of overall survival for erlotinib compared with best supportive care for people with stable, non-squamous disease were associated with considerable uncertainty.

4.7 The Committee observed that all of the survival benefit for erlotinib compared with best supportive care in patients with non-squamous disease occurred in the progression-free period, however only 40% of survival benefit in patients with squamous disease was assumed to occur in the progression-free period. The Committee considered that there was no convincing explanation for the fact that most of the apparent survival benefit for erlotinib in the squamous disease group came after treatment had been discontinued. The Committee therefore regarded the overall survival benefit for erlotinib in the squamous disease group with caution, agreeing that this estimate was highly uncertain.

4.8 The Committee was aware that EGFR mutation status was not recorded in half of the patients in the SATURN trial. It noted however that of the patients who were tested, a small proportion (11%) had activated EGFR mutations, and that this subgroup gained substantially more benefit from erlotinib than the trial population as a whole. It also heard from clinical specialists that these patients have a better prognosis with treatment than other patients with non-small-cell lung cancer. However, the Committee was aware that patients with EGFR mutations were unlikely to receive erlotinib maintenance treatment in UK clinical practice because NICE recommends gefitinib as an option for the first-line treatment in this group rather than chemotherapy ('Gefitinib for the first-line treatment of locally advanced or metastatic non-small-cell lung cancer' [NICE technology appraisal guidance 192]). The Committee was therefore concerned that some of the survival benefit of maintenance treatment with erlotinib demonstrated in the SATURN trial would not be seen in clinical practice because patients with EGFR mutations would usually receive gefitinib and would therefore not be eligible for maintenance treatment with erlotinib.

4.9 The Committee acknowledged that the manufacturer had provided additional information about the patient characteristics of the squamous disease population in the SATURN trial during consultation. The Committee accepted that these data showed that the number of patients with squamous disease with an activated EGFR mutation or who were of Asian origin or who had never smoked was small and therefore it agreed that these prognostic factors were unlikely to significantly bias the estimate of overall survival for this subpopulation. By implication, however, the Committee concluded that the numbers of patients with stable, non-squamous disease who had these baseline characteristics which may lead to better prognosis were even higher than previously thought.

4.10 The Committee discussed the first-line treatments received by patients in the SATURN trial. It considered that about 64% of patients received first-line chemotherapy regimens that are commonly used in the UK. The Committee noted that no one in the SATURN trial had received first-line treatment with pemetrexed and cisplatin, a regimen that is now commonly used as combination chemotherapy for patients with non-squamous disease because of its superiority to the regimens used in the SATURN trial. The Committee considered that patients in the SATURN trial were fitter than patients seen in UK clinical practice, noting that patients with stable disease after four cycles of platinum-based chemotherapy still had a good performance status and approximately 60% of patients went on to receive further systemic therapy after the SATURN trial, some of which would not be routinely used in the UK. It also observed that only a small proportion of patients in the placebo group had received erlotinib after progression. It considered that the post-progression treatments and the small proportion of patients in the placebo group who had received erlotinib after progression would affect the estimates of overall survival in the erlotinib and placebo groups. The Committee was aware that it is unclear whether patients would benefit more from receiving erlotinib as maintenance treatment or for treatment of relapsed disease. The Committee concluded that there was very considerable uncertainty that the benefit of erlotinib seen in the trial would be translated into routine practice.

4.11 The Committee discussed the RECIST criteria for determining disease response in the SATURN trial, taking into account the marketing authorisation for erlotinib, which includes patients with stable disease only. It heard from clinical specialists that some centres, particularly those involved in clinical trials, use the RECIST criteria routinely but that centres less involved in research may not use RECIST criteria on a day-to-day basis. The Committee noted that the RECIST criteria used in the SATURN trial were based on 6-weekly CT scans and considered that such frequent scans were not likely in the routine care of lung cancer patients in the UK. The Committee therefore concluded that it was likely that the duration of erlotinib maintenance treatment in clinical practice would exceed that observed in the SATURN trial as CT scans would be performed less frequently.

4.12 In summary, the Committee agreed that the benefit of maintenance treatment with erlotinib seen in the SATURN trial was likely to be lower in routine clinical practice when considering that the trial population represented patients who are likely to have a better prognosis than the average patient treated in the UK. In addition, the Committee considered that there were several factors that led to considerable uncertainty about the magnitude of overall survival gain expected from erlotinib maintenance treatment in the stable population and in the squamous and non-squamous disease subpopulations. These included the small numbers of patients in the post hoc subgroup analyses informing the survival estimates for the squamous and non-squamous disease groups and the use of post-progression treatments in the SATURN trial, which are not routinely used in the UK; and the lack of explanation as to why most of the survival benefit for erlotinib in the squamous disease group occurred after treatment was discontinued (in the post-progression period).

Cost effectiveness

4.13 The Committee was aware that a patient access scheme had been agreed between the manufacturer and the Department of Health. It noted that this is a simple scheme in which erlotinib is supplied to the NHS at a discount of 14.5% of the list price. The Committee concluded that it was appropriate to appraise the cost effectiveness of erlotinib maintenance treatment on the basis of ICERs that include this discount.

4.14 The Committee discussed the evidence for the cost effectiveness of erlotinib compared with best supportive care derived the manufacturer's new economic analyses provided during consultation on the first appraisal consultation document. The Committee noted that the costs used in the analyses were based on those used in previous NICE technology appraisals of treatments for non-small-cell lung cancer and considered them to be appropriate. The Committee noted the manufacturer's ICERs for erlotinib compared with best supportive care of £40,800 per QALY gained for all patients with stable disease, £35,500 per QALY gained for patients with stable, squamous disease, £40,000 per QALY gained for patients with stable, non-squamous disease. The Committee noted that the manufacturer's ICER for all patients with stable disease was greater than both the ICERs presented for the squamous and non-squamous groups and acknowledged that the factors that had the greatest effect on the manufacturer's new ICERs were assumptions about how much time a patient spent on treatment in the progression-free health state, and the costs attributed to best supportive care.

4.15 The Committee considered the ERG's critique of the manufacturer's economic analysis. It noted the ERG's comment that it was more appropriate to consider the cost effectiveness of erlotinib in the subgroups of patients with squamous disease and non-squamous disease separately, rather than in the stable disease population as a whole, because of heterogeneity between the subgroups. The Committee agreed with the ERG, but was concerned about the subgroup analyses because the trial population had not been stratified by histology and analyses for these histological subgroups and for the stable disease population as a whole had not been predefined, which added uncertainty to the survival estimates and therefore also to the ICERs. Overall, the Committee concluded that it was justified in considering the squamous and non-squamous populations separately on clinical grounds.

4.16 The Committee discussed the ERG's comments on the methods used by the manufacturer to model progression-free survival and overall survival, in particular that post-progression survival had been calculated as the difference between overall survival and progression-free survival. The Committee concluded that the ERG's approach to estimating survival was more appropriate because it was based as much as possible on data directly from the trial and used modelling only when necessary. It noted that the ERG's modelling approach resulted in lower estimates of overall survival than the manufacturer's method, and that there were wide confidence intervals around these estimates, indicating a high degree of uncertainty. The Committee observed that the ERG's analyses resulted in significantly higher ICERs for erlotinib compared with best supportive care than those estimated by the manufacturer.

4.17 The Committee noted that the ERG's revisions to the manufacturer's model (including correcting the pemetrexed costs and using an alternative survival modelling method) increased the ICERs for erlotinib compared with best supportive care to £44,800 per QALY gained for treatment of stable, squamous disease and £68,100 per QALY gained for treatment of stable, non-squamous disease. The Committee considered that the most plausible ICERs would be considerably higher than those estimated by the ERG, and likely to be above £50,000 per QALY gained even for treatment of stable, non-squamous disease when taking into account the fact that the survival benefit observed in the SATURN trial was likely to be reduced in clinical practice where patients are less fit and have different prognostic characteristics from those seen in the trial population.

4.18 The Committee discussed the evidence for the cost effectiveness of erlotinib compared with pemetrexed. It noted that this was based on the manufacturer's new economic analysis submitted in response to the first appraisal consultation document, in which various relative efficacy scenarios were modelled because of the lack of data for erlotinib compared directly with pemetrexed. The Committee was aware that erlotinib was less costly than pemetrexed. The Committee was aware that the manufacturer considered that the JMEN and SATURN trials were not directly comparable and that a robust estimate of the relative effectiveness of erlotinib and pemetrexed was not possible to establish. However, it noted that the ERG's indirect analysis of the JMEN and SATURN trials showed that erlotinib was less effective than pemetrexed. The Committee also observed the ERG's concerns not only about the comparability of these two trial populations but also about their generalisability to UK practice. In the light of these issues, the Committee concluded that it had not been presented with a plausible estimate of the cost savings per QALY lost that would be associated with the use of erlotinib maintenance compared with pemetrexed and that therefore, erlotinib could not be specifically recommended compared with pemetrexed.

4.19 The Committee noted the manufacturer's claim that pemetrexed maintenance treatment had been recommended in TA190 for patients with non-squamous disease despite ICER estimates from the ERG exceeding £50,000 per QALY gained. However, the Committee understood that many considerations were taken into account by the Committee when finalising its appraisal of pemetrexed maintenance treatment, which subsequently decreased the ICER below £50,000 per QALY gained.

4.20 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:

In addition, when taking these into account the Committee must be persuaded that the estimates of the extension to life are robust and the assumptions used in the reference case economic modelling are plausible, objective and robust.

4.21 The Committee noted that the median survival duration of patients in the UK with non-small-cell lung cancer who receive first-line chemotherapy is between 7 and 11 months. The Committee discussed the size of the patient population and was aware that the erlotinib marketing authorisation includes monotherapy for maintenance treatment of patients with locally advanced or metastatic non-small-cell lung cancer with stable disease after four cycles of standard platinum-based first-line chemotherapy, but also the treatment of patients with locally advanced or metastatic non-small-cell lung cancer after failure of at least one prior chemotherapy regimen. During consultation, the manufacturer estimated that about 4100 patients would be suitable for treatment with erlotinib in the UK according to its current marketing authorisations. The Committee noted that the manufacturer had indicated that 6700 patients receive first-line chemotherapy in the UK. Some of these patients would receive erlotinib as maintenance treatment rather than as a second-line therapy. The Committee also noted that the erlotinib marketing authorisation includes the treatment of patients with metastatic pancreatic cancer in combination with gemcitabine. Most of the 7000 patients with pancreatic cancer present with metastatic disease and erlotinib would potentially be indicated for this population. The Committee discussed written evidence from a previous NICE technology appraisal appeal and noted that the Appeal Panel recognised that the criterion in the supplementary advice for end-of-life treatments for small patient populations indicated that 'Sufficient regard should be given to recognition of the desirability of developing new treatments in smaller disease areas and that higher prices, and therefore reduced cost effectiveness, were more likely to be justified given the need to recoup costs of development of the product from more limited licences'.The Appeal Panel had concluded that it was appropriate, according to the supplementary advice, to add together the potential patient populations covered by the marketing authorisation for different indications rather than on the basis of actual or recommended use. The Committee therefore considered that the true size of the cumulative population potentially eligible for treatment with erlotinib according to its UK marketing authorisations was not small and was considerably higher than the manufacturer's estimate.

4.22 The Committee then discussed the extension to life offered by erlotinib for patients with stable disease. It also considered whether mean or median survival was a more appropriate measure for evaluating the end-of-life criteria. The Committee agreed with comments from the ERG that the mean survival figures were more informative because they were based on all available data for all patients across the whole trial period. The Committee also heard from the clinical specialists that some patients have significantly longer responses to treatment with erlotinib, which was another reason to consider the mean rather than the median values. It noted that in the new analyses provided during consultation, the manufacturer estimated the mean overall survival benefit of erlotinib compared with best supportive care to be 3.3 months in the whole stable disease population, 4.2 months in the stable, squamous disease population and 4.5 months in the stable, non-squamous disease population. It also noted that the ERG estimated the mean overall survival benefit to be 3.4 months and 2.2 months in the populations of patients with squamous and non-squamous disease respectively. The Committee was concerned that no rationale could be provided to explain why both the manufacturer's median and mean survival estimates for each subpopulation were greater than for the whole population, which cast uncertainty over the validity of the analysis. During consultation the manufacturer explained that this was because of the different prognostic baseline characteristics of the patients in the squamous and non-squamous disease groups. Although the ERG did not provide an overall survival estimate for the whole stable disease population, the Committee heard from the ERG that this figure was likely to be closer to the mean overall survival estimate for patients in the non-squamous disease group (that is, 2.2 months) than the mean overall survival estimate for patients in the squamous disease group (that is, 3.4 months). The Committee had previously concluded that the overall survival benefit of erlotinib in clinical practice was uncertain and likely to be less than the ERG's estimates. The Committee did not consider that robust evidence had been provided to demonstrate an extension to life of at least 3 months and, taken together with the consideration on population size, therefore concluded that the end-of-life criteria were not met in this appraisal.

4.23 In summary, the Committee considered that the most plausible ICERs for erlotinib compared with best supportive care would be higher than those estimated by the ERG (£44,800 and £68,100 per QALY gained for treatment of patients with stable, squamous disease and with stable, non-squamous disease respectively) and considerably above £50,000 per QALY gained for treatment of the whole stable disease population. The Committee agreed that the end-of-life criteria were not met in this appraisal, but it noted that even if they were taken into account, the most plausible ICERs were higher than those normally considered to be associated with cost effective treatments. The Committee concluded that erlotinib was likely to be associated with cost savings per QALY lost compared with pemetrexed in patients with stable, non-squamous disease, but that it was not possible to establish a robust estimate. It therefore agreed that no specific recommendation could be made related to the use of erlotinib compared with pemetrexed. The Committee concluded that erlotinib could not be considered a cost-effective use of NHS resources when used as monotherapy for maintenance treatment in patients with locally advanced or metastatic non-small-cell lung cancer who have stable disease following platinum-based first-line chemotherapy.

4.24 The Committee discussed whether NICE's duties under the equalities legislation required it to alter or add to its recommendations in any way. It noted that in response to the second appraisal consultation document the manufacturer stated that the preliminary recommendations mean that patients with squamous disease will not have a maintenance treatment option, whereas those with non-squamous disease currently have access to pemetrexed maintenance treatment through TA190. The manufacturer further stated that the histological mix of non-small-cell lung cancer shows a gender imbalance with squamous disease making up a substantially larger proportion of non-small-cell lung cancer in men. It was the manufacturer's view therefore that having no maintenance treatment option for people with squamous disease has a greater impact on men with non-small-cell lung cancer, and that this was particularly concerning given that men have an inherently worse prognosis than women. The Committee noted that no data on gender distribution based on histology were provided by the manufacturer and therefore this assertion was impossible to substantiate. However, the Committee noted that any possible differences in maintenance treatment access referred to by the manufacturer were related to TA190, rather than this appraisal. The Committee agreed that its decision about erlotinib maintenance treatment needed to be based on the evidence seen in this appraisal. Furthermore, the final decision not to recommend erlotinib maintenance treatment was made because erlotinib was not cost-effective in either of the squamous or non-squamous subgroups compared with best supportive care. The Committee concluded that its recommendations do not make it more difficult in practice for a specific group to access erlotinib maintenance treatment compared with other groups. In addition the Committee noted that, following the publication of TA181, the proportion of patients who would be eligible to receive pemetrexed maintenance treatment was declining quickly over time (because they are receiving pemetrexed as a first-line treatment instead) and therefore the manufacturer's concern that pemetrexed is currently only available as a maintenance option for non-squamous disease was becoming less relevant.

Summary of Appraisal Committee's key conclusions