4 Committee discussion

The appraisal committee reviewed the data available on the clinical and cost effectiveness of ceritinib, having considered evidence on the nature of anaplastic lymphoma kinase (ALK) positive non‑small‑cell lung cancer (NSCLC) and the value placed on the benefits of ceritinib by people with the condition, those who represent them, and clinical experts. It also took into account the effective use of NHS resources.

4.1 The committee heard from the clinical and patient experts about the effect of advanced or metastatic ALK‑positive NSCLC on people with the condition. It noted that the most common effects of the disease are persistent cough, chest pain, breathlessness, and fatigue, and that complications include chest infections and metastases to the brain and elsewhere. It also heard that currently there is no targeted treatment available for ALK‑positive NSCLC when the disease progresses after treatment with crizotinib. The patient expert suggested that ceritinib gives hope because it has the potential to extend life and improve quality of life. The committee concluded that additional treatment options would be of value to people with ALK‑positive NSCLC.

4.2 The committee discussed the treatment pathway for advanced or metastatic ALK‑positive NSCLC and the relevant comparators for ceritinib. It heard from the clinical experts that most people with advanced or metastatic ALK‑positive NSCLC would first have platinum-based chemotherapy. People whose disease progresses, and who have a confirmed ALK mutation, may have crizotinib, which is available only through the Cancer Drugs Fund (NICE does not recommend crizotinib in its technology appraisal guidance on crizotinib for previously treated non-small-cell lung cancer associated with an anaplastic lymphoma kinase fusion gene). The committee noted that the population relevant to this appraisal has ALK‑positive NSCLC that has progressed after crizotinib. The committee noted that both the company, and the clinical experts, advised that currently in the NHS there are no active treatments available, and that best supportive care (BSC) is usually offered. It heard from the clinical experts that a few people who are relatively fit are offered chemotherapy, but the clinical experts were not aware of evidence that chemotherapy at this stage of treatment improves outcomes. After consultation, the company suggested that systemic chemotherapy was a relevant comparator for ceritinib for fitter patients and for patients who have not yet had chemotherapy. The committee noted that no additional evidence was presented to support chemotherapy as a comparator. The committee was also aware of the comments from the clinical experts about the lack of evidence for chemotherapy in people whose disease had progressed after crizotinib. The committee concluded that the relevant comparator for ceritinib was BSC, which would not include any active chemotherapy.

4.3 The committee discussed whether testing for the ALK mutation is established practice in the NHS. It heard from the clinical experts that currently there are differences across England. It understood that the summary of product characteristics for ceritinib states that, before starting treatment, clinicians should assess the person's ALK status. It noted that, according to the marketing authorisation, ceritinib can only be used after crizotinib, which is also an ALK inhibitor. The committee was aware that ALK mutation testing would be done before starting crizotinib, so the relevant population for this appraisal would have been tested already. It concluded therefore, that the costs and availability of ALK mutation testing were not a consideration for this appraisal.

Clinical effectiveness

4.4 The committee discussed the clinical evidence presented by the company and its critique by the evidence review group (ERG). It noted that the company presented efficacy data from the single-arm ASCEND‑1 and ASCEND‑2 studies. It noted that data on the efficacy of BSC came from 2 separate studies; to assess overall survival the company used Ou et al. (2014), and to assess progression-free survival the company used Shepherd et al. (2005). The company conducted a naive indirect comparison of the results from the ASCEND‑1 and ASCEND‑2 studies, Ou et al., and Shepherd et al.

4.5 The committee discussed the differences in overall survival between ceritinib and BSC:

  • The committee understood that the results of the ASCEND trials were combined without adjustment, but the ERG did not consider that this greatly affected the clinical-effectiveness results.

  • The committee was aware that median overall survival with ceritinib was 16.7 months in ASCEND‑1 (data cut-off April 2014, duration of study 40 months) and 14.9 months in ASCEND‑2 (data cut-off August 2014, duration of study 21 months), and that the pooled median overall survival was 15.6 months. The committee considered the number of patients on which these data were based and noted the company's submission stated that 59 people had died in ASCEND‑2 (n=130). The committee agreed that the data were immature.

  • The committee was aware that median overall survival with BSC was 2.2 months in Ou et al. (2014). The committee noted that the BSC results were based on a small sample of patients, and so were uncertain.

  • The committee discussed the risk of confounding in the overall-survival analysis by considering whether people in the ceritinib studies had a different underlying risk of dying than people in the Ou et al. (2014) study. The committee was aware that the ASCEND studies only included people who had a life expectancy of at least 12 weeks, which was not true for the Ou et al. cohort. It was aware that in Ou et al., treatment was determined by clinician choice and people offered BSC may have been more unwell than those offered active treatment. The committee understood from the ERG that the BSC group in Ou et al. may have been sicker than patients in the ASCEND studies, because these patients had higher Eastern Cooperative Oncology Group (ECOG) values at disease progression than patients in the ASCEND trials. The committee was aware that differences in baseline patient characteristics, such as age and ECOG status, were not statistically significant between the studies, but recognised the challenges of showing statistical significance with a small study population. The committee acknowledged that if those in the Ou et al. study were sicker than those in the ASCEND studies, the overall-survival benefit of ceritinib compared with BSC would be overestimated. The committee heard from the company that because of limited data it could not compare the ASCEND and Ou et al. studies for other potential confounders, such as disease burden, which the clinical experts noted would reasonably be associated with mortality. So, the committee considered that the results of the naive indirect comparison, and specifically the size of benefit, were uncertain because there was a high risk of bias from confounding.

    The committee concluded that ceritinib was likely to prolong life but the extent of treatment benefit was highly uncertain.

4.6 The committee discussed the differences in progression-free survival between ceritinib and BSC:

  • The committee noted that median progression-free survival with ceritinib was 6.9 months in ASCEND‑1 and 7.0 months in ASCEND‑2, and that the pooled median progression-free survival estimate was 7.0 months. For comparison, median progression-free survival with BSC was 1.8 months in Shepherd et al. (2005).

  • The committee discussed whether the difference in progression-free survival could be attributed to differences between the studies rather than to the benefit of treatment with ceritinib itself. The committee was aware that the company used the BSC arm from Shepherd et al. (2005), which although limited to NSCLC, was not limited to patients with ALK‑positive tumours. The committee was not presented with data on whether the disease in people with ALK‑positive NSCLC progresses faster or slower than in people whose tumours are not ALK positive, but heard from the clinical experts that ALK‑positive NSCLC may have a natural history that differs from other types of NSCLC. The committee also heard that the Shepherd et al. trial was not limited to people having third‑line treatment. The committee learned from the company during the third committee meeting that none of the patients in the Shepherd et al. study had previously had treatment with crizotinib. In addition, patients in Shepherd et al. had lower (that is, better) scores for ECOG performance status than patients in the ASCEND trials, so they might have been fitter and their disease less likely to progress.

    The committee concluded that ceritinib was likely to delay disease progression, but that the size of the difference in progression-free survival is highly uncertain.

4.7 The committee considered whether the evidence for the clinical effectiveness of ceritinib could be generalised to people with advanced ALK‑positive NSCLC in England. It noted that the company included only about 50% of the patients from ASCEND‑1, but that the clinical experts felt that it represented the relevant population in England. The committee was aware that the analysis included only about 20% of patients from Ou et al. (2014), making it difficult to determine whether they represented patients who might be offered BSC in England. The committee also heard from the clinical experts that Shepherd et al. (2005) enrolled patients with all types of NSCLC, including genetic mutations that differed from the ALK‑positive mutation; the proportion of patients with ALK‑positive mutation is unknown; and none of the patients had previously had crizotinib. The committee concluded that the ASCEND studies were generalisable to people with ALK‑positive tumours in England, but the Shepherd et al. study was not.

4.8 The committee was aware that the marketing authorisation for ceritinib states that treatment should continue as long as clinical benefit is seen. It discussed whether, in clinical practice in England, people are likely to continue ceritinib after disease progression. It heard from the clinical experts that this was done for other targeted treatments for NSCLC (such as epidermal growth factor receptor [EGFR] inhibitors) and so it would be reasonable to expect ceritinib treatment to continue after progression. The committee was aware from Ou et al. (2014) that crizotinib as second‑line treatment is continued after disease progression. The committee noted that in ASCEND‑2, ceritinib was taken for a median of 1.6 months after disease progression. The committee noted a comment received after consultation that in clinical practice, treatment after disease progression might be even longer but it was aware that no other real-world evidence is currently available. The committee concluded that, in clinical practice, treatment with ceritinib could plausibly continue after disease progression and the best estimate of the duration of treatment came from ASCEND‑2.

4.9 The committee discussed the ongoing studies of ceritinib. It noted that the ASCEND‑5 randomised controlled trial compares ceritinib with chemotherapy (pemetrexed or docetaxel) in people with ALK‑positive NSCLC previously treated with crizotinib. The committee heard from the company that the results of ASCEND‑5 should be available in the second quarter of 2016. The committee acknowledged that the control group in the ASCEND‑5 trial, having been given chemotherapy rather than BSC, was not consistent with the decision problem for this appraisal. But, the committee concluded that the ASCEND‑5 trial may give useful data about the clinical effectiveness of ceritinib to inform the evidence network.

4.10 The committee discussed the adverse events associated with ceritinib. It noted that in the ASCEND studies, all patients had adverse events and many had a dose reduction or interrupted treatment. It also considered the comments from the patient and clinical experts (both in their submissions and during the first meeting) that the adverse events associated with ceritinib are tolerable and manageable. The committee concluded that, although many people had adverse events while taking ceritinib, these events were manageable.

Cost effectiveness

4.11 The committee discussed the company's economic model, noting that it used clinical evidence from 4 different sources: ASCEND‑1 and ASCEND‑2, Ou et al. (2014), and Shepherd et al. (2005). The committee was aware that the model used data from a naive (unadjusted) indirect comparison and noted that this had weaknesses (see sections 4.5 and 4.6). However, it concluded that this was the best evidence available and the model was sufficient for current decision-making.

4.12 The committee discussed the methods used by the company for modelling overall survival. It noted that the company used parametric curves to extrapolate overall survival over the 10‑year time horizon of the model. For ceritinib, the company used pooled results from the ASCEND‑1 and ASCEND‑2 studies. For BSC, it used results from the BSC-only subgroup of patients from Ou et al. (2014). The committee noted that the company chose the Weibull curve to extrapolate overall survival for both arms of the model whereas the ERG's exploratory analyses used a different curve for the BSC arm, the better-fitting log-normal curve. The committee heard from the ERG that the log-normal curve predicts that an extremely small proportion of patients would be alive after 10 years, whereas the Weibull curve predicts that no patients would be alive after 10 years. The clinical experts advised that they would be surprised if people with ALK‑positive NSCLC that had progressed after crizotinib were alive after 10 years. The committee concluded that, for extrapolating overall survival with BSC, the log-normal curve was a better fit for the data whereas the Weibull curve gave results that better reflected clinical experience. So, the committee considered both in its decision-making.

4.13 The committee discussed the assumptions about how long patients take ceritinib. It noted that the company's base case assumed treatment until disease progression. The committee recognised that the ERG presented analyses exploring the impact of treatment continuing after progression, in which it assumed 1.6 months of treatment after progression based on the median seen in ASCEND‑2. The committee concluded that, in clinical practice, treatment could plausibly continue after progression, and the best estimate of the duration of treatment came from ASCEND‑2 (see section 4.8), so it preferred the ERG's approach to modelling treatment duration.

4.14 The committee discussed the assumptions about the duration of treatment benefit. It noted that the company's model assumed that the benefits of treatment with ceritinib persist beyond the study period and after stopping treatment. The committee heard from the clinical experts at the meeting that it was unlikely that ceritinib would offer a benefit beyond the end of treatment, and if it did, it would not be as long as 2 years. It noted that the company's revised model had explored 2 scenarios, which reduced the treatment benefit to 18 months or 24 months, and that this had little impact on the company's base-case incremental cost-effectiveness ratio (ICER; see section 3.32). The committee heard from the ERG that this was because most quality-adjusted life year (QALY) benefits and costs are accrued in the first few years of treatment. The committee was not given evidence that the treatment benefit from ceritinib would continue after the end of treatment, but concluded that the company had shown that this had minimal impact on cost effectiveness.

4.15 The committee considered the utility values in the company's model. It noted the inconsistency between the model and the company's submission and that the company used different utility values for ceritinib and BSC in the progression-free health state. It also noted the ERG's critique, that using the same utility value would have been more appropriate and that the company did not justify its choice in its submission (see section 3.28). The committee concluded that the ERG's approach was more appropriate, and that the same utility values should be applied to both arms of the model; that is, the utility should depend on the health state rather than the treatment. The committee noted that there was additional uncertainty with the utility value for the progression-free health state because it was derived using a UK‑specific algorithm developed in a different disease area (multiple myeloma) and that the value used appeared higher than might be expected for people with ALK‑positive NSCLC. The committee discussed whether the model should include utility decrements associated with adverse events. It considered the ERG's approach reasonable, that is, increasing the utility value for patients having BSC for the progression-free health state because these patients would not experience adverse events associated with ceritinib. The committee concluded that collecting EQ‑5D data from people with ALK‑positive NSCLC would be desirable and would address the uncertainties related to the modelled utility values.

4.16 The committee discussed the cost of ceritinib treatment, noting that the company's model assumed that patients do not take all of the licensed dose of ceritinib and so the NHS would pay for only 82.8% of the licensed dose. It was aware that the dose intensity in the company's model was based on data from ASCEND‑1 and ASCEND‑2. It heard from the clinical experts that, for a short‑term reduction in dose, people would continue to have a 30‑day supply of their usual dose of ceritinib and unused tablets would be wasted. In contrast, for a long‑term dose reduction, the lower dose would be prescribed and tablets were unlikely to be wasted. The committee heard from the clinical experts that people who stop ceritinib because of adverse reactions cannot return unused tablets to the NHS.It also heard from the company that the recommended dose of ceritinib is 750 mg per day, which would be given in 5 doses of 150‑mg capsules. This allows people to easily reduce doses, which may mean less wastage. Based on this advice, the committee agreed that on average in clinical practice the NHS would not pay for the full dose, but it was likely to pay for more than 82.8%, because of wastage. The committee concluded that the dose intensity in the model should be lower than 100% but higher than the estimate of 82.8% used by the company. The committee noted that the company had taken this into account in its revised analysis (submitted with the patient access scheme discount; see section 3.32) by incorporating a dose intensity of 90%, which the committee accepted.

4.17 The committee also discussed administration costs related to having ceritinib. It noted that the company assumed there were no administration costs for ceritinib because it is taken orally. However, the committee noted a comment received during consultation that ceritinib would be available only through cancer centres, and so the company should have included pharmacy costs for a specialist cancer centre in the modelling. By contrast, BSC is currently shared between GPs and specialists. The committee acknowledged that the company had subsequently included these costs in a revised analysis (see section 3.32 and section 4.18).

4.18 The committee considered whether ceritinib was a cost-effective use of NHS resources compared with BSC for people with ALK‑positive NSCLC. It noted that the company's base case without the discount resulted in an ICER of £62,500 per QALY gained (incremental costs £51,952; incremental QALYs 0.83). It noted that the ERG's preferred parameters resulted in an ICER of £79,500 per QALY gained (incremental costs £63,281; incremental QALYs 0.80). The committee noted that the company submitted a patient access scheme and had also revised the model to address some of the committee's key concerns (see section 3.32). The committee recognised that the company had:

  • extrapolated overall survival for BSC by using the log-normal normal distribution, as suggested by the ERG (see section 4.12)

  • estimated treatment duration and costs by including an extra 1.6 months of time on treatment in line with the trial data (see section 4.8 and section 4.13)

  • estimated duration of treatment benefit by assuming that the treatment benefit for ceritinib persists (for progression-free and overall survival) until a patient stops treatment; it presented 2 scenarios that reduced treatment benefits for ceritinib to 18 and 24 months from starting treatment showing minimal impact on the ICER (see section 3.32 and section 4.14)

  • modelled utility by updating the model so that the same utility values were used for both ceritinib and BSC in all health states (see section 4.15)

  • estimated dose intensity by including a value of 90% for ceritinib, as well as updating costs for giving ceritinib, blood tests and all grade 3 and 4 adverse events, which the company had previously excluded.

    The committee noted that by implementing these changes to the model, the company's base‑case ICER without the patient access scheme discount increased from £62,500 to £86,400 per QALY gained. But, including the patient access scheme discount reduced the ICER substantially. The committee accepted the company's changes to the model, although it remained concerned about large uncertainty around some of the parameters in the model. The committee also heard from the ERG that the cost-effectiveness acceptability curve did not reflect this uncertainty and lacked face validity. It noted that this could be improved with better data. Taking these factors into account, the committee concluded that, on balance, the most plausible ICER was likely to be lower than £50,000 per QALY gained when including the patient access scheme discount and the company's revisions to the model.

4.19 The committee considered supplementary advice from NICE that should be taken into account when appraising treatments that may extend the lives of patients who have 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:

  • The treatment is indicated for patients with a short life expectancy, normally less than 24 months.

  • There is sufficient evidence to indicate that the treatment offers an extension to life, normally of at least an additional 3 months, compared with current NHS treatment.

  • The treatment is licensed or otherwise indicated for small patient populations.

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

4.20 The committee discussed whether ceritinib for ALK‑positive NSCLC met the end-of-life criteria. It noted that the clinical evidence in the company's submission showed that the life expectancy for people with ALK‑positive NSCLC is a median of 2.2 months with the currently available BSC. However, because there was significant uncertainty around this value, the committee also considered the life expectancy for other types of NSCLC treated with BSC, noting that this was 4.7 months (Shepherd et al. 2005). It agreed that this life expectancy is significantly less than 24 months, and therefore it concluded that the life expectancy criterion was met. The committee also discussed the size of the patient population eligible for ceritinib and noted that about 66 or 98 patients would be eligible for ceritinib treatment each year in England and Wales. It concluded that ceritinib is licensed for a small patient population and that the population-size criterion was met.

4.21 The committee then discussed whether ceritinib is likely to extend life by an additional 3 months, compared with BSC. It was aware that in its submission, the company stated that ceritinib prolonged life by a median of 10 months compared with BSC, an approximation based on a naive indirect comparison using the results of ASCEND‑1, ASCEND‑2 and Ou et al. (2014). The committee noted the company's comment that this was the best available evidence, and that trials could not compare ceritinib with BSC because of ethical considerations. The committee recalled its conclusion that this naive indirect comparison was at high risk of bias because of confounding. It was also aware that the mean survival estimates for both ceritinib and BSC were very uncertain because the ceritinib estimate came from interim analyses and the BSC estimate came from a very small number of patients. The committee discussed all the estimates for overall survival without treatment with ceritinib, the risk factors for dying, the differences in these risk factors in the patients in the studies, and whether the degree of confounding, if accounted for, was likely to reduce the estimates to a mean difference in overall survival between ceritinib and BSC care of 3 months or less. The committee concluded that controlling for confounding was unlikely to reduce the mean difference to less than 3 months. The committee recognised the uncertainty, and called for future appraisals of ceritinib compared with BSC to show objectively and robustly a mean difference of greater than or equal to 3 months, but considered that it was reasonable to conclude that ceritinib offers an average extension to life of at least 3 months.

4.22 The committee discussed whether ceritinib is an innovative treatment providing additional benefits to patients. The committee was aware that the company and the patient expert considered ceritinib innovative. The committee also acknowledged that ceritinib had a Promising Innovative Medicine designation from the Medicines and Healthcare products Regulatory Agency (MHRA). It noted further benefits of ceritinib: clinical experts advised that it may control brain metastases; and the patient expert advised that it allows people to continue to work and live a more normal life. However, the committee noted that it had not been presented with evidence about the extent to which these benefits were realised in practice, compared with BSC. It also noted the comments from consultation that in the ASCEND‑2 trial, symptoms or quality of life did not worsen in patients having ceritinib. The committee concluded that ceritinib may be innovative, but it had not been presented with any additional evidence of benefits that were not captured in the measurement of QALYs.

4.23 The committee considered whether it should take into account the consequences of the Pharmaceutical Price Regulation Scheme (PPRS) 2014, and in particular the PPRS payment mechanism, when appraising ceritinib. The appraisal committee noted NICE's position statement, and accepted the conclusion 'that the 2014 PPRS payment mechanism should not, as a matter of course, be regarded as a relevant consideration in its assessment of the cost effectiveness of branded medicines'. The committee heard nothing to suggest that there is any basis for taking a different view on the relevance of the PPRS to this appraisal of ceritinib. It therefore concluded that the PPRS payment mechanism was not applicable for this appraisal.

4.24 During its third meeting, the committee noted that the company's ICERs for ceritinib now included the patient access scheme discount. Based on the available evidence, the committee considered that although there was some uncertainty, it was satisfied that the end-of-life criteria had been met. The committee noted that when the patient access scheme discount was applied, the ICERs were likely to be within the range usually considered a cost-effective use of NHS resources for an end-of-life treatment. The committee considered that there was a high degree of uncertainty associated with the size of effectiveness of certinib compared with BSC because the company based the estimates on a naive indirect comparison with a high risk of bias from confounding (see section 4.6). The committee considered these uncertainties with respect to the small size of the population noting section 6.2.14 of NICE's guide to the methods of technology appraisal, which states that in general, the committee will want to be increasingly certain of the cost effectiveness of a technology as the impact of adopting the technology on NHS resources increases. The committee agreed that although more robust evidence on the clinical and cost effectiveness of ceritinib would be desirable, the impact on NHS resources was not expected to be large. Therefore, because of the large unmet need in the small number of patients, who have a very poor prognosis and for whom no active therapy is available, the committee recommended ceritinib for advanced ALK‑positive NSCLC previously treated with crizotinib. The committee also asked that the review date of this guidance be brought forward to address emerging evidence (see section 4.9) and the conditional nature of the marketing authorisation received from the European Medicines Agency.

Summary of appraisal committee's key conclusions

TA395

Appraisal title: Ceritinib for previously treated anaplastic lymphoma kinase positive non‑small‑cell lung cancer

Section

Key conclusion

Ceritinib is recommended, within its marketing authorisation, as an option for treating adults with advanced anaplastic lymphoma kinase (ALK) positive non‑small‑cell lung cancer (NSCLC) previously treated with crizotinib, only when the company provides the drug with the discount agreed in the patient access scheme.

The clinical-effectiveness data were based on single-arm phase I and II studies for ceritinib and on 2 observational studies for best supportive care. The committee concluded that ceritinib was likely to prolong life and delay disease progression compared with best supportive care, but the extent of treatment benefit was highly uncertain because there was a high risk of bias from confounding.

The key drivers of cost effectiveness were the survival functions used to extrapolate overall survival, assumptions about whether ceritinib treatment continues after disease progression, and assumptions about the duration of treatment benefit. The company's base-case incremental cost-effectiveness ratio (ICER) was £62,500 per quality-adjusted life year (QALY) gained and this increased to £79,500 per QALY gained when the evidence review group (ERG) incorporated its preferred estimates. However, the ICERs were likely to be within the range usually considered a cost-effective use of NHS resources for an end-of-life treatment when applying the patient access scheme discount

The committee accepted the company's changes to the model, although it remained concerned about large uncertainty around some of the parameters in the model. The committee concluded that, on balance, the most plausible ICER was likely to be lower than £50,000 per QALY gained when including the patient access scheme discount.

Based on the available evidence, the committee considered that although there was some uncertainty, it was satisfied that the end-of-life criteria had been met. The committee noted that when the patient access scheme discount was applied, the ICERs were likely to be within the range usually considered a cost-effective use of NHS resources for an end-of-life treatment. Because of the large unmet need in the small number of patients, who have a very poor prognosis and for whom no active therapy is available, the committee recommended ceritinib for advanced ALK‑positive NSCLC.

1.1, 4.4– 4.6, 3.30, 3.31, 4.18, 4.24

Current practice

Clinical need of patients, including the availability of alternative treatments

Currently there is no targeted treatment available for ALK‑positive NSCLC that has progressed after crizotinib; most patients have best supportive care.

4.2

The technology

Proposed benefits of the technology

How innovative is the technology in its potential to make a significant and substantial impact on health-related benefits?

Ceritinib has the potential to extend life, improve quality of life and give people hope for the future.

The committee concluded that ceritinib may be innovative, but it had not been presented with evidence of benefits that were not captured in the measurement of QALYs.

4.1, 4.22

What is the position of the treatment in the pathway of care for the condition?

Most people with advanced or metastatic ALK‑positive NSCLC first have platinum-based chemotherapy. People whose disease progresses, may have crizotinib which is only available through the Cancer Drugs Fund (NICE does not recommend crizotinib in its technology appraisal guidance on crizotinib for previously treated non-small-cell lung cancer associated with an anaplastic lymphoma kinase fusion gene). Ceritinib would be used after the disease progresses after treatment with crizotinib.

4.2

Adverse reactions

The committee concluded that, although many people had adverse events while taking ceritinib, these events were manageable.

4.10

Evidence for clinical effectiveness

Availability, nature and quality of evidence

The company presented efficacy data from the single-arm ASCEND‑1 and ASCEND‑2 studies of ceritinib. Evidence on the efficacy of best supportive care came from 2 separate studies; data on overall survival from Ou et al. (2014) and data on progression-free survival from Shepherd et al. (2005). The company conducted a naive indirect comparison of the results from these studies.

4.4

Relevance to general clinical practice in the NHS

The committee concluded that the ASCEND studies were generalisable to people with ALK‑positive tumours in England. The Shepherd et al. (2005) study was not generalisable because it included patients with all types of NSCLC, rather than ALK‑positive NSCLC.

4.7

Uncertainties generated by the evidence

The studies in the indirect comparison differed in eligibility criteria and patient characteristics. The committee concluded that the results of the naive indirect comparison were uncertain because there was a high risk of bias due to confounding.

Regarding progression-free survival, the data for best supportive care came from Shepherd et al. (2005), which included patients with all types of NSCLC. The committee was not presented with data on whether the disease in people with ALK‑positive NSCLC progresses faster or slower than in people whose tumours are not ALK positive. It concluded that the size of the difference in progression-free survival is highly uncertain.

Regarding overall survival, the data for ceritinib were immature and the data for best supportive care came from only 20% of patients from Ou et al. (2014).

The committee concluded that ceritinib was likely to prolong life compared with best supportive care, but the extent of treatment benefit was highly uncertain.

4.5– 4.7,

Are there any clinically relevant subgroups for which there is evidence of differential effectiveness?

No subgroups were considered.

Estimate of the size of the clinical effectiveness including strength of supporting evidence

The median overall survival with ceritinib was 16.7 months in ASCEND‑1, 14.9 months in ASCEND‑2, and the pooled result was 15.6 months. The median overall survival with best supportive care was 2.2 months in Ou et al. (2014).

The median progression-free survival with ceritinib was 6.9 months in ASCEND‑1, 7.0 months in ASCEND‑2, and the pooled result was 7.0 months. The median progression-free survival with best supportive care was 1.8 months in Shepherd et al. (2005).

4.5, 4.6

Evidence for cost effectiveness

Availability and nature of evidence

A Markov model was developed, which used data from ASCEND‑1, ASCEND‑2, Ou et al. (2014), and Shepherd et al. (2005). The committee was aware that the model used data from a naive indirect comparison and noted that this had weaknesses. However, it concluded that this was the best evidence available and the model was sufficient for decision-making.

3.12, 4.11

Uncertainties around and plausibility of assumptions and inputs in the economic model

For extrapolating overall survival with best supportive care, the committee concluded that the log-normal curve (chosen by the ERG) was a better fit for the data whereas the Weibull curve (chosen by the company) gave results that better reflected clinical experience. So it considered both in its decision-making.

The company's base case assumed treatment with ceritinib continued until disease progression, whereas the ERG assumed a median of 1.6 months of treatment after progression, based on ASCEND‑2. The committee concluded that treatment could plausibly continue after progression.

The company assumed that the benefits of ceritinib persist after stopping treatment, but the committee heard from clinical experts that this was unlikely. The Committee was not given evidence that the treatment benefit from ceritinib would continue after the end of treatment, but concluded that the company had shown that this had minimal impact on cost effectiveness.

The company's model assumed that patients take 82.8% of the licensed dose of ceritinib based on data from ASCEND‑1 and ASCEND‑2. The committee concluded that on average in clinical practice, the NHS would not pay for the full dose, but it was likely to pay for more than 82.8% because of wastage. So, the dose intensity in the model should be between 82.8% and 100%.

4.12–4.14, 4.16

Incorporation of health-related quality-of-life benefits and utility values

Have any potential significant and substantial health-related benefits been identified that were not included in the economic model, and how have they been considered?

The company used different utility values for ceritinib and best supportive care in the progression-free health state. The committee concluded that the ERG's approach (applying the same utility value to both arms of the model) was more appropriate.

Clinical experts advised that ceritinib may control brain metastases. The patient expert advised that ceritinib allows people to continue to work and live a more normal life. The committee concluded that it had not been presented with evidence that these benefits were realised in practice, nor had it seen evidence that these benefits were not captured in the measurement of QALYs.

4.15, 4.22

Are there specific groups of people for whom the technology is particularly cost effective?

No subgroups were considered.

What are the key drivers of cost effectiveness?

The key drivers of cost effectiveness were the survival functions used to extrapolate overall survival, assumptions about whether ceritinib treatment continues after disease progression, and assumptions about the duration of treatment benefit.

3.30, 3.31

Most likely cost-effectiveness estimate (given as an ICER)

The committee concluded that, on balance, the most plausible ICER was likely to be lower than £50,000 per QALY gained when including the patient access scheme discount and the company's revisions to the model.

4.18

Additional factors taken into account

Patient access schemes (PPRS)

None

End-of-life considerations

The committee concluded that the life expectancy for people with ALK‑positive NSCLC is short and the size of the population is small. It also concluded that although the naive indirect comparison was at high risk of bias because of confounding, it was reasonable to conclude that ceritinib offers an average extension to life of at least 3 months.

4.20, 4.21

Equalities considerations and social value judgements

No equality issues were raised.