Pembrolizumab for treating advanced melanoma after disease progression with ipilimumab

Clinical effectiveness

4.1 The Committee discussed the current clinical management of advanced (unresectable or metastatic) melanoma in the NHS and the place of pembrolizumab within the treatment pathway. It was aware of the disparity between the wording of the marketing authorisation for pembrolizumab, the definition of the population in the scope and the decision problem addressed by the company in its submission. The Committee noted the broad wording of the marketing authorisation for pembrolizumab (that is, the treatment of advanced [unresectable or metastatic] melanoma in adults), and the narrower population in the scope (that is, people with advanced [unresectable stage III or IV] melanoma whose disease has progressed after previous treatment with ipilimumab). It further noted that the decision problem addressed by the company in its submission was even narrower, based on the population included in the clinical trial evidence available for pembrolizumab from KEYNOTE‑002 (that is, people with unresectable or metastatic melanoma whose disease has progressed after ipilimumab and, if the disease is BRAF V600 mutation positive, a BRAF or MEK inhibitor). The Committee heard from the company that, when preparing the submission, the marketing authorisation was expected to specify that pembrolizumab would be indicated after progression with ipilimumab, and if BRAF V600 mutation‑positive disease, after a BRAF or MEK inhibitor, and that this had been the approach taken when pembrolizumab was made available via the early access to medicines scheme in the NHS. The Committee questioned whether in clinical practice pembrolizumab would always be used after a BRAF or MEK inhibitor in people with BRAF V600 mutation positive disease, as in KEYNOTE‑002 and the company submission, or whether pembrolizumab would sometimes be considered as an alternative to a BRAF or MEK inhibitor. It heard from the clinical experts that, after disease progression with ipilimumab, most people with BRAF V600 mutation‑positive disease would have a BRAF or MEK inhibitor but that, in a few people, pembrolizumab might be preferred because of slow‑growing disease and the expectation of a longer survival. However, the Committee was aware that the company had not submitted any evidence for the efficacy of pembrolizumab compared with BRAF inhibitors. It considered that it could not make recommendations for a population for whom there was no evidence of the relative clinical effectiveness of pembrolizumab. Therefore, the Committee accepted the company's approach to the decision problem. It concluded that, on the basis of the evidence submitted, it could only make recommendations for pembrolizumab after treatment with ipilimumab and, for BRAF V600 mutation‑positive disease, after both ipilimumab and a BRAF or MEK inhibitor.

4.2 The Committee discussed the relevant comparators for pembrolizumab and noted that the company considered conventional chemotherapy to be the appropriate comparator for pembrolizumab. The Committee heard from the clinical experts that conventional chemotherapy, including dacarbazine, remained the only treatment option after treatment with ipilimumab and, for BRAF V600 mutation‑positive disease, a BRAF or MEK inhibitor. It also heard that there is no proven survival benefit associated with these therapies. The Committee concluded that conventional chemotherapy was an appropriate comparator for pembrolizumab at this stage of the disease.

4.3 The Committee discussed the clinical need of people with advanced melanoma after progression with ipilimumab and, for BRAF V600 mutation‑positive disease, a BRAF or MEK inhibitor. It heard from the patient expert that metastatic melanoma is associated with severe emotional stress and anxiety about the future for the patient and their family, and a reduced quality and length of life. The patient expert and clinical experts also explained that ipilimumab, which is used earlier in the pathway, can be associated with severe side effects that affect normal activities and that these may be so severe the patient sometimes needs hospital admission. The patient expert also noted that, in contrast, pembrolizumab offers a much more manageable side effect profile that means people with metastatic melanoma can continue their normal lives and activities, including employment. The patient expert also highlighted that, once the disease progresses after ipilimumab, the lack of further therapy options that can extend survival is devastating, so having an additional option such as pembrolizumab provides hope for the future. The Committee concluded that the availability of a new treatment that slows disease progression and improves quality of life when other therapies have failed is very important to patients and their families.

4.4 The Committee considered that the key clinical evidence came from KEYNOTE‑002. The Committee noted that KEYNOTE‑002 had not included any UK sites but had included centres in the USA, Argentina, and some European countries. It heard from the clinical experts that historically there had been a difference in outcomes in people with advanced melanoma in the UK compared with some other countries, possibly related to later diagnosis. However, the clinical experts explained that, because of recent advances in managing malignant melanoma in the NHS, including earlier diagnosis and the availability of new treatments, this has changed. The Committee also noted that 23% of people in KEYNOTE‑002 had BRAF V600 mutation‑positive disease. It heard from the clinical experts that this was lower than the overall prevalence of BRAF V600 mutation‑positive disease in the UK, which is about 45%. The clinical experts noted that this difference related mainly to patient selection in the trial and that the results would still be generalisable to the BRAF V600 mutation‑positive population eligible to have pembrolizumab in clinical practice. Therefore, the Committee concluded that KEYNOTE‑002 was generalisable to clinical practice in the NHS for those people whose disease has progressed after ipilimumab and, for BRAF V600 mutation‑positive disease, a BRAF or MEK inhibitor.

4.5 The Committee discussed the progression‑free survival results from KEYNOTE‑002. It was aware that follow‑up in the trial was short and therefore, the results were immature. It was aware that the company presented results based on an independent central review and an investigator assessment, and that both results were based on an interim analysis. It was also aware that the median progression‑free survival for pembrolizumab was higher using the investigator assessment, which it considered could relate to the lack of blinding in the trial. However, the Committee noted that both methods showed small but statistically significant improvements in median progression‑free survival with pembrolizumab compared with conventional chemotherapy. It also noted that there was a separation of the progression‑free survival curves that began around 12 weeks, after which a larger difference in progression‑free survival was seen between chemotherapy and pembrolizumab. The Committee heard from the clinical experts that treatment response is best assessed around 12 weeks and, after this, the progression‑free survival benefit starts becoming apparent. The Committee concluded that the evidence from KEYNOTE‑002, although immature, suggested that pembrolizumab improved progression‑free survival compared with conventional chemotherapy, and that the benefit became most apparent after 12 weeks.

4.6 The Committee discussed the overall survival results and was aware that there was no statistically significant difference seen between pembrolizumab and chemotherapy. It considered that this could relate to immaturity of the data given the short follow‑up in the trial, and particularly to the fact that 48% of people in the chemotherapy group had switched to pembrolizumab. The Committee noted that the company had explored different methods of adjusting for treatment switching, and had concluded that the 2‑stage adjustment method was the most appropriate. It noted that the Evidence Review Group (ERG) had agreed with the company's approach and that, when using this method, there was a statistically significant difference in median overall survival between pembrolizumab and chemotherapy of 3.5 months. The Committee concluded that, although the results were immature and there was uncertainty about the true survival benefit associated with pembrolizumab compared with conventional chemotherapy, the best available evidence from KEYNOTE‑002, using the 2‑stage adjustment method, suggested a difference in median overall survival of 3.5 months.

4.7 The Committee discussed the adverse events associated with pembrolizumab. It noted that, in KEYNOTE‑002, pembrolizumab was associated with fewer adverse events than chemotherapy. It also recalled the comments from the clinical and patient experts that the adverse events of pembrolizumab were usually manageable and allowed people to continue with their normal activities (see section 4.3). The Committee concluded that the adverse events of pembrolizumab were manageable, and favourable when compared with chemotherapy.

Cost effectiveness

4.8 The Committee considered the company's model, which compared pembrolizumab with best supportive care in people with advanced (unresectable or metastatic) melanoma after progression with ipilimumab and, for BRAF V600 mutation‑positive disease, a BRAF or MEK inhibitor. The Committee considered that a 3‑state model structure was appropriate for decision‑making. It also accepted the company's use of data from the chemotherapy group in KEYNOTE‑002 as a proxy for best supportive care based on the Committee's previous conclusion that conventional chemotherapy was an appropriate comparator for pembrolizumab in its expected place in the treatment pathway (see section 4.2).

4.9 The Committee noted that the company's economic analysis resulted in a deterministic incremental cost‑effectiveness ratio (ICER) for pembrolizumab compared with best supportive care of about £43,000 per quality‑adjusted life year (QALY) gained, and a probabilistic ICER of about £68,000 per QALY gained. The Committee expressed concerns about the substantial difference between these figures and discussed possible reasons for the difference. It understood that the variables with the biggest impact on the results were the parameters used to extrapolate progression‑free survival and the hazard ratio for overall survival from the 2‑stage adjustment method used to extrapolate the overall survival results.

4.10 The Committee was aware that the company used a Gompertz distribution for extrapolating progression‑free survival from the time of assessment (week 12) onwards in the pembrolizumab group, and that it assumed all people having chemotherapy had progressed or died by the end of the trial (week 67). The Committee noted the comments from the ERG that the Gompertz distribution is associated with a long tail in the progression‑free survival curve and that this tends to overestimate progression‑free survival in the long‑term. It also noted that the confidence intervals were wide because there was a very small proportion of the cohort still at risk at the end of the tail in the progression‑free survival curve, which would have a particular impact on the probabilistic cost‑effectiveness estimate. The ERG also considered that the company's assumption that all people in the chemotherapy group had disease progression or died at week 67 overestimated the relative progression‑free survival benefit associated with pembrolizumab. It also noted that it was more appropriate to use exponential models for extrapolation in both treatment groups. The Committee was concerned that the progression‑free survival results were immature, and that it was uncertain how many and for how long people would have progression‑free disease. However, it accepted that, when using a Gompertz distribution, the progression‑free survival results appeared too optimistic because it was unlikely that people would have life‑long progression‑free disease.

4.11 The Committee noted that the company had used a 3‑stage approach to modelling overall survival based on different data sources. It also noted the comments from the ERG that this provided clinically implausible results such as: a 4‑year period of zero mortality risk followed by a sudden increase in the mortality risk occurring after 10 years, and an overall survival benefit of pembrolizumab compared with best supportive care that persisted indefinitely. The Committee noted that the ERG had implemented amendments to the company's model in line with its preferred assumptions. It also noted that it provided an exploratory analysis incorporating a different approach to modelling overall survival using trial results, together with a mixed‑exponential model based on registry data. The Committee expressed the view that the ERG's overall approach was generally more clinically plausible than the company's model using 3 separate sources.

4.12 The Committee noted that, even when incorporating multiple amendments such as changes to the utility values and costs, the cumulative effect of all ERG amendments on the ICER was modest, increasing the ICER for pembrolizumab compared with best supportive care by about £4000, to £47,000 per QALY gained. The Committee was aware that the ERG had not presented probabilistic cost‑effectiveness results and recalled its previous concerns about the wide difference between the company's deterministic and probabilistic results (see sections 4.9 and 4.10). The Committee heard from the ERG that, because its approach for modelling progression‑free survival used an exponential model instead of a Gompertz model, it was likely that the ERG's deterministic and probabilistic analyses would be much more similar. The Committee therefore considered that, despite the differences between the company's model and the ERG's preferred approach, the ICERs were not very different. It concluded that the most plausible ICER for pembrolizumab compared with best supportive care was likely to be less than £50,000 per QALY gained.

4.13 The Committee discussed the innovative nature of pembrolizumab. It noted that the company stated that pembrolizumab was innovative and a step‑change in the management of advanced melanoma because it treats a life‑threatening and seriously debilitating condition, meets a high unmet need and provides a significant advantage over other treatments used in the UK. The Committee agreed with the company that pembrolizumab is innovative because it meets a high unmet medical need, and because of its low toxicity and favourable adverse effects profile compared with other treatments for metastatic melanoma. However, it could not identify any specific health‑related benefit that had not been already captured in the QALY calculation.

4.14 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 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.

4.15 The Committee discussed whether pembrolizumab met all the criteria to be considered a life‑extending, end‑of‑life treatment and whether the evidence presented was plausible, objective and robust enough to support it. The Committee agreed that the life expectancy of people with advanced melanoma after progression with ipilimumab and, for BRAF V600 mutation‑positive disease, a BRAF or MEK inhibitor, who do not have any other treatments available apart from conventional chemotherapy, is normally less than 24 months. For example, life expectancy in people with metastatic melanoma was up to 9.0 months based on data from Balch et al. (2001), Korn et al. (2008) and Thirlwell and Nathan (2008), 20.9 months in the company's base‑case analysis, and 17.2 months in the ERG's exploratory analyses. The Committee also agreed that, although the overall survival data are immature, the best available data from KEYNOTE‑002 and using the 2‑stage adjustment method for treatment switching suggested that pembrolizumab offers an extension to life compared with conventional chemotherapy of 3.5 months (see section 4.6). Finally, the Committee discussed the company's estimate for the number of people eligible to have pembrolizumab (that is, people with unresectable or metastatic melanoma whose disease has progressed after ipilimumab and, for BRAF V600 mutation‑positive disease, a BRAF or MEK inhibitor), noting that this was about 600 in 2015, and about 300 annually thereafter. It heard from the clinical experts that these estimates were plausible and in line with the number of people who have had pembrolizumab through the early access to medicines scheme in the NHS. It concluded that this represents a small patient population. The Committee considered that, although the evidence for pembrolizumab in terms of progression‑free survival and overall survival is immature and based on a short follow‑up, the estimates and assumptions applied can be considered plausible, objective and robust enough to conclude that pembrolizumab meets all the criteria to be considered a life‑extending, end‑of‑life treatment.

4.16 Having accepted that the supplementary advice for appraising a life‑extending, end‑of‑life treatment applies, the Committee discussed whether pembrolizumab could be considered a cost‑effective use of NHS resources. The Committee recalled its previous conclusion that, despite the differences between the company's model and the ERG's amendments, the results were not very different and that it was likely that the most plausible ICER for pembrolizumab compared with best supportive care was less than £50,000 per QALY gained. Therefore, the Committee concluded that, on balance, pembrolizumab could be considered a cost‑effective use of NHS resources for people with advanced (unresectable or metastatic melanoma) after progression with ipilimumab and, for BRAF V600 mutation‑positive disease, a BRAF or MEK inhibitor.

4.17 The Committee was aware of NICE's position statement about the Pharmaceutical Price Regulation Scheme (PPRS) 2014, and in particular the PPRS Payment Mechanism. It acknowledged '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 about the relevance of the PPRS to this appraisal of pembrolizumab. It therefore concluded that the PPRS Payment Mechanism was irrelevant for the consideration of the cost effectiveness of pembrolizumab.

Summary of Appraisal Committee's key conclusions