Vemurafenib for treating locally advanced or metastatic BRAF V600 mutation‑positive malignant melanoma

Clinical effectiveness

4.2 The Committee considered the results presented by the manufacturer on the clinical effectiveness of vemurafenib. It noted that the manufacturer derived efficacy data primarily from the BRIM3 trial. This showed that treatment with vemurafenib led to a statistically significant increase in median progression‑free survival of 5.3 months (HR 0.38; 95% CI 0.32 to 0.46), and an increase in median overall survival of approximately 3.3 months (uncensored HR 0.76; 95% CI 0.63 to 0.93) based on the February 2012 data cut‑off, compared with dacarbazine for people with previously untreated advanced or metastatic disease. The Committee also took particular note of the increase in response rate to treatment with vemurafenib over time and acknowledged the number of people whose disease responded completely to treatment with vemurafenib was 5.6% at the February 2012 data cut‑off compared with 0.9% in December 2010. The Committee accepted that there could be a long‑term benefit for some people and concluded that vemurafenib is a highly effective treatment for locally advanced or metastatic BRAF V600 mutation‑positive malignant melanoma.

4.3 The Committee discussed whether the BRIM3 study is generalisable to UK clinical practice. The Committee noted that patients with an ECOG performance status of 0 or 1 were included in the BRIM3 study, and discussed whether people with an ECOG performance status of 2 or 3 are likely to receive vemurafenib treatment in UK clinical practice. The Committee heard from the clinical specialists that the use of vemurafenib is unlikely to be restricted to people with a good performance status, because case studies have demonstrated that even people with the poorest prognosis can still benefit from treatment. The Committee concluded that the results of the BRIM3 study were generalisable to UK clinical practice.

4.4 The Committee considered the 4 different data cut‑off points from the BRIM3 study presented by the manufacturer. It acknowledged that the Data and Safety Monitoring Board ended the study early and allowed patients to switch from dacarbazine to vemurafenib (or another treatment) on disease progression based on the evidence for the efficacy of vemurafenib after an interim analysis in December 2010. The Committee noted that vemurafenib, irrespective of the data cut‑off, was superior to dacarbazine with respect to the primary endpoints of progression‑free survival and overall survival, and that this was statistically significant at all data‑cut offs. The Committee acknowledged that the March 2011 cut‑off data included 15% of participants who switched from dacarbazine to vemurafenib on disease progression, and that although the October 2011 and February 2012 data cut‑offs provided an additional 7 months and 11 months of data respectively, they also included a greater proportion of patients who switched from dacarbazine to vemurafenib (bringing the switching rate to 24% in October 2011 and 34% in February 2012). The Committee was cautioned by the clinical specialists and the manufacturer that the data on overall survival from the later data cut‑offs were confounded not only by switching from dacarbazine to vemurafenib, but also by the fact that patients whose disease did not show an objective response were able to receive a range of other therapies including ipilimumab (another treatment for metastatic melanoma) and other investigational BRAF inhibitor treatments. The Committee acknowledged the clinical specialists' concerns but were minded to accept that more information on the long‑term clinical effectiveness of vemurafenib at the February 2012 data cut‑off outweighed concerns about the robustness of the data compared with the earlier data cut‑offs. However, the Committee also acknowledged that in the situation in which significant numbers of trial participants switched from a drug with limited efficacy to one with much higher efficacy, it would be reasonable to consider the effect that this would have on the results in the uncensored trial arms.

4.5 The Committee discussed the issue of switching in the BRIM3 trial. It was aware that people from the dacarbazine arm could receive treatment with vemurafenib on disease progression, and recognised that this change to a more effective treatment could have confounded the calculation of overall survival benefit from vemurafenib. The Committee agreed that it was appropriate to adjust the overall survival results from the February 2012 data cut‑off (which gave a median overall survival estimate of 13.6 months for vemurafenib and 10.3 months for dacarbazine, HR 0.76 [95% CI 0.63 to 0.93]) to control for switching using statistical modelling or other techniques. However, the Committee agreed that any estimate of overall survival obtained using these techniques would be subject to uncertainty.

4.6 The Committee discussed the manufacturer's approach to adjusting the survival estimate for the dacarbazine arm of the BRIM3 study (February 2012 data cut‑off) to account for people who switched to vemurafenib on disease progression using the RPSFT method. The Committee considered the key assumptions outlined by the manufacturer that underpin the RPSFT method and their validity in relation to the BRIM3 study. The Committee noted that both the manufacturer and the ERG agreed that the effect of vemurafenib treatment on mortality changes over time (see section 3.22) and that applying a single acceleration factor may therefore be an oversimplification, and that the results should be viewed with caution. To evaluate the plausibility of the ICER obtained using the RPSFT method, the Committee noted that the hazard ratio for overall survival (using February 2012 data and RPSFT adjustment) was 0.64 (95% CI 0.53 to 0.78; p<0.0001) and acknowledged that this was in line with the overall survival hazard ratio using the October 2011 data cut‑off (censored HR 0.62; 95% CI 0.49 to 0.77). The Committee then considered the scenario in which external data (Bedikian et al. [2011]) were used to represent the clinical effectiveness of dacarbazine, and acknowledged similarities to the BRIM3 trial population. It expressed caution about using an external trial over data available in the BRIM3 trial, but accepted that this gave an ICER similar to, but lower than the RPSFT method, providing reassurance that the RPSFT method was a reasonable approach in this case. The Committee accepted that there was evidence that vemurafenib increased overall survival compared with dacarbazine and concluded that, of the various methods to adjust the BRIM3 trial data for crossover, the RPSFT method was the most plausible because it gave results in line with those obtained using an alternative indirect method (Bedikian et al. [2011]) for removing the effect of crossover.

4.7 The Committee discussed whether the benefit from vemurafenib over dacarbazine was likely to continue once treatment was stopped, or conversely whether there may be accelerated disease progression. It noted the concerns of the ERG that people may experience accelerated disease progression once treatment with vemurafenib is stopped. The Committee heard from the clinical specialists that people whose disease progresses after treatment with vemurafenib may have a smaller tumour burden compared with those treated with dacarbazine because of the higher disease response rate seen with vemurafenib. Therefore, on disease progression they may have a survival advantage. The clinical specialists also explained that some people will continue to experience a benefit after treatment is stopped because they have discordant progression (if only 1 area of their tumour, which can be potentially resected or treated separately, has progressed but other parts have not). The Committee acknowledged that the existence or magnitude of continued benefit from vemurafenib after treatment is stopped is uncertain, but recognised there is no evidence currently available to suggest that people who stop vemurafenib treatment will experience accelerated disease progression compared with those who have been treated with dacarbazine.

4.8 The Committee considered the adverse events associated with treatment with vemurafenib. It noted that cutaneous events were commonly reported in the BRIM3 study, with 61 people (18%) needing treatment for grade 3 cutaneous squamous‑cell carcinoma, keratocanthoma or both. The Committee heard from the clinical specialists and patient experts that people being treated with vemurafenib can have significant skin toxicities, but these are manageable with dose reductions, topical treatments or local excision of lesions. The Committee concluded that treatment with vemurafenib had an acceptable adverse event profile when taking into account the potential benefits.

Cost effectiveness

4.9 The Committee discussed the assumptions underpinning the estimate of overall survival in the manufacturer's revised economic model. It noted that the hazard ratio estimates from the February 2012 data cut‑off for up to 14 months were used, and no further beneficial effect of vemurafenib on the risk of death compared with dacarbazine was assumed after 14 months (hazard ratio of 1). The Committee heard from the manufacturer that it considered its approach to modelling survival beyond 14 months to be conservative because the RPSFT method does not take into account post‑progression use of ipilimumab or BRAF inhibitors other than vemurafenib, which both favour the dacarbazine arm. The clinical specialists stated that treatment with vemurafenib could alter the biology of the disease and could potentially have a beneficial impact on the post‑progression survival of patients. They also stated that there was no trial evidence or clinical evidence that patients who progressed after vemurafenib treatment did so any faster than those who had received dacarbazine as was suggested both in the ERG's original exploratory approach, and in the scenario analysis requested by the Committee. The Committee therefore accepted the manufacturer's assumption that patients in both treatment arms would have an equal risk of death beyond 14 months. The Committee also noted that the ERG, and to a lesser extent the manufacturer, relied on register data to predict long‑term survival in the dacarbazine arm. These data are historical, may be subject to selection or treatment centre bias, and are pooled data that are not specific for BRAF‑positive patients. The Committee considered that this further contributed to the uncertainty about the most robust approach to modelling long‑term survival, but found no evidence that accelerated disease progression occurred in vemurafenib‑treated patients compared with those treated with dacarbazine beyond 14 months.

4.10 The Committee discussed the ERG's exploration of an alternative approach to modelling overall survival. It noted that the ERG used the October 2011 data cut‑off and assumed that, after 14 months, the benefit of vemurafenib decreased, and the rate of death exceeded that in the dacarbazine arm until the 2 survival arms converged by 4 years, after which, the risk of death in the vemurafenib and dacarbazine arms would be equal. This resulted in a calculated mean overall survival benefit of 97 days for patients treated with vemurafenib compared with those who received dacarbazine, which the Committee noted was less than the median overall survival benefit of 3.6 months demonstrated in the October 2011 data cut‑off in the trial. The Committee reiterated its conclusion that there was significant uncertainty about the magnitude of the survival benefit attributable to vemurafenib but concluded that there was no evidence to support disease acceleration after vemurafenib treatment relative to dacarbazine, as occurred in the ERG's exploratory analysis. The Committee therefore accepted the manufacturer's approach to modelling overall survival and its assumption that there was no further beneficial effect of vemurafenib on the risk of death.

4.11 The Committee discussed the costs associated with supporting BRAF V600 mutation testing. It heard from the clinical specialists that the test is currently being used in selected reference centres around the UK, with the cost of implementation supported by the manufacturer. The clinical specialists said that, although it can take weeks to provide a test result, the main time‑limiting factor is accessing and preparing the tumour blocks, and transporting them, rather than performing the test. The Committee recognised the additional burden on pathology laboratories associated with vemurafenib treatment, but it was satisfied that BRAF V600 mutation testing is likely to become part of routine management for people with advanced melanoma and that it would not impose a significant resource impact on the NHS in the future.

4.12 The Committee noted that utility values in the manufacturer's model were sourced from the literature in the absence of robust data from the BRIM3 study. The Committee agreed with the manufacturer's assumption of a higher utility value for progression‑free survival for vemurafenib, given its improved clinical profile, including oral administration compared with intravenous administration for dacarbazine. The Committee noted that a utility of 0.59 was applied to the progressed disease state, which is lower than utilities for the same state accepted previously by the Committee for ipilimumab. It heard from the ERG that people who survive in the long term are likely to have a higher quality of life than those with rapidly progressive disease and therefore a higher utility for long‑term survival (that is, survival greater than 5 years estimated to be 0.767) is justified. The Committee acknowledged the logic of this approach and was therefore persuaded that an improved utility value for the progressed disease state after 5 years of survival was justified.

4.13 The Committee considered the results of the manufacturer's updated cost‑effectiveness analysis, which incorporated the ERG's suggested amendments (see sections 3.15 and 3.17) and the further errors identified in the economic model. The revised base‑case deterministic ICER for vemurafenib compared with dacarbazine was £51,800 per QALY gained when the February 2012 data cut‑off point (after RPSFT adjustment for switching) was used. The Committee was minded to compare the results of the RPSFT analysis with a scenario in which external data from trials of dacarbazine without switching to another drug was used (that is, Bedikian et al. [2011]). The Committee agreed that it was both a potential alternative method to evaluate the magnitude of the effect of switching and also a consistency check on direct methods such as the RPSFT. When the Bedikian trial data were used to model long‑term survival with dacarbazine, rather than the RPSFT‑adjusted BRIM3 trial data, the ICER was £44,400 per QALY gained using the February 2012 data cut‑off point. The Committee concluded that, using the manufacturer's approach to modelling overall survival and the February 2012 data cut‑off, the ICERs for vemurafenib compared with dacarbazine ranged between £44,000 and £51,800 per QALY gained.

4.14 The Committee considered the cost‑effectiveness estimates for the additional scenario it had requested (see section 3.24). It was aware of the manufacturer's and the clinical specialists' opinions that such an analysis was inappropriate because the modelling approach gave a post‑progression survival after treatment with vemurafenib that they considered to be implausible. The Committee noted that the ERG had undertaken the analysis as requested by the Committee, and this had resulted in an ICER for vemurafenib compared with dacarbazine of £121,000 per QALY gained using the February 2012 data cut‑off. The Committee then considered whether the ICERs presented by the manufacturer or the ICER from the Committee's requested scenario analysis were the more plausible. The Committee noted that it was unlikely that a single acceleration factor would capture the benefit of vemurafenib, which meant that the resulting ICER of £51,800 presented by the manufacturer should be interpreted with caution, although it accepted that the RPSFT method had been used in previous NICE technology appraisals. The Committee was satisfied with the use of external data, although there were concerns with reliability, and noted that the ICER calculated in this way was not inconsistent with, but was lower than, the RPSFT analysis that gave an ICER of £44,000 per QALY gained. The Committee was not satisfied that the alternative ICER (based on the Committee's requested scenario) of £121,000 per QALY gained represented the true benefit of vemurafenib because it accepted that it was implausible that post‑progression survival after vemurafenib was shorter than after dacarbazine. The Committee therefore concluded that the most plausible ICER was in the range of £44,000 to £51,800 per QALY gained.

4.15 The Committee considered whether it would be appropriate to take into account sensitivity analyses on the discount rates used in the model and their effects on the revised ICER. It noted that a sensitivity analysis on the original base‑case ICER, which included 3.5% discounting of costs and 1.5% discounting of benefits, gave an ICER of £48,200 per QALY gained. The Guide to the methods of technology appraisal (2008) clarification issued by the Board of NICE states that 'where the Appraisal Committee has considered it appropriate to undertake sensitivity analysis on the effects of discounting because treatment effects are both substantial in restoring health and sustained over a very long period (normally at least 30 years), the Committee should apply a rate of 1.5% for health effects and 3.5% for costs'. Having referred to this clarification, the Committee considered that substantial restoration of health for a very long period equated to restoration of health to the extent that the person could be considered as having effectively been cured of their condition. The Committee noted that although there were patients in the vemurafenib arm of the BRIM3 study who showed a complete disease response, there was considerable uncertainty about how prolonged the benefit from the treatment would be. The Committee heard from the clinical specialists that there is no evidence at present to suggest that advanced melanoma is curable, or that vemurafenib would have substantial benefits beyond 30 years. The Committee therefore concluded that there was no case for differential discounting to be applied.

4.16 The Committee considered supplementary advice from NICE that should be taken into account when appraising treatments which may extend the life of people 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 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.

4.17 The Committee heard from the clinical specialists that the average life expectancy for people with locally advanced or metastatic BRAF V600 mutation‑positive malignant melanoma, particularly for those with distant metastases, as reflected in the trial population, was 3 to 9 months, and was unlikely to be greater than 24 months. The Committee also agreed that there was sufficient evidence from the BRIM3 study to indicate that treatment offers an extension to life of at least an additional 3 months, compared with current NHS treatment. The Committee heard from the manufacturer and the clinical specialists that the total number of people who would be eligible for treatment with vemurafenib was fewer than 1000 each year in England and Wales, which the Committee accepted represents a small patient population. Therefore, the Committee was satisfied that vemurafenib met all the criteria for being a life‑extending, end‑of‑life treatment and that the trial evidence presented for this was robust.

4.18 The Committee was aware that NICE's Guide to the methods of technology appraisal (2008) states that a strong case should be identified for accepting an ICER that is higher than £30,000 per QALY gained. The Committee noted that, in these circumstances, the NICE methods guide states that judgements about the acceptability of the technology as an effective use of NHS resources will specifically take account of:

4.19 The Committee discussed whether the assessment of the change in health‑related quality of life had been adequately captured in the economic analysis. It heard from a patient expert that successfully treated people could lead an active and fulfilling life and were able to contribute to society. The Committee accepted that vemurafenib represents a valuable new therapy and that its mechanism of action is novel. It acknowledged that few advances had been made in the treatment of advanced melanoma in recent years and vemurafenib could be considered a significant innovation for a disease with a high unmet clinical need. It also heard from the clinical specialists that vemurafenib had advanced the understanding of this disease and opened the way to new treatments. The Committee considered that the symptomatic improvement attributable to vemurafenib had been captured in the higher utility value assigned to the progression‑free survival health state for people receiving vemurafenib compared with dacarbazine. It considered that vemurafenib represents a highly effective new therapy in an area of unmet need; it has a novel mechanism of action, is a targeted therapy, is administered orally, is life‑extending and meets the criteria for an end‑of‑life treatment. The Committee concluded that the combined value of these factors meant that vemurafenib could be considered a cost‑effective use of NHS resources.

Summary of Appraisal Committee's key conclusions