Bevacizumab in combination with gemcitabine and carboplatin for treating the first recurrence of platinum-sensitive advanced ovarian cancer

Clinical effectiveness

4.5 The Committee considered that the main source of evidence for the clinical effectiveness of bevacizumab plus gemcitabine and carboplatin was the OCEANS trial that had been conducted in the USA. The Committee agreed with the Evidence Review Group's (ERG's) comments that overall, this was a well-designed double-blind, randomised, placebo-controlled trial. The Committee understood from the clinical specialists that there were no clinical differences between the patients in the trial and patients in the UK with recurrent ovarian cancer, apart from body weight and body surface area. The Committee heard from the clinical specialists that the comparator used in the trial, gemcitabine and carboplatin, is not the most widely used treatment option for recurrent advanced ovarian cancer in the NHS. However, it also heard that gemcitabine and carboplatin could be considered to have a similar efficacy to other treatment options currently used in the NHS, particularly in terms of PFS. The Committee concluded that the results from the OCEANS trial were generalisable to UK clinical practice.

4.6 The Committee discussed PFS results reported in the manufacturer's submission based on the OCEANS trial. It noted that the results for the intention-to-treat population at the September 2010 cut-off date gave a difference in median PFS of 4 months in favour of bevacizumab and this was statistically significant. The Committee noted that in the OCEANS trial, there was a statistically significant difference of approximately 20% in response rate with bevacizumab plus gemcitabine and carboplatin compared with gemcitabine and carboplatin, indicating that bevacizumab is an active drug. Nevertheless, it also acknowledged the ERG's concerns about the issue of censoring. The Committee noted that the data from approximately 30% of the patients had been censored and it was unclear whether these data had been censored because of patients stopping treatment because of adverse events or patients being lost to follow-up. It heard from the manufacturer that information on the number of patients for whom data were censored and the reason why, was not available at the time of the submission. The Committee concluded that, although the trial showed an increase in PFS for bevacizumab plus gemcitabine and carboplatin compared with gemcitabine and carboplatin, it was unclear what effect censoring might have had on these results.

4.7 The Committee considered the most relevant overall survival results for the clinical effectiveness of bevacizumab plus gemcitabine and carboplatin. It explored the 3 interim analyses presented by the manufacturer and noted that none of the analyses showed a statistically significant increase in overall survival in the bevacizumab-treated group. Although the first interim analysis showed a trend towards increased overall survival in the bevacizumab arm (35.5 and 29.9 months in the bevacizumab and placebo groups respectively), in the second and third interim analyses, the difference in median overall survival favoured placebo (1.9 months and 0.3 months respectively). The Committee agreed with the manufacturer's comments that the lack of statistically significant differences between bevacizumab and placebo could have been affected by confounding effects of post-progression treatments. It noted that 18.1% of patients in the bevacizumab arm and 34.7% of patients in the placebo arm received bevacizumab post progression, but also noted that bevacizumab is not licensed for this stage in the treatment pathway because its licence is for first recurrence only. The Committee noted that more than 85% of the patients in both study arms had 3 or more lines of anti-cancer therapy post progression, and it heard from the clinical specialists that it would therefore be very difficult to see any overall survival benefit from bevacizumab with this high level of post-progression treatment without a very much larger trial population. The Committee also heard from the clinical specialists that, although the third interim analysis (March 2012) may be the most reliable because at this stage 59% of patients had died, there could be a bigger issue with confounding. In contrast, the first interim analysis (September 2010) contained overall survival data for only 29% of patients, but may be less confounded by post-progression treatments. The Committee expressed a preference for the more mature and complete overall survival data, but acknowledged that the argument about which were the most reliable data was finely balanced. The Committee concluded that no overall survival benefit for bevacizumab plus gemcitabine and carboplatin had been shown in the OCEANS trial, but the results could have been confounded by post-progression therapies.

4.8 The Committee considered reasons for the discrepancy between the PFS and overall survival results in the OCEANS trial for bevacizumab plus gemcitabine and carboplatin for treating recurrent advanced ovarian cancer. It noted that there were 3 possible underlying causes for the differences:

4.9 The Committee considered the adverse events reported in the OCEANS trial and noted that more patients in the bevacizumab arm (19.8%) stopped treatment because of adverse events than in the placebo arm (4.7%). It heard from the clinical specialists that the discontinuation rate in UK clinical practice would be expected to be lower than in the clinical trial and that most adverse events can be satisfactorily managed. The Committee also heard from one of the patient experts that they had experienced gastrointestinal problems during and after chemotherapy. However, it heard from the patient experts that these problems are usually well managed by clinicians and do not necessarily disrupt a patient's daily life or quality of life. The Committee concluded that the adverse events related to treatment with bevacizumab plus gemcitabine and carboplatin were similar to those related to other chemotherapy regimens and that these events were manageable.

Cost effectiveness

4.10 The Committee discussed the manufacturer's cost-effectiveness estimates, derived from the manufacturer's economic model based on data from the OCEANS trial, and the assumptions in the model. The Committee noted the ERG's comments that it considered the manufacturer's model structure to be generally appropriate, well constructed and transparent. The Committee concluded that the model adhered to the NICE reference case for economic analysis and was acceptable for assessing the cost effectiveness of bevacizumab plus gemcitabine and carboplatin for treating recurrent advanced ovarian cancer.

4.11 The Committee did, however, acknowledge that there were potential shortcomings with some of the assumptions used in the manufacturer's economic model. It noted that health-related quality-of-life data were not collected in the OCEANS trial. The Committee agreed that health-related quality-of-life data collected in the trial would have been preferable for deriving the utilities for the economic model. It noted that the estimates of utility for the PFS and progressed-disease health states were derived from a previous model submitted to NICE (Trabectedin for the treatment of relapsed ovarian cancer [NICE technology appraisal guidance 222]) and that the difference between the utilities for PFS and progressed disease was relatively small (0.718 and 0.649 respectively). The Committee heard from the patient experts that patients may experience a good health-related quality of life while they are progression free. It also noted the comments received from a consultee in response to the appraisal consultation document that reiterated the importance of PFS to patients, and the Committee therefore agreed that it may be plausible for a larger decrement in utility to occur when a person moves from the progression-free health state to a progressed-disease health state and that the difference in utility between the PFS state and progressed state used by the manufacturer could be an underestimate. The Committee also noted that a disutility associated with adverse events was not applied and that there were more serious adverse events in the bevacizumab arm than in the placebo arm. It also discussed how the PFS results were incorporated in the manufacturer's economic model. The Committee noted the ERG's comments on the extrapolation of PFS results by fitting a log-logistic distribution when the Kaplan–Meier data were available. It acknowledged the manufacturer's and ERG's sensitivity and scenario analyses, and concluded that taking all these relevant issues into account (that is, using a higher utility value for the PFS state, including a disutility for adverse events or using the Kaplan–Meier data for PFS) would not be likely to have a significant effect on the incremental cost-effectiveness ratio (ICER).

4.12 The Committee discussed the overall survival data used in the model and noted the ERG's comments that 90% of the quality-adjusted life years (QALYs) gained in the model were a function of the overall survival. It noted that the overall survival data from the first interim analysis (September 2010), in which bevacizumab showed a non-statistically significant increase in overall survival compared with placebo, had been used by the manufacturer in the model with a resulting ICER of £149,000 per QALY gained. The Committee acknowledged its earlier discussion about the uncertainty around the overall survival estimates (see section 4.7). It noted that the manufacturer was unable to provide the ERG with the March 2012 overall survival data and noted that the ERG scenario analysis, which assumed an equivalent overall survival gain for patients in both treatment arms, had resulted in an ICER of over £1.7 million per QALY gained. The Committee concluded that overall survival was the biggest driver of the cost-effectiveness estimate and that, in principle, it would have liked to have seen a sensitivity analysis from the manufacturer that used the March 2012 data, which would have resulted in a higher ICER than the base case.

4.13 The Committee noted the cost-effectiveness results based on the network meta-analysis presented by the ERG. It noted that there were no significant differences in the ICERs for any of the other comparators listed in the scope. The Committee acknowledged that these analyses were exploratory and the underlying assumption was that all comparators had an efficacy similar to that of gemcitabine and carboplatin. It considered this to be a reasonable assumption (see section 4.5) and therefore concluded that the ICER for other comparators was unlikely to be significantly different from that calculated for gemcitabine and carboplatin.

4.14 The Committee considered the most plausible ICER from the model based on the OCEANS trial presented by the manufacturer and by the ERG in their exploratory analyses. It agreed that the manufacturer's base-case ICER, using the September 2010 overall survival data of £149,000 per QALY gained, was likely to be an optimistic cost-effectiveness estimate and that the most plausible ICER could be much higher than this. The Committee noted that the cost-effectiveness estimates for bevacizumab plus gemcitabine and carboplatin were outside the range normally considered to be a cost-effective use of NHS resources. It therefore concluded that bevacizumab plus gemcitabine and carboplatin would not be a cost-effective use of NHS resources for treating the first recurrence of platinum-sensitive advanced ovarian cancer compared with gemcitabine and carboplatin alone.

4.15 The Committee considered supplementary advice from NICE that should be taken into account when appraising treatments that 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:

4.16 The Committee discussed whether bevacizumab plus gemcitabine and carboplatin for treating the first recurrence of platinum-sensitive advanced ovarian cancer fulfilled the criteria for a life-extending, end-of-life treatment. It noted that bevacizumab is licensed for a relatively large population across a range of indications in the treatment of breast, colorectal, renal and non-small-cell lung cancers. Therefore, it does not meet the criterion of the supplementary advice that the treatment should be licensed for small populations. Having established that bevacizumab did not meet the population criterion, the Committee decided it was not necessary to make a decision about the life-expectancy or extension-to-life criteria. It concluded that, on this basis, bevacizumab plus gemcitabine and carboplatin did not fulfil the criteria for being a life-extending, end-of-life treatment.

4.17 The Committee noted the manufacturer's opinion that bevacizumab was an innovative treatment. It acknowledged that advanced recurrent ovarian cancer is a disease with limited treatment options, and that bevacizumab represented a novel biological approach to therapy. It also noted the patient expert comment (see section 4.3). However, the Committee concluded that all substantial benefits related to treatment with bevacizumab plus gemcitabine and carboplatin had been captured in the QALY calculation.

Summary of Appraisal Committee's key conclusions