Bevacizumab in combination with capecitabine for the first-line treatment of metastatic breast cancer

Clinical effectiveness

4.3 The Committee considered the generalisability of the RIBBON-1 trial to UK clinical practice. The Committee was aware of NICE clinical guideline 81 for advanced breast cancer in which capecitabine follows both anthracycline and taxane therapy in the care pathway. It noted the design of the RIBBON-1 trial, which had 2 cohorts of patients; those to be treated with anthracyclines or a taxane and those to be treated with capecitabine. The Committee noted that 63% of the capecitabine cohort had received prior anthracycline therapy and 40% had received prior taxane therapy. Because there was likely to have been considerable overlap between these 2 groups (and perhaps total overlap, as assumed by the manufacturer), a significant percentage of patients (up to 37%) could have received capecitabine for first-line treatment of their metastatic breast cancer as their first ever chemotherapy. The Committee considered that this meant that this group of patients was not representative of the typical UK metastatic breast cancer population. The Committee also noted that the 60% of patients in the capecitabine cohort of the RIBBON-1 trial who had not received prior taxane therapy had good performance status and yet taxane treatment was not considered appropriate for them. In addition, the Committee observed that 30% of the patients in the prior taxane subgroup for whom taxanes were not considered appropriate (as indicated by the entry criteria for the capecitabine arm of the RIBBON-1 trial) subsequently received taxanes after disease progression. However, the Committee was aware that the decision to treat with capecitabine or a taxane could also be based on other factors which may be important to patients, such as the lack of hair loss with capecitabine. The Committee concluded that there were still some issues about the generalisability of the RIBBON-1 trial to clinical practice in the UK. This was because a significant proportion of patients in the trial had not received previous chemotherapy, and taxanes had not been considered for a significant proportion despite their young age and good performance status.

4.4 The Committee also noted that the dose of capecitabine in the trial was 1000 mg/m² rather than the licensed dose of 1250 mg/m². The Committee was aware that the dose of capecitabine used in UK practice was often lower in older patients and those with poor performance status, but observed that all patients in the RIBBON-1 trial were of ECOG performance status 0 or 1 and the median age was 56 years. However, it noted the comments during consultation that some clinicians in the UK start at a dose lower than the licensed dose (often 1000 mg/m²) even in fitter patients. The Committee therefore concluded that the dose of capecitabine used in the trial may have some relevance to clinical practice in the UK.

4.5 The Committee considered the clinical-effectiveness data from the capecitabine cohort of the RIBBON-1 trial for the comparison of bevacizumab plus capecitabine with capecitabine plus placebo. The Committee noted that the results from the ITT population demonstrated a statistically significant median investigator-assessed progression-free survival benefit of 2.9 months for bevacizumab plus capecitabine compared with capecitabine plus placebo. However, the Committee noted this improvement in progression-free survival did not translate into a statistically significant improvement in overall survival. The Committee was aware that patients from both arms of the trial could receive treatment with bevacizumab after disease progression as well as other subsequent treatments and that all these subsequent therapies could have confounded the relative treatment effect in terms of overall survival. The Committee also noted that no quality of life data had been collected in the trial. The Committee considered quality of life to be an important outcome measure in advanced cancer and that this was an omission from the trial. Without quality of life data and a statistically significant improvement in overall survival, the Committee explored the value of an increase in progression-free survival. The Committee was aware of a statement from the clinical specialist that the most important outcome for patients with metastatic breast cancer is prolonging disease-free survival. However, the Committee heard from the patient expert that patients would value an increase in progression-free survival when it is accompanied by an improvement in quality of life that would allow them to carry out normal daily activities. The Committee concluded that bevacizumab plus capecitabine improved progression-free survival relative to capecitabine plus placebo, but that there was no robust evidence that it improved overall survival and that its effects on health-related quality of life had not been captured.

4.6 The Committee discussed the adverse event profile associated with bevacizumab plus capecitabine compared with capecitabine plus placebo. The Committee noted that grade 3–5 adverse events were higher with bevacizumab plus capecitabine (36.6%) compared with capecitabine plus placebo (22.9%). In addition, the number of patients with hypertension, proteinuria, sensory neuropathy and venous thromboembolic events was higher with bevacizumab plus capecitabine compared with capecitabine plus placebo. The Committee considered that the adverse event profile of bevacizumab plus capecitabine was particularly important because if people opt for capecitabine instead of taxane-based treatment it may be an indication that a better adverse event profile was important to them. The Committee concluded that bevacizumab plus capecitabine had a less favourable adverse event profile than capecitabine plus placebo.

4.7 The Committee noted that no clinical evidence of the effectiveness of bevacizumab plus capecitabine compared with vinorelbine was presented by the manufacturer as specified in the final scope issued by NICE. The Committee noted the manufacturer's statement that capecitabine is generally preferred to vinorelbine in in UK clinical practice and that vinorelbine had been included as a comparator as part of the scenario analysis in the economic modelling. The Committee concluded that, without studies that would allow for an indirect comparison of bevacizumab plus capecitabine with vinorelbine, and without evidence to suggest that vinorelbine was superior to capecitabine, it was appropriate for capecitabine to be presented as the main comparator.

4.8 The Committee examined the subgroup analysis conducted by the manufacturer comparing bevacizumab plus capecitabine with capecitabine plus placebo in patients who had previously received a taxane. The Committee was aware that the differences in progression-free and overall survival between the bevacizumab plus capecitabine and capecitabine plus placebo arms were statistically significantly greater in this subgroup of patients. However, the Committee noted that previous taxane therapy was not a stratification factor at randomisation and that this subgroup was specified after the trial had begun but before the analysis was completed. The Committee also noted that the overall survival results were based on small numbers of events: 70 deaths in the bevacizumab plus capecitabine arm and 44 deaths in the capecitabine plus placebo arm. In addition, the Committee was aware that no statistical adjustments were made to control for multiple testing, thus increasing the risk of chance findings. The Committee considered the manufacturer's original submission and their consultation comments about an increased benefit observed in the prior taxane subgroup compared with the ITT population in other metastatic breast cancer trials of bevacizumab as part of first-line chemotherapy (the AVADO and E2100 trials). The Committee noted that although the prior taxane subgroups showed significant benefits with bevacizumab in the AVADO and E2100 trials, the benefits were not replicated in the prior taxane subgroup in the anthracycline/taxane cohort of RIBBON-1. The Committee heard from the clinical specialist that the prior taxane subgroup was a clinically relevant subgroup given the current treatment pathway of metastatic breast cancer. The Committee also noted the comment from the consultees and clinical specialist that women with triple negative breast cancer for whom there are limited treatment options are the most clinically relevant subgroup, with a realistic chance of benefitting from bevacizumab plus capecitabine treatment. However, the Committee was aware that the RIBBON-1 trial results did not show an advantage for bevacizumab plus capecitabine in this triple negative subgroup. The Committee was concerned about the robustness of the data from the AVADO and E2100 trials because of the small patient numbers in the prior taxane subgroups in the AVADO and E2100 trials and the unblinding and non-stratification for prior taxane use in the E2100 trial. It noted that although the progression-free survival and overall survival benefits were higher in the prior taxane subgroups in the E2100, AVADO and the capecitabine cohort of the RIBBON-1 trials, there was no biologically plausible reason why bevacizumab plus capecitabine would be more effective in this subgroup than in the ITT population. The Committee therefore considered that a formal study would be needed to confirm these benefits, as had been recognised by the manufacturer. The Committee concluded that the results from the prior taxane subgroup of the RIBBON-1 trial were not sufficiently robust to use for the development of guidance.

Cost effectiveness

4.9 The Committee considered the manufacturer's economic model and the ERG's critique of this model. The Committee was aware that the manufacturer had based the economic evaluation on the subgroup of patients who had previously received a taxane rather than the Committee's preferred choice of the whole capecitabine cohort (the ITT population). The Committee noted the manufacturer's and ERG's statements that an analysis of the ITT population would result in a larger ICER than the £77,000 per QALY gained for the subgroup included in the manufacturer's base-case analysis. The Committee agreed with this assessment. The Committee considered the ERG critique and explorations of the manufacturer's model to determine the most plausible ICER for the subgroup previously treated with a taxane. The aim was to establish a benchmark for the incremental cost per QALY gained for the ITT population.

4.10 The Committee noted the explorations made by the ERG to the economic model:

4.11 The Committee noted the ERG's concerns around the rank preserving structural failure time method used by the manufacturer to account for the effect of crossover to open-label bevacizumab in the modelling of survival in the progressed disease state. The Committee discussed the ways in which the analyses were adjusted for crossover by the manufacturer and the ERG. The Committee was unclear as to whether these were appropriate (or whether any other method would be appropriate) or would introduce potential bias because about half the patients in both arms of the trial crossed over to have open-label bevacizumab after disease progression. The Committee also noted that the subsequent treatments received had not been modelled, which in combination with the impact of crossover, could have led to confounding of the overall survival results. The ERG confirmed that it had not been possible to estimate the effect of these factors on overall survival. The Committee noted that 60% of the manufacturer's base-case QALY gain was from the progressed disease phase and it was unsure of the impact of crossover on this finding. The Committee concluded that given these uncertainties, the manufacturer's modelled overall survival results could not be considered robust.

4.12 The Committee noted that in the manufacturer's model, the costs of administration and pharmacy time from the second cycle onwards was £255 in the capecitabine alone arm and £206 in the bevacizumab plus capecitabine arm. The Committee discussed that it was unexpected that the costs associated with bevacizumab plus capecitabine would be lower than the costs for capecitabine alone. The Committee accepted that although these costs were based on NHS reference costs, it would have been possible to generate more plausible values. The Committee concluded that despite the incorporation of NHS tariffs, the discrepancy in the costs of administration and pharmacy time contributed to the uncertainty associated with the results of the manufacturer's economic model.

4.13 The Committee noted that the ERG had carried out an exploratory analysis of the progressed disease trial data to explore survival during this phase, assuming equal survival in each arm of the model (common projection scenario) as well as assuming different survival in each arm of the model depending on what the first treatment was (different projections scenario). The Committee noted that this analysis, in combination with the rest of the ERG changes, resulted in an ICER of £182,000 per QALY gained for the common projection model and £98,000 per QALY gained for the different projections model for bevacizumab plus capecitabine compared with capecitabine alone. In addition, the Committee was also aware that a disutility for adverse events had not been applied in the manufacturer's model, despite utility estimates being available in the literature to account for adverse events, and it was likely that this could have resulted in underestimated ICERs. The Committee concluded that given all of the uncertainties, it was not possible to determine the most plausible ICER for bevacizumab plus capecitabine compared with capecitabine alone for the subgroup of patients who were previously treated with a taxane. However, it was convinced that the ICER would be higher than the most optimistic ICER of £82,000 per QALY gained resulting from the ERG explorations. The Committee considered that the ICER for bevacizumab plus capecitabine compared with capecitabine alone in the ITT population would be even higher (see section 4.9). The Committee noted the comments received during consultation, but considered that there was no evidence to alter its conclusion that the ICER for bevacizumab plus capecitabine compared with capecitabine alone would be higher than £82,000 per QALY gained. The Committee concluded that given the lack of robust evidence of survival benefit supplemented by the high ICER, bevacizumab plus capecitabine as a first-line treatment for metastatic breast cancer was not a cost-effective use of NHS resources.

4.14 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:

4.15 The Committee discussed whether bevacizumab plus capecitabine for the first-line treatment of metastatic breast cancer fulfilled the criteria for a life-extending, end-of-life treatment. The Committee 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. The Committee concluded on this basis that bevacizumab plus capecitabine did not fulfil the criteria for being a life-extending, end-of-life treatment.

4.16 The Committee recognised the novel mode of action of bevacizumab, which may benefit breast cancer patients whose treatment options are limited. However, it considered that there were no additional gains in health-related quality of life over those already included in the QALY calculations. The Committee therefore concluded that the innovative aspects of bevacizumab were already incorporated in the economic model and it did not alter its decision on the cost effectiveness of bevacizumab in combination with capecitabine.

Summary of Appraisal Committee's key conclusions