Bevacizumab (originator and biosimilars) with fluoropyrimidine-based chemotherapy for metastatic colorectal cancer

mCRC is a tumour arising from the lining of the large intestine (colon and rectum) that has spread beyond the large intestine, most often to the liver, lungs or peritoneum. The aim of treatment for mCRC is to prolong survival and improve quality of life. There are targeted treatment options and immunotherapies for mCRC, which include:

• First line:

  • pembrolizumab (see the NICE technology appraisal guidance on pembrolizumab for untreated mCRC with high microsatellite instability or mismatch repair deficiency)

  • cetuximab and panitumumab (see the NICE technology appraisal guidance on cetuximab and panitumumab for previously untreated mCRC)

  • nivolumab plus ipilimumab (see the NICE technology appraisal guidance on nivolumab plus ipilimumab for untreated unresectable or mCRC with high microsatellite instability or mismatch repair deficiency)

• Second line:

  • encorafenib plus cetuximab (see the NICE technology appraisal guidance on encorafenib plus cetuximab for previously treated BRAF V600E mutation-positive mCRC)

  • nivolumab with ipilimumab (see the NICE technology appraisal guidance on nivolumab with ipilimumab for previously treated mCRC with high microsatellite instability or mismatch repair deficiency)

  • pembrolizumab (see the NICE technology appraisal guidance on pembrolizumab for previously treated endometrial, biliary, colorectal, gastric or small intestine cancer with high microsatellite instability or mismatch repair deficiency)

• Third line and later:

  • trifluridine–tipiracil with bevacizumab (see the NICE technology appraisal guidance on trifluridine–tipiracil with bevacizumab for treating mCRC after 2 systemic treatments; TA1008)

  • trifluridine–tipiracil alone (see the NICE technology appraisal guidance on trifluridine–tipiracil for previously treated mCRC)

  • regorafenib (see the NICE technology appraisal guidance on regorafenib for previously treated mCRC)

  • fruquintinib (see the NICE technology appraisal guidance on fruquintinib for previously treated metastatic colorectal cancer).
    
    The choice of first- or second-line targeted therapy or immunotherapy depends on the presence or absence of 3 molecular markers: BRAF 600, RAS wild type, and microsatellite instability or mismatch repair deficiency. If targeted treatments are not suitable, fluoropyrimidine-based chemotherapy is usually offered as first- or second-line treatment (see NICE's guideline on colorectal cancer). The EAG noted that bevacizumab could be added on to chemotherapy that is currently used as a first- or second-line treatment option.
    
    The EAG's assessment focused on 2 populations:

• people with untreated metastatic carcinoma of the colon or rectum who would be offered fluoropyrimidine-based chemotherapy, and

• people with metastatic carcinoma of the colon or rectum who have previously had fluoropyrimidine-based chemotherapy and would be offered second-line fluoropyrimidine-based chemotherapy.
  
  The committee noted the NICE scope restricted the population (at first and second line) to people for whom targeted treatments or immunotherapy are not suitable, and to whom chemotherapy would otherwise be offered. This did not include everyone who bevacizumab is licensed for (see section 2.1). The clinical experts stated that adding bevacizumab to chemotherapy gave additional clinical benefit. The patient expert noted that bevacizumab has the ability to shrink tumours to make surgery viable for people whose tumours were previously inoperable. They described this as having a significant impact on their overall outcome. The committee concluded that people having chemotherapy at first line, second line or both for mCRC would welcome a new treatment option.

The EAG did not specifically assess the RAS wild type, BRAF V600 mutation, or microsatellite instability or mismatch repair molecular marker subgroups, which have targeted treatment options. The patient experts highlighted the importance of considering the KRAS mutation status in mCRC and noted that bevacizumab has been seen as an advantageous treatment in this subgroup. The committee noted that some people in these groups would be having chemotherapy rather than targeted treatments. In particular:

• KRAS mutation: people would be having chemotherapy at first and second line because there are no NICE-recommended first- or second-line targeted treatments for tumours with this mutation.

• RAS wild type: cetuximab or panitumumab are the only first-line treatment options. Furthermore, the clinical experts explained that people with right-sided tumours may be having chemotherapy and so bevacizumab may be beneficial in the first-line setting. This is because cetuximab and panitumumab do not appear to be beneficial for right-sided tumours.

• BRAF 600 mutation: people may be having first-line chemotherapy because encorafenib plus cetuximab is only recommended as a second-line treatment.

• Microsatellite instability or mismatch repair: there may be people who have chemotherapy after initial immunotherapy, because immunotherapy can only be used once in the treatment pathway. This would be a small number because of the rarity of this subgroup and because most people in this subgroup (approximately 60% to 70%) do not appear to have disease progression after first-line immunotherapy.
  
  The committee acknowledged that the EAG's analysis focused on people without known mutations, who make up most people with mCRC. A clinical expert highlighted that mCRC was not routinely tested for molecular markers in the studies reviewed by the EAG, because they were older. This meant that although not specifically identified, people with these molecular markers may have been included in the trials. The clinical experts noted that prognosis differs across the subgroups based on molecular markers, but they expected a benefit of bevacizumab added on to chemotherapy across all groups. The clinical experts noted that even if data was available, subgroup analyses would be difficult because the small number of people (only 5% to 10% of people with mCRC) have BRAF 600 mutations, microsatellite instability or mismatch repair. The committee concluded that it would appraise bevacizumab as an add on to fluoropyrimidine-based chemotherapy for first- and for second-line treatment of mCRC, when targeted treatments and immunotherapy are not suitable and chemotherapy would otherwise be offered. This evaluation includes people whose tumours have molecular markers. But the committee noted that it would consider the uncertainty because no evidence had been provided for the clinical or cost effectiveness for the subgroups with molecular markers.

The comparators listed in the final NICE scope were:

• FOLFOX (folinic acid, fluorouracil, oxaliplatin)

• FOLFIRI (folinic acid, fluorouracil, irinotecan)

• CAPOX (capecitabine, oxaliplatin)

• capecitabine monotherapy.
  
  The clinical expert explained that in UK clinical practice, both FOLFIRI and FOLFOX are commonly used as first-line treatments, with the CAPOX regimen typically used at second line. They noted the choice of chemotherapy is influenced by a range of individual patient factors and most people would have both oxaliplatin- and irinotecan-based regimens over the course of treatment. The clinical expert emphasised a preference for flexibility in clinical practice and, if recommended, bevacizumab should be able to be used with either oxaliplatin- or irinotecan-based regimens at first and second line. The NHS England clinical lead (from here, NHSE lead) and patient expert supported the need for flexibility in how bevacizumab is used with different chemotherapy regimens. The clinical expert stated that frailer, older people may be more likely to have bevacizumab with a single-agent fluoropyrimidine (such as capecitabine monotherapy) rather than combination regimens because of toxicity.
  
  The EAG did not identify any NICE technology appraisals or clinical studies evaluating bevacizumab plus capecitabine compared with capecitabine alone for first- or for second-line treatment of mCRC. Previous NICE appraisals were informed by the regulatory trials informing the licence for bevacizumab with fluoropyrimidine-based chemotherapy, in which the chemotherapies were FOLFIRI, CAPOX and FOLFOX. So, the EAG could not consider capecitabine monotherapy as a comparator in its assessment. A clinical expert explained that a common strategy in clinical practice is maintenance therapy. This typically involves using first-line treatments such as oxaliplatin for 4 to 6 months, then switching to a maintenance drug such as oral capecitabine to reduce toxicity and improve progression-free survival. The clinical expert noted that combining bevacizumab with oral capecitabine is common outside the UK and is increasingly used in colorectal cancer care because there is a clinical benefit of this strategy. The regulatory trials had not assessed using bevacizumab in addition to maintenance capecitabine. But the marketing authorisation allows bevacizumab to be used with fluoropyrimidine-based chemotherapy and does not specify whether the chemotherapy should be fluoropyrimidine in combination or fluoropyrimidine monotherapy. After the committee meeting, the NHS England national specialty adviser for cancer drugs stated that it was appropriate for the recommendation to cover bevacizumab added on to any fluoropyrimidine-based chemotherapy, including capecitabine, to align with the marketing authorisation for bevacizumab and clinical practice.
  
  The committee concluded that the chemotherapy regimens outlined in the scope were used in clinical practice, and that it would consider the available evidence for all comparators in the first- and second-line settings. The committee acknowledged the lack of evidence from the regulatory trials for the clinical effectiveness of bevacizumab added on to capecitabine. But it noted that the clinical experts stated that bevacizumab has an observed benefit when added to capecitabine, offering flexibility for the choice of fluoropyrimidine-based chemotherapy backbone. The committee also thought that use with capecitabine monotherapy fell within the marketing authorisation for bevacizumab. It therefore agreed it would reach a decision for bevacizumab added on to any of the fluoropyrimidine-based chemotherapies used in clinical practice.

The EAG's main evidence on the clinical effectiveness of bevacizumab plus fluoropyrimidine-based chemotherapy came from different sources. The EAG identified 2 clinical studies related to the first-line treatment of mCRC with bevacizumab plus fluoropyrimidine-based chemotherapy. These were:

• Study NO16966 (the primary source of clinical-effectiveness data for bevacizumab plus FOLFOX or CAPOX, and FOLFOX or CAPOX alone in TA212)

• Study AVF2107g (the primary source of clinical-effectiveness data for bevacizumab plus FOLFIRI, and FOLFIRI alone in TA118).
  
  Study NO16966 was a phase 3, multicentre, multinational, randomised, open-label study. Data was from a 2-by-2 factorial part of the trial assessing the superiority of bevacizumab plus chemotherapy compared with placebo plus chemotherapy. The chemotherapy regimen was either FOLFOX-4 (that is, the FOLFOX regimen given every 2 weeks, with 2 long infusions in the first 48 hours) or CAPOX. In TA212, the committee concluded it was appropriate to assume CAPOX and FOLFOX were clinically equivalent and that data from the CAPOX and FOLFOX arms could be combined, as could data from the bevacizumab plus CAPOX and bevacizumab plus FOLFOX arms. Results from the 2-by-2 part of the trial (secondary pooled analysis) showed that bevacizumab plus chemotherapy (bevacizumab plus CAPOX and bevacizumab plus FOLFOX-4 combined) significantly improved progression-free survival compared with chemotherapy alone (placebo plus CAPOX and placebo plus FOLFOX-4 combined). For the intention-to-treat population, median progression-free survival was 9.4 months in the bevacizumab plus chemotherapy group, and 8.0 months in the chemotherapy group (a difference of 1.4 months). The hazard ratio for remaining free of disease progression was 0.83, and median follow up was 28 months. Bevacizumab plus chemotherapy also improved overall survival compared with chemotherapy alone. For the intention-to-treat population, median overall survival was 21.3 months in the bevacizumab plus chemotherapy group, and 19.9 months in the chemotherapy group (a difference of 1.4 months). The hazard ratio for death was 0.89 at a median follow up of 28 months. Excluding data from people in the trial who had prior adjuvant therapy (the committee's preference in TA212, which was used in the EAG's model for the current evaluation) decreased the hazard ratio for progression-free survival and overall survival.
  
  Study AVF2107g was a multicentre, international, phase 3 randomised controlled trial comparing first-line bevacizumab plus FOLFIRI compared with placebo plus FOLFIRI. Bevacizumab plus FOLFIRI improved progression-free survival compared with placebo plus FOLFIRI. Median progression-free survival was 10.6 months in the bevacizumab plus FOLFIRI group, and 6.2 months in the placebo plus FOLFIRI group (a difference of 4.4 months). The hazard ratio for remaining free of disease progression was 0.54. Median overall survival was 20.3 months for bevacizumab plus FOLFIRI and 15.6 months for placebo plus FOLFIRI (improving median overall survival by 4.7 months). The hazard ratio for death was 0.66. The committee concluded that bevacizumab plus FOLFOX, CAPOX or FOLFIRI offered better overall and progression-free survival than chemotherapy alone.

The EAG identified Study E3200, which was the only clinical evidence informing the effectiveness of bevacizumab plus FOLFOX and FOLFOX alone for second-line treatment of mCRC in TA212. It also identified a meta-analysis by Mocellin et al. (2017). The EAG did not identify any previous NICE technology appraisals or relevant clinical studies for the UK population evaluating bevacizumab plus FOLFIRI compared with FOLFIRI alone for the second-line treatment of mCRC.

Study E3200 was a phase 3, multicentre, 3-arm, randomised, open-label study. It compared bevacizumab plus FOLFOX-4, FOLFOX-4 alone, and bevacizumab alone in adults with advanced or mCRC that had been treated with a fluoropyrimidine and irinotecan, either separately or in combination. The bevacizumab-alone arm was stopped early because of poor efficacy. The primary endpoint was overall survival, with additional determinations of progression-free survival, response, and toxicity. Median progression-free survival was 7.3 months in the bevacizumab plus FOLFOX group, and 4.7 months in the FOLFOX alone group (a difference of 2.6 months). The hazard ratio for remaining free of disease progression was 0.61. Median overall survival was 12.9 months in the bevacizumab plus FOLFOX group, and 10.8 months in the FOLFOX alone group (a difference of 2.1 months). The hazard ratio for death was 0.75. The committee noted that the studies included in the Mocellin et al. (2017) meta-analysis, which included studies differing in fluoropyrimidine chemotherapy regimens, study location and bevacizumab dose, gave similar hazard ratios for progression-free survival and overall survival to those from Study E3200. People in Study E3200 had not previously had bevacizumab, which the EAG noted was a limitation, because the model for second-line treatment did not consider previous treatments (see section 3.9). But the EAG noted that data reported by Masi et al. (2015), a trial included in the Mocellin meta-analysis, indicated that previous bevacizumab use may not influence the efficacy of second-line bevacizumab. The committee concluded that bevacizumab plus FOLFOX offered better overall and progression-free survival than FOLFOX alone.

Studies NO169665 and AVF2107g were the primary sources of first-line clinical-effectiveness evidence for bevacizumab plus FOLFOX or CAPOX compared with FOLFOX or CAPOX alone, and bevacizumab plus FOLFIRI compared with FOLFIRI alone, respectively (see section 3.5). E3200 was the primary source of second-line clinical-effectiveness evidence for bevacizumab plus FOLFOX compared with FOLFOX alone (see section 3.6) The EAG did not have access to individual patient-level data from these studies, so it generated pseudo-individual patient data for progression-free survival and overall survival. The EAG independently fitted standard parametric survival models to the pseudo-individual patient data to extrapolate beyond the period for which there was trial data. To select the most appropriate parametric distributions to model progression-free survival and overall survival, the EAG assessed:

• the statistical fit to the Kaplan–Meier data using the Akaike information criterion and Bayesian information criterion

• the visual fit of fitted survival curves against the trial data

• the clinical plausibility of the survival model predictions (based on hazard plots and input from clinical experts). Clinical expert advice to the EAG was that the hazards should be consistently increasing for both progression-free survival and overall survival. So, the EAG prioritised models with consistently increasing hazard trends in its base case and explored alternative models in scenario analyses.
  
  The same distributions were applied across treatment arms. The EAG also preferred to use the same distribution for both progression-free survival and overall survival within each pairwise comparison. The EAG selected scenarios of alternative parametric distributions that were plausible and provided the widest possible range of plausible incremental cost-effectiveness ratios (ICERs). These were:

• for first-line treatment:

  • the gamma distribution for progression-free survival and overall survival in models 1 and 3 (with the log-logistic used in scenario analyses)

  • the Weibull distribution for progression-free and overall survival in model 2 (with generalised gamma in scenario analyses)

• for second-line treatment: the gamma distribution for progression-free survival and overall survival in models 1 and 3 (with the log-logistic used in scenario analyses).
  
  The choice of distribution had a moderate impact on the ICER. The committee noted that, in principle, it would prefer the distributions for progression-free survival and overall survival not to be restricted to using the same distribution. But after reviewing the distributions used, it noted that using different distributions for progression-free survival and overall survival made very little difference to the results. The committee concluded it agreed with the EAG's base-case parametric distributions for extrapolating progression-free survival and overall survival.

A range of confidential prices for bevacizumab originator and biosimilars is listed in the Medicines Procurement and Supply Chain (MPSC; see section 4.4.4 in NICE's health technology evaluations manual). NICE originally advised the EAG to use the unweighted mean MPSC price of bevacizumab, and the EAG was not provided with data on market share for the originator and for each biosimilar. The EAG's report noted that because of the range of prices for products in the MPSC, it would prefer to use a weighted mean based on market share to reflect usage in the NHS. After further consideration, NICE asked the EAG to use the mean MPSC price weighted by market share (based on bevacizumab usage across all indications) in its analyses. For this pilot approach, NICE asked the committee to use the weighted mean price in its decision making. The committee noted that regional variation in procurement practices may lead to differences in access to bevacizumab biosimilars across the NHS, and the cheapest product may not always be selected. The committee noted that although mean biosimilar prices would be expected to remain stable or reduce over time, this was uncertain. The prices of bevacizumab products in the MPSC are confidential so the weighted and unweighted mean prices cannot be reported here. The committee concluded that it would use the weighted mean price of bevacizumab products in the MPSC as indicative of the cost of bevacizumab, and take into account the uncertainty about this estimate in its decision making.

The NHSE lead asked whether the EAG had incorporated vial sharing in its analyses. They noted that colorectal cancer is common, and bevacizumab is also used in other tumour types (such as hepatocellular carcinoma and gynaecological cancers). So if bevacizumab were recommended, there would be significant opportunity for vial sharing across the NHS. The clinical expert agreed and highlighted that vial sharing is standard practice in many cancers and is done routinely by hospital pharmacies. The EAG explained that vial sharing was explored in a scenario analysis, which showed a minimal reduction in the ICER, but it was not included in its base case. The committee acknowledged that vial sharing is standard practice and is likely to be feasible for bevacizumab. It acknowledged that the EAG had explored vial sharing in a scenario analysis and concluded the impact was minimal.

The EAG considered utility values reported from previous NICE technology appraisals for first-line treatment (TA118 and TA212) and second-line treatment (TA1008) to select its preferred estimates for this evaluation. It also included disutilities related to adverse events, which had not been considered in these appraisals (see section 3.14). For the first-line setting, the EAG followed the external review group (ERG)'s model assumptions from TA212 by assuming a utility value of 0.77 for the progression-free health state and 0.68 for the progressed-disease health state in its base case. For the second-line setting, the EAG followed the committee's preferences from TA1008 by assuming a utility value of 0.73 for the progression-free health state and 0.64 for the progressed-disease health state in its base case. All health-state utility values were adjusted for ageing using Hernandez Alava et al. 2022 (see section 4.3.16 in NICE's health technology evaluations manual). The EAG explored uncertainty in the utility values in the first- and second-line settings by arbitrarily increasing and decreasing them by 5% in scenario analyses.

The committee noted that the utility values in the previous appraisals were not based on data collected in the bevacizumab trials. It also noted that the utility values used in the second-line setting were similar to those used in the first-line setting. The clinical expert highlighted that people considered for second-line treatment are generally similar to those having first-line treatment and therefore considered it reasonable for the utility values to be comparable. The committee acknowledged that the utility values were sourced from previous technology appraisals. It would have preferred to use trial-based data but concluded that the EAG's approach was pragmatic given the limitations in the available evidence. The committee noted that the utility values chosen by the EAG for the first- and second-line settings were similar, but concluded this was reasonable.

The EAG applied adverse event-related disutilities in the model to estimate the reduction in health-related quality of life for the duration of the adverse events. In the first-line setting, adverse event frequencies for bevacizumab plus FOLFOX or CAPOX and FOLFOX or CAPOX alone were based on Study NO16966. Adverse event data for bevacizumab plus FOLFIRI and FOLFIRI alone came from Study AVF2107g. In the second-line setting, the frequencies of adverse events were taken from Study E3200. The EAG applied a one-off disutility for adverse events by multiplying the estimated disutility of each adverse event by its respective duration and frequency. It assumed that each adverse event lasted for 14 days, based on clinical opinion. The EAG did a targeted literature search for disutility values, and the values used in the model were taken from the literature, previous NICE technology appraisals and assumptions relating to proxy conditions. The QALY losses for all adverse events were added to estimate a one-off QALY loss which was applied in the first model cycle. The committee noted that the previous technology appraisals did not include disutilities for adverse events. The patient expert highlighted that adverse events are a routine part of chemotherapy and are managed through a joint approach between the healthcare professionals and the patient. They explained that management is about balancing factors such as dose adjustments and treatment breaks. They emphasised that adverse events should be taken into consideration, noting that bevacizumab has its own adverse event profile. The EAG highlighted that it explored increasing adverse event disutilities by 10 times in scenario analyses, which had a minimal impact on the ICER. The committee acknowledged the views of the patient and clinical experts. The committee concluded that adverse event disutilities should be included in the model.

Unit costs and resource use for each treatment were based on regimens from Study NO16966 (first- and second-line) and Study AVF2107g (first-line). The EAG highlighted that current practice uses modified de Gramont regimens (for FOLFIRI-containing regimens only) and stops oxaliplatin after 6 months because of toxicity (FOLFOX and CAPOX regimens). These adaptations were explored in sensitivity analyses, assuming the same efficacy as in the base case. The EAG noted the results from these may better represent current clinical practice. Results from the scenario analyses showed that stopping oxaliplatin at 6 months had minimal impact on the ICERs for bevacizumab plus FOLFOX or CAPOX, and using a modified de Gramont regimen for FOLFIRI-containing regimens had a moderate impact. The committee acknowledged that the trial regimens informing the economic model were based on older protocols that differ from those currently used in NHS practice. It noted that the second-line trial used a double dose of bevacizumab, and using the FOLFIRI regimen (used in the NHS) reduced costs. It concluded that it was appropriate to use dosing regimens used in the NHS to estimate costs rather than those used in the trials.

NICE's health technology evaluations manual notes that above a most plausible ICER of £20,000 per QALY gained, judgements about the acceptability of a technology as an effective use of NHS resources will take into account the degree of certainty around the ICER. The committee will be more cautious about recommending a technology if it is less certain about the ICERs presented. But it will also take into account other aspects including uncaptured health benefits. The committee acknowledged that there is an unmet need for people with mCRC. The committee noted the uncertainties in the costs of bevacizumab and the economic model assumptions, and those that came with the EAG's pragmatic approach. Because of these uncertainties, the committee concluded that an acceptable ICER would be towards the lower end of the range NICE considers a cost-effective use of NHS resources (£20,000 to £30,000 per QALY gained).

The exact cost-effectiveness estimates cannot be reported here because of confidential discounts for bevacizumab. For the first-line setting, the EAG's base-case ICERs were within the range that NICE considers an acceptable use of NHS resources for the models in which bevacizumab was an add on to FOLFOX, FOLFIRI or CAPOX. For the second-line setting, with a 1.2 weighting applied for severity, the EAG's base-case ICERs were also within the range that NICE considers an acceptable use of NHS resources, for the models in which bevacizumab was an add on to FOLFOX or CAPOX. The committee accepted the EAG's base-case assumptions. But it noted that scenarios reflecting the dosing regimen of chemotherapy used in the NHS should be included in the base case. In the absence of a specific model for bevacizumab plus FOLFIRI in the second-line setting, the committee thought it appropriate to use the ICERs from the FOLFOX and CAPOX models to indicate the cost effectiveness of bevacizumab with chemotherapy, regardless of the chemotherapy regimen. The committee could not determine the cost effectiveness of bevacizumab with capecitabine, because no evidence was provided. But the committee concluded that it was reasonable not to restrict the recommendation to bevacizumab being added to combination fluoropyrimidine-based chemotherapy. This was because the clinical experts had stated that bevacizumab would have a clinical benefit being added on to any fluoropyrimidine-based chemotherapy, and there was a need for flexibility in the backbone chemotherapy offered in clinical practice.

The committee concluded that bevacizumab plus fluoropyrimidine-based chemotherapy improves progression-free and overall survival in mCRC compared with fluoropyrimidine-based chemotherapy alone. The most likely cost-effectiveness estimates for bevacizumab plus fluoropyrimidine-based chemotherapy are below what NICE considers a cost-effective use of NHS resources. So, bevacizumab (originator and biosimilars) plus fluoropyrimidine-based chemotherapy can be used in the NHS as first- and second-line treatment when targeted treatments or immunotherapy are unsuitable, and chemotherapy would otherwise be offered.