Avelumab with axitinib for untreated advanced renal cell carcinoma

Most types of renal cell carcinoma (RCC) are histologically categorised as clear cell. Other RCC subtypes are heterogeneous and include papillary, chromophobe and collecting duct. Collectively, this group is called non-clear cell. The patient expert submission said that some subtypes of RCC (including non-clear cell) may not respond well to treatments and have a poor prognosis. The clinical experts explained that survival and quality-of-life outcomes for people with non-clear-cell RCC are worse than with clear-cell RCC.

From the original NICE technology appraisal guidance on avelumab with axitinib for untreated advanced RCC (from here, TA645), the committee recalled that first-line advanced RCC treatment options were defined by the International Metastatic RCC Database Consortium (IMDC) risk groups (favourable, intermediate or poor). In the updated submission for the managed access review (from here, 'this review'), the clinical experts explained that the IMDC has limitations. They said that it does not capture tumour biology and is not a good tool for predicting tumour progression. They noted that the favourable-risk group is heterogeneous. For example, the group can include people with brain or bone metastases who may have worse clinical outcomes than others in the same group. The NHS England Cancer Drugs Fund clinical lead (from here, the Cancer Drugs Fund lead) explained that NICE had to continue to use the IMDC risk group classification. This was because the marketing authorisations and technology appraisal guidance recommendations for other advanced RCC treatments were based on these groups. The committee acknowledged the concerns from the clinical experts and concluded that the IMDC classification system would need to be used for this review. The committee asked the clinical experts how the IMDC risk status corresponds with clear-cell and non-clear-cell RCC histology. They explained that clear and non-clear-cell RCC can be split according to IMDC risk groups. But non-clear-cell RCC tends not to be classified as being favourable risk.

In TA645, the company positioned avelumab plus axitinib as a first-line treatment option for both favourable-risk and intermediate- or poor-risk groups (see section 3.2). In this review, the company's proposed positioning was for untreated advanced RCC in the favourable-risk group only. The company explained that additional treatment options, including 3 immunotherapies, have become routinely available on the NHS for the intermediate- or poor-risk group since TA645 was published. It added that there is a high unmet need in the favourable-risk group, for whom there are no immunotherapies.

The Cancer Drugs Fund lead said that the company's positioning reflected most of the current use of avelumab plus axitinib in the NHS. They noted that, for the intermediate- or poor-risk group, very few people have avelumab plus axitinib. The clinical experts agreed and said that avelumab plus axitinib is primarily used in the favourable-risk group. This is because it is more clinically effective for this risk group than for the intermediate- and poor-risk group. The committee noted this and also acknowledged that the marketing authorisation for avelumab plus axitinib does not restrict its use based on clear-cell and non-clear-cell histology. It recalled that past NICE technology appraisals in advanced RCC did not differentiate the treatment pathway between clear-cell and other types of RCC. The committee concluded that any recommendations from this evaluation would only apply for the favourable-risk group, which would include both the clear-cell and non-clear-cell RCC subtypes.

In TA645, the committee accepted that the first-line treatment options for the favourable-risk group included pazopanib, sunitinib and tivozanib. These are all tyrosine kinase inhibitors (TKIs). For this review, the company explained there are no UK‑specific RCC guidelines. UK real-world evidence from McGrane et al. (2024) showed that sunitinib is the most commonly used TKI for the favourable-risk group, followed by pazopanib and tivozanib. The EAG said, for the favourable-risk group, all comparators have been included, as in the NICE scope. The clinical expert submissions said that, for the favourable-risk group, treatment options include surveillance, a TKI alone or combination immunotherapy.

At the meeting, 1 clinical expert said that sunitinib is the main TKI used in the UK. Large RCC trials also used sunitinib as the main comparator arm, so most of the clinical-effectiveness data is for this treatment. Tivozanib is used in people who are frailer. Another clinical expert thought that tivozanib, rather than sunitinib, is the preferred TKI at some centres. This was because it is considered equally efficacious but with a better tolerability profile. Both clinical experts agreed that pazopanib is less commonly used in the NHS. They also agreed that the treatment options for clear and non-clear-cell RCC are similar. The committee concluded that sunitinib and tivozanib were the main comparators for this review for the favourable-risk group across both clear and non-clear RCC subtypes.

The patient expert submission said that there is high unmet need for first-line treatments for advanced RCC. The clinical experts at the committee meeting agreed. They said that axitinib with avelumab is the only first-line immunotherapy for the favourable-risk group with a TKI combination, which has been available to the NHS through the Cancer Drugs Fund. They explained that the priority is to stop tumour progression. They further explained that immunotherapies (in this case avelumab) provide the best chance for a durable response for both the clear and non-clear-cell subtypes of RCC. They flagged that non-clear-cell RCC subtypes are more complex to treat (see section 3.1). They explained that combined immunotherapy and TKI mechanisms of action have shown better clinical outcomes than single-agent TKIs. One clinical expert highlighted the McGrane et al. (2024) UK real-world evidence study that showed, after taking account of people still on treatment, that:

• 32.4% of the favourable-risk group did not get second-line treatment

• 51.5% of people who died did not get immunotherapy.
  
  This meant that, if first-line axitinib with avelumab were not available, most people would miss out on the opportunity to get an immunotherapy that could improve their quality of life. The clinical experts thought that not having an immunotherapy and TKI combination as a first-line treatment option would further disadvantage the favourable-risk group. The committee understood the unmet need for this risk group and took this into consideration for its decision making.

JAVELIN Renal 101 is a phase 3 randomised controlled trial of avelumab plus axitinib (442 people) compared with sunitinib (444 people) in advanced RCC. In TA645, the committee considered the clinical evidence from the all-risk overall population. For this review, the company provided data from the final analysis of JAVELIN Renal 101. Median follow up was 73.2 months in the avelumab plus axitinib arm and 73.0 months in the sunitinib arm. The clinical data for this review was for the favourable-risk group. Median progression-free survival (PFS) was 20.7 months in the avelumab plus axitinib arm and 13.8 months in the sunitinib arm (stratified hazard ratio [HR] 0.75; 95% confidence interval [CI] 0.53 to 1.07; p=0.1109). Median overall survival (OS) was 79.4 months in the avelumab plus axitinib arm and 65.5 months in the sunitinib arm (stratified HR 0.73; 95% CI 0.48 to 1.10; p=0.1290).

The EAG said that the OS data was mature. It also said that the median survival time had been reached for the overall population and for both the favourable-risk, and intermediate- or poor-risk groups. The committee acknowledged this and noted that the trial was not powered to determine statistical significance in the IMDC risk groups. It further noted that, in the all-risk overall population, statistical significance was reached for difference in PFS but not for OS. The committee noted that the OS hazard ratio indicated a survival benefit but with a wide confidence interval that included 1. It also concluded that avelumab plus axitinib arm had been shown to improve PFS compared with sunitinib.

In TA645, the committee noted that data collection through the Systemic Anti-Cancer Therapy (SACT) dataset could be used to collect evidence on clinical outcomes for avelumab plus axitinib. For this review, the company presented real-world SACT data for avelumab plus axitinib. Sensitivity analyses was done for OS, in which outcomes were compared across IMDC defined risk groups and RCC histology (clear cell and non-clear cell; see section 3.1). For the overall Cancer Drugs Fund cohort, median follow up was 20.5 months and median OS was 33.9 months. In the favourable-risk group, median OS was not reached. The median OS was 36.4 months for clear-cell RCC and 15.6 months for non-clear-cell RCC. The SACT dataset did not provide data on comparators, and RCC histology data was not split by IMDC risk groups. The committee understood that the SACT data could not inform the treatment-effectiveness outcomes for avelumab plus axitinib compared with the comparators for the favourable-risk group (including the non-clear-cell RCC subtype). But it concluded that the SACT data did provide further evidence on clinical outcomes and real-world evidence relevant to UK clinical practice for avelumab plus axitinib.

In TA645, the committee concluded that avelumab plus axitinib prolonged PFS compared with sunitinib. But it added that the immature trial data meant that there was uncertainty about whether avelumab plus axitinib prolonged OS compared with sunitinib and, if so, by how much. In this review, the company submitted the longer-term follow-up data from JAVELIN Renal 101. The company said it provided the longest follow up for an immunotherapy with TKI combination treatment from an RCC trial. It noted that, for the favourable-risk group, median OS was 14 months longer in the avelumab plus axitinib arm than the sunitinib arm. Also, there was a clear separation between the OS curves from about 30 months, which continued to the end of follow up.

The clinical experts explained that the JAVELIN Renal 101 favourable-risk data showed that the combined partial and complete cancer response rates (objective response) to treatment were 30 percentage points higher for avelumab plus axitinib (n=71; 75.5%; 95% CI 65.5 to 83.8) compared with sunitinib (n=44; 45.8%; 95% CI 35.6 to 56.3). One expert commented that the response rate was higher than with any other immunotherapy with TKI combination in advanced RCC. The committee asked the clinical experts why the benefit of response rates did not translate to statistical significance for OS outcomes. They explained that statistical significance in the trial was related to the sample size and the statistical power of the analysis. One clinical expert referenced McGrane et al. (2024). For the favourable-risk group, this study concluded that TKIs (50.5% sunitinib, 31.6% pazopanib, 15% tivozanib) were associated with shorter median PFS and OS compared with immunotherapy TKI combination therapies (95.5% avelumab plus axitinib). The hazard ratio was 0.42 (95% CI 0.18 to 0.99). This showed a statistically significant survival benefit of avelumab plus axitinib compared with other TKIs in a UK real-world setting. The EAG noted in its report that the evidence was not based on randomised and matched comparisons, and had a high risk of bias. Both clinical experts concluded that they were not concerned about the lack of a statistically significant difference in OS from JAVELIN Renal 101. This was because the response rate evidence was compelling. The committee acknowledged that there was uncertainty about the long-term OS benefit of avelumab plus axitinib compared with sunitinib. It concluded that the pivotal trial showed a statistically significantly extended period of PFS for all population subgroups.

For TA645, the committee noted that all participants in JAVELIN Renal 101 had clear-cell disease. But people with non-clear-cell RCC are also seen in NHS practice and SACT data could provide this clinical evidence. In this review (section 3.7), the committee understood that comparative effectiveness data was not available from the SACT dataset. The clinical experts noted that there was limited data for non-clear-cell RCC. In their experience, TKI and immunotherapy with TKI combinations have shown clinical effectiveness in non-clear-cell RCC. They considered that it is possible to generalise relative treatment-effectiveness evidence from the clear-cell RCC population to the non-clear-cell population. But the outcomes for non-clear-cell RCC subtypes are likely worse for all treatments compared with clear-cell RCC because of different prognostic factors. The committee recalled from TA645 that the lack of comparative effectiveness data in non-clear-cell RCC was similar to other first-line treatments for advanced RCC. The committee concluded that, because of a lack of clinical evidence in non-clear-cell RCC, it was uncertain whether avelumab plus axitinib was clinically effective compared with sunitinib for this RCC subtype.

The company presented a partitioned survival model to estimate the cost effectiveness of avelumab plus axitinib. The model included the following health states:

• progression-free (on and off treatment)

• post-progression

• death.
  
  The model structure was accepted by the committee as part of TA645 and was updated with the JAVELIN Renal 101 trial data (see section 3.6). The committee concluded that the model was appropriate for decision making.

For this evaluation, there were no head-to-head trials of avelumab plus axitinib compared with tivozanib or pazopanib. In the model, efficacy data for the comparison with sunitinib came from the JAVELIN Renal 101 trial. The company assumed that tivozanib and pazopanib PFS, time to treatment discontinuation (TTD) and OS outcomes were equivalent to those with sunitinib. The company said that this was in line with TKI equal efficacy assumptions from past NICE technology appraisals. The committee understood that TKIs may have a similar efficacy profile to each other, but tivozanib is better tolerated compared with sunitinib (see section 3.4). The committee recalled from TA645 that the equal efficacy assumption for PFS, TTD and OS outcomes in the economic model had been accepted in NICE's technology appraisal guidance on tivozanib for treating advanced RCC and NICE's technology appraisal guidance on nivolumab with ipilimumab for untreated advanced RCC. The committee concluded that it was reasonable to assume equal efficacy between sunitinib, tivozanib and pazopanib because they all have the same mechanism of action.

The company fitted independent parametric distributions to the JAVELIN Renal 101 OS favourable-risk group data to extrapolate beyond the trial period. The company noted that, based on the hazard plots, the proportional hazards assumption was not met, and independent parametric distributions were needed. The company's preferred choice of base-case parametric distributions was an independent log-normal distribution for avelumab plus axitinib, and an independent generalised-gamma distribution for sunitinib. The distributions were selected based on visual inspection and clinical expert opinion. The EAG noted the statistical goodness-of-fit scores showed similar fits for the log-logistic, Weibull and log-normal distributions for avelumab plus axitinib. The log-logistic, Weibull and log-normal distributions for sunitinib also had similar statistical fits. The EAG agreed with the company's methods and selection of base-case extrapolations for OS. But it said there was uncertainty in the parametric distributions chosen for OS.

The EAG flagged that there were several alternative parametric survival distributions with a good statistical fit to the trial data (see section 3.12) giving different survival projections at 10 years and beyond. At the clarification stage, the EAG noted that the hazard plots provided by the company were on a log-cumulative scale. These plots were unable to show whether and how the OS hazards changed over time. The EAG provided scenarios choosing alternative distributions that had a large impact on the cost-effectiveness estimates. For the base case, the EAG selected the same distributions as the company with the caveat that, to better inform the OS parametric distributions beyond 10 years, it would have preferred:

• more clinical evidence

• independent clinical expert opinion

• individual hazard plots for mortality over time.
  
  The committee asked the clinical experts what proportion of people would be expected to be alive at 10 years after having treatment with a TKI. One clinical expert said that it was difficult to estimate but 10% to 15% for a single-agent TKI in clinical practice could be considered reasonable. The second expert thought under 20% seemed reasonable. The committee noted this and acknowledged that survival outcomes would be different between clinical practice and a clinical trial. The committee questioned why the Weibull distribution was not selected for both arms, because it had the best statistical fit. The clinical expert cautioned against using the same parametric distribution to model survival outcomes for an immunotherapy and a TKI. This was because they would expect the immunotherapy outcomes to improve over time compared with TKIs because of their different mechanisms of action. The committee understood that the log-normal distribution was chosen to take account of an immunotherapy's anticipated long-term durable benefits in some people. The committee said there was no strong justification to choose the same distribution for each arm. It thought that, for the avelumab plus axitinib arm, the generalised-gamma distribution predicted similar results to the log-normal distribution at 5 and 10 years, with slightly less optimistic longer-term survival. It concluded that both these distributions were plausible for avelumab plus axitinib.

To calculate treatment costs, the company used TTD data from JAVELIN Renal 101 for avelumab, axitinib and sunitinib separately. In the trial, TTD was defined as the difference in duration between when the treatment was started and when it was stopped because of progression or side effects. Similar to OS, the company fitted independent parametric distributions to the JAVELIN Renal 101 TTD favourable-risk group data to extrapolate beyond the trial period. The EAG said that data for TTD was mature and all the parametric distributions provided a good fit to the trial data. It noted that cost-effectiveness results were not sensitive to the choice of distributions. The clinical expert explained that there are no stopping rules for avelumab plus axitinib and, if it is well tolerated, its use is continued. The EAG said there may be uncertainty in a real-world setting about how long avelumab plus axitinib is used. The committee noted that the PFS curve was above the treatment discontinuation curve. This meant that there was a difference between the extrapolated PFS and TTD curves for avelumab plus axitinib, especially after 5 years. This suggested that treatment benefit was potentially being modelled for this period without any additional treatment costs for avelumab plus axitinib. The committee concluded there was uncertainty in how treatment costs were modelled using TTD.

NICE's manual on health technology evaluations notes that, above a most plausible incremental cost-effectiveness ratio (ICER) of £20,000 per quality-adjusted life years (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 noted the uncertainty in this evaluation, specifically about:

• the lack of statistically significant OS benefit of avelumab plus axitinib compared with sunitinib (see section 3.6 and section 3.8)

• issues of the generalisability of JAVELIN Renal 101 to NHS practice because of the lack of comparative effectiveness evidence for the non-clear-cell RCC subtype (see section 3.9)

• justification for choice of parametric distributions for long-term survival (see section 3.12)

• potential underestimation of the long-term treatment cost of avelumab plus axitinib (see section 3.14).
  
  But the committee also noted:

• the unmet need in the favourable-risk group (see section 3.5)

• the lack of treatment options with multiple mechanisms of action (such as an immunotherapy and TKI combination) in the NHS for first-line treatment in the favourable-risk group (see section 3.5).
  
  So, the committee concluded that an acceptable ICER would be around the middle of the range NICE considers a cost-effective use of NHS resources (£20,000 to £30,000 per QALY gained).

The committee recalled its preferences for the cost-effectiveness modelling, which were to:

• apply the decision to the favourable-risk group across both clear and non-clear-cell RCC subtypes (see section 3.3)

• model sunitinib, tivozanib and pazopanib as comparators (see section 3.4)

• assume equal efficacy between sunitinib, tivozanib and pazopanib (see section 3.11)

• use the generalised-gamma distribution to model OS for the sunitinib arm (see section 3.13)

• assume that both the generalised-gamma and log-normal distributions were plausible for modelling OS for the avelumab plus axitinib arm (see section 3.13).
  
  The exact cost-effectiveness results cannot be reported here because of confidential discounts for avelumab, axitinib, the comparators and subsequent treatments. Using the company's and EAG's base-case models and applying the committee's preferred assumptions, the probabilistic cost-effectiveness estimates comparing avelumab plus axitinib with sunitinib were within the range that NICE considers an acceptable use of NHS resources.

The committee did not identify any equality issues.

The patient experts explained that the side-effect profile of avelumab plus axitinib was better than that of other available treatment options. The clinical experts noted that avelumab plus axitinib has a shorter 'half-life' than TKIs. This means it takes less time for it to leave the body, which explains the better tolerability profile. The committee took this into consideration but said that this was captured in the economic modelling. So the committee concluded that all additional benefits of avelumab plus axitinib had already been taken into account.

The committee concluded that avelumab plus axitinib is an effective treatment in terms of PFS compared with sunitinib. With the committee's preferred assumptions, the cost-effectiveness estimates were within the range NICE considers an acceptable use of NHS resources. So, avelumab plus axitinib is recommended for adults who have untreated advanced RCC and a favourable-risk status.