Atezolizumab for untreated PD-L1-positive advanced urothelial cancer when cisplatin is unsuitable

3.1 Urothelial cancer causes a number of symptoms, including haematuria (blood in the urine), and increased frequency, urgency and pain associated with urination. Surgical treatments such as urostomy can have a substantial impact on quality of life and restrict daily activities. The patient experts explained that chemotherapy is associated with unpleasant side effects such as nausea and has limited effectiveness. In the original appraisal the committee recognised that many people with locally advanced or metastatic urothelial cancer are older and may have comorbidities, which can affect treatment decisions. The committee concluded that locally advanced or metastatic urothelial cancer has a significant impact on quality of life.

3.2 Initial treatment is usually with a cisplatin-containing chemotherapy regimen. However, cisplatin can be damaging to the kidneys, so is not suitable for some people with impaired kidney function or a poor performance status. People for whom cisplatin is unsuitable will usually be offered carboplatin plus gemcitabine. If they are not well enough to tolerate this or they choose not to have it, they will be offered best supportive care. Patient and clinical experts explained that none of the current treatments offer lasting benefit, and the prognosis is poor. The patient experts explained that the disease has a substantial impact on the ability to work and travel, and that the side effects of chemotherapy can have a major negative impact on quality of life. In the committee meeting, the patient experts noted that there have been no new treatments recommended for locally advanced or metastatic urothelial cancer for a number of years. The patient experts also noted that pembrolizumab is no longer recommended for untreated PD‑L1‑positive locally advanced or metastatic urothelial cancer when cisplatin is unsuitable, and so there are limited treatment options for this disease. The committee concluded that there is an unmet need for effective treatment options for people with locally advanced or metastatic urothelial cancer.

3.3 The company submitted clinical- and cost-effectiveness analyses comparing atezolizumab with carboplatin plus gemcitabine. Although it was included in the NICE scope, the company did not submit a comparison with best supportive care. It considered that there was not enough data for comparison with best supportive care. In the original appraisal, the committee heard that, in clinical practice, carboplatin plus gemcitabine may not be suitable for a significant proportion of people for whom cisplatin is unsuitable, and this group of people therefore have best supportive care. It understood that because atezolizumab is an immunotherapy with a different side effect profile to carboplatin plus gemcitabine, there may be some people for whom atezolizumab is suitable who would otherwise have best supportive care. For this review, the committee noted that collecting best supportive care data was not part of the managed access agreement. It also acknowledged that there was a lack of evidence in the literature for people who may have best supportive care in the population for whom cisplatin is unsuitable and who have high levels of PD‑L1. The committee concluded that best supportive care was an appropriate comparator for the population with untreated disease for whom cisplatin is unsuitable, but as in the original appraisal, acknowledged the lack of data would make a comparison difficult.

3.4 In the original appraisal the committee concluded that most people with untreated locally advanced or metastatic urothelial cancer would stop treatment with atezolizumab when their disease progresses and a 2‑year stopping rule was not appropriate. The committee did not hear any relevant new evidence and so concluded that its view from the original appraisal had not changed.

3.5 In the original appraisal the clinical-effectiveness evidence for atezolizumab came from IMvigor 210, a phase 2 single-arm study. The company did a simulated treatment comparison and network meta-analysis to compare atezolizumab with gemcitabine plus carboplatin. The committee in the original appraisal was concerned that the simulated treatment comparison was not robust and had concerns about the reliability and robustness of the network meta-analysis. To address the committee's concerns from the original appraisal, the company provided the clinical-effectiveness evidence for atezolizumab from IMvigor 130, a phase 3 randomised controlled trial. The trial included 1,213 adults with previously untreated locally advanced or metastatic urothelial cancer, who were in the investigators' judgement eligible to receive platinum-based therapy. Only a subgroup of 93 people, which included people with untreated PD‑L1‑positive (tumour expression of 5% or more) locally advanced or metastatic urothelial cancer and who were ineligible to have treatment with cisplatin, was relevant to this appraisal. This was the result of the restricted EMA marketing authorisation, which stated atezolizumab should only be used in adults with high levels of PD‑L1. The median overall survival was 18.6 months for atezolizumab and 10.0 months for platinum-based chemotherapy. The stratified hazard ratio was 0.50 (95% confidence interval [CI] 0.29 to 0.87, p=0.0125), showing atezolizumab was associated with a statistically significant improvement in overall survival compared with platinum-based chemotherapy. The median progression-free survival for atezolizumab was 6.4 months compared with 6.0 months for platinum-based chemotherapy. The stratified hazard ratio was 0.56 (95% CI 0.34 to 0.93, p=0.0235), showing atezolizumab was associated with a statistically significant improvement in progression-free survival compared with platinum-based chemotherapy. The ERG noted that there was uncertainty in the treatment effects as they were based on an interim data analysis of a small subgroup of the trial's total population. Also, the hazard ratio confidence intervals were wide (but did not cross 1). The committee recalled that the small sample size was a result of the restricted marketing authorisation for atezolizumab for people with high levels of PD‑L1. The committee also noted that the survival data was 87% mature. The committee concluded that the data from IMvigor 130 was the most appropriate evidence for decision making.

3.6 Within the IMvigor 130 subgroup there were baseline differences between trial arms in terms of gender and racial characteristics. The ERG explained that some of the imbalances were likely to bias treatment effects but that the direction and magnitude of the bias were unclear. The committee noted that the gender and racial characteristics may bias in favour of atezolizumab but the Bajorin risk factor and Eastern Cooperative Oncology Group Performance Status scores may bias in favour of platinum-based chemotherapy. The committee concluded that the small sample size and opposing influences meant it was not possible to determine the magnitude of direction of any potential bias and the differences in baseline characteristics were acceptable.

3.7 NHS England also provided data from the SACT dataset. It was collected from 64 people who had atezolizumab through the Cancer Drugs Fund between July 2018 and August 2020. The clinical expert explained that the clinical experience with atezolizumab is positive. The committee noted that the median overall survival in the trial was longer (18.6 months, 95% CI 14.0 to not evaluable) than in the SACT dataset (12.4 months, 95% CI 8.3 to 20.1). However, the 95% confidence intervals overlapped. Both the clinical expert and the lead for the Cancer Drugs Fund explained that people included in the SACT dataset were older and had a poorer performance status, which may have contributed to the differences in the median survival estimates. The clinical expert also noted that there may have been selection bias of good prognostic features for people enrolled in IMvigor 130, who were eligible to have chemotherapy, compared with the SACT dataset in which people were not necessarily able to have chemotherapy. The lead for the Cancer Drugs Fund explained that the COVID‑19 pandemic may have affected people's choice to continue treatment, even if their disease was responding, which would have impacted the SACT dataset. But the IMvigor 130 trial would likely not have been impacted by the COVID‑19 pandemic. The committee concluded that these population differences and impact of the COVID‑19 pandemic likely contributed to the differences in treatment effects between the IMvigor 130 trial and SACT dataset.

3.8 In the original appraisal the clinical experts explained that, in their experience of using atezolizumab, it is well tolerated and associated with fewer adverse events than chemotherapy. However, the committee in the original appraisal also understood that atezolizumab was associated with some unpleasant and potentially serious adverse events. A clinical expert in this review stated that toxicities are well managed by specialist hospitals in collaboration with other specialities. The committee concluded that atezolizumab is reasonably well tolerated in clinical practice.

3.9 The company used data from the subgroup of people with untreated PD‑L1‑positive locally advanced or metastatic urothelial cancer, who were ineligible to have treatment with cisplatin. It used the Kaplan–Meier overall survival curve from the clinical trial until 20% of people were at risk and extrapolated the tail using the exponential distribution. The ERG proposed an approach in which it used an exponential parametric function over the whole time period. The ERG explained that it proposed this approach as there was considerable uncertainty with the small sample size in the cisplatin-ineligible, PD‑L1‑positive subgroup from IMvigor 130. This approach resulted in a 5‑year survival of 16%. The clinical expert estimated that it was plausible that between 5% and 30% of people having atezolizumab would be alive at 5 years. In its response to technical engagement, the company updated its approach to use the Kaplan–Meier curve to model the early part of the overall survival curve until 24 (48%) people were at risk, to maintain consistency with the original appraisal. This resulted in a 5‑year survival estimate of 15%. The committee concluded that the company and ERG's approaches were both clinically plausible and estimated similar 5‑year survival estimates and cost-effectiveness results.

3.10 Treatment in the platinum-based chemotherapy arm was restricted to 6 cycles, so curve selection had a negligible impact on costs, and the discussion around time to treatment discontinuation curve selection focused on the atezolizumab arm. The company used the Kaplan–Meier time to treatment discontinuation curve for atezolizumab from the clinical trial, until 48% of people were at risk, and extrapolated the tail using the exponential distribution. This resulted in a 5‑year time to treatment discontinuation estimate of 1% for atezolizumab. The clinical experts estimated 5‑year time to treatment discontinuation for atezolizumab would be between 0% and 2%. The committee heard that this approach was clinically plausible, provided a good statistical fit to the data, and was a conservative extrapolation that aligned with the SACT data. The ERG noted that this approach assumed a constant probability of treatment discontinuation over time, whereas this probability decreased in the trial. The ERG explained that using this approach overestimated the probability of treatment discontinuation and underestimated the costs. The ERG preferred to use the Weibull parametric function over the whole time period, resulting in a 5‑year time to treatment discontinuation estimate of 7%. The committee noted that using the Weibull parametric function to extrapolate progression-free survival, the approach proposed by the ERG and accepted by the company in its response to technical engagement, resulted in 5‑year progression-free survival estimates of 5%. Using the ERG's approach to extrapolate time to treatment discontinuation would result in more people having treatment than were progression free, which the committee considered to be implausible. The committee heard that the SACT dataset showed that no one was on treatment after 2 years, but the lead for the Cancer Drugs Fund cautioned that the maximum follow up for the SACT dataset was 25 months. The committee recalled that there is no stopping rule for atezolizumab (see section 3.4). The committee concluded that the most plausible estimate of time to treatment discontinuation would be less than 5% of people having treatment at 5 years to align with the 5‑year progression-free survival estimate and clinical expectations.

3.11 In the original appraisal, no health-related quality-of-life data was collected in IMvigor 210 and the company used alternative utility values. The committee was concerned that the utility value of 0.71 used for the progressed-disease state was too high and concluded that the most plausible value for post-progression utility was likely to be between 0.5 and 0.71. The company submitted new health-state utility values for the atezolizumab and platinum-based chemotherapy arms, based on the IMvigor 130 trial. The company provided naive utility estimates that did not consider the longitudinal nature of the data. To overcome this, the company used a mixed-effects model that accounted for changes in utility over time as well as correlation among observations within people in the trial to estimate the base-case utilities. In its response to technical engagement, the company updated the selection of the mixed-effects model used to estimate the utility values and explained the updated approach was more robust. The ERG agreed that the updated values were more appropriate and clinically plausible. The company's updated base-case utilities were lower than the naive values also presented by the company and led to higher cost-effectiveness estimates. The committee concluded that the updated base-case utilities were acceptable.

3.12 The committee considered the advice about life-extending treatments for people with a short life expectancy in NICE's guide to the methods of technology appraisal. In the original appraisal, the data showed that life expectancy for people with urothelial cancer was less than 24 months for people having treatment with any standard care. The new evidence from IMvigor 130 also showed that mean overall survival was less than 24 months for people having treatment with UK standard care. The committee concluded that atezolizumab met the short life expectancy criterion.

3.13 The company's economic model also predicted that atezolizumab extended life by at least 3 months, but the exact results are confidential and cannot be reported here. The committee concluded that atezolizumab would extend life by more than 3 months, and therefore met the end of life criteria.

3.14 NICE's guide to the methods of technology appraisal notes that judgements about the acceptability of a technology as an effective use of NHS resources will take into account the degree of uncertainty around the incremental cost-effectiveness ratios (ICERs). It states that the committee will be more cautious about recommending a technology if it is less certain about the ICERs presented. The committee noted the level of uncertainty, specifically around the size of the relevant population from IMvigor 130 and the time to treatment discontinuation (see section 3.5 and section 3.10). But it acknowledged that the small sample size was a result of the restricted EMA marketing authorisation and concluded that the most plausible estimate of time to treatment discontinuation would be less than the 5‑year progression-free survival estimate and clinical expectations. For a life-extending treatment at the end of life, the upper limit of the range usually considered to represent a cost-effective use of NHS resources is £50,000 per quality-adjusted life year (QALY) gained. The committee noted that the company's base-case ICER for atezolizumab, including a patient access scheme and corrected by the ERG who found a minor difference in the results estimated by the company, was £32,235 per QALY gained. The committee's preferred assumptions for decision making at the appraisal committee meeting were to use:

3.15 A patient expert questioned whether there is an equality issue regarding gender. The clinical expert confirmed that women tend to present later so are more likely to have advanced disease and worse cancer-specific mortality. The committee heard from the Cancer Drugs Fund clinical lead that data collected by Public Health England from NHS patients in England showed that more men had taken atezolizumab while it was available in the Cancer Drugs Fund. The committee concluded that its recommendation applies equally, regardless of gender, so this difference is not in itself an equality issue.