Osimertinib for adjuvant treatment of EGFR mutation-positive non-small-cell lung cancer after complete tumour resection

3.1 Surgical removal of tumours is the preferred treatment for many people with early‑stage non‑small‑cell lung cancer (NSCLC) because it is potentially a cure. But despite the curative intent of complete resection, the disease recurs within about 5 years of surgery in 45% of patients with stage 1b, 62% with stage 2, and 76% with stage 3 disease. In the UK, around 13% of people with stage 1b NSCLC and up to about 50% of people with stage 3a NSCLC have adjuvant chemotherapy after resection. Because it provides only a relatively small benefit in overall and disease‑free survival compared with no chemotherapy over 5 years, many people decline adjuvant therapy. The patient experts explained that people with fully resected stage 1b to 3a epidermal growth factor receptor (EGFR) mutation‑positive NSCLC would welcome new effective adjuvant treatments that reduce the risk of recurrence. They highlighted that people with EGFR mutation‑positive NSCLC tended to be younger than people with other types of NSCLC, so having a treatment option that delays or prevents recurrence or central nervous system (CNS) metastases is important. Being disease free allows people to spend more time working or with their families. Patient experts stated that osimertinib is also better tolerated than other tyrosine kinase inhibitors (TKIs). The committee concluded that patients and their families would welcome new, effective treatments that reduce the risk of recurrence.

3.2 The only treatment currently available in England as adjuvant therapy for NSCLC (including for EGFR mutations) after complete resection is adjuvant chemotherapy, which provides a small benefit in overall survival. Treatment options for people with resectable EGFR mutation‑positive NSCLC are therefore only those that are generally available and are non‑targeted. The clinical and patient experts explained that osimertinib is well tolerated with manageable side effects. The patient experts explained that having an oral option would be welcomed because it would not require a visit to hospital. The committee acknowledged that positioning osimertinib as an adjuvant treatment may address an unmet need for people who have had a resection. It concluded that osimertinib is an oral treatment in a new place in the pathway.

3.3 The company assumed that everyone who develops distant metastases within 5 years of starting adjuvant osimertinib treatment would have pemetrexed plus cisplatin followed by docetaxel. It assumed that after this 5-year timepoint, 50% of people who develop distant metastases would have retreatment with osimertinib as first-line therapy (see NICE's technology appraisal guidance on osimertinib for untreated EGFR mutation-positive NSCLC) followed by pemetrexed plus cisplatin, with the remaining 50% having pemetrexed plus cisplatin followed by docetaxel. The ERG explained that in its base case, it had included atezolizumab, bevacizumab, carboplatin and paclitaxel as a second-line treatment in both groups and that it had excluded retreatment with osimertinib in the adjuvant osimertinib group. The clinical experts suggested that people are likely to have chemotherapy or atezolizumab, bevacizumab, carboplatin and paclitaxel if their disease progresses during treatment with osimertinib, while osimertinib would be offered to people whose disease progresses after adjuvant treatment with osimertinib. The Cancer Drugs Fund clinical lead agreed that if disease relapsed after treatment with osimertinib stopped, then retreatment would be commissioned in the NHS. They explained that this would depend on the time since finishing osimertinib and the onset of metastatic disease. If this time gap was short then there may not be much benefit, but they noted that the time gap would not need to be as long as that assumed by the company (at least 2 years, depending on when adjuvant osimertinib is taken). The Cancer Drugs Fund clinical lead also said that atezolizumab, bevacizumab, carboplatin and paclitaxel would be offered first-line if treatment with TKIs is inappropriate. The committee was concerned that the 50% split used in the company model is arbitrary. After consultation, the company provided additional sensitivity analyses where the proportion of people who had retreatment with osimertinib was varied between 40% to 60%. The committee noted that although the proportion of people having retreatment with osimertinib remained uncertain, it made no significant difference to the cost‑effectiveness estimates since the cure timepoint used in the company's model coincided with when retreatment is allowed. The committee concluded that retreatment with osimertinib would be offered to some people whose disease had progressed after having osimertinib as an adjuvant treatment.

3.4 The clinical‑effectiveness evidence for osimertinib is based on the ADAURA randomised controlled trial. This is a phase 3 randomised, double‑blind, placebo-controlled, multicentre trial in adults with completely resected stage 1b to 3a NSCLC (stratified by tumour stage, race [Asian versus non‑Asian] and EGFR [exon 19 deletions or L858R] status). ADAURA compared adjuvant osimertinib 80 mg (n=339) with placebo (n=343) over a follow‑up period of 12 and 24 weeks. Some people in both arms of the trial also had adjuvant chemotherapy. The planned treatment duration was 3 years. However, the trial results were released 2 years early after determination of overwhelming efficacy with osimertinib. In the overall trial population, treatment with osimertinib resulted in significantly longer disease‑free survival, with a lower risk of disease recurrence (hazard ratio: 0.20; 99.12% confidence intervals: 0.14 to 0.30; p<0.001). However, the disease‑free survival data is immature and there have been very few events from which to calculate overall survival.

3.5 Because results of the ADAURA trial have been released 2 years early, overall survival data is immature. The company explained that even though overall survival data from ADAURA is very immature, adjuvant osimertinib is expected to have a long‑term survival benefit. This is because of the size of the disease-free survival benefit, a significant reduction in CNS metastases, and a consistent overall survival benefit when it is used to treat metastatic disease. Both the ERG and clinical experts agreed that disease‑free survival is a clinically relevant end point. The clinical experts also emphasised the important benefits of a reduction in CNS metastases. However, the ERG explained that because of the immaturity of the overall survival data from ADAURA, the size of any potential overall survival benefit is uncertain. The committee acknowledged that uncertainty remains around the extent to which adjuvant osimertinib prevents disease recurrence compared with delaying disease recurrence. Very few patients had reached 3 years of treatment with osimertinib and data on recurrence after stopping treatment was not presented. Therefore, it is uncertain what will happen after stopping treatment. The committee was also aware of recent publications by Gyawali (2021) and Uprety (2021), which noted that other adjuvant TKIs showed disease‑free survival benefits that have not translated to an overall survival benefit. The committee was concerned that the experience with earlier generation TKIs such as erlotinib suggested that disease often recurred after stopping treatment. However, a clinical expert cautioned against placing too much weight on this because erlotinib does not have the same brain penetration as osimertinib. In response to consultation, the company stated that the comparison to earlier TKI data in the adjuvant setting was not appropriate. It noted that the benefits of osimertinib, particularly around reducing CNS metastases, were greater than with the earlier TKIs. However, the committee noted there is currently no evidence to show that after stopping treatment with osimertinib, the hazards, and therefore the hazard ratios, for disease‑free survival do not increase (as with the other TKIs). Therefore, the committee reiterated its concern over the immaturity of the disease‑free survival and overall survival data as well as the uncertainty around the extent to which disease‑free survival can accurately predict overall survival. The committee concluded that it was not certain to what extent a benefit in disease‑free survival translates into a benefit in overall survival.

3.6 The company used a state transition, semi‑Markov model with 5 health states: disease free, loco-regional recurrence, first‑line treatment for distant metastases, second‑line treatment for distant metastases, and dead. In the company's model, retreatment with osimertinib for distant metastases is assumed for 50% of people, with the remaining 50% having pemetrexed plus cisplatin. The committee recalled that chemotherapy or atezolizumab, bevacizumab, carboplatin and paclitaxel would be offered to people whose disease had progressed during treatment with osimertinib. Retreatment with osimertinib would be offered to people whose disease had progressed after adjuvant treatment with osimertinib (see section 3.3). In its model, the company had also assumed that 100% of people in the active monitoring arm have osimertinib as their first treatment for metastatic disease (see section 3.10). The ERG explained that the company's model therefore did not reflect the expected treatment pathway. After consultation, the company did additional scenario analyses:

3.7 The company included a 3‑year treatment stopping rule in its model. This is based on the trial design of ADAURA, where the maximum possible treatment duration was 3 years. It is also stated in the summary of product characteristics that treatment for more than 3 years was not studied. The clinical experts said that adjuvant treatment could not be indefinite and that the 3‑year time period is appropriate. However, the patient experts said they would prefer to continue treatment beyond 3 years if the disease had not progressed. They explained that some people would find stopping treatment difficult because they would fear the disease coming back. The committee noted that in ADAURA, 12% of patients in the intervention arm and 10% in the active monitoring arm had reached 3 years of treatment. The committee concluded that a 3‑year treatment stopping rule, in line with the clinical and cost-effectiveness evidence, was acceptable but the impact of stopping treatment at 3 years is uncertain.

3.8 The predicted overall survival gain is a function of all transitions included in the model (see section 3.6), most of which are informed by external data and the company's structural cure assumption (a reduction in risk by 95% for people without disease recurrence at 5 years in both arms). The company's choice of survival models was based on a visual inspection of the combined disease‑free survival and overall survival curves, with input from its clinical experts. In line with advice in the NICE Decision Support Unit Technical Support Document 14, the company applied the same parametric curves across both treatment arms. For the transition from disease free to loco‑regional recurrence, the company applied log‑normal curves, whereas generalised gamma curves were applied for the transition from disease free to distant metastatic NSCLC. The committee was aware that disease‑free survival was a key driver of the company's economic model. It was concerned that most of the disease‑free and overall survival benefits were gained during the extrapolated period, so the choice of extrapolation has a significant effect on the results. For the transition from disease free to distant metastases, the generalised gamma chosen by the company had the best statistical fit for the placebo arm. However, the log‑normal had the best fit for the osimertinib arm. The ERG explained that because of the cure assumption, the choice of extrapolation has little effect beyond the cure timepoint. So, in this situation it is appropriate to give more weight to the statistical fit to the observed data. The ERG did additional sensitivity analyses in which it applied alternative parametric survival models to represent the transition from disease free to distant metastatic NSCLC. These used a log‑normal distribution in:

3.9 The company originally applied a 5‑year cure timepoint in its modelling based on information from its clinical experts. Clinical expert advice to the ERG was that, for the active monitoring arm of the model, a 5‑year cure timepoint may be appropriate, but a potential cure timepoint for the intervention arm is uncertain. The ERG did exploratory analyses to assess the effect of changing the timepoint at which the cure assumption is applied in the company's economic model. The ERG's optimistic analysis retained the company's original approach, whereas the pessimistic analysis applied a later timepoint for cure in the adjuvant osimertinib group of 8 years (5‑year cure timepoint in the active monitoring group plus the 3‑year osimertinib treatment period). In response to technical engagement, the company proposed a 6‑year cure timepoint, which was supported by its clinical experts. The committee was aware that the maximum follow‑up period in ADAURA was 4 years, so the company's cure assumption was uncertain. Very few patients had reached 3 years of treatment with osimertinib so it is also uncertain what will happen after stopping treatment. The committee was concerned that osimertinib may only delay rather than prevent recurrence. Taking into account that there was no data on people who have stopped osimertinib treatment, and the evidence from other TKIs used as adjuvant treatment (see section 3.5), the committee noted that the ERG's pessimistic analysis may also be plausible. After consultation, the company did a scenario analysis where the cure assumption was removed. The ERG explained that this analysis may not be meaningful because it applies parametric survival models, which do not explicitly allow for the potential of cure, to a population in whom cure is expected for a proportion of people. The committee concluded that there was significant uncertainty about the company's cure assumptions, and it would consider both of the ERG's approaches in its decision making.

3.10 NICE recently recommended osimertinib for untreated EGFR mutation-positive NSCLC for metastatic disease. The company base case assumes that 100% of people in the active monitoring arm will have osimertinib as their first treatment for metastatic disease. The committee recognised that people in the ADAURA trial are being actively monitored and disease may be identified at an earlier stage of progression than in current practice. Therefore, more people could be fit enough to have treatment, so outcomes in advanced disease could be better than seen in the FLAURA (a randomised double blind, phase 3 controlled trial comparing osimertinib with erlotinib or gefitinib for the first-line treatment of EGFR mutation-positive advanced NSCLC) trial data. The ERG presented a scenario analysis using a different mix of TKIs. This was based on the latest TKI prescribing data as presented by the company. The committee considered that the proportion of people having osimertinib is likely to increase over time but may not reach 100%. After consultation, the company provided additional sensitivity analyses where 80% to 90% of people in the active monitoring arm had osimertinib as their first treatment for metastatic disease. It noted that the latest prescribing data shows this figure to be around 80%. The committee concluded that it was appropriate to base its decision making on the latest available prescribing data and that 80% was therefore appropriate to use in the analyses.

3.11 The company included utility values based on EQ-5D-3L estimates from ADAURA, EQ‑5D‑3L estimates from FLAURA and published EQ‑5D‑3L estimates from the literature (Labbé et al. 2017). Disutilities associated with adverse events were based on published literature (Nafees et al. 2008, standard gamble) and on a previous appraisal of osimertinib used second line for metastatic disease (see NICE's guidance on osimertinib for EGFR T790M mutation-positive advanced NSCLC). The ERG was concerned that the utility values applied in the disease free, loco‑regional recurrence and distant metastatic NSCLC health states may be implausibly high compared with the general population. The ERG was also concerned that the model does not include health‑related quality of life decrements for late effects of adjuvant treatment or downstream adverse events. However, it suggested that although the utility values may have been overestimated, they did not necessarily favour osimertinib. The ERG explained that it did an additional sensitivity analysis using utility values from a study by Andreas et al. (2018). This had a limited effect on the cost-effectiveness estimates. The committee concluded that the company's utility values were acceptable for decision making.

3.12 Because of confidential discounts for subsequent therapies, the cost‑effectiveness results are commercial in confidence and cannot be reported here. However, the company's base case including all discounts was less than £20,000 per quality‑adjusted life year (QALY) gained. The ERG made several changes to the company's base case and presented 2 analyses. The first was based on a 5‑year cure point in both arms and produced a similar ICER to the company's base case. The second was based on an 8‑year cure point in the osimertinib arm and produced an ICER greater than £20,000 per QALY gained. At the second committee meeting, the committee considered several modelling assumptions plausible:

3.13 Because results from ADAURA were released 2 years early, the disease‑free survival and overall survival data for osimertinib is immature. After considering the uncertainty with the clinical evidence along with its preferred assumptions, the committee considered that the upper end of the most plausible ICER range may not be within the range usually considered a cost‑effective use of resources. The committee concluded it could not recommend osimertinib for the adjuvant treatment of stage 1b to 3a NSCLC after complete resection in adults whose tumours have EGFR exon 19 deletions or exon 21 (L858R) substitution mutations for routine use in the NHS.

3.14 Having concluded that osimertinib could not be recommended for routine use, the committee then considered if it could be recommended for treating stage 1b to 3a NSCLC within the Cancer Drugs Fund. The committee discussed the arrangements for the Cancer Drugs Fund agreed by NICE and NHS England in 2016, noting NICE's Cancer Drugs Fund methods guide (addendum). The committee acknowledged that the disease-free survival and overall survival data from ADAURA was not mature and that further data collection may help address uncertainty. After consultation, the company stated that it would welcome a recommendation in the Cancer Drugs Fund if the committee recognised plausible uncertainties that would result in ICERs higher than what NICE normally considers a cost‑effective use of NHS resources. The committee was aware that, although a period of time in the Cancer Drugs Fund may not produce enough mature overall survival data for a robust mixture‑cure model, there would still be benefits:

3.15 The company said that osimertinib is innovative because there has been little innovation in adjuvant treatment for stage 1b to 3a EGFR mutation-positive NSCLC, aside from adjuvant chemotherapy, in 20 years. Osimertinib has been reviewed as part of Project Orbis because it is considered an innovative adjuvant treatment. In response to consultation, the company stated that not all additional benefits associated with osimertinib had been captured in the economic analysis. In particular, the company highlighted the effect osimertinib has on reducing CNS metastases. It stated that a utility decrement had not been applied in the modelling for people whose quality of life had declined because of CNS metastases. It therefore suggested that the ICERs were a conservative estimate. The ERG stated that it did not accept that any relevant aspects of the value of adjuvant osimertinib had clearly been omitted from the company's economic analyses. The committee recognised osimertinib as an innovative therapy in the adjuvant setting but concluded that it did not consider there were any additional benefits associated with osimertinib that had not been captured in the economic analysis.

3.16 The clinical experts explained that EGFR mutation‑positive NSCLC tends to be more common in women and people with a Chinese family background. The committee noted that the issue of different disease prevalence cannot be addressed in a technology appraisal.

3.17 The only EGFR mutations considered within the scope of this appraisal are EGFR exon 19 deletions or exon 21 (L858R) substitution mutations. This is in line with osimertinib's current marketing authorisation. The Cancer Drugs Fund clinical lead explained that if NICE recommends osimertinib for these mutations, NHS England would consider commissioning adjuvant osimertinib treatment for other less common EGFR mutations. The committee noted that NICE can only appraise a medicine within its marketing authorisation and welcomed the comments from the Cancer Drugs Fund clinical lead.

3.18 The company did not make a case for osimertinib meeting NICE's end of life criteria. NICE's advice about life‑extending treatments for people with a short life expectancy did not apply.