Cost‐effectiveness analysis of first‐line treatments for advanced epidermal growth factor receptor‐mutant non‐small cell lung cancer patients

Abstract Objectives Recent studies showed prolonged survival for advanced epidermal growth factor receptor (EGFR)‐mutant non‐small cell lung cancer (NSCLC) patients treated with both monotherapies and combined therapies. However, high costs limit clinical applications. Thus, we conducted this cost‐effectiveness analysis to explore an optimal first‐line treatment for advanced EGFR‐mutant NSCLC patients. Materials and Methods Survival data were extracted from six clinical trials, including ARCHER1050 (dacomitinib vs. gefitinib); FLAURA (osimertinib vs. gefitinib/erlotinib); JO25567 and NEJ026 (bevacizumab +erlotinib vs. erlotinib); NEJ009 (gefitinib +chemotherapy vs. gefitinib); and NCT02148380 (gefitinib +chemotherapy vs. gefitinib vs. chemotherapy) trials. Cost‐related data were obtained from hospitals and published literature. The effect parameter (quality‐adjusted life year [QALY]) was the reflection of both survival and utility. Incremental cost‐effectiveness ratio (ICER), average cost‐effectiveness ratio (ACER), and net benefit were calculated, and the willingness‐to‐pay (WTP) threshold was set at $30828/QALY from the perspective of the Chinese healthcare system. Sensitivity analysis was performed to explore the stability of results. Results We compared treatment groups with control groups in each trial. ICERs were $1897750.74/QALY (ARCHER1050), $416560.02/QALY (FLAURA), ‐$477607.48/QALY (JO25567), ‐$464326.66/QALY (NEJ026), ‐$277121.22/QALY (NEJ009), ‐$399360.94/QALY (gefitinib as comparison, NCT02148380), and ‐$170733.05/QALY (chemotherapy as comparison, NCT02148380). Moreover, ACER and net benefit showed that the combination of EGFR‐TKI with chemotherapy and osimertinib was of more economic benefit following first‐generation EGFR‐TKIs. Sensitivity analyses showed that the impact of utilities and monotherapy could be cost‐effective with a 50% cost reduction. Conclusion First‐generation EGFR‐TKI therapy remained the most cost‐effective treatment option for advanced EGFR‐mutant NSCLC patients. Our results could serve as both a reference for both clinical practice and the formulation of medical insurance reimbursement.


| BACKGROUND
Lung cancer remains the leading cause of cancer-related mortality worldwide. Non-small cell lung cancer (NSCLC) is a major component, accounting for approximately 85% of cases. [1][2][3] However, the 5-year survival rate was poor and ranged 10%-20% in most countries. 4 Epidermal growth factor receptor (EGFR) mutations are a common molecular therapeutic target occurring in approximately 40%-60% of Asian patients and 10%-20% of Caucasian patients with lung adenocarcinomas. 5 Currently, first-generation EGFR-tyrosine kinase inhibitors (EGFR-TKIs, [gefitinib, erlotinib, and icotinib]) and second-generation EGFR-TKI (afatinib)-targeted therapy have demonstrated improved efficacy over conventional chemotherapy for advanced EGFR-mutant patients and is generally accepted as the standard first-line treatment, followed by sequential chemotherapy or third-generation EGFR-TKIs with EGFR T790 M mutation after disease progression. [6][7][8] However, the median progression-free survival (PFS) for first-generation EGFR-TKIs was less than a year. 9 In addition, considering the impact of drug resistance to EGFR-TKI, adverse events (AEs), physical state, psychology, and costs of drugs, the proportion of patients who benefit from sequential treatment strategies is limited. 10,11 Thus, first-line treatment strategies for EFGR-mutant NSCLC patients were further explored and remained challenging. Current studies have mainly focused on monotherapies including second-and third-generation EGFR-TKIs (dacomitinib and osimertinib) as well as combined therapies including combination of EGFR-TKIs with chemotherapy and anti-angiogenic therapy, respectively. However, despite the efficacy of multiple treatment strategies, high costs limit their clinical application. 12,13 Optimal treatment decisions should be made after balancing symptom burdens, survival outcomes, tolerability, quality of life, costs, and reimbursement issues comprehensively. 14 Therefore, we conducted a cost-effectiveness analysis that measured different treatment schemes via multiple dimensions from the perspective of the Chinese health system, aiming to provide guidance for treatment decisions in clinical practice. Considering the accessibility and comparability of data, six clinical trials were extracted in order to represent different first-line treatment strategies for advanced EGFR-mutant NSCLC patients, including two phase 3 clinical trials. These trials include ARCHER 1050 trial and FLAURA trial for monotherapy; phase 2 JO25567 trial and phase 3 NEJ026 trial for the combination of targeted therapy and antiangiogenic therapy; phase 2 NCT02148380 trial and  23 The detailed inclusion criteria, administration, survival outcomes, and AE rates are listed in Table 1. Quality-adjusted life year (QALY) was measured as an effect parameter, which represented a combination of survival and utility. Utility is the reflection of a patient's quality of life, which ranges from 0 to 1. The utilities of administration methods (including oral and intravenous administration) as well as utilities of different EGFR-TKIs, were distinguished. 15,27 In addition, utilities of PD state and AE ≥grade 3 were included. 27,28 Both costs and utilities are displayed in Table 2, and the discount rate was set at 3%.

| Outcomes
We conducted cost-effectiveness analyses for various treatment groups and comparative groups in different clinical trials using both the 5-year horizon and 10-year horizon. The willingness-to-pay (WTP) threshold was set at 3 per gross national product (GDP), $30828/QALY in China. The primary result was expressed as incremental cost-effectiveness ratio ([ICER], ICER=incremental cost/ incremental effect), representing the results of intra-group cost-effectiveness analyses. Besides, we further explored the ideal percentage of cost adjustments based on the WTP threshold, incremental effect, and current costs using a positive ICER. Secondary results included the average cost-effectiveness ratio ([ACER], ACER =cost/ effect) and net benefit (net benefit=QALY* willingness-to-pay [WTP] threshold-cost), focusing on inter-group cost-effectiveness analyses.
To detect the influence of various parameters and the stability of results, both one-way sensitivity analysis and probabilistic sensitivity analysis were performed by Treeage. In the one-way sensitivity analysis, cost parameters were assumed with a 30% range and a 20% range for both utilities and probabilities 28 (Table S1). Results are shown in Tornado Diagrams. For probabilistic sensitivity analysis, a Monte Carlo simulation with 1000 iterations was conducted, and cost parameters were assumed to fit the gamma distribution, while other parameters fitted the beta distribution, 29 and the results would be shown in cost-effectiveness acceptability curves.   Table 3. The parameters of the transition probability are listed in Table S2. The cost-effectiveness analysis curves are shown in Figure S1.
Finally, in one-way sensitivity analyses, the utility of the combination group showed a greater impact in the NCT02148380 and NEJ009 trials over the 5-year horizon. Meanwhile, in other cases, utility parameters in first-generation EGFR-TKI groups had a greater impact on the results. Results are shown in Figure S2 and Figure S3. For probabilistic sensitivity analyses, cost-effectiveness acceptability curves showed that at a WTP threshold of $1600000/QALY, dacomitinib could be cost-effective based on the ARCHER1050 trial, and at a WTP threshold of approximately $350000/QALY -$500000/ QALY, osimertinib could be cost-effective (Figure 2).

| DISCUSSION
Despite the rapid development of anti-tumor therapies, the prognosis of advanced NSCLC patients remains poor. Thus, choices of first-line treatment strategies are of high clinical concern, and it is urgent to weigh treatment options from multiple dimensions in order to provide guidance for clinical practice. 30 To the best of my knowledge, this is the first cost-effectiveness analysis to comprehensively evaluate first-line treatment strategies from multiple dimensions for advanced EGFR-mutant patients. Six clinical trials were extracted representing targeted monotherapy (ARCHER 1050 trial, FLAURA trial), the combination of targeted therapy and antiangiogenic therapy (JO25567 trial, NEJ026 trial), and the combination of targeted therapy and chemotherapy (NCT02148380 trial, NEJ009 trial). We analyzed the intragroup cost-effectiveness ratio between various treatment regimens and corresponding control groups based on clinical trials. Results showed that with the current WTP threshold ($30828/QALY), none of the emerging therapies was costeffective compared with standard first-generation EGFR-TKIs for advanced EGFR-mutant patients. As for inter-group comparison, when considering ACER and net benefit, the combination of EGFR-TKI with chemotherapy and osimertinib showed more economic benefit following the application of first-generation EGFR-TKIs. Probabilistic sensitivity analysis suggested that monotherapy could be cost-effective at a higher WTP threshold.
Common methods of cost-effectiveness analyses included both the construction of the Markov model and the calculation of QALY through the area under survival curves (AUC). 31,32 However, the AUC calculation was highly influenced by the heterogeneity among trials, especially different follow-up periods, while the calculation of transition probabilities and horizon settings in the Markov model could reduce the heterogeneity. Since, the current study involved inter-group comparisons among multiple trials, we chose the Markov model method. In addition, Treeage software is a commonly used and specialized software for cost-effectiveness analyses. Thus, we built the decision tree and Markov model according to relevant literature. [33][34][35] The results of the cost-effectiveness analysis were mainly affected by the following aspects. First, the prolongation of survival period was an important influencing factor. Although the prolongation of PFS was significant in all six trials, only the investigations of monotherapy (dacomitinib and osimertinib) and the combination therapy of targeted therapy with chemotherapy extended the survival benefit in OS analysis. Therefore, it was suggested that although clinical trials were often compared with PFS, the OS benefit was important in cost-effectiveness analysis. Second, the patient's quality of life was also an important factor in measuring economic benefits. Considering the lower health state utilities of intravenous administration than oral administration as well as the disutility of AEs, intravenous combination therapy had lower health state utilities than oral monotherapies. As a result, only patients administered EGFR-TKI monotherapy had higher QALYs than the standard first-generation EGFR-TKI group. Third, high costs also affected the clinical application of drugs. In China, first-generation EGFR-TKIs, afatinib, and chemotherapeutic drugs were covered by medical insurance reimbursement, while dacomitinib, first-line osimertinib, and bevacizumab were still under negotiation. In addition to clinical data, adjustments to the costs of drugs and health insurance policies are also important factors for the cost-effectiveness of various treatment regions. Thus, we further explored the ideal cost concessions for treatment regimens with high ICERs in order to achieve economic benefits. The results showed that with a cost reduction of 49.75% and 51.98%, osimertinib and dacomitinib could be cost-effective, respectively.
Previous cost-effectiveness analyses have mainly focused on different generations of EGFR-TKI monotherapy. Studies have demonstrated that first-generation EGFR-TKIs are cost-effective compared with conventional chemotherapy. 36 In the comparison between afatinib and first-generation EGFR-TKIs, current cost-effectiveness studies showed controversial results, the prognosis and economic benefits were relatively similar, and treatment decisions could be relied on personal characteristics of patients in a comprehensive view. [37][38][39] However, none of the cost-effectiveness analyses exploring first-line osimertinib showed economic benefit, similar to our results. 31 inclusion criteria, FLAURA, NCT02148380, and NEJ026 trials included patients with neurologically stable central nervous system (CNS) metastases, while patients with CNS metastases were excluded from other trials. Clinical characteristics of eligible patients, subsequent treatments after disease progression, items and frequency of examinations, proportions of cross-over patients, and follow-up time also varied. Further head-to-head trials are expected to explore the optimal treatment strategy for advanced EGFR mutant NSCLC patients, including the conventional strategy that sets first-generation EGFR-TKIs as first-line administration followed by second-line osimertinib for T790 M mutation-positive patients. Second, several phase 3 clinical trials were not included, mainly due to the lack of data. Phase 3 RELAY trial focused on the comparison between ramucirumab plus erlotinib versus erlotinib; however, despite the absence of OS data, ramucirumab has not yet been widely used in Chinese clinical practice. 42 The CTONG1509 trial was the first randomized phase 3 trial comparing the combination of bevacizumab and erlotinib versus erlotinib in a Chinese population with advanced EGFR-mutant NSCLC. However, since only PFS data from oral reports were available, we did not include this study. 43 Cost-effectiveness analyses could be further evaluated by updating the data. Finally, studies showed a distinction between EGFR mutation subtypes, which influenced the survival outcomes of first-line treatments. 44 Further cost-effectiveness analyses should also be based on distinct grognoses of EGFR mutation subtypes for choosing the optimal treatment strategies for advanced EGFR-mutant NSCLC patients.

| CONCLUSION
Based on survival data from clinical trials, patients' quality of life and current costs of medical resources, standard firstgeneration EGFR-TKI therapy remained the most cost-effective treatment option for advanced EGFR-mutant NSCLC patients comprehensively.