Background
Despite the declining incidence of gastric cancer, it remains the second leading cause of cancer deaths worldwide, with approximately 736,000 deaths and 988,000 new cases each year [
1]. East Asia, including China, Korea, and Japan, has one of the highest incidences and mortality rates of gastric cancer [
2‐
4]. D2 gastrectomy is the most widely used surgical treatment for localized gastric cancer, and long-term follow-up has demonstrated a reduction in gastric cancer-related deaths in patients who have undergone D2 gastrectomy compared with D1 gastrectomy [
5‐
7]. As a result, D2 gastrectomy is preferred in Asia for patients presenting with resectable gastric cancer [
8].
Although surgery is the most efficient treatment for operable cancer, recurrence may result in cases with poor prognosis. As an important component of resectable gastric cancer therapy, adjuvant chemotherapy could improve patient outcomes, although no consensus about the preferred treatment has been reached [
9‐
11]. According to the guidelines of the National Comprehensive Cancer Network (NCCN), both capecitabine plus oxaliplatin and S-1 are recommended as adjuvant treatments for gastric cancer [
12].
Capecitabine is a new oral drug derived from fluorouracil (FU), which is widely used in the therapy of breast, gastrointestinal, and head and neck cancers [
13‐
15]. A regimen consisting of oxaliplatin plus capecitabine is very effective and tolerable in patients with gastric cancer [
16]. By contrast, S-1 is an orally active combination of tegafur, gimeracil, and oteracil at a molar ratio of 1:0.4:1 and is used as a novel mode of neoadjuvant chemotherapy [
2,
17,
18]. As relatively new adjuvant chemotherapies, both modes of therapy have the potential to decrease recurrence rates and achieve survival benefits for patients compared with surgery alone [
19‐
21]. However, both modes of therapy markedly increase the cost of the entire treatment approach for gastric cancer, and widespread use of these modes would be limited, particularly in health resource-poor countries such as China [
22].
Cost-effectiveness analyses can improve resource allocation efficiency by identifying therapies that provide the greatest health benefits at acceptable cost. However, clinical trials that feature health economics assessments are scarce. Therefore, we have used a mathematical modeling approach to conduct health economics analyses.
Within this study, we have developed a health economics model to evaluate the long-term cost-effectiveness of two adjuvant chemotherapies (the adjuvant S-1 and XELOX strategies) compared with surgery alone in patients presenting with gastric cancer and undergoing D2 gastrectomy in China. The model integrates the best available evidence in terms of costs and clinical outcomes resulting from the use of adjuvant therapies to determine whether such strategies truly represent a health budgetary advantage in the context of the Chinese healthcare system.
Discussion
The current analysis is the first to evaluate the health and economic outcomes of different adjuvant regimens in gastric cancer patients treated with D2 gastrectomy. We determined that adjuvant therapies provide substantial health benefits relative to surgery only by increasing the QALYs (Table
4). This increase in QALYs, in turn, could be contributed to improvements in the RFS and OS rates [
2,
23]. The total costs associated with the use of adjuvant chemotherapy for the S-1 and XELOX strategies were $24,503.1 and $18,379.6, respectively, both of which are significantly higher than the cost of surgery only ($13,638.2). However, our cost-effectiveness analysis revealed that the ICERs of the S-1 and XELOX strategies versus the Surgery only strategy are $6,837 and $ 3,502, respectively, per additional QALY gained. According to the WHO recommendation for the cost-effectiveness threshold, both adjuvant strategies are cost-effective because their ICERs are lower than the threshold of $18,300 per additional QALY gained (which represents three times the per capita GDP of China in 2012) [
34‐
36]. In particular, the ICER of the XELOX strategy in the base-case analysis was less than one times the per capita GDP of China in 2012, indicating that adjuvant therapy with the XELOX strategy would be very cost-effective in the Chinese setting based on the WHO recommendation.
As shown in Table
4 and Figure
3, the XELOX strategy may provide greater health benefits and relatively lower costs compared with the S-1 strategy, indicating that the S-1 strategy would be dominant. These results were further confirmed by probabilistic sensitivity analyses and cost-effectiveness acceptability curves (see Figures
5 and
6).
Only three other economic analyses of adjuvant chemotherapy for the treatment of patients with stage II-IIIB gastric cancer with D2 gastrectomy have been conducted, all in Japanese or Chinese settings [
37‐
39]. These reports determined that adjuvant chemotherapy with S-1 or capecitabine plus oxaliplatin after D2 gastrectomy for patients with resectable gastric cancer was a favorable recommendation in accordance with long-term cost-effectiveness compared with D2 gastrectomy alone. On the basis of the current clinical trial and from the perspective of the Chinese healthcare system, the results of these studies are consistent with ours. However, the result reported by Tan C
et al. suggest that adjuvant treatment with XELOX strategy is a cost-saving strategy over the long term, despite the higher total cost of the XELOX strategy compared with that of surgery alone. Tan C
et al. determined that the total costs of the surgery-only strategy were $87,004 and $65,894, respectively, far higher than our evaluation ($13,638.2) and the finding ($9,346) reported by Hisashige A
et al. [
37‐
39]. This discrepancy may be due to the considerable assumption made by Tan C
et al. that in the situation of tumor recurrence or new occurrences of gastric cancer, cycles of intravenous paclitaxel (at 80 mg/m
2, three times per week) would be administered every 4 weeks as a first-line chemotherapy for advanced gastric cancer [
37,
38]. However, according to the gastric cancer guidelines of the National Comprehensive Cancer Network, clinical trials and our assumptions in previous work, paclitaxel (80 mg/m
2) is administered as a second-line chemotherapy for advanced gastric cancer and is repeated weekly for 3 of every 4 weeks [
12,
40,
41].
As reported by Tan C
et al. the cost of oxaliplatin per 50 mg is a substantial consideration [
37]. When brand-name oxaliplatin (Eloxatin®, produced by Sanofi-Aventis) was used in the XELOX strategy, the ICER of the XELOX strategy increased to $10,469 per additional QALY gained. The cost of generic oxaliplatin is only one-quarter that of brand-name oxaliplatin, and the generic is now widely accepted and prescribed in Chinese clinical practice. Although the sensitivity analysis indicated that the XELOX strategy was cost-effective (using brand-name oxaliplatin), we suggest the use of generic oxaliplatin to further conserve limited healthcare resources. Another important influential factor was RFS, which would improve the ICER of the XELOX strategy by decreasing the HR. This finding indicates that it is more cost-effective to treat subgroups with more favorable prognostic factors, such as nodal status 1 or 2, with the XELOX strategy compared with the Surgery strategy [
23]. The cost-effectiveness threshold of $18,300/QALY was robust and revealed that treatment with an adjuvant therapy using the XELOX strategy was cost-effective.
The results of this analysis must be interpreted carefully within the limitations of the data and study design. First, we used a two-parameter Weibull survival model to extend the tails of survival beyond the follow-up time horizon [
42]. Table
3 and Figure
2 show the estimated survival rates fitted to the nonparametric Kaplan-Meier survival rates from the trials, which support the validity of our model. However, there are no long-term (>5 years) RFS and OS data available for patients receiving adjuvant chemotherapy, which could influence the results. Although one-way sensitivity analyses were conducted to evaluate the uncertainty in model outcomes arising from the parameters, this lack of long-term data represents another limitation of our research approach. The current analysis must be updated when long-term outcomes are reported. Second, new therapies are rapidly being developed for managing gastric cancer, including treatment with trastuzumab for HER2-positive gastric cancer; this approach improved the survival of patients with HER2-positive advanced gastric or gastro-esophageal junction cancer [
43]. However, these new agents tend to be more expensive than current therapies. Although the current analysis could not trace all medical resources associated with potential new agents in the future, the findings from the one-way sensitivity analyses indicate that the ICER of adjuvant chemotherapy would be improved by increased resource utilization after disease relapse. Third, owing to the absence of head-to-head trials for both of the adjuvant strategies for the adjuvant therapy of gastric cancer patients following D2 gastrectomy compared in this study, an indirect comparison was conducted, another inevitable weakness of the present analysis. The patient characteristics in the CLASSIC trial and in the trial reported by the ACTS-GC Group were assumed to be similar in our indirect comparisons, and the results of the indirect comparison were imputed into the analytical model. Nevertheless, as no data directly comparing the effectiveness of the S-1 and XELOX strategies in large RCTs are available, many investigators worldwide accept indirect comparisons using robust methods. Further studies will be required to directly determine the clinical efficacy of these adjuvant strategies. Fourth, because of the absence of RFS data in the CLASSIC trial, we replaced the RFS data with DFS data to compare the outcomes of the XELOX and S-1 strategies. It could be inferred that the XELOX strategy would have more a favorable RFS than the S-1 strategy because there are more DFS than RFS events; events such as the development of a second primary cancer would lead to more favorable economic outcome [
44]. Fifth, the results of this analysis should be carefully interpreted because several factors such as the costs associated with death, the patterns of clinical practice and the availability of health care resources limit the transferability of economic evaluations across jurisdictions. Finally, owing to the nature of the study design, we did not measure exact costs, such as the costs associated with adverse events and palliative care. A cost-of-illness study should be conducted in the future. We believe that our results have theoretical and reference value and provide valuable policy-making data to guide the allocation of health resources in China.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
WB and ZG adapted the model, conducted the analyses, interpreted the results and wrote the manuscript. Dr. LT, JC and YX contributed to data collection and manuscript preparation. Dr. ZG supervised the study, contributed to the interpretation of the results, reviewed the manuscript and is the guarantor of the overall content. All authors read and approved the final manuscript.