Introduction
Thyroid cancer (TC) is the tenth most common cancer worldwide, with more than 580,000 new cases diagnosed and more than 43,000 deaths (1). More than 190,000 new cases have been reported in China (1). Differentiated TC (DTC) is the most prevalent, accounting for more than 90% of all TCs (2). The probability of recurrence or metastasis disease was close to 60% (3). Radioactive iodine therapy was the primary treatment for patients with advanced DTC, but 30% of patients become radioactive iodine-refractory (RAIR) cancers (4, 5), which have a 10-year survival rate of 10% (6).
Two phase III trials, DECISION (NCT00984282) and SELECT (NCT01321554), demonstrated that sorafenib and lenvatinib extend progression-free survival (PFS) by a significant 10.3 months and 18.3 months, respectively, compared with placebo in the whole population of RAIR-DTC patients (7, 8). Subsequently, sorafenib and lenvatinib have been approved by Chinese Society of Clinical Oncology (CSCO) and were added to their guidelines in 2017 and 2020, respectively, as standard treatments for Chinese patients with RAIR-DTC (9). Unfortunately, no significant increase in overall survival (OS) was observed in either study, which included populations from multiple countries (7, 8). Therefore, as China is a country of major RAIR-DTC prevalence, the limited treatment regimens cannot meet the demand. Treatment strategies for Chinese patients with RAIR-DTC need to be improved and widely applied in clinical practice.
Apatinib is a small-molecule angiogenic inhibitor of vascular endothelial growth factor receptor (VEGFR)-2 of high selectivity. Lenvatinib is a multitargeted tyrosine kinase inhibitor (TKI) specific for VEGFR-1, -2, and -3. Based on Chinese people with RAIR-DTC, two studies showed the significant clinical benefits of TKIs. The REALITY trial (NCT03048877) showed that apatinib significantly extended the median PFS (22.2 months; HR, 0.26; 95% CI, 0.14–0.47; P < 0.001) and OS (HR, 0.42; 95% CI, 0.18–0.97; P = 0.04) of Chinese patients in advanced stages of RAIR-DTC compared with placebo (10). Study-308 (NCT02966093) demonstrated that lenvatinib significantly improved the median PFS (23.9 months; HR, 0.16; 95% CI, 0.10–0.26; P < 0.0001) of Chinese patients with RAIR-DTC compared with placebo. However, there was no significant benefit in terms of OS (HR, 0.42; 95% CI, 0.18–0.97; P = 0.04) (11). Because of these findings, both apatinib and lenvatinib are recommended as advanced RAIR-DTC treatment in the 2021 CSCO guidelines (9). With the remarkable results of the two Chinese-patient-based studies, the concomitant cost-effectiveness of the two TKI types has become the focus of attention. To answer this question, we compared the cost-effectiveness of apatinib and lenvatinib for patients with advanced RAIR-DTC from the perspective of Chinese payers.
Discussion
TC is endocrine system malignancy (23). The morbidity, mortality, and burden of the disease continue to rise worldwide (24–26). China has one of the highest TC burdens in the world, with an average cost of $11,560 per patient in the first year after diagnosis (27, 28). Because of rising healthcare costs, treatment cost evaluations are necessary. TKI is currently on the radar of clinicians and patients with RAIR-DTC. Two previous studies, DECISION (NCT00984282) and SELECT (NCT01321554), demonstrated the significant clinical efficacy of TKIs (7, 8). Therefore, three economic assessments have been published based on two major studies. Huang et al. and Tremblay et al. showed that the ICER generated by lenvatinib and sorafenib in patients with advanced RAIR-DTC were $103,925 per QALY and $95,695 per QALY, respectively, and that lenvatinib is cost-effective compared with sorafenib at a WTP of $150,000 per QALY and $100,000 per QALY, respectively, from the US perspective (29, 30). Leslie et al. showed lenvatinib to be more cost-effective than sorafenib (ICER = $25,275 per QALY) or placebo (ICER = $40,869 per QALY) and that sorafenib was also cost-effective compared to placebo (ICER = $64,067 per QALY) (15). The final analysis demonstrated lenvatinib to be the most cost-effective option for RAIR-DTC at a WTP of 100,000 per QALY, although both lenvatinib and sorafenib were more cost-effective than placebo (15). These studies evaluated the cost performance of the two TKIs based on the perspective of US patients and obtained consistent results that show lenvatinib may be more cost-effective. With the development of new drugs, Chinese physicians and patients are gradually paying attention to the cost and efficacy of TKIs. Therefore, using the Markov model and the clinical efficacy and safety data from two large, randomized phase III clinical trials, we estimated the cost-effectiveness over a 20-year time horizon for apatinib and lenvatinib as therapies for RAIR-DTC. This study of RAIR-DTC patients in China showed that apatinib therapy provided a 0.837 improvement in QALY and $6,975 reduction in costs compared with lenvatinib, resulting in ICER value is definitely lower than the WTP value. Deep meaning indicates that apatinib is a superior treatment strategy compared to lenvatinib, achieving higher efficacy as well as a lower healthcare cost. Therefore, apatinib was more cost-effective compared to lenvatinib in the first-line treatment of Chinese RAIR-DTC patients. Although, it is worth considering that the additional costs associated with lenvatinib are mainly due to follow-up, meaning that the same follow-up plan should be set up for the same cancer type.
We used sensitivity analysis to confirm model uncertainty. From the one-way sensitivity analysis, we deduced that the model’s most influential parameter was the HR of PFS (apatinib versus lenvatinib), underscoring the need for robust head-to-head clinical data. It was also sensitive to the utility of PFS, and the analysis found that, for patients with a lower utility of PFS, apatinib had a more favorable economic outcomes compared with lenvatinib, but apatinib had a worse economic outcome for patients with a higher utility. Another important influencing variable was the price of the TKIs: apatinib’s price increase of more than 56% and lenvatinib’s price decrease of more than 46% mean lenvatinib is more cost-effective. Changing other parameters had virtually no influence on our results, and due to the high cost of TKIs, reducing the prices of apatinib and lenvatinib was considered the most practical measure in the context of cost-effectiveness and optimal logistics.
The results were consistent with baseline in the subgroup analysis, showing that the ICER (apatinib versus lenvatinib) was lower than the WTP. In the sensitivity analysis, apatinib was more cost-effective than lenvatinib 55%. It is worth noting that in the subgroups of Chinese patients of ≤65 years, male, and PTC, apatinib generation was associated with a higher efficacy at a lower cost. Changes in the QALYs of PFS in PTC were widely observed, leading to the possibility that TKI was more effective among many pathological types of PTC. This is consistent with the results of two previous retrospective studies, the multivariate analysis of which identified histological grade as a favorable prognostic factor. Lars et al. reviewed 173 patients with PTC, and their analysis showed that the 10-year survival rates of patients with low-and high-grade PTC were 95.3% and 75.1%, respectively (31). Allen et al. reviewed 37,858 cases of PTC and found the prognosis of moderately differentiated and poorly differentiated PTC was positively correlated with OS compared with highly differentiated PTC (HR, 1.21; 95% CI, 1.04-1.41; P = 0.02 and HR, 2.62; 95%CI, 2.23-3.08, P < 0.001) (32). Therefore, under the influence of cost-effectiveness, pathological typing to predict treatment prognosis needs to be performed.
In China, those approving innovative drugs to maintain our healthcare system should not only take into account the huge clinical and economic benefits but also realize the importance of prognostic marker biomarker analysis. Unfortunately, we did not analyze predictive markers. However, studies have found that thyroglobulin (Tg) and anti-thyroglobulin (TgAb) are important prognostic factors for guiding clinical treatment and are valuable parameters for long-term monitoring of DTC patients. Patients with Tg of <0.2 ng/mL or thyroid-stimulating hormone-stimulated Tg of <1 ng/mL responded well to treatment and had minimal levels of recurrence and an almost complete absence of disease-specific deaths (33–36). However, patients with higher than normal Tg levels (inhibition of Tg of ≥1 ng/mL or stimulation of Tg of ≥10 ng/mL) or elevated TgAb values after treatment had a low mortality rate, but a significant proportion of this group developed structural disease recurrence (33, 37). A prospective study involving serum from 249 patients showed that VEGF and VEGFR may have prognostic value for RAIR-DTC (38). Levels of VEGF showed a clear link to a lower risk of recurrence (overall response [OR], 0.08; 95% CI, 0.01-1.43; P = 0.018 and OR, 0.08; 95% CI, 0.01-1.37; P = 0.016) (38). In addition, molecular markers such as BRAF, EGFR, Ki67, and P53 were also potentially effective prognostic factors (39–42).
This study had several significant advantages. Firstly, we ascertained the cost-effectiveness of two TKI-based RAIA-DTC treatments over a 20-year time range using an economic model. As input for our model, clinical efficacy and safety data were extracted from high-quality phase III clinical trial datasets, and the costs were from the perspective of Chinese payers. We concluded that our model provides long-term cost and effectiveness predictions that are easily translatable to clinical practice. Secondly, we considered the disutility generated by AEs. We used the average health utility of PFS in patients with advanced RAIR-DTC and corrected it using the disutility generated by AEs. When evaluating the economic benefits of the two TKIs, only the negative utility generated by severe AEs had a strong correlation with the quality of life (QoL) (43). Finally, we compared the cost-effectiveness of the two TKIs and added three subgroups that might be useful in clinical practice.
The study had some limitations. Firstly, the survival data were obtained from the interim analysis results of the phase III clinical trials REALITY and Study-308. The survival data will mature with the extension of follow-up time, and the model will become more stable. Secondly, these two TKI schemes were not directly evaluated in any of the trials. Therefore, we compared the two TKIs schemes indirectly using the NMA findings from two phase III clinical trials with similar research content and characteristics of the included patients. However, this method comes with potential uncertainties. Thirdly, to simplify the calculation, we assumed that patients in both groups only received BSC after treatment with TKIs for recurrence, so the analysis may have underestimated the cost of PD. However, we discovered from the sensitivity analysis that the economic burden of PD had little effect on the model’s outcomes. Fourthly, since neither the REALITY nor Study-308 reports provided QoL data, the health state utility in this model was obtained from previously published data and was based on patients in the US or UK. As the QoL of Chinese patients with RAIR-DTC has not yet been reported, this was an essential deviation. Including the QoL of Chinese patients in future studies means the economic results will be updated in time. Fifthly, because of the lack of subgroup survival curves in both studies, we were unable to run a complete model for each subgroup, and the original group equilibrium generated by NMA analysis may not apply to the subgroups. Therefore, the results of the subgroup analysis should be interpreted with caution. Finally, in this model, we only considered the cost and corresponding disutility of treatment-related grade 3/4 and ≥5% AEs, which may have had some influence on the overall cost and utility. However, sensitivity analysis showed that the incidence and disutility of major AEs had little effect on the results.