Cost-Effectiveness Analysis of Hepatic Arterial Chemotherapy for Advanced Hepatocellular Carcinoma in China: A Comparative Analysis of HAIC-FO and Sorafenib

Introduction

Hepatocellular carcinoma (HCC) represents a prevalent type of primary liver cancer, with approximately 800 000 cases reported worldwide annually [1]. In China, there were nearly 430 000 new cases of liver cancer in 2022, with HCC accounting for 85% to 90% [2]. A significant portion of patients with HCC are typically diagnosed at advanced stages, and approximately 50% of them receive systemic therapies [3]. The introduction of molecular targeted therapies has revolutionized HCC management, with sorafenib being the inaugural Food and Drug Administration approved molecular targeted therapy in 2007, establishing a benchmark for a decade (SHARP trial) [4]. Lenvatinib received further approval in 2018 for the first-line treatment of advanced HCC (REFLECT trial). The IMbrave150 study in 2020 demonstrated that the atezolizumab and bevacizumab regimen outperformed sorafenib, establishing it as the preferred first-line treatment globally [5]. However, the benefit of atezolizumab and bevacizumab in high-risk (Vp4 portal vein invasion, the bile duct invasion, and/or tumor occupancy of ≥50% of the liver) patients is limited, with a median overall survival (OS) of 7.6 months. Additionally, concerns persist regarding the health economic value of the atezolizumab and bevacizumab regimen [6].

Recently, Mohammadnezhad et al conducted a comprehensive search across 3 scientific databases to identify atezolizumab- and bevacizumab-relevant economic studies [6]. Out of 315 records identified, 12 cost-effectiveness analyses met the criteria for inclusion in the systematic review. Across all studies, treatment costs were notably higher for atezolizumab and bevacizumab, compared with sorafenib and nivolumab, ranging from $61,397 to $253,687 per patient, respectively. The collective findings underscored the lack of cost-effectiveness associated with atezolizumab and bevacizumab. Given these analyses, the current pricing structure indicates limited potential for cost-effectiveness, particularly over the long term, in the treatment of advanced HCC. Thus, substantial discounts on prices are imperative for this innovative approach to achieve cost-effectiveness and widespread adoption.

Hepatic arterial infusion chemotherapy (HAIC), owing to its liver-specific delivery, demonstrates enhanced antitumor efficacy and safety, compared with systemic therapy [7]. It was evidenced that HAIC utilizing infusional fluorouracil, leucovorin, and oxaliplatin (FOLFOX) regimens (HAIC-FO) served as a superior first-line therapeutic option, compared with sorafenib for advanced HCC [8]. The findings of the FOHAIC-1 trial unveiled that HAIC-FO treatment resulted in enhanced survival outcomes when compared with sorafenib in patients with advanced HCC (median OS: 13.9 vs 8.2 months, P<0.001; and median progression-free survival [PFS]: 7.8 vs 4.3 months, P<0.001) [9]. Meta-analyses demonstrated that HAIC was associated with significantly improved survival, compared with alternative therapies, including systemic chemotherapy, sorafenib, transarterial chemoembolization (TACE), and others [10,11].

Considering the substantial population of patients with HCC and the constrained medical resources in China [12], it is imperative to undertake a pharmacoeconomic assessment of treatment modalities for HCC. Despite the conceivable advantages of HAIC-FO in extending patient survival, there has been a lack of prior investigations assessing its cost-effectiveness in managing advanced HCC. Hence, our study aimed to compare the pharmacoeconomics of HAIC-FO and sorafenib in patients with advanced HCC, incorporating the perspective of the Chinese healthcare system. This comparison was facilitated through the use of Markov models.

Material and Methods

PATIENTS AND INTERVENTION:

Our study was based on data from the FOHAIC-1 trial [9], which enrolled patients diagnosed with locally advanced or unresectable HCC in China. The primary inclusion criteria were as follows: presence of a dominant liver mass with or without extrahepatic oligometastasis, defined as up to 3 metastatic lesions in up to 2 organs, with a largest diameter of ≤3 cm; disease deemed unsuitable for surgical, ablative, or TACE interventions, or progression of the disease following such therapies; absence of prior systemic treatments; Child-Pugh grade ≤7; Eastern Cooperative Oncology Group Performance Status of 0–2; and adequate hematologic, hepatic, coagulation, and renal function. Detailed criteria were delineated on clinicaltrials.gov with registration ID NCT03164382. The essential characteristics of the sample have been previously outlined. Among the participants, 90.8% were male, 89.3% tested positive for hepatitis B virus infection, 69.1% had Child-Pugh A classification, with a mean tumor diameter of 11.2 cm, while 82.8% displayed macrovascular invasion. These patients were randomized to receive either HAIC-FO or sorafenib. The HAIC-FO regimen comprised oxaliplatin (130 mg/m2), leucovorin (200 mg/m2), fluorouracil (400 mg/m2), and fluorouracil (2400 mg/m2). This treatment protocol was administered sequentially via a catheter every 3 weeks, for a maximum of 8 cycles. Conversely, sorafenib was administered at a dosage of 400 mg twice daily. Patients continued their respective treatment regimen until disease progression or intolerable toxicity occurred.

MODEL CONSTRUCTION:

We developed a Markov model utilizing TreeAge Pro software (TreeAge, Williamstown, MA) to perform a cost-effectiveness analysis comparing HAIC-FO and sorafenib, taking into account the perspective of the Chinese healthcare system [14]. The economic evaluation was conducted between July 2023 and February 2024, spanning a 10-year investment horizon. The model included 3 health states: progression-free disease (PFD), progressed disease (PD), and death (Figure 1). Patients in the PFD state received either HAIC-FO or sorafenib until disease progression, unmanageable toxicity, or death. Upon progression or unmanageable toxicity, patients would undergo second-line treatment until death, assuming that within each cycle of the model, patients had the capability to either maintain their assigned health status or transition to another. This presumption was consistent with the dynamic nature of health conditions, where an individual’s health status can undergo changes over time. By integrating the potential for transitions between various health states, the model clarified the likelihood of patients’ health conditions over time.

Cycle-specific transition probabilities were derived from data obtained in the FOHAIC-1 trial. To obtain these probabilities, we initially extracted survival rates using the GetData Graph Digitizer online tool (https://apps.automeris.io/wpd/). Subsequently, we applied the method proposed by Hoyle et al to generate pseudo-individual patient data [15]. These data were then used to fit various survival functions, including Weibull, log-normal, log-logistic, exponential, Gompertz, and generalized gamma models, based on the Akaike information criterion. Ultimately, the log-logistic distribution was determined to be the most suitable function for extrapolating the PFS and OS of the HAIC-FO group. For the sorafenib group, log-normal and Weibull distributions were determined to be the most appropriate functions for PFS and OS, respectively. The monthly transition probabilities were calibrated to optimize alignment with the survival Kaplan-Meier curves observed in the trial (Figure 2). The background mortality rate was obtained from the Chinese life tables.

COSTS:

The model encompassed direct medical expenses, encompassing medication costs, hepatic artery catheterization, subsequent treatment, and the management of adverse events (AEs), as delineated in Table 1. Costs related to initial treatment were sourced from published literature. Dosing calculations assumed a body surface area of 1.72 m2. As observed in the FOHAIC-1 trial, approximately 19.2% of patients in the HAIC-FO group and 31.1% of patients in the sorafenib group received second-line treatment. Second-line therapy options included TACE, HAIC-FO, regorafenib, and nivolumab. The costs associated with these second-line regimens are detailed in Table 1. Costs from previous sources were adjusted to USD 2024 using the Consumer Price Index. The exchange rate used in this study was 1 US dollar=7.24 Chinese yuan (2024).

UTILITIES:

A health state utility of 0.76 was presumed for the PFD state, and a utility of 0.68 was presumed for the PD state, in line with previously published literature [16]. Grade 1–2 minor toxicities were regarded as intrinsic to the cancer state and were excluded from the model. Grade 3–4 AEs with an incidence exceeding 5% were modeled as a one-time disutility.

STATISTICAL ANALYSIS:

To assess cost-effectiveness, we calculated the incremental cost-effectiveness ratio (ICER), which reflects the additional cost per 1 quality-adjusted life years (QALY). In our primary analysis, a willingness-to-pay (WTP) threshold of $30,492/QALY was applied using 3 times the GDP per capita [16]. A discount rate of 3% per annum was applied to adjust for inflation, consistent with the pharmacoeconomics guidelines [13]. A base case analysis was performed using the basic estimates of cost, health utilities, and transition probabilities to ascertain the total cost and QALY for the HAIC-FO and sorafenib cohorts. One-way sensitivity analyses were conducted to investigate the impact of uncertainty on treatment factors, utility, and cost. Moreover, probabilistic sensitivity analysis was executed to explore the stochastic effects of parameter uncertainty on the outcomes, by sampling 10 000 times from the probability distributions of input parameters and running the model for each sampled set of parameter values to generate a distribution of the outcomes. Cost uncertainty was modeled using a Gamma distribution, while utility uncertainty was modeled using a Beta distribution. Additionally, we examined the subgroup of patients with high-risk factors from the FOHAIC-1 trial [9].

Results

BASE CASE RESULTS:

The findings revealed that the HAIC-FO group incurred a total cost of $22,781, whereas the sorafenib group totaled $18,795. Regarding outcomes, the HAIC-FO group attained 1.06 QALYs, while the sorafenib group achieved 0.65 QALYs. Despite incurring an additional cost of $3,985, the HAIC-FO group also yielded an increase of 0.41 QALYs, compared with sorafenib. Consequently, this resulted in an ICER of $9,720 per QALY (Table 2).

SENSITIVITY ANALYSIS:

The tornado diagram depicted in Figure 3 illustrates the top 10 parameters exerting the most significant influence on the ICER. The cost of sorafenib proved to be the most impactful factor. At a sorafenib price of $1,880, the corresponding ICER was $15,605 per QALY; whereas at $2,820, the corresponding ICER was $3,831 per QALY. It is noteworthy that all ICER values remained below the WTP threshold of $30,492 per QALY. Specifically, AE parameters had relatively minimal impact on the ICER. In probabilistic sensitivity analysis, HAIC-FO yielded a mean cost of $18,814±$1,536 and an effectiveness of 0.65±0.05 QALYs, while sorafenib resulted in a mean cost of $22,758±$3,702 and an effectiveness of 1.06±0.17 QALYs. The cost-effectiveness acceptability curve derived from the probabilistic sensitivity analysis (Figure 4) indicated a 99.8% probability that the ICER for HAIC-FO vs sorafenib would fall below $30,492/QALY.

SUBGROUP ANALYSIS:

In patients with high-risk factors, the utilization of HAIC-FO resulted in 0.96 QALYs, compared with 0.49 QALYs for sorafenib. Furthermore, the overall cost projection for HAIC-FO was $19,612, exceeding that of sorafenib at $13,885. Consequently, this led to an ICER of $12,186 per QALY (as shown in Table 3). Therefore, similar to the overall population, HAIC-FO remained cost-effective in the subgroup of patients with high-risk factors, compared with sorafenib.

Discussion

Liver cancer-related healthcare costs constitute a significant portion of China’s overall cancer-related healthcare expenditure, underscoring the importance of judicious allocation of public health resources [17]. In recent years, HAIC-FO has garnered substantial attention from researchers [18]. The clinical benefits of HAIC-FO were observed in the FOHAIC-1 trial [9]. Our study is the first to conduct a cost-effectiveness analysis comparing HAIC-FO with sorafenib as a primary treatment for advanced HCC, drawing on data from the FOHAIC-1 trial. Our analysis demonstrated that HAIC-FO represented a cost-effective treatment approach for advanced HCC, even within the subgroup of patients with high-risk factors, considering the current Chinese health economic system.

Atezolizumab and bevacizumab stand as preferred first-line treatments for advanced HCC. However, the cost-effectiveness of atezolizumab and bevacizumab vs that of sorafenib appears to fluctuate depending on the perspective of individual countries. Several studies have indicated that atezolizumab and bevacizumab may not be cost-effective, compared with sorafenib. Su et al and Zhang et al utilized a partitioned survival model from the US payer perspective, reporting ICERs of $169,223/QALY and $322,500/QALY, respectively [19,20]. Similarly, Chiang et al, also adopting the US payer perspective, employed a Markov model and reported an ICER of $179,729/QALY [21]. From the standpoint of China and the United States, Wen et al reported ICERs of $145,546.21 per QALY in China and $168,030.21 per QALY in the United States [22]. In contrast, recent studies have suggested that atezolizumab and bevacizumab could be a cost-effective option from the perspective of the Chinese healthcare system. Zhao et al found an ICER of $27,630.63/QALY in China [23], while Tseng et al identified an ICER of $75,192 per QALY, which fell below the predefined WTP threshold in Taiwan [24]. The latest research indicates that atezolizumab and bevacizumab may be cost-effective within a Chinese subpopulation. Potential factors contributing to this include the superior efficacy of atezolizumab and bevacizumab in the Chinese subpopulation, compared with the global population, lower standard of care costs, and other relevant considerations.

Several economic evaluations have been conducted on the first-line treatment of unresectable or advanced HCC in China. Among these, 2 studies investigated the cost-effectiveness of sorafenib in combination with other treatments compared with sorafenib monotherapy in patients with advanced HCC [25,26]. Ho et al concluded that the combination of sorafenib with treatments such as surgical resection, percutaneous ethanol injection, TACE, and radiotherapy were not cost-effective options, with an ICER of $89,758 per QALY [25]. Another study showed that sorafenib plus HAIC (SoraHAIC) was not cost-effective compared with sorafenib alone, with an ICER of $77,132/QALY [26]. Additionally, 2 studies performed a cost-effectiveness analysis of oxaliplatin plus infusional-fluorouracil/leucovorin (FOLFOX4) in patients with advanced HCC [27,28]. The results indicated that FOLFOX4 was a cost-effective option compared with sorafenib [27,28]. Cai et al concluded that lenvatinib was a cost-effective alternative to sorafenib in patients with unresectable HCC [29]. Furthermore, 3 economic evaluations focused on immunotherapy for advanced HCC. Peng et al analyzed the cost-effectiveness of sintilimab plus bevacizumab biosimilar compared with that of sorafenib in the treatment of unresectable HCC, revealing an ICER of $23,352/QALY [30]. Two other studies, based on the IMbrave150 trial, concluded that atezolizumab plus bevacizumab was not a cost-effective first-line option compared with sorafenib in patients with unresectable HCC [19,20]. Previously, Chen et al found HAIC-FO was a cost-effective strategy for patients with high-risk factors (the ICER of HAIC-FO was $10190.41/QALYs, compared with sorafenib) [16]. Consistently, our study showed the cost-effectiveness of HAIC-FO vs that of sorafenib in advanced HCC.

As the economy develops and the WTP increases in China [12], HAIC-FO becomes more likely to be cost-effective compared with sorafenib. The sensitivity analysis revealed that the costs of sorafenib and HAIC had the most substantial impact. Expenses related to anticancer treatments can present significant challenges in clinical settings, and patients can experience considerable financial strain if they lack medical insurance coverage. Fortunately, under the trend of medical reform, China has made substantial efforts to address the pressing issue of high financial burdens faced by patients within its healthcare system, as proposed in the “Healthy China 2020” initiative [31]. There are disparities in economic and medical development across different regions in China. In comparison with the use of sorafenib, the use of HAIC demands more advanced medical equipment and highly skilled personnel, thus limiting its availability in certain medical facilities, particularly primary hospitals [12]. The findings of the present study contribute valuable insights for making informed choices regarding treatment regimens from a health economics standpoint. The cost-effectiveness of HAIC-FO has the potential to enhance the fairness of healthcare access. Accessible and efficacious treatment options are essential to ensure that all patients with HCC, regardless of their socioeconomic status, can benefit from the latest advancements in cancer therapy [24]. This approach may aid in reducing health disparities within the HCC population.

However, our study has several limitations that warrant consideration. First, the WTP threshold in China was not explicitly defined, prompting us to adopt the approach of using 3 times the GDP per capita based on WHO recommendations. Second, our study directly compared HAIC-FO with sorafenib based on the FOHAIC-1 trial, without conducting an indirect comparison with other potential first-line treatments for advanced HCC, due to the absence of convincing and robust head-to-head trial data. Third, our model did not encompass the cost of all AEs, due to their low incidence and negligible cost. However, we did conduct a sensitivity analysis on grade 3–4 AEs, demonstrating minimal impact on the ICER. Although AEs had a minor impact on the ultimate cost-effectiveness, they still need to be emphasized, as they could affect patients’ healthcare experiences and treatment adherence. Fourth, the health utility values employed in our study were derived from data sources other than patients enrolled in the FOHAIC-1 trial, potentially introducing bias into the results. Sensitivity analyses were also conducted for these parameters. Finally, this study was based on the healthcare context in China. Despite these limitations, our study offers valuable insights into the current treatment landscape for advanced HCC patients in China and provides a certain reference value for Chinese clinicians and policymakers.

Conclusions

In conclusion, our modelling analysis, from the perspective of the Chinese health system, indicated that HAIC-FO offers a health economic benefit, compared with sorafenib, in the overall population of patients with advanced HCC and the subgroup of patients with high-risk factors.