Adebrelimab plus chemotherapy vs. chemotherapy for treatment of extensive-stage small-cell lung cancer from the US and Chinese healthcare sector perspectives: a cost-effectiveness analysis to inform drug pricing

Purpose: The aim of this study was to evaluate the cost-effectiveness of a recently approved first-line therapy (adebrelimab plus chemotherapy vs. chemotherapy alone) for patients with extensive-stage small-cell lung cancer (ES-SCLC) in the US and China, and to estimate the reasonable range of adebrelimab price from the decision-makers. Methods: Several partitioned survival models were built to compare the cost and effectiveness of adebrelimab plus chemotherapy vs. chemotherapy alone over a 10-year time horizon. Clinical efficacy and safety data were extracted from the CAPSTONE-1 trial. Costs and utilities were obtained from previously published studies. Sensitivity, scenario and subgroup analyses were performed to explore the uncertainty of the model outcomes. Price simulation was conducted at three thresholds of willingness-to-pay (WTP), including WTP of $100,000 in the US and of $37,422 in China, 0.5WTP of $50,000 in the US and of $18,711 in China, and 1.5WTP of 150,000 in the US and of $56,133 in China. Findings: Base-case analysis at $1382.82/600 mg of adebrelimab price indicated that adebrelimab plus chemotherapy would be cost-effective in the US at the WTP threshold of $100,000, but not in China at the WTP threshold of $37,422. If PAP was taken into account, the regimen would be cost-effective in China at the given WTP. The results of price simulation indicated that adebrelimab plus chemotherapy was completely favored in the US if adebrelimab price was less than $8894.98/600 mg (total quality-adjusted life years [QALYs] were calculated with progression-based utility [PB-utility]) or $8912.51/600 mg (total QALYs were calculated with time-to-death utility [TTD-utility]) at the WTP threshold of $100,000; if adebrelimab price was reduced by at least $202.03/600 mg (total QALYs were calculated with PB-utility) or $103.06/600 mg (total QALYs were calculated with TTD-utility), the regimen was also cost-effective in China without PAP at the WTP threshold of $37,422. The above results were stable in the sensitivity analyses. Subgroup analysis found that the subgroup with better survival benefits tended to have a higher probability of cost-effectiveness, which was also associated with adebrelimab price. Implications: First-line adebrelimab plus chemotherapy represented a dominant treatment strategy comparing with chemotherapy alone in the US and also did in China with PAP at $1382.82/600 mg of adebrelimab price. Decision-makers could benefit from pricing strategy provided by this study in making optimal decisions. More evidences were needed to verify and improve the results.

cytologically confirmed ES-SCLC, without previously systemic treatment, with an Eastern Cooperative Oncology Group (ECOG) performance status score of 0 or 1, assumed area under the concentration curve of 5 mg/mL/min, serum creatinine of 1, weight 65 kg for Chinese and 70 kg for American, and body surface area (BSA) 1.72 m 2 for Chinese and 1.79 m 2 for American.

Costs of subsequent treatment regimens
The CAPSTONE-1 reported about 30 subsequent treatment regimens, but most regimens were used by few patients, and some price information for drugs cannot be accessed through public databases. Additionally, the CAPSTONE-1 trial was conducted in China, and some drugs used in the study had not yet been approved for sale in the US, resulting in missing price information. To reduce the uncertainty in the estimation of adebrelimab price due to this portion of uncertain information, we only included drugs with a usage percentage of 10% or greater in the models. The dosage and frequency of each drug were based on the drug monograph (eTable 2).
The drugs usage percentages included their usage percentages in monotherapy and combination therapies. The usage percentages of other drugs in combination therapy were relatively low and had not been included. We assumed that the costs of treating progressive disease (PD) were the total costs of using each of the abovementioned drugs for subsequent treatment, and the cost of the drug was equal to the cost of the drug multiplied by the patient percentage who used it. Only 67.53% of patients in the study reported subsequent treatment, and the specific treatment regimens for the other patients were unknown. We assumed that these patients received best supportive care (BSC) (38.70% in adebrelimab group and 26.29% in chemotherapy group). Since the CAPSTONE-1 did not report the duration of subsequent treatment, we assumed that patients received the above-mentioned drugs from PD until 1 month before death.

Cost of each drug per cycle
The cost of the drug per cycle was equal to the product of the unit packaging cost of the drug and the minimum package number. The minimum package number should be the smallest positive integer that contains the average dosage of the drug. For carboplatin, for example, the average dosage used by patients with adebrelimab plus chemotherapy was 485mg per cycle, and the minimum package number required was 10 (9 < 485 50 < 10). Therefore, the cost of carboplatin for this portion of patients each cycle was $26.7 ($2.67 × 10 = $26.7). The costs of placebo-related drug acquisition and administration were not included.

Treatment duration of drugs
Changes in the duration of regimens with different costs may affect the calculation of total treatment-related costs. According to the protocol, patients could continue to receive adebrelimab or placebo for up to 2 years, unless disease progression or intolerable drug toxicity was observed. However, the median cycle number of adebrelimab or placebo reported in the trial was 8 cycles. To reduce the impact of uncertain assumptions about the treatment duration on results, we simultaneously considered the treatment duration for 2 years and 8 cycles. The following assumptions were made in the base-case analysis: (1) model 1: patients recovered after receiving 8 cycles of adebrelimab or placebo, or discontinued treatment due to intolerable drug toxicity, or disease progression, or death; (2) model 2: patients recovered after receiving adebrelimab or placebo treatment for 2 years, or discontinued treatment due to intolerable drug toxicity, or disease progression, or death.

Costs of follow up
The costs of follow-up mainly consisted of two aspects: laboratory tests and imaging exams. Laboratory tests, including electrocardiograms, haematological examination, liver and renal function, coagulation, and thyroid function tests, were performed every 6 weeks. Imaging exams, including CT or MRI, were performed every 6 weeks for the first year, and then every 9 weeks after that. We assumed that all patients were followed up at the above-mentioned frequency until disease progression or discontinued treatment. After the first year, the follow-up frequency was changed to every 9 weeks for the next 2 years, every 3 months for the next 3 years, every 6 months for the next 3 years, and then annually after that.

Methodology of external validation and the final selection of extrapolation models
The survival results of long-term follow-up were not provided in the CAPSTONE-1, which would be obtained from the fitted models. To ensure the credibility of models fitting results, we considered selecting models by two steps. Firstly, the best survival model was selected based on the minimum akaike information criterion (AIC) and maximum Log likelihood. Then, external data were used to verify the clinical rationality of models' extrapolation results and to calibrate the models' selection. In this case, mean squared errors (MSE) were chosen as the evaluation index, the smaller the MSE value, the better the model performance. CASPIAN had the longest follow-up period in published clinical studies related to immunotherapy (27 months of PFS and 42 months of OS), and it was selected as an external data source. We evaluated the rationality of PFS extrapolation results within the range of 6 to 27 months, and the OS extrapolation results within the range of 15 to 42 months.
For the OS data of patients with adebrelimab plus chemotherapy, we selected "RP(royston-parmar spline)-hazard-1" as the survival model, which had no significant difference in fitting and extrapolation performance comparing with the best model (RCS1), but had significant advantages in the clinical rationality of extrapolation results. The other three groups selected the survival models with the best fitting and extrapolation performance, and the extrapolation results generated by them all had good clinical rationality.