Relacorilant plus nab-paclitaxel for recurrent, platinum-resistant ovarian cancer: a cost-effectiveness study

Abstract

Objective

The phase 2, open-label, randomized, 3-arm study (NCT03776812) found that intermittently dosed relacorilant + nab-paclitaxel improved progression-free survival, duration of response, and overall survival compared with nab-paclitaxel monotherapy with minimal additional toxicity. This study analyzed the cost-effectiveness of intermittent relacorilant + nab-paclitaxel (IN), continuous relacorilant + nab-paclitaxel (CN) and nab-paclitaxel monotherapy (N) from the payer’s perspective over a 5-year time horizon.

Methods

The health outcome is expressed in quality-adjusted life years (QALYs). IN, CN and N were evaluated using QALYs and incremental cost-effectiveness ratio (ICER). The data for this model come from the NCT03776812 trial and other published literature. The impact of the variables is studied using a one-way deterministic sensitivity analysis and a probabilistic sensitivity analysis based on a second-order Monte Carlo simulation.

Results

N was the least costly strategy, at $ 4606.05, followed by IN ($22,597.75) and CN ($44,276.86). Based on a willingness-to-pay threshold of $100,000/QALY, IN was cost-effective compared with N, with an ICER of $21,418.69 per QALY gained for N, whereas CN was ruled out by extended dominance (less effective, more costly), compared with N. The incremental benefits of IN compared to CN and N were 0.72 QALYs and 0.84 QALYs.

Conclusion

From a US health care system perspective, IN may be cost-effective compared to CN for patients with recurrent platinum-resistant ovarian cancer, and IN is also better than CN and N in terms of efficacy. Therefore, IN is a high-quality regimen for clinicians to treat patients with recurrent platinum-resistant ovarian cancer.

Trial Registration

ClinicalTrials.gov Identifier: NCT03776812

Synopsis

Intermittent relacorilant + nab-paclitaxel (IN) for recurrent platinum-resistant ovarian cancer is cost-effective and improves survival and tolerability while reducing adverse effects in the US healthcare system. Sensitivity analyses confirmed its robustness, making IN the best choice to optimize clinical efficacy and economic benefit.

INTRODUCTION

Globally, ovarian cancer (OC) is the 11th most common cancer among women in the USA and 6th deadliest, with 19,680 new diagnoses and 12,740 deaths forecast in 2024 alone [1,2]. In addition, early diagnosis of OC is particularly challenging, as it has few clinical symptoms in the early stages and is prone to rapid progression to advanced stage [3]. First-line treatment for advanced OC consists of tumor resection and platinum-based chemotherapy, and while platinum-based chemotherapy is effective for most patients with OC, their response to platinum therapy is unknown until chemotherapy is completed, and some patients may develop resistance to it [4]. Platinum tolerance in OC may be closely related to genomic expression, dysregulation of drug influx and efflux pathways, DNA repair, epigenetic alterations, and tumor microenvironment [5]. The platinum-free interval (PFI) is a reliable indicator of treatment efficacy and patient prognosis because it can assess whether OC patients respond to platinum drugs and their recurrence. The Gynecologic Cancer Intergroup (GCIG) classifies the response to platinum chemotherapy according to the duration of PFI into 4 categories (platinum-refractory: <1 month, platinum-resistant: 1–6 months, partial platinum response: 6–12 months, platinum response: >12 months) [6]. About 20% to 30% of OC patients will relapse or progress within 6 months after completing platinum-based chemotherapy. This state is called recurrent platinum-resistant OC. The median overall survival (OS) of these patients is only 12–18 months [7].

The main treatment for platinum-resistant OC is single-agent, non-platinum chemotherapy, including polyethylene glycol liposome adriamycin, gemcitabine, topotecan and paclitaxel. The response rate reported in recent trials for platinum-resistant OC was 10% to 15%, and the median progression-free survival (PFS) was approximately 3.5 months [8]. For paclitaxel, due to its low water solubility, polyethylene castor oil and ethanol are usually used as solvents for administration, which causes hypersensitivity reactions in about 10% of cases [9]. Compared with solvent-based paclitaxel, nab-paclitaxel uses human serum albumin as a carrier. It is characterized by the following: no need for anti-allergic treatment before administration, short infusion time, higher dose administration, higher serum concentration of paclitaxel, and effective delivery of paclitaxel to tumor tissue through endocytosis, resulting in high concentrations in the tumor [9,10]. In clinical practice, nab-paclitaxel demonstrates superior efficacy compared to conventional formulations, particularly due to its enhanced drug delivery and reduced toxicities, making it a valuable treatment option for platinum-resistant/refractory disease [11]. Nab-paclitaxel shows promising efficacy in OC treatment by achieving higher intra-tumoral concentration, improved response rates, and a favorable safety profile [12]. Compared with conventional paclitaxel, nab-paclitaxel enhances the efficacy of breast cancer treatment by increasing intra-tumor drug concentration and retention, thereby increasing the efficiency of breast cancer treatment and prolonging PFS, and has been approved by the U.S. Food and Drug Administration for the treatment of metastatic breast cancer (2008) and non-small cell lung cancer (2011) [13,14]. Meanwhile, nab-paclitaxel has also shown better efficacy and safety than solvent-based paclitaxel in gastrointestinal tumors [15]. Glucocorticoids can cause chemotherapy resistance by inhibiting the apoptosis pathways (BCL2 and FOXO3a) that cytotoxic agents, including paclitaxel, rely on, and reduce the efficacy of cytotoxic drugs. Relacorilant is a selective glucocorticoid receptor modulator that can compete with glucocorticoids, reverse the anti-apoptotic effects of glucocorticoids to restore chemosensitivity and enhance the efficacy of platinum and paclitaxel [16]. A 3-arm, randomized, controlled, open-label phase II study (NCT03776812) in patients with recurrent platinum-resistant OC showed that intermittent relacorilant + nab-paclitaxel (IN) significantly improved OS in patients with non-primary platinum-resistant recurrent OC compared to nab-paclitaxel alone, and the regimen is undergoing phase III trials [17].

However, new treatment options are often more costly than previous standard options, so evaluating the cost-effectiveness of new treatments is critical to providing policymakers, providers and patients with reliable evidence on the value of implementing new treatment interventions. Therefore, in this study, we compared IN with continuous relacorilant + nab-paclitaxel (CN), and IN with nab-paclitaxel monotherapy (N), from the perspective of the medical system, to evaluate the cost-effectiveness and safety of recurrent platinum-resistant OC. This study uses a Markov model to perform a cost-effectiveness analysis, using clinical data from the NCT03776812 trial and other published literature on drug prices and utility values to evaluate quality-adjusted life years (QALYs) and incremental cost-effectiveness ratio (ICER) of patients to make better decisions.

MATERIALS AND METHODS

1. Compliance with ethics guidelines

The models used in this analysis are based on previously conducted studies and other economic models, and no human participants or animals were studied by any of the authors.

2. Model building

We used TreeAge Pro Suite 2022 (TreeAge Software, Williamstown, MA, USA) to construct decision trees and Markov models to analyze the cost-effectiveness of IN versus CN and IN versus N for treating patients with platinum-resistant recurrent OC. We designed the patient population based on the clinical trial (NCT01945775) to include eligible patients in our model and were randomly assigned to receive treatment with either IN, CN, or N for a cost-effectiveness analysis using overall OS, PFS, duration of remission (DOR), and relative risk (RR) as metrics. After receiving treatment, patients go through 4 health states: stable disease (SD), progressive disease (PD), remission (RE), and death (DE) until they enter the final state of death, which spans a period of 5 years (Fig. S1). Patients accumulate costs and QALYs during each period in a specific health state. Due to the time value of money, the discount rate has a significant impact on cost calculation. In the baseline analysis, a 5% discount rate is used for calculating costs, and a willingness-to-pay (WTP) threshold of $100,000 per QALY is adopted [18]. The study set the entire cycle at 28 days, which corresponds to the duration of one treatment cycle. The results are presented as an ICER, the cost per QALY gained. To perform a probabilistic sensitivity analysis (PSA), we simulated the model using second-order Monte Carlo analysis (1,000 simulations) and constructed confidence intervals (1,000 replications) by bootstrap replication.

3. Patients and treatment plans

This trial (NCT03776812) included 178 patients as subjects. Inclusion criteria: 1) Female, age 18 years and older; 2) Pathological examination to high-grade serous or endometrioid OC, primary peritoneal cancer or fallopian tube cancer, or OC sarcoma; 3) Has received at least first-line treatment and has shown platinum resistance with disease progression within 6 months of the last platinum-based therapy; or has shown platinum refractory disease with disease progression during or immediately after the first platinum-based therapy [17]. It was divided into 3 groups, with 60 people in the IN program group, 58 people in the CN program group and 60 people in the N program group. For IN, patients took 150 mg of relacorilant orally the day before, on, and the day after the treatment days (days 1, 8, and 15) and concurrently received an intravenous infusion of nab-paclitaxel at 80 mg/m² on these days. For CN, patients consistently took 100 mg of relacorilant daily and received an intravenous infusion of nab-paclitaxel at 80 mg/m² on days 1, 8, and 15 of each 28-day cycle. N is to receive nab-paclitaxel 100 mg/m² injection on the 1st, 8th, and 15th day of each cycle.

4. Model inputs for transition probabilities

Model Probability Markov state model transition probability refers to the probability that a patient will move from one of four states to another during a treatment cycle in a pharmacoeconomic evaluation. This study uses the DEALE method to convert the length of time into a rate index, and then converts the rate index into a probability, analyzes the incidence rate and various survival data, and uses a simple exponential function to approximate life expectancy [19]. The 4 clinical efficacy indicators RR, OS, PFS, and DOR are important bases for calculating the probability of metastasis. The data required for the probability of metastasis come from the NCT03776812 experiment, and the calculation method is shown in Table 1 [17,20,21,22,23].

Table 1. The key parameters of platinum-resistant ovarian cancer model with IN, CN and N.

Variable		Formula	Best estimate	TP range	Distribution	Reference
IN						[17]
	RR		0.60	0.54–0.66
	OS		13.9	12.51–15.29
	PFS		5.60	5.04–6.16
	DOR		5.55	5.00–6.11
	Stable → stable (Iss)	1-Isp-Isr	0.46	0.414–0.506	Beta
	Stable → remission (Isr)	1-exp (−RR/3)	0.18	0.162–0.198	Beta
	Stable → progress (Isp)	Irp*4	0.36	0.324–0.396	Beta
	Remission → remission (Irr)	1-Irp	0.91	0.819–1.00	Beta
	Remission → progress (Irp)	1−exp{−0.75*In(2)/(DOR)}	0.09	0.081–0.099	Beta
	Progress → progress (Ipp)	1−Ipd	0.94	0.846–1.034	Beta
	Progress → death (Ipd)	1−exp{−0.75*In(2)/(OS−PFS)}	0.06	0.054–0.066	Beta
CN						[17]
	RR		0.53	0.48–0.58
	OS		11.3	10.17–12.43
	PFS		5.30	4.77–5.83
	DOR		3.79	3.41–4.17
	Stable → stable (Css)	1−Csp−Csr	0.32	0.288–0.352	Beta
	Stable → remission (Csr)	1−exp (−RR/3)	0.16	0.144–0.176	Beta
	Stable → progress (Csp)	Crp*4	0.52	0.468–0.572	Beta
	Remission → remission (Crr)	1−Crp	0.87	0.783–0.957	Beta
	Remission → progress (Crp)	1−exp{−0.75*In(2)/DOR)}	0.13	0.117–0.143	Beta
	Progress → progress (Cpp)	1−Cpd	0.92	0.828–1.012	Beta
	Progress → death (Cpd)	1−exp{−0.75*In(2)/(OS−PFS)}	0.08	0.072–0.088	Beta
N						[17]
	RR		0.69	0.62–0.76
	OS		12.2	10.89–13.42
	PFS		3.8	3.42–4.18
	DOR		3.65	2.89–5.09
	Stable → stable (Nss)	1−Csp−Csr	0.27	0.243–0.297	Beta
	Stable → remission (Nsr)	1−exp(−RR/3)	0.21	0.189–0.231	Beta
	Stable → progress (Nsp)	Crp*4	0.52	0.468–0.572	Beta
	Remission → remission (Nrr)	1−Crp	0.87	0.783–0.957	Beta
	Remission → progress (Nrp)	1−exp{−0.75*In(2)/DOR)}	0.13	0.117–0.143	Beta
	Progress → progress (Npp)	1−Cpd	0.94	0.846–1.034	Beta
	Progress → death (Npd)	1−exp{−0.75*In(2)/(OS−PFS)}	0.06	0.054–0.066	Beta
	Discount rate for costs and QALYs		5% per year
	Health state utilities
	Utility of PFS		0.84	0.672–1.008	Beta	[20]
	Utility of PD		0.79	0.632–0.948	Beta	[20]
	Disutility of leukopenia		0.09	0.072–0.108	Beta	[21]
	Disutility of fatigue		0.17	0.136–0.204	Beta	[22]
	Body weight (kg)		70	56–84	Normal	[23]
	Body surface area (m2)		1.67	1.50–1.84	Normal	[23]

RR=(OS−PFS)/OS; OS refers to the time of patient death from the randomized grouping value; PFS refers to the time from when the subject enters the experiment to when the tumor progresses or dies; DOR is the time from the first assessment of CR or PR to the first assessment of progression or death.

CN, continuous relacorilant + nab-paclitaxel; DOR, duration of remission; IN, intermittent relacorilant + nab-paclitaxel; N, nab-paclitaxel monotherapy; OS, overall survival; PD, progressive disease; PFS, progression-free survival; QALY, quality-adjusted life year; RR, relative risk; TP, test probability.

5. Model inputs for utility

In the field of health care, the health utility value is a key indicator used to assess the degree to which different treatments satisfy a particular health condition and the individual’s preference for that condition. It indicates the relative importance of a specific health state relative to a completely healthy state. It is an important indicator for evaluating the satisfaction of treatment results and is also a comprehensive index reflecting the overall health status of an individual [24]. The value ranges from 0 to 1, with 0 representing death and 1 representing complete health. See Table 1 for data on health utility values.

6. Model inputs for costs

The costs considered include drug costs, hospitalization and administration costs, concomitant medication costs, serious adverse events and management costs, and this study considers the cost of follow-up testing of PFS status and is calculated throughout the treatment process. Calculate the total cost of treatment and obtain the ICER. To calculate the drug dosage for a drug, we assume a typical patient surface area of 1.67 m². According to clinical data, a total of four common adverse events were included. The cost of these adverse events (AEs) was calculated by multiplying the estimated incidence of each AE by the corresponding unit treatment cost. The prices of various medical expenses were derived from relevant public literature, local hospitals, and pharmaceutical information websites. The cost-related data is shown in Table S2 [23,25,26]. The cost of Relacorilant was calculated using the median price of four glucocorticoid receptor-related therapies: mifepristone, osilodrostat, ketoconazole, and metyrapone, following the median price approach.

7. Sensitivity analysis

We performed a one-way deterministic sensitivity analysis (DSA) to assess the robustness of our results to assumptions about the response rate, toxicity, cost, and survival rate associated with each drug [27]. We studied the impact of each parameter on the ICER by varying all uncertain parameters, such as the probability of transfer and the discount rate, within ±10% or 95% confidence intervals (CIs) of their base values, and determined the 5 parameters that had the greatest impact on the results and whether their changes affected our decision-making on chemotherapy regimens. The results of the single-factor DSA are presented in the form of a tornado diagram. The various bars displayed in the tornado diagram represent the model output results obtained through the parameters within a certain range. The longer the cylindrical bar, the greater the impact on the result.

We also conducted a PSA to express the range of variation in ICERs by assuming a distribution of parameters, which can reduce the potential bias caused by the analyst’s determination of parameters and their range of variation to a certain extent [28]. We determine the most suitable distribution for each parameter based on its type, using the beta distribution to represent uncertainty in utility, probability and proportion, and the gamma distribution to represent uncertainty in cost. A Monte Carlo simulation with 1,000 iterations is performed by using the pre-specified distribution and changing all parameters at the same time. The PSA results are presented in the form of a probability scatter plot and a cost-effectiveness acceptability curve to visually display the ratio of cost to effect of the 2 options under the premise of satisfying the WTP. However, sometimes there are negative QALYs and negative costs, which makes it difficult to interpret the average and median ICERs.

Finally, we also conducted a net monetary benefit (NMB) analysis and plotted a cost-effectiveness acceptance curve based on the NMB analysis results, because the NMB is more effective if there are negative ICERs. The cost-effectiveness threshold for NMB analysis settings is US $100,000/QALY.

8. Ethics approval and consent to participate

This article is based on prior research and does not involve any new studies with human participants or animals conducted by the authors. Therefore, it does not require approval from an independent ethics committee.

RESULTS

1. Base-case treatment analyses

Our model calculated the cumulative lifetime costs, QALYs, incremental QALYs, incremental costs, ICERs, and mortality rates for IN, CN and N. The Markov model calculated four distribution proportions for the IN scenario: SD (0.0%), PD (1.7%), RE (30.4%), and DE (67.9%). The four distribution ratios for the CN and N scenarios are SD (0.0%), PD (0.0%), RE (0.0%), and DE (100.0%). The Markov probability analysis is shown in Fig. S2.

Under our baseline assumptions, the model results obtained indicate that IN is the best strategy for the treatment of metastatic OC. IN not only has a lower average price of $21,679.11 than CN, but IN was estimated to result in an incremental 0.73 QALYs. Applying incremental analysis principles, IN compared to CN has an ICER of $30109.88/QALY. Compared to N, IN has a certain price advantage, but IN will generate an incremental 0.84 QALYs. Since the WTP set in this study is 100,000 USD, IN is the optimal plan for this study (Table 2, Fig. 1).

Table 2. Cost and effect of 3 treatment options within 5 years.

Strategy	Cost	Incr. cost	PFS/%	DE/%	QALY/year	Incr. QALY	ICER
IN	22,597.75	0	5.60	67.9	2.01	0	0
CN	44,276.86	21,679.11	5.30	100	1.29	−0.72	−30,109.88
N	4,606.05	−17,991.7	3.80	100	1.17	−0.84	21,418.69

CN, continuous relacorilant + nab-paclitaxel; DE, death; ICER, incremental cost-effectiveness ratio; IN, intermittent relacorilant + nab-paclitaxel; N, nab-paclitaxel monotherapy; PFS, progression-free survival; QALY, quality-adjusted life year.

Fig. 1. Results of the cost-effectiveness analysis of platinum-resistant recurrent ovarian cancer IN versus CN and IN versus N. The vertical axis shows the cumulative cost of life, and the horizontal axis shows the QALYs gained.

(A) Cost-effectiveness analysis of platinum-resistant recurrent ovarian cancer IN versus CN. (B) Cost-effectiveness analysis of platinum-resistant recurrent ovarian cancer IN versus N.

CN, continuous relacorilant + nab-paclitaxel; IN, intermittent relacorilant + nab-paclitaxel; N, nab-paclitaxel monotherapy; QALY, quality-adjusted life year.

2. General safety

IN has better tolerance than CN, and most AEs have a lower incidence in IN than in CN. Grade 3 to 4 anemia was recorded in 8 cases (13.3%) in IN, which was the most common and serious AE of IN, while 11 patients (19.3%) in CN recorded grade 3 to 4 anemia. The tolerability of IN was similar to that of N, with 1 patient (1.7%) in IN and 1 patient (1.7%) in N recording nausea of grade 3–4. All levels of abdominal discomfort had 25 cases (41.7%) in both IN and N, and the other adverse reactions were similar in both groups. The incidence of adverse reactions is shown in Table S1.

3. Sensitivity analyses

Different parameters have different effects on the model results. A single-factor sensitivity analysis shows that the five most influential parameters for IN compared to CN are: the probability of CN in the initial state of progressive phase (CPpd), the probability of IN in the initial state of progressive phase (IPpd), the probability of CN in the initial state of stable phase (CPsd), the transition probability of CN from stable to stable (Css), and the transition probability of IN from stable to stable (Iss). The five most influential parameters for comparing IN to N are: the transition probability of IN from remission to remission (Irr), IPpd, Iss, the probability of IN in the initial state of stable phase (IPsd), and the transition probability of IN from stable to remission (Isr). Other parameters have little effect on the robustness of the model. When the threshold is set at $100,000/QALY, no factor affects the ICER result within the sensitivity analysis range. This means that the model provides consistent results and can be considered a stable model, and the IN scenario is still our preferred scenario (Fig. 2).

Fig. 2. Tornado diagram. The tornado diagram shows the incremental cost per QALY gained in the single-factor sensitivity analysis for IN versus CN and IN versus N. The width of the bars shows the range of results when the variable is varied within the sensitivity analysis. The vertical dotted line shows the result for the base case.

(A) The tornado diagram shows the incremental cost per QALY gained in the single-factor sensitivity analysis for IN versus CN. (B) The tornado diagram shows the incremental cost per QALY gained in the single-factor sensitivity analysis for IN versus N.

CN, continuous relacorilant + nab-paclitaxel; IN, intermittent relacorilant + nab-paclitaxel; N, nab-paclitaxel monotherapy; QALY, quality-adjusted life year.

In the probabilistic sensitivity analysis, 1,000 Monte Carlo simulations were randomly performed for each parameter within the determined range of variation, and the scatter plot (Fig. 3) was obtained. The horizontal axis represents QALYs obtained, the vertical axis represents incremental costs, and the diagonal line represents the WTP threshold. Each scatter plot represents ICER of the three treatment options for patients with platinum-resistant recurrent OC. As can be seen from the figure, 100.00% of the generated scenarios (represented by green dots) are more inclined to IN than CN. Compared with N, we still prefer IN in most generated scenarios. The ICER of all cases in IN is negative, indicating absolute effectiveness. None of the ICERs exceeds the WTP threshold, so IN has a relative advantage over CN. In addition, as can be seen in Fig. 3, the scatter points are concentrated within the ICER (ellipse), indicating that the ICER analysis results for IN are relatively stable.

Fig. 3. Scatter plot. Probability results for the incremental cost-effectiveness difference between IN versus CN and IN versus N treatment in 1,000 breast cancer patients.

(A) Probability results for the incremental cost-effectiveness difference between IN versus CN. (B) Probability results for the incremental cost-effectiveness difference between IN versus N.

CN, continuous relacorilant + nab-paclitaxel; IN, intermittent relacorilant + nab-paclitaxel; N, nab-paclitaxel monotherapy.

In the NMB analysis, 1,000 simulations were performed to obtain 2 cost-effectiveness curves for platinum-resistant recurrent OC patients, one comparing IN with CN and the other comparing IN with N. In the first comparison, within the WTP range, the probability of IN being acceptable is always much higher than that of CN within the scope of our analysis, regardless of how the WTP changes. In the comparison between IN and N, when the WTP is greater than $25,000, the level of the IN acceptable curve will exceed that of N. Since we set the WTP at $100,000, IN has the highest probability of being acceptable in both comparisons and is the optimal treatment for platinum-resistant recurrent OC (Fig. 4).

Fig. 4. The curves show the probability of obtaining net income for each strategy for different WTP thresholds in 2 groups of comparisons: IN versus CN and IN versus N. The vertical axis shows the probability of cost-effectiveness. The horizontal axis shows the WTP threshold for an additional QALY.

(A) The probability of obtaining net income for each strategy for different WTP thresholds in 2 groups of comparisons: IN versus CN. (B) The probability of obtaining net income for each strategy for different WTP thresholds in 2 groups of comparisons: IN versus N.

CE, cost-effectiveness; CN, continuous relacorilant + nab-paclitaxel; IN, intermittent relacorilant + nab-paclitaxel; N, nab-paclitaxel monotherapy; QALY, quality-adjusted life year; WTP, willingness-to-pay.

DISCUSSION

OC is the deadliest gynecological malignancy in high-income countries [29]. More than 50% of patients with advanced OC who achieve complete remission after chemotherapy will still relapse and die, due to resistance to chemotherapy, especially platinum resistance [30]. In this study, we performed the first cost-effectiveness analysis of IN versus CN and IN versus N as treatment options for platinum-resistant recurrent OC based on the latest research data, with the aim of determining the long-term costs and long-term effects of the three treatment strategies for platinum-resistant recurrent OC. The results show that IN (22,597.75 USD) is more cost-effective than CN (44,276.86 USD). Compared with CN, the use of IN can generate 0.72 QALY/year of additional utility. Although IN is not cost-effective compared to N, it can generate 0.84 QALY/year more utility under the condition of a WTP of 100,000 US dollars, so IN is more cost-effective than N. The results of the one-way DSA and PSA at the 100,000 WTP threshold show that our findings are reliable. The Monte Carlo simulation produced 1,000 scenarios, which showed that IN is more cost-effective.

Most patients with advanced epithelial OC will die within three years of developing platinum resistance, even if they initially respond well to chemotherapy [31]. Studies have shown that high glucocorticoid receptor (GR) expression is significantly associated with chemotherapy resistance and increased recurrence. It inhibits chemotherapy-induced cell death by regulating the transcription of anti-apoptotic proteins, shortening the PFS of patients and reducing chemotherapy sensitivity, regardless of BRCA mutation status [32,33,34]. Neutropenia was the more frequent adverse reaction in the 3 groups, so granulocyte colony-stimulating factor (G-CSF) agents were used in all cases within the relacorilant + nab-paclitaxel group for primary prevention in NCT03776812 [17], thus reducing the neutropenia frequency [16]. We considered the cost of G-CSF among the adverse effects of Neutropenia, and the study showed that the incidence of neutropenia IN the IN group was the lowest of the 3 groups, which further supports the superiority of IN over CN and n as a preferred option [25,35]. Relacorilant is an orally administered selective GR modulator that restores sensitivity to chemotherapy and enhances its efficacy by effectively binding to GR, inhibiting GR in cells, and not binding to androgen or progesterone receptors [16,36]. Studies have also shown that relacorilant can inhibit the anti-apoptotic genes that are upregulated by GC and restore the anti-apoptotic genes induced by paclitaxel [32]. Patients with recurrent platinum-resistant OC have few effective options. The emergence of relacorilant 1, an albumin-bound paclitaxel drug, provides patients with more options and may have better response rates and survival rates. However, these new drugs are often expensive, and in today’s era of escalating medical costs, the cost of these drugs has become an important consideration. This article clearly shows that IN has better efficacy and lower cost than CN and N, improves patient PFS, DOR and OS, and has less additional toxicity [17]. Studies have also compared the effects of intermittent and continuous administration of prostate cancer treatments, and the results show that intermittent therapy is better at reducing side effects and improving patients’ quality of life. Its potential advantages in terms of improving the killing of cancer cells and reducing serious side effects provide scientific evidence for comparing the effects of intermittent and continuous administration strategies in clinical practice and explaining the biological mechanisms [37].

In this study, we introduced duration of response as a new evaluation index in addition to the traditional OS and PFS analyses. Duration of response refers to the time from the first evaluation of complete response (CR) or partial response (PR) to PD or death from any cause. This index reflects the long-term benefits of the drug to patients and is a key indicator for evaluating the long-term efficacy of the drug. In previous pharmacoeconomic studies, few have used duration of response as an analysis indicator. The introduction of this indicator can help to make treatment decisions more accurately.

However, there are some limitations to our study: 1) The health utility values and some key model inputs in this study are mainly based on specific clinical trial data (NCT01945775), which may not fully represent the reality of all regions or populations, which limits the general applicability of the results. Any small deviation in this experiment may lead to a deviation in our research results [38]. 2) Model construction often relies on fixed assumptions, such as treatment cycles and metastasis probabilities, and inaccuracies in these assumptions can lead to errors in the results. Although Monte Carlo simulation increases the credibility of the results, model uncertainty still exists. Univariate sensitivity analysis can help identify key parameters, but it cannot capture the interactive effects of simultaneous changes in multiple parameters, so it may limit the interpretability of the model output [39]. 3) The regional dependence of costs limits the generalizability of the results. Due to the use of specific exchange rates and regional prices, the cost estimates may not be applicable to all countries or regions. Therefore, when these results are applied to different economic contexts, the differences in economic and medical systems should be considered carefully and necessary adjustments should be made [40].

Overall, this study is the first to analyze the cost-effectiveness of IN, CN, and N for the treatment of platinum-resistant recurrent OC. We established a Markov decision tree model and applied Markov cohort analysis to determine that IN is more cost-effective for patients with platinum-resistant recurrent OC than CN and N from a US social perspective. In terms of efficacy, IN is significantly better than CN and N. Therefore, IN is the preferred regimen. Our analysis will help clinicians make the best treatment decisions for patients with platinum-resistant recurrent OC. The study of this economic evaluation suggests that with an ICER less than $100,000 per QALY gained, IN in eligible patients represents a good value at its 2023 price.

SUPPLEMENTARY MATERIALS

Table S1

Adverse events

Table S2

Model input for costs

Fig. S1

TreeAge and Markov model for patients with advanced, platinum-resistant/refractory ovarian cancer.

Fig. S2

Markov cohort analysis. These curves show the results of IN, CN and N models. The horizontal axis shows the time (in years) and the vertical axis shows the proportion of people.