Cost-effectiveness of adding quizartinib to induction chemotherapy for patients with FLT3-mutant acute myeloid leukemia

Abstract

The FLT3 inhibitor quizartinib has been shown to improve overall survival when added to intensive induction chemotherapy (“7+3”) in patients 18–75 years old with newly diagnosed AML harboring a FLT3-ITD mutation. However, the health economic implications of this approval are unknown.

We evaluated the cost-effectiveness of quizartinib using a partitioned survival analysis model. One-way and probabilistic sensitivity analyses were conducted.

In the base case scenario, the addition of quizartinib to 7+3 resulted in incremental costs of $289,932 compared with 7+3 alone. With an incremental gain of 0.84 quality-adjusted life years (QALYs) with quizartinib + 7+3 induction vs. 7+3 alone, the incremental cost-effectiveness ratio for the addition of quizartinib to standard 7+3 was $344,039/QALY. Only an 87% reduction in the average wholesale price of quizartinib or omitting quizartinib continuation therapy after completion of consolidation therapy and allogeneic hematopoietic cell transplant would make quizartinib a cost-effective option.

Introduction:

Mutations in FMS-like tyrosine kinase 3 (FLT3) are among the most common molecular abnormalities detected in de novo AML and are present in about one third of cases.^1–4 The majority of FLT3 mutations in AML occur as internal-tandem duplications (ITD) with a smaller subset of mutations in the tyrosine kinase domain (TKD).⁵ The prognostic implications of FLT3 mutations in AML depend on the type of FLT3 mutation and the presence of concurrent mutations especially in NPM1, with FLT3-ITD mutations in NPM1 wild-type disease considered to confer the highest risk scenario.^2,6,7 With its high incidence and clear role in leukemogenesis, FLT3 mutations have served as a potential target for the development of small-molecule inhibitors. In fact, three FLT3 inhibitors (midostaurin, gilteritinib, and quizartinib) have received approval by the United States (US) Food and Drug Administration (FDA) for the treatment of patients with FLT3-mutant AML and several others were or are currently being studied.^5,8–10

The approval of quizartinib for the treatment of patients with FLT3-ITD-mutant AML was based on the randomized, placebo-controlled, double-blind QuANTUM-First trial, which evaluated the addition of quizartinib to standard 7+3 induction chemotherapy and consolidation as well as maintenance therapy after completion of consolidative therapy including post-allogeneic hematopoietic cell transplantation (allo-HCT). The addition of quizartinib improved overall survival (OS) compared to placebo (hazard ratio [HR] 0.78, 95% confidence interval [CI] 0.62 – 0.98, p=0.032) with a generally comparable safety profile between both arms.¹⁰ However, this trial has been criticized for the use of a placebo arm despite the approval of midostaurin in 2017.

Although quizartinib was administered as continuation therapy both after allo-HCT and after completion of consolidation chemotherapy in patients who did not proceed to allo-HCT in the QuANTUM-First trial, the US FDA approval explicitly excludes quizartinib continuation therapy after allo-HCT from the drug label.^10,11 This restriction on the indication for quizartinib was based on the absence of an OS benefit attributable to the continuation phase.¹¹ Despite the high risk of disease relapse of FLT3-ITD-mutant AML even after allo-HCT there are currently no FDA-approved therapies in this setting although randomized trials of other FLT3 inhibitors have shown an improvement in relapse-free survival.^12–14 To date, the role of ongoing therapy with a FLT3 inhibitor after allo-HCT in patients with FLT3-ITD-mutant AML remains uncertain and questions related to optimal patient selection and treatment duration remain.

While all of the currently approved FLT3 inhibitors have shown an OS benefit in patients with FLT3-mutant AML in either the frontline setting when combined with intensive induction chemotherapy (midostaurin and quizartinib) or in patients with relapsed or refractory (R/R) disease,^8–10 these agents are associated with a significant economic burden on the health care system. Limited data support the cost-effectiveness of midostaurin and gilteritinib in AML, however, no such data exist for quizartinib to date and the aforementioned studies of midostaurin and gilteritinib were conducted with support by the drug manufacturers themselves.^15–17 We performed a cost-effectiveness analysis of quizartinib in the treatment of patients with FLT3-ITD mutant AML based on the original QuANTUM-First trial to evaluate the health economic implications of this recent approval.

Methods:

Construction of survival curves and parameterization:

This cost-effectiveness analysis was performed using a partitioned survival analysis methodology and was based on data reported in the original QuANTUM-First trial.¹⁰ As such, patients with newly diagnosed FLT3-ITD mutant AML who were 18–75 years of age entered the model and were treated with either quizartinib + 7+3 induction or 7+3 induction chemotherapy alone. In line with the original trial, patients who achieved a complete remission (CR) or CR with incomplete hematologic recovery (CRi) after induction +/− reinduction therapy proceeded to receive up to 4 cycles of consolidation therapy with high-dose cytarabine + quizartinib or placebo with or without allo-HCT.¹⁰ After completion of consolidation chemotherapy or allo-HCT, patients could receive continuation therapy with quizartinib for up to 36 cycles.¹⁰

Patient-level data derived from the Kaplan-Meier curves and at-risk tables for event-free survival (EFS) and OS from the original trial were used to reconstruct the survival function for both treatment arms using TreeAge Pro HealthCare 2023 (TreeAge LLC, Williamstown, MA, USA). Figure 1 shows a comparison of the survival curves used in our partitioned survival analysis model and the reconstructed survival curves from the QuANTUM-First trial.

Figure 1: Reconstructed survival curves.

Figure 1 provides a comparison of the reconstructed survival curves used for the model and the original survival curves as published in the QuANTUM-First trial.¹⁰ (A) and (C) show the event-free survival (EFS) and overall survival (OS) curves used for the model with the placebo arm (i.e., standard 7+3) and the quizartinib arm shown in red and blue, respectively. Starting at month 60, patients who were event-free were considered to experience a mortality rate similar to an age-matched US, AML patient population.³³ (B) and (D) show an overlay of the reconstructed (shown in grayscale) and original EFS and OS curves (shown in red and blue) for both the placebo and quizartinib arm, respectively.

Model inputs:

Table 1 provides an outline of the parameters included in the model. We used the average wholesale price (AWP) as a cost estimate for quizartinib.¹⁸ As prior studies have shown that the average sales price (ASP) is 26–30% lower than the AWP due to rebates granted by drug manufacturers, we applied a 28% discount off the AWP in the base-case scenario, which was further varied during sensitivity analyses between 18% and 38%.^19–21 Rates of dose modification, treatment interruptions and treatment discontinuation due to adverse events were included in the model as reported in the QuANTUM-First trial.¹⁰ Costs for induction, re-induction and consolidation chemotherapy as well as allo-HCT were derived from the previously published literature.^22,23 Patients could receive consolidation chemotherapy either in an inpatient (75%) or outpatient (25%) setting in line with prior real-world studies reporting on resource utilization among AML patients.²² In line with the original QuANTUM-First trial, patients could receive maintenance therapy with quizartinib after completion of consolidation chemotherapy or allo-HCT for up to 36 28-day cycles.¹⁰

Table 1:

Clinical variables, costs, and utilities included in model

Model variable	7+3+quizartinib		7+3		Ref.
	Base-case scenario	Range	Base-case scenario	Range
Primary treatment
Cost quizartinib (per 17.7mg tablet; AWP): • Induction: 35.4 mg orally once daily on Days 8–21 • Consolidation: 35.4 mg orally once daily on Days 6–19 of HiDAC for up to 4 cycles • Maintenance: 26.5 mg orally once daily on Days 1 to 14 and 53 mg once daily, thereafter, for up to thirty-six 28-day cycles.	655.20	N/A	N/A	N/A
AWP discount	0.28	0.18 – 0.38	N/A	N/A	19–21
Induction therapy
Cost 7+3 induction for 1 cycle	$135,851	$67,926 – $203,778	$135,851	$67,926 – $203,778	22
Cost 7+3 induction for 2 cycles	$222,435	$111,218 – $333,653	$222,435	$111,218 – $333,653	22
Rate of reinduction chemotherapy	0.20	0.101 – 0.302	0.21	0.103 – 0.310	10
Consolidation therapy
Rate of consolidation treatment	0.65	0.323 – 0.968	0.65	0.323 – 0.969	10
Cost per consolidation cycle chemo inpatient	$31,226	$15,613 – $46,839	$31,226	$15,613 – $46,839	22
Cost per consolidation cycle chemo outpatient	$12,508	$6,254 – $18,762	$12,508	$6,254 – $18,762	22
Probability of outpatient consolidation	0.25	0.125 – 0.375	0.25	0.125 – 0.375	22
Median number of consolidation cycles	2	1 – 4	2	1 – 4	10
Allogeneic hematopoietic cell transplant (allo-HCT)
Probability of undergoing allo-HCT	0.37	0.183 – 0.549	0.33	0.165 – 0.495	10
Cost of allo-HCT	$665,358	$332,679 – $998,037	$665,358	$332,679 – $998,037	23
Continuation therapy
Rate of continuation therapy	0.43	0.216 – 0.649	N/A	N/A	10
Median number of cycles of continuation therapy	16	8 – 24	N/A	N/A	10
Subsequent treatment
Gilteritinib
Probability of receiving gilteritinib salvage therapy	0.28	0.14 – 0.42	0.28	0.14 – 0.42	Expert opinion
Duration of gilteritinib salvage therapy	4.5	2 – 7	4.5	2 – 7	9
Cost of gilteritinib salvage therapy per month	$49,145	$24,572 – $73.718	$49,145	$24,572 – $73.718	24
Venetoclax-based regimens
Probability of receiving venetoclax-based salvage therapy	0.28	0.14 – 0.42	0.28	0.14 – 0.42	Expert opinion
Duration of venetoclax-based salvage therapy	2	1 – 4	2	1 – 4	25
Cost of venetoclax-based salvage therapy per month	$80732	$40,366 – $121,098	$80732	$40,366 – $121,098	24
Best supportive care
Probability of receiving best supportive care as 2nd line therapy	0.44	0.22 – 0.66	0.44	0.22 – 0.66	39
Cost of best supportive care per month	$5302	$2,651 – $7,953	$5302	$2,651 – $7,953	26
Additional resource utilization during induction, consolidation, and HCT until 24 months post-allo
Number of inpatient admissions per month	0.35	0.175 – 0.525	0.35	0.175 – 0.525	22
Cost per inpatient admission	$50592	$25,296 – $75,888	$50592	$25,296 – $75,888	22
Number of outpatient visits per month	7.6	3.8 – 11.4	7.6	3.8 – 11.4	22
Cost per outpatient visit	$1335	$668 – $2,003	$1335	$668 – $2,003	22
Number of ED visits per month	0.15	0.075 – 0.225	0.15	0.075 – 0.225	22
Cost per ED visit	$2251	$1,126 – $3,377	$2251	$1,126 – $3,377	22
Additional costs during continuation phase for patients in long-term remission
Cost of supportive care per month	$4961	$1,981 – $5,945			26
Cost of terminal care	$95930	$47,054 – $141,162	$94,108	$47,054 – $141,162	27
Utilities
During induction chemotherapy (months 1 & 2)	0.40	0.36 – 0.44	0.40	0.36 – 0.44	29
AML in early remission (months 3–6)	0.66	0.60 – 0.72	0.66	0.60 – 0.72	29
AML in long-term remission (>6 months)	0.83	0.75 – 0.91	0.83	0.75 – 0.91	30
Relapsed AML	0.53	0.48 – 0.53	0.53	0.48 – 0.53	30

Model assumptions were derived preferentially from the original QuANTUM-First trial.¹⁰ If not available, costs and utilities were derived from the literature reporting cost and treatment patterns from a US perspective. All costs were adjusted for inflation to 2022 USD. Costs and utilities were varied by 50% and 10%, respectively, during sensitivity analyses and as shown in the table.

Patients who experienced disease relapse or had refractory disease received either gilteritinib, venetoclax-based therapy or best supportive care. Costs and duration of salvage therapy were derived from the literature.^9,24,25 Patterns of healthcare resource utilization, supportive and terminal care were based on previous studies.^22,26,27

All costs were adjusted for inflation to 2022 US dollars using the personal consumption expenditure health index.²⁸ As utilities and quality of life data from the QuANTUM-First trial are not available, we used data from the literature and assumed equal utilities for both arms.^29,30

Costs and utilities were modeled over a 30-year time horizon and discounted by 3% annually as recommended by the second panel on cost-effectiveness in health and medicine.³¹ Utilities were measured in quality-adjusted life years (QALYs). Although the risk of disease relapse is small among patients in remission for at least 5 years, all-cause mortality remains higher in long-term AML survivors compared to an age-matched control population. Using a recent analysis of Surveillance, Epidemiology, and End Results (SEER) data and data published by the US Social Security Administration, we calculated a time-dependent, AML-attributable background mortality rate.^32,33 During long-term follow-up patients were assumed to not incur additional AML-related healthcare expenses.

Model outputs are presented as the incremental cost-effectiveness ratio (ICER) for 7+3+quizartinib, which is the added cost in 2022 US dollars (USD) per QALY gained compared to 7+3 alone.

Sensitivity analyses:

We performed one-way sensitivity analyses to evaluate the impact of individual variables on the model. Variables were varied across a 50% range except for utility values, which were varied across a 10% range. The results of one-way sensitivity analyses are shown as a tornado diagram showing the 10 variables with the greatest influence on the ICER.

While the original QuANTUM-First trial included maintenance therapy with quizartinib following completion of consolidation chemotherapy or post-allo-HCT, quizartinib has not been approved by the US FDA for maintenance therapy after allo-HCT given the absence of an OS benefit in this setting. Thus, we developed an alternative model without quizartinib maintenance therapy following allo-HCT as well as a model in which any continuation therapy with quizartinib (i.e., both after consolidation chemotherapy and after allo-HCT) was omitted.

To evaluate the influence of differences in clinical efficacy on the ICER, we performed a sensitivity analysis in which we used the HRs for EFS and OS reported by the QuANTUM-First trial. The reconstructed EFS and OS curves of the standard 7+3 arm were used to derive the respective curves of the quizartinib arm.¹⁰ In this sensitivity analysis we varied the HRs of EFS and OS across the full width of the 95% CI as reported by the QuANTUM-First trial.

Probabilistic sensitivity analysis was performed using 10,000 Monte Carlo simulations, each time randomly sampling from the distributions of model inputs. In this analysis each parameter was defined by a distribution: beta distributions were used to describe probabilities and utilities, while gamma distributions were used for costs. Results of probabilistic sensitivity analyses are shown as cost-effectiveness acceptability curves.

Results:

Our reconstructed survival curves resulted in a median EFS and OS of 12 months and 32 months in the quizartinib arm and 6 months and 16 months in the standard 7+3 arm, respectively. These results were comparable to the original QuANTUM-First trial (median EFS: 11.9 months with quizartinib and 5.7 months with placebo; median OS: 31.9 months with quizartinib and 15.1 months with placebo) supporting the validity of our modelling approach.¹⁰ Figure 1 provides an illustration of the survival curves used in the model.

In the base case scenario, the addition of quizartinib to intensive induction chemotherapy resulted in life-time costs of $1,135,136. Compared to the lifetime costs in the standard 7+3 arm of $845,204, quizartinib use resulted in incremental costs of $289,932. With an incremental gain of 0.84 QALYs with quizartinib + 7+3 induction (quizartinib + 7+3 induction: 4.70 QALYs; 7+3 induction alone: 3.86 QALYs), the ICER of the addition of quizartinib to standard 7+3 induction chemotherapy compared to 7+3 induction chemotherapy alone was estimated at $344,039/QALY. Key model outputs of the base-case scenario as well as 95% confidence intervals that were derived from the probabilistic sensitivity analysis are shown in Table 2.

Table 2:

Key model outputs of base-case scenario

Strategy	Cost (USD; 95% CI)	Incremental cost (USD; 95% CI)	Effectiveness (QALY; 95% CI)*	Incremental effectiveness (QALY; 95% CI)	ICER (USD/QALY; 95% CI)
7+3 induction	$845,204 ($739,204 – $965,816)	N/A	3.86 (3.67– 4.05)	N/A	N/A
Quizartinib + 7+3 induction	$1,135,136 ($965,445 – $1,349,584)	$289,932 ($135,147 – $486,920)	4.70 (4.46 – 4.94)	0.84 (0.80 – 0.89)	$344,039 ($160,125 – $579,330)

CI – confidence interval; ICER – incremental cost-effectiveness ratio; QALY – quality-adjusted life years; USD – US dollar

^*QALYs are reported to 2 decimal places.

We next performed one-way sensitivity analyses to identify variables with the greatest influence on the ICER. Figure 2 shows a tornado diagram of the ten variables with the largest impact on the ICER. When varied by +/−50% none of the variables included in the model was able to achieve an ICER of less than $150,000/QALY. Only a reduction in the AWP of quizartinib by 86.6% from $655.20 to $87.67 per 17.7 mg tablet would lower the ICER to below $150,000/QALY.

Figure 2: Tornado diagram of the ten most influential variables on the ICER in one-way sensitivity analyses.

Figure 2 shows a tornado diagram of the ten variables with the greatest influence on the ICER. Variables were varied by 50% for costs and probabilities and by 10% for utilities as outlined in table 1. Bars shown in blue and red in the figure represent lower and higher values in the range, respectively.

In a probabilistic sensitivity analysis using 10,000 Monte Carlo simulations standard 7+3 induction was favored in 98.1% of 10,000 iterations at a willingness-to-pay threshold of $150,000/QALY (Figure 3).

Figure 3: Probabilistic sensitivity analyses.

Figure 3 (A) shows a probabilistic sensitivity analysis based on 10,000 Monte Carlo simulations. The willingness-to-pay threshold is shown on the x-axis and the percentage of cost-effective iterations is shown on the y-axis. The arm with the addition of quizartinib to standard 7+3 induction is shown in blue, the standard 7+3 arm is shown in red. Costs and probabilities were varied by 50%, while utilities were varied by 10%. Beta distributions were used to describe probabilities and utilities, while gamma distributions were used for costs. (B) shows an incremental cost-effectiveness scatterplot of 7+3 + quizartinib vs. standard 7+3 with each circle representing an iteration in the probabilistic sensitivity analysis. Iterations yielding an ICER of <$150,000/QALY are shown in green, iterations with an ICER of ≥$150,000/QALY are shown in red. The 95% confidence interval is shown as a green eclipse.

Finally, we modelled three additional scenarios. As continuation therapy with quizartinib after allo-HCT did not receive regulatory approval by the US FDA due to the absence of a survival benefit attributable to maintenance therapy in this setting, we modelled a scenario without quizartinib continuation therapy after allo-HCT. Of note, patients who did not proceed to allo-HCT could receive quizartinib continuation therapy after completion of consolidation chemotherapy. Under the assumption that both the EFS and OS curves as well as other model parameters remained constant, the total lifetime costs of therapy in the quizartinib arm dropped to $995,268 with the omission of quizartinib continuation therapy after allo-HCT. With constant cost in the standard 7+3 arm ($845,204) and effectiveness (quizartinib + 7+3 arm: 4.70 QALYs; standard 7+3 arm: 3.86 QALYs), the ICER was $178,069/QALY. If quizartinib continuation after both allo-HCT and after completion of consolidation chemotherapy was omitted, the ICER dropped to $147,566/QALY consistent with cost-effectiveness of quizartinib against a conventional willingness-to-pay threshold of $150,000/QALY.

Lastly, we analyzed the effect of differences in clinical effectiveness of quizartinib on the ICER by using a variable hazard ratio to modulate the EFS and OS curves of the quizartinib + 7+3 arm. In this scenario, a HR for OS of ≤0.587 would be required to lower the ICER for the addition of quizartinib to <$150,000/QALY if all other model parameters remain constant.

Discussion:

While the FLT3 inhibitor quizartinib has been shown to improve both EFS and OS when added to intensive induction chemotherapy among patients with newly diagnosed FLT3-ITD-mutant AML, the health economic implications of this approval are unknown. In our cost-effectiveness analysis, we showed that the use of quizartinib under the current pricing model is unlikely to be cost-effective against the standard willingness-to-pay threshold of $150,000/QALY compared with conventional 7+3 induction chemotherapy. However, in multiple sensitivity analyses we found that a variety of changes would be able to lower the ICER to less than $150,000/QALY.

First, a reduction in the AWP of quizartinib by 86.6% from $655.20 to $87.67 per 17.7 mg tablet would be required to lower the ICER to below $150,000/QALY. The extent of this AWP reduction exceeds what has been reported for the required AWP reduction in prior studies evaluating the cost-effectiveness of novel, oral AML therapeutics.^34,35 In light of the continuing increase in health care expenses including drug costs in oncology, studies like ours can potentially provide the scientific support for efforts to curb drug prices in the US and globally.^36,37

The second aspect to consider with quizartinib is the continuation therapy phase for patients after completion of consolidation chemotherapy or allo-HCT. While oral azacitidine has been approved for maintenance therapy in AML based on a survival benefit in randomized clinical trials, no FLT3 inhibitor has been shown to improve OS when used for maintenance therapy after allo-HCT although sorafenib and gilteritinib demonstrated improvements in relapse-free survival in recent randomized trials.^13,14,38 In fact, the FDA label of quizartinib explicitly excludes maintenance therapy after allo-HCT from the approval. With monthly drug costs of $23,900, quizartinib maintenance therapy for up to 36 cycles can add a significant burden on the health care system. In a model that did not include quizartinib continuation after allo-HCT, we found an ICER of $178,069/QALY. Furthermore, if any continuation therapy with quizartinib is omitted, the ICER would be further reduced to $147,566/QALY consistent with cost-effectiveness against a conventional willingness-to-pay threshold of $150,000/QALY.

As it is difficult to isolate any effect of quizartinib continuation therapy on survival in the absence of a dedicated clinical trial, our assumption that the survival curves are not affected by the omission of continuation therapy may limit the interpretation of these models. However, from a health economic perspective and with uncertainty about the clinical effectiveness, restricting the use of quizartinib continuation therapy to patients at highest risk of disease relapse (e.g., patients with persistent morphologic or measurable residual disease [MRD] at the time of allo-HCT) appears justified. Unselected use of maintenance therapy after allo-HCT is also being questioned by the results of the recent MORPHO trial, which showed that relapse-free survival and OS were similar among patients with FLT3-ITD-mutant AML who received gilteritinib or placebo for maintenance therapy after allo-HCT.¹² However, in a prespecified subset analysis, MRD-positive patients had a statistically significant relapse-free survival benefit when treated with gilteritinib suggesting that MRD status can be an option to select patient who might benefit the most from maintenance therapy approaches.¹²

Prior to the approval of quizartinib, midostaurin had been the only FDA-approved FLT3 inhibitor for the combination with intensive chemotherapy in the frontline setting. This approval was based on the RATIFY trial, a randomized, placebo-controlled phase III trial with several key differences compared with the QuANTUM-First study.^8,10 In contrast to QuANTUM-First, RATIFY enrolled patients 18–59 years-old with both FLT3-ITD and FLT3-TKD mutations.⁸ Due to the differences in clinical trial design and patient populations, we were unable to compare the cost-effectiveness of midostaurin and quizartinib. Indirect comparisons across cost-effectiveness studies are also limited by differences in modelling characteristics and patient-level data would be required to identify the most cost-effective FLT3 inhibitor.

Potential limitations of our study include possible differences between clinical trials and real-world practice patterns related to health care resource utilization and patient characteristics. Similarly, we did not have access to individual patient-level data from the QuANTUM-First trial and had to rely on previously published estimates from the literature. However, we addressed these uncertainties in several sensitivity analyses, which did not reveal any model inputs other than the cost of quizartinib and the number of quizartinib continuation cycles that would have an effect on the ICER large enough to lower it below $150,000/QALY. Additionally, we used the AWP of quizartinib for the model which – despite the use of a previously published discount rate – might overestimate the actual costs of quizartinib. However, given that a reduction in AWP by ~87% would be required to make quizartinib cost-effective, it is unlikely that such substantial discount rates would be granted in real-world practice. Finally, our study was conducted from a US health care perspective and the results might not be generalizable to other countries given differences in drug availability and health care resource utilization.

In conclusion, the addition of quizartinib to standard 7+3 induction chemotherapy is unlikely to be cost-effective under the current pricing model against a willingness-to-pay threshold of $150,000/QALY. Only a reduction in the AWP of quizartinib by 87% or the omission of quizartinib continuation therapy would make quizartinib a cost-effective option with the latter being in line with the current FDA label.

Data sharing statement:

Original data can be requested from the corresponding author (amer.zeidan@yale.edu)