BACKGROUND: A randomized phase 3 trial (MONALEESA-7) has established that adding of Ribociclib, a CDK4/6 inhibitor, to endocrine therapy can significantly improve progression-free-survival (PFS) versus placebo and is safe and well tolerated in premenopausal HR+/HER2- advanced breast cancer. This paper analyses the cost-effectiveness of ribociclib addition to endocrine therapy versus placebo in premenopausal women.
OBJECTIVE: To evaluate the cost-effectiveness of ribociclib plus endocrine therapy versus placebo in for premenopausal women with hormone-receptor-positive, advanced breast cancer.
METHODS: A markov model including 3 health states (progression free, progressed disease and death) simulated costs and outcomes over a 20-year time horizon with a cycle time of one month. Clinical efficacy data (PFS and overall survival [OS]) from a phase III trial of ribociclib plus endocrine therapy (MONALEESA-7; NCT02278120) were used to calculate the transition probabilities. Health care costs were derived from iMTA-costing-tool and the ZIN report. The utility values were based on the NICE report. Effectiveness was measured with health-state utility values in quality-adjusted life-years (QALY’s).
The impact of uncertainty in model parameters on results was examined by a one-way deterministic sensitivity analysis and a probabilistic sensitivity analysis was performed.
RESULTS: Compared with placebo, ribociclib plus endocrine therapy was associated with an incremental cost of €96,70 and an incremental QALY of 0.77, equating to an incremental cost-effectiveness ratio of €125.468 per QALY. The probability that ribociclib plus endocrine therapy versus placebo was cost-effective at threshold €80.000 per QALY gained was 0,0%.
CONCLUSION: In the Netherlands, ribociclib plus endocrine therapy is not a cost-effective alternative to placebo for premenopausal women with hormone-receptor-positive, advanced breast cancer. The incremental cost-effectiveness ratio of €125.468 per QALY is much higher than the willingness-to-pay threshold of €80.000.
Funding Utrecht University
Copyright ©2018, Pharmaceutical Weekly. All rights reserved.
Table of contents
Table of contents 2
1. Background 3
2. Alternative treatments 4
3. Objective 5
4. Cost-analysis 6
5. Analysis of effect-elements 7
6. Pharmacoeconomic analysis 8
6.1 Model description 8
6.2 Calculation of the ICER or ICUR 9
6.3 Sensitivity analysis 10
7. Discussion 11
8. Appendix: References 12
9. Appendix: Project planning table 13
Introduction
Among women worldwide, breast cancer is the most frequently diagnosed type of cancer. Breast cancer is also the leading cause of cancer-related mortality amongst women worldwide. Approximately 500,000 women are diagnosed with breast cancer annually in Europe. In 2012 breast cancer accounted for 131,259 deaths which was a third of all cancer related deaths (1). Due to the aging population in Western countries and the progressive nature of the disease, the incidence and prevalence of breast cancer increased over the past decades.
Advanced breast cancer is in most cases observed in patient with hormone receptor-positive (HR+) breast cancer. In the Netherlands, about 14,000 women (and 103 men) are diagnosed with invasive breast cancer each year and about 1,900 with in-situ breast cancer. During the life of a woman the chance of getting breast cancer is 12-13%. This makes breast cancer the most common form of cancer in women in the Netherlands.
Treatment options for premenopausal women with HR+ breast cancer include aromatase inhibitors (e.g., letrozole, anastrole), selective estragon receptor modulators (e.g., tamoxifen) and a gonadotropin-releasing hormone agonist (e.g., goserline). The cyclin-dependent kinase (CDK) 4 and 6 inhibitors palbociclib and abemaciclib have been tested in a subset of premenopausal patients in the PALOMA-3 and MONARCH-2 studies, and ribociclib has been tested in the MONALEESA-7 trial with only premenopausal patients. In the MONALEESA-7 trial addition of ribociclib to endocrine therapy prolonged progression-free survival (PFS) by 10.8 months which means a 45% lower relative risk of progression compared to placebo 9hazard ratio [HR]=0,55; 95% confidence interval [CI]= 0.44-0.69; p<0,0001).
This also means that the use of ribociclib versus placebo may increase drug and healthcare costs due to longer PFS and life expectancy. With a cost-effectiveness analysis the relationship between these extra costs and the health benefits can be evaluated. A cost-effectiveness study of ribociclib plus letrozole versus letrozole monotherapy in the first-line treatment of postmenopausal women with HR+/HER2- advanced or metastatic breast cancer has shown that the addition of ribociclib is cost-effective compared with letrozole, resulting in gains in quality-adjusted life years and cost savings.
To make a statement about the situation in premenopausal women, this cost-effectiveness analysis was conducted.
Methods
The basis of this cost-effectiveness analysis is the MONALEESA-7 trial. The MONALEESA-7 trial did a research on premenopausal women aged 18-59 years who had HR-positive advanced breast cancer. 672 women were randomly assigned in the ribociclib group or the placebo group. 335 women received oral ribociclib and endocrine therapy, 337 women received placebo and endocrine therapy. The endocrine therapy included tamoxifen and a non-steroidal aromatase inhibitor (NSAI), like letrozole or anastrozole. The placebo was a mixture of microcrystalline cellulose, starch and magnesium stearate. By introducing ribociclib to the therapy there will be additional costs. Therefore we chose the social perspective.
A treatment cycle started on day 1 and was initiated to day 28. The time horizon that we used is 20 years. The assumption here is that premenopausal women are younger than postmenopausal women. Both groups received tamoxifen (20 mg) once daily or NSAI (letrozole 2,5mg or anastrozole 1mg). ) Ribociclib and placebo had a different schedule. Ribociclib (600mg/day) and placebo were administered on a 3 weeks-on, 1-week-off schedule. The treatment was continued until disease progression, unacceptable toxicity, death or discontinuation for any other reason.
The tumour was screened with CT or MRS every 8 weeks, during the first 18 months. Thereafter the patients were screened every 12 weeks. The biochemical and haematological laboratory tests were also screened; cycle 1 day 1 and 15, cycle 2 day 1, cycle 3 day 15 and day 1 till the end of the treatment. The adverse events were monitored continuously throughout the study.
Model structure
Markov model was used to estimate the lifetime costs and the quality- adjusted life- years (QALY) for premenopausal women with advanced HR-positive breast cancer, see figure ?. The markov model consists of 3 states: asymptomatic disease, progressive disease and death. Each state consists of a number of patients. The patients can move from one state to another in time. Number of patients in the 3 states were calculated with the formulas:
-Asymptomatic disease= Disease free survival (DFS)= DFS(t-1) – DFS(t-1)*TP1 – DFS(t-1)*TP2
-Progressive disease=Recurrent disease (RD)= RD(t-1) + DFS(t-1)*TP1 – RD(t-1)*TP3
-Death= Death(t-1) + DFS(t-1)*TP2 + RD(t-1)*TP3
Figure ?: Markov model, health-state transitions.
TP1, 2 en 3 are the transition rates. These rates are needed to move from one state to another. Transition rates were calculated with the formulas:
-Instantaneous rate (r)= -1/t *Ln (1-p); t=time, p=%death
-Annual TP=1-e^(-r*t)
TP2 was calculated with values from Centraal Bureau Statistiek (CBS), see table?.
r: 0,00124395395 TP2: 1,0373E-4. This value is TP2 for the ribociclib group and also for the placebo group.
Table ?: Number of deaths in year 2016.
Age Death Total % (death/total *100)
15-20 57 501060 0,011375883
20-25 104 524383 0,019832832
25-30 136 534138 0,025461585
30-35 186 509234 0,036525448
35-40 274 507293 0,054012178
40-45 441 550928 0,080046757
45-50 888 639862 0,138779924
50-55 1659 636649 0,260583147
55-60 2467 590207 0,417988943
Total 6212 4993754 0,124395395
Figure ?: Kaplan- Meier curve. Blue is ribociclib group, red is placebo group.
Costs Data
Drug costs
In our study we assumed two situation. Situation A is the treatment without ribociclib and in situation B ribociclib is added to the treatment. For the medicine costs we used the ribociclib report by the care institute from the Netherlands (ZIN). The ZIN report shows the costs for two different situations. In the first situation, the dosages as normally prescribed are assumed. In the second situation, the doses were adjusted as a result of the side effects of ribociclib. For our report, we assumed the first situation. The dosage frequency and dosage unit of MONALEEA-7 has been followed in this study. Ribociclib 600mg/day is given on a 3-weeks-on and 1-week-off schedule. The medicinal costs is shown in table 1.
Ribociclib 3dd1t van 200mg verpakking van 15 stuks = 817,71 euro gem kosten per dag = 163,54 euro per cycle (28 dagen)
4579,12 euro
tamoxifen 1dd20mg verpakking van 15 stuks=
4,71 euro gem kosten per dag =0,16 4,48
letrozole 1dd2,5mg verpakking van 15 stuks =
9,55 euro “= 0,64 17,92
anastrozole 1dd1mg verpakking van 15 stuks=
8,06 euro “= 0,54 15,12
gem= 16,52
gosereline 3,6mg dag 1 van elke cyclus 113,60
medicijnkosten per cycli
ribociclib + tamoxifen+ goserlin 4697,20
ribociclib + NSAI + goserlin 4709,24
tamoxifen + goserlin 118,08
NSAI + goserlin 130,12
gemiddelde genomen van 1 en 2 ., 3 en 4
DISEASE RELATED COSTS
The iMTA-costing-tool is used for the disease related costs. The tool contains reference prices for diagnostics, nursing days, travel costs, which have been used for this.
Verpleegdag, gewogen gemiddelde 476
Referentieprijzen diagnostiek
Auto, kosten per kilometer
pakeerkosten 0,19
3,00
The following assumptions were made:
1) The families only travels by car and do not use public transport or taxi.
2) The dose of ribociclib was not adjusted due to side effects
QALY
The QALY is needed to calculate the ICER. The formula used to calculate the QALY is
QALY DFS= DFS * U(DFS) * DR; DFS= fraction in DFS state, U= utility, DR= discount rate 1,5% (DR= 1/(1+DR)^t).
The QALY was also calculated for the RD and Death state with the same formula. Total QALY is the sum of QALY DFS, QALY RD and QALY Death.
Utility values were derived from the NICE report
-U(DFS)= 0,774
-U(RD)=0,505
-U(Death)=0
Analyses
The analysis has the outputs incremental costs and QALYS gained with ribociclib plus endocrine therapy versus placebo. Costs were discounted at 4% and QALY with 1,5% annually, as recommended by Dutch National Health Care Institute (ZIN).
Deterministic sensitivity analysis
To examine the impact of variation in parameters values on the incremental cost-effectiveness ratio (ICER) of ribociclib plus endocrine therapy versus placebo, a one-way deterministic sensitivity analysis was executed. The parameters used in this sensitivity analyses are the costs, drug costs, utility and transition probabilities. However, these parameters were not reported with a standard error or 95% CI. Therefore a standard range of +/-20% of the base value was used.
Probabilistic sensitivity analyses
For the probabilistic sensitivity analyses (PSA; n=1.000 iterations) a template in Microsoft Excell was used. In this template cost data and utility data were filled. Beta distribution was applied for the utilities and gamma distribution for the costs.
Results
Base case
The costs and QALY’s were calculated for 20 years, for results see table?. The difference in values between the ribociclib and placebo group results in
-QALY= 0,77
-Costs= €96.701
With these values the ICER was calculated, the ICER is €125,468/QALY (Costs/QALY).
Table ?: Costs and QALY’s of the ribociclib and placebo group.
QALY Costs
Ribociclib 3,49 €178.020
Placebo 2,72 €81,318
Tornado-plot
The tornado-plot shows that the utility of progression free survival and the costs of ribociclib were the key model drivers.
According to the PSA, ribociclib plus endocrine therapy is not cost-effective compared to placebo. With the majority of the probable ICER’s in the upper right quadrant, this means there is a high probability that the intervention will indeed gain quality adjusted life years but it will also increase the community costs.
Transition probabilities Base case Low High Distribution Source
Tp1 0,00774 0,0064 0,0096 Beta (1)
Tp2 0,00010373 0,0001016554 0,000124476 Beta (2)
Tp3 0,00486 0,003888 0,005832 Beta (1,2)
Tp1* 0,0323 0,02584 0,03876 Beta (1)
Tp2* 0,00010373 0,000082984 0,000124476 Beta (2)
Tp3* 0,00547 0,004376 0,006564 Beta (2)
Costs Base case Low High Distribution
Casymp 3236 2589 3883 Gamma
Csymp 3569 2855 4283 Gamma
Cribociclib 4703,22 3762,58 5643,86 Gamma
Ccomp 124,1 99,28 148,92 Gamma
Utility Base case Low High Distribution
uDFS 0,774 0,619 0,929 Beta
uRD 0,505 0,404 0,606 beta
Discussion
Limitations
Conclusion