Cost-Effectiveness Analysis of Breast Cancer Treatments

Cost-Effectiveness Analysis of Pertuzumab With Trastuzumab and Chemotherapy Compared to Trastuzumab and Chemotherapy in the Adjuvant Treatment of HER2-Positive Breast Cancer in the United States

Introduction

Throughout the past decades, burden of cancer has increased globally. As per WHO report in 2018, the burden has risen to 18 million new cases diagnosed and almost 9.6 million deaths. From each six females, one would develop cancer and from each 11 females, one would die from cancer. The most commonly diagnosed cancer in females is breast cancer (24.2%), as well as the leading cause of cancer deaths in females (15%).

Since the introduction of anti-HER2 therapies, there has been noticed a great improvements in survival for patients with HER2 positive breast cancer in both early and advanced states. Use of combinations of anti-HER2 therapies has been studied in different clinical trials, and it is suggested that these combinations can potentially improve outcomes however, in some patients may also allow de-escalation of therapy and sparing needless therapies with their related side effects and costs. In this essay, we will review a recently published economic evaluation for combination of Pertuzumab with Trastuzumab and chemotherapy against Trastuzumab and chemotherapy (Pernas & Tolaney, 2019).

Study Overview and Recommendation:

There are four HER-2 targeting therapies and Pertuzumab is one of them. When used alone, it has limited activity with Trastuzumab-resistant patients. When used in combination with Trastuzumab and chemotherapy, complementary effect is noticed. Based on that, it is recommended by the National Comprehensive Cancer Network (NCCN) guidelines to use dual HER-2 targeting therapies (Pertuzumab + Trastuzumab) (Garrison et al., 2019).

Several studies have been conducted to demonstrate the extra benefit of using duo HER-2 therapies. In Cleopatra, using the duo plus docetaxel (TPH) resulted in improved overall survival by 15.7 months versus the use of Trastuzumab with chemotherapy (TH). In APHINITY, treatment of HER-2 positive in early breast cancer was evaluated using TPH and TH. When 4-year invasive-disease-free survival, was evaluated, it was shown to be 90.6% in TPH arm compared to 92.3% in the TH arm. This study had three subgroups; ITT population, node-positive tumors and hormone receptor negative tumor, where there was risk reduction of recurrence or death compared to control arm (Swain et al., 2015).

The objective of this economic evaluation was clearly mentioned in the title “Cost-Effectiveness Analysis of Pertuzumab With Trastuzumab and Chemotherapy Compared to Trastuzumab and Chemotherapy in the Adjuvant Treatment of HER2-Positive Breast Cancer in the United States”. The perspective is also clearly mentioned targeting US payers and different stakeholders to facilitate decision making of the new treatment coverage and reimbursement. The Markov model was built to cover six different health states: invasive-disease-free survival, non-metastatic recurrence, remission, first line metastatic, subsequent line metastatic and death. Primary outcomes of the model were life-years (LY), quality-adjusted life years (QALYs) and ICERs (Garrison et al., 2019)

For the results, they detailed them into the three subgroups; ITT population, where the analysis projected improved outcomes by 0.5 LYs and 0.45 QALYs and increased costs for ICERs of 147,774$/LY gained and 167,185$/QALY gained for THP versus TH patients respectively. Second group was the node positive patients, where the analysis projected improved outcomes by 0.86 LYs and 0.76 QALYs and increased costs for ICERs of 77,684$/LY gained and 87,929$/QALY gained for THP versus TH patients respectively. Third group was the hormone receptor negative patients, where the analysis projected improved outcomes by 0.52 LYs and 0.47 QALYs and increased costs for ICERs of 147,022$/LY gained and 166,518$/QALY gained for THP versus TH patients respectively (Garrison et al., 2019).

They used 50,000$ as the lower limit and 200,000$ as the upper limit for cost-effectiveness, and mentioned briefly that these results were most sensitive to time horizon. Accordingly, they concluded that addition of Pertuzumab to standard therapy of TH in high risk of recurrence HER-2 positive early breast cancer patients would “likely” be cost-effective (Garrison et al., 2019).

Looking at clinical evidence, it is clear that there is an added benefit when adding Pertuzumab to TH, however there is an uncertainty about the size of this clinical benefit. The trial shows that 1.7% fewer patients had invasive disease at 4 years using Pertuzumab as adjuvant (in the ITT group). Whereas there was much better evidence for patients with node positive disease with 3.2% fewer patients had invasive disease at 4 years using Pertuzumab as adjuvant. With these results, it would be difficult to assume whether overall survival is better when using Pertuzumab as adjuvant (NICE appraisal, 2019)

There is an argument whether disease-free survival can be a substitute to overall survival in those high-risk patients, where in most recent studies the use of invasive-disease-free survival has been widely accepted. Despite the fact that there might be a 10% difference in outcomes between invasive-disease-free and disease-free survival, it is acknowledged that obtaining overall survival data would be challenging for adjuvant treatments. Based on the APHINITY, patients with node positive HER-2 as well as patients with residual disease after neo-adjuvant chemotherapy are likely the most benefited from addition of Pertuzumab, despite the trial was not intended to evaluate treatment outcomes in different subgroups of interest. It is also worthwhile mentioning that these benefits shown in APHINITY trial are very similar to that in ALLTO trial, when Lapatinib was added to adjuvant Trastuzumab and chemotherapy, which showed a hazard ratio of 0.84 (Stanton & Davidson, 2017).

Additional information needed:

Apart from APHINITY, it would have been useful to use other studies like Cleopatra in the analysis, to cover different populations and to consider different permutations of second and third line therapies. In this analysis, Pertuzumab was compared to placebo, which would have been more robust if compared with other HER-2 targeting therapies (Durkee et al., 2016). Basing all assumptions on a single trial would fail to take into consideration different factors like selection bias, different demographics, patients’ variations and probability in staying in each stage (Miller, 2017).

There was un-clarity in the model on the cost effectiveness threshold used on US. Back in 2010, Permera Blue cross has developed a guidance on a value based drug formulary tiers based on ICERs. This was developed because of rising premiums and concerns about affordability of higher value therapies and it has four Tiers where Tier 4: (ICER >$150,000/QALY) – insufficient evidence. Rare conditions with no effective treatments were considered special cases (Dubois, 2016).

Based on literature, policymakers will have to consider the research; however, they would also need to balance population level assessments with individual patients’ needs. The question is whether cost-effectiveness thresholds are likely to come to US and be of wider use. One of the points that was not covered in this study was the mode of administration of Pertuzumab, which is administered intravenously, while Trastuzumab is administered subcutaneosly. Accordingly, Trastuzumab has to be administered intravenously and patients has to stay more time at the hospital and incur extra costs. Furthermore, the use of bio-similar of Trastuzumab would steeply affect the results and show much better cost-effectiveness. (NICE appraisal, 2019)

Key data sources used:

In this model, different data sources were considered, however results were mainly based on APHINITY trial. They referenced National Comprehensive Cancer Network (NCCN) guidelines, for the use of dual HER-2 targeting therapies. They also used Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC), in addition to conformity tothe original Kaplan-Meier survival curves, to assess the log-logistic distribution for the base-case analysis. Recurrence rates were adjusted using HERA and BCIRG-006 studies (Garrison et al., 2019).

To make sure model has resulted in valid death rates matching the general population mortality, they used US life tables. Hamilton et al. study was used to measure the probabilities of transition between the remission state and the first line metastatic breast cancer. Cleopatra trial was used to evaluate the monthly probability of progression in metastatic breast cancer. They used Average Sale Price (ASP)(Genentech data) for Pertuzumab and Trastuzumab costs while costs of other chemotherapy drugs were gathered from Centers for Medicare and Medicaid (CMS) Average Sales Price in 2017(ASP). They have assumed that patients who advanced into local recurrence will receive an additional round of adjuvant treatment, for the purpose of cost estimation. They used published literature to estimate end of life healthcare costs (Garrison et al., 2019).

Quality of Life Data:

The main objective of the analysis is to compare the outcomes of the TPH arm compared to TH arm and reflect that on life years (LYs) and quality adjusted life years (QALYs). Utility values were estimates from APHINITY trial where they used EQ-5D to gather patient reported outcomes. Utilities estimates were calculated separately for patients whether on or off treatments and across different disease states. They also used published literature to generate utilities for metastatic states (Garrison et al., 2019).

These estimates were then used to calculate utility values for patients and health states based on a specific scoring mechanism, which was gathered from a survey on the non-institutionalized general population in the US. Furthermore, the model took into consideration, long-term utilities of patients in general population to account for increased comorbidity with age. A limitation for this survey is that it is not a robust source (Garrison et al., 2019).

Comparators, outcomes and Model:

In this analysis, despite the fact that the dual HER-2 therapies is recommended by guidelines, there are some other therapies like Lapatinib were not mentioned in the analysis, which may have had very similar outcomes. In general, comparators and outcomes were well detailed either clinical outcomes for the APHINITY or economic outcomes of cost-effectiveness per subgroup of patients.

Markov model was used and an illustrative figure was presented to show the six states and transition in each state. Authors did not explain the reason behind choosing Markov model neither they explained other modelling choices. Of course, all models are just simplification of the reality and the main objective of such models is to facilitate decision-making. May be it would have been more robust if they used a Markov model using TreeAge Pro software (TreeAge,Williamstown, MA), like the one used in the paper of CEA of Cleopatra (Durkee et al., 2016).

Authors had a conservative assumption, assuming that Pertuzumab treatment effect would last for 7 years and totally fade away at 10 years. It was considered conservative since authors neglected the fact of observed efficacy of Trastuzumab in adjuvant setting and Pertuzumab in metastatic setting. Accordingly predicting treatment effect duration would be impossible given the uncertainty in estimation of treatment benefit (Garrison et al., 2019).

Uncertainty:

In the model, based on the trial extrapolation that the risk of recurrence would drop to 10% if the patients were disease-free after 120 months, in both treatment arms. As mentioned by the authors, when performing a univariate sensitivity analysis, ICERs for ITT patients, node-positive patients or hormone-negative population were most sensitive to time horizon of the model (variation from 25-52 years) as well as the timing of onset of increasing cure proportion (variation from 48-120 months). For example, at a time horizon of 25 years, the ICER of node-positive patients increased by almost 35% (Garrison et al., 2019).

Authors have also worked on a probabilistic sensitivity analysis to show proportion of simulations that treatment with PTH was cost-effective. At a willingness to pay of $162,500, cost-effectiveness acceptability curve for the ITT patients in PHT and TH arms were cost-effective. If a threshold of $150,000/QALY was considered, only 45% of simulations would show that PHT is more cost-effective than TH. Based on that, it would be very difficult to draw a conclusion, because of the uncertainty of clinical effectiveness across the three subgroups, where the maximum benefit was shown in node-positive patients (Garrison et al., 2019).

Discounting:

Costs and outcomes was discounted at 3% per annum, and they referred to Cortes J paper, which has not shown a clear rationale why these figures were selected. The US Panel and Cost-Effectiveness in Health and medicine proposed 3% as a discount rate given the fact that QALY is not stable over time. On the other hand, Paulden et al argued that 3% discount rate is quite high, especially from healthcare perspective (Attema, Brouwer, & Claxton, 2018).

If discount rates for both costs and outcomes were not the same, it could have impacted the overall study results. Discount rates were not considered in sensitivity analysis, which is a limitation for this model. Based on Discounting article, that department of health is recommending that health benefits should not be discounted however many others argue that discounting health benefits should be as low as 1.5-2.5%. In most economic evaluations, choice of different discount rates will not only affect the final result of the analysis, it might also mislead decision makers. It is a good practice to evaluate the outcomes of using different discount rates using a sensitivity analysis (which is a limitation of this analysis) (Attema, Brouwer, & Claxton, 2018).

Future research:

As mentioned by the authors, a limitation of the analysis is the extrapolation of clinical trial data to a lifetime horizon, where there is uncertainty whether there is an additional benefit after the follow up period. In other words, the median follow up period in the clinical trial was 45.4 months, however the model projected a treatment effect of up to 84 months and reaching zero effect at 120 months where they relied on data from HERA and BCIRG-006. Accordingly, long-term efficacy data is needed from ongoing follow up of the APHINITY to estimate the cost-effectiveness of Pertuzumab as well as longer follow-up on the adverse-events profile (Garrison et al., 2019).

Another direction is focusing more on less-intensive therapies and searching for predictive biomarkers (including immunological markers) that would identify patient population benefiting from therapies, through randomized larger studies (e.g. ongoing trial (NCT01037117) to evaluate value of PET-CT in prediction of PCR to Peruzumab and Trastuzumab in HER2+ patients). Up until now, there is very limited data on identifying biomarkers in response to anti-HER therapies. Furthermore, with the satisfying outcomes of the standard therapy, additive therapies are expected to show limited benefits in general population (Stanton & Davidson, 2017).

External Validity:

(Mahmoud) Based on how authors presented the results, this model can be applied for US payers, with very limited external validity in other countries. The model setup, design, methodology followed and perspective can be valid in other countries, however other countries cannot take decisions or build assumptions based on these results unless they considered localizing some elements. Interventions vary in price, so costs should be localized (treatments, administrations, adverse events and utility costs). Which standard of care used and whether they use the brand or its biosimilar can be an element. EQ-5D estimates should be obtained from a more robust source since benefits and costs accrue differently to different constituents (patients, physicians, caregivers and society) (Durkee et al., 2016) 

Conclusion:

In conclusion, there is a massive role on oncologists on how to advise their patients on different treatment options given their clinical effectiveness, safety and tolerability profile, economic burden and patients’ preferences for risk tolerance (Garrison et al., 2019). Based on the data presented, Pertuzumab seems to be cost-effective only in node-positive HER2+ patients with debatable cost-effectiveness in other subgroups. Further long-term data is needed with a follow up on safety and tolerability profile to be able to draw a conclusion of the most appropriate approach to use Pertuzumab.

References:

1. Attema, A. E., Brouwer, W. B. F., & Claxton, K. (2018). Discounting in Economic Evaluations. PharmacoEconomics, 36(7), 745–758. https://doi.org/10.1007/s40273-018-0672-z
2. Dubois, R. W. (2016). Cost–effectiveness thresholds in the USA: are they coming? Are they already here? Journal of Comparative Effectiveness Research, 5(1), 9–12. https://doi.org/10.2217/cer.15.50
3. Durkee, B. Y., Qian, Y., Pollom, E. L., King, M. T., Dudley, S. A., Shaffer, J. L., … Horst, K. C. (2016). Cost-Effectiveness of Pertuzumab in Human Epidermal Growth Factor Receptor 2–Positive Metastatic Breast Cancer. Journal of Clinical Oncology, 34(9), 902–909. https://doi.org/10.1200/JCO.2015.62.9105
4. Garrison, L. P., Babigumira, J., Tournier, C., Goertz, H.-P., Lubinga, S. J., & Perez, E. A. (2019). Cost-Effectiveness Analysis of Pertuzumab With Trastuzumab and Chemotherapy Compared to Trastuzumab and Chemotherapy in the Adjuvant Treatment of HER2-Positive Breast Cancer in the United States. Value in Health, 22(4), 408–415. https://doi.org/10.1016/j.jval.2018.11.014
5. Miller, K. D. (2017). Questioning Our APHINITY for More. New England Journal of Medicine, 377(2), 186–187. https://doi.org/10.1056/NEJMe1706150
6. NICE appraisal, Pertuzumab for adjuvant treatment of HER2-positive early stage breast cancer. Technology appraisal guidance, 20 March 2019 www.nice.org.uk/guidance/ta569
7. Pernas, S., & Tolaney, S. M. (2019). HER2-positive breast cancer: new therapeutic frontiers and overcoming resistance. Therapeutic Advances in Medical Oncology, 11, 175883591983351. https://doi.org/10.1177/1758835919833519
8. Stanton, S. E., & Davidson, N. E. (2017). What lies beyond APHINITY for HER2-positive breast cancer? Nature Reviews Clinical Oncology, 14(12), 715–716. https://doi.org/10.1038/nrclinonc.2017.125
9. Swain, S. M., Baselga, J., Kim, S.-B., Ro, J., Semiglazov, V., Campone, M., … Cortés, J. (2015). Pertuzumab, Trastuzumab, and Docetaxel in HER2-Positive Metastatic Breast Cancer. New England Journal of Medicine, 372(8), 724–734. https://doi.org/10.1056/NEJMoa1413513
10. WHO report, 2018, Online GLOBOCAN 2018 database, accessible at http://gco.iarc.fr/, part of IARC’s Global Cancer Observatory.