Breast Cancer Research and Treatment

, Volume 136, Issue 1, pp 221–229

Cost effectiveness of new breast cancer radiotherapy technologies in diverse populations

Authors

    • New York University School of Medicine and Cancer Institute
  • Mary Katherine Hayes
    • Department of Radiation OncologyWeill Cornell Medical College
Epidemiology

DOI: 10.1007/s10549-012-2242-y

Cite this article as:
Gold, H.T. & Hayes, M.K. Breast Cancer Res Treat (2012) 136: 221. doi:10.1007/s10549-012-2242-y

Abstract

Accelerated partial breast radiotherapy (RT) strategies (3-D conformal external-beam RT (3-D CRT) and brachytherapy with balloon catheter) reduce time and transportation burdens of whole breast RT for breast cancer. Long-term clinical trial evidence is unavailable for accelerated modalities, but uncertainty might be acceptable for patients likely to receive suboptimal whole breast RT. The objective of this study is to assess the cost effectiveness of accelerated partial breast RT compared to on-time and delayed whole breast RT. The design used in this study is decision analytic Markov model. The data sources are published literature; and national/federal sources. The target population of this study is a hypothetical cohort of 60 years old women previously treated with breast-conserving surgery for node-negative, estrogen receptor-positive breast cancer with tumors <1 cm. The time horizon is 15 years, and the perspective is societal. The interventions are whole breast RT, 3-D CRT, and brachytherapy breast irradiation. The outcome measures are costs (2008 US$), quality-adjusted life years (QALYs), and incremental cost-effectiveness ratios. The base-case results were: 3-D CRT was the preferred strategy, costing on average $10,800 and yielding 11.21 QALYs. On-time whole breast RT costs $368,000/QALY compared to 3-D CRT, above the $100,000/QALY WTP threshold. 3-D CRT was also preferred over delayed whole breast RT. Brachytherapy was never preferred. Sensitivity analysis indicated that the results were sensitive to the rate of recurrence outside the initial tumor quadrant (“elsewhere failure”) in one-way analysis. Probabilistic sensitivity analysis indicated that results were sensitive to parameter uncertainty, and that the elsewhere-failure rate and treatment preferences may drive results. The limitation of this study is that efficacy estimates are derived from studies that may not fully represent the population modeled. As a conclusion, 3-D CRT was preferred to whole breast RT and for women likely to delay RT, indicating that 3-D CRT could be targeted more efficiently before randomized trial evidence.

Keywords

Accelerated partial breast radiotherapyCost-effectiveness analysisWhole breast radiotherapy

Background

Over 100,000 women are diagnosed with Stage I, early invasive breast cancer annually in the United States [1]. The goal of definitive treatment for these patients is to control local disease and prevent later recurrence and subsequent systemic disease. A proven treatment is breast-conserving surgery (BCS) with whole breast radiotherapy (RT), which requires daily therapy for 4–7 weeks [25]. Although over half of early invasive breast cancer patients receive BCS [6], not all receive subsequent whole breast RT, with delayed or incomplete whole breast RT in common [711]. Approximately 16 % of women ages 65+ diagnosed with Stage I breast cancer in fee-for-service Medicare (controlling for receipt of chemotherapy) and 14 % of elderly women diagnosed with Stages I–II breast cancer in health maintenance organizations (excluding subjects receiving chemotherapy) received whole breast RT delayed 8+ weeks following definitive surgery. Such delays can lead to worse outcomes [11, 12]. Patients receiving suboptimal care are more likely to be of lower socioeconomic status, older age, non-White race, and live in areas with lower availability of radiation oncologists [711]. Obtaining RT is associated with transportation, distance to treatment, and the amount of time required for whole breast RT [1316].

New RT technologies, such as accelerated partial breast RT, may play an important role especially for patients in whom optimal therapy cannot be assured. Accelerated partial breast RT strategies aim to deliver RT more efficiently to the tumor bed itself, where most ipsilateral (i.e., same side) breast recurrences occur [3, 1724]. Accelerated partial breast RT requires only twice-daily treatment for 5 days, greatly reducing the time and transportation burden required for whole breast RT [17, 25]. Whole breast RT may cause dismay or worry to some breast cancer patients due to the 4–7 week time burden; however, a mere week of treatment might seem simpler and, therefore, more attractive to start. Partial breast RT is available in several forms, including intracavitary brachytherapy (e.g., single brachytherapy-based partial breast RT using MammoSite™ or Contura™ balloon catheter technology) and three-dimensional conformal external-beam RT (3-D CRT, or intensity-modulated radiotherapy). Accelerated partial breast RT has other potential advantages. It may reduce the toxicity of whole breast RT, improve quality-of-life and cosmetic outcome, and allow for future breast RT should it become necessary [26].

Randomized clinical trials of brachytherapy-based partial breast RT and 3-D CRT compared to whole breast RT are ongoing [24, 27, 28], and the major trial (NSABP B-39/RTOG 0413) is powered to test for equivalent risk of ipsilateral breast tumor recurrence between accelerated partial breast RT and whole breast RT [24]. A small randomized study (n = 126) of brachytherapy versus whole breast RT with 30 months of follow-up showed no significant difference between locoregional tumor control or cancer-specific or relapse-free survival [29]. Short-term, non-randomized studies of accelerated partial breast RT have indicated low local recurrence rates and good cosmetic outcomes [17, 26, 30], but no study has more than 5 years of follow-up [31]. Even without longer term follow-up data, many patients in the United States are already undergoing accelerated partial breast RT [32]. However, with partial breast RT, there is theoretically a higher recurrence risk due to failure to treat elsewhere in the breast.

Brachytherapy-based partial breast RT costs up to $12,000 more per case for initial treatment compared to whole breast RT [33], so additional questions arise about whether this cost is justified [34], especially if the technology does not reach those who could benefit the most. In contrast, 3-D CRT treatment costs are similar to whole breast RT [33]. These new techniques, even if not quite as effective as whole breast RT, may still be of value to increase RT participation, especially for those who would have difficulty managing a full course of whole breast RT. We, therefore, developed a decision analytic model to evaluate whether new technologies can improve outcomes for women who might otherwise receive suboptimal whole breast RT and whether they are cost effective. We compared two accelerated partial breast RT approaches to optimal and delayed whole breast RT, using data from available studies and a large observational cohort that included patients receiving less-than-optimal care delivery [11], to answer the following questions: (1) is whole breast or one of the accelerated partial breast RT strategies more cost effective for patients expected to complete whole breast RT on time? (2) if a patient is expected to delay whole breast RT by 8+ or 12+ weeks, is a different strategy more cost effective? and (3) how effective should accelerated partial breast RT be to be cost effective?

Methods

Model overview

We developed a decision analytic Markov model to evaluate several RT scenarios following BCS for Stage I breast cancer. Hypothetical patients underwent one of the following RT techniques: traditional whole breast RT, balloon brachytherapy, or 3-D CRT. We also evaluated the effects of whole breast RT delayed by 8+ or 12+ weeks as it occurs in real-world situations [11, 12]. A hypothetical cohort of 75,000 women aged 60 years old was followed, starting with no evidence of local disease (NED) after BCS, and then proceeding to the health states of (1) having an ipsilateral breast recurrence in either the same quadrant as the original tumor or an “elsewhere” (i.e., other) quadrant, (2) distant metastases, (3) death from other causes, or (4) death from breast cancer (only possible following distant metastases) (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-012-2242-y/MediaObjects/10549_2012_2242_Fig1_HTML.gif
Fig. 1

Decision analytic model. NED no evidence of disease, IBTR ipsilateral breast tumor recurrence, Q quadrant

Consistent with patient populations eligible for accelerated partial breast RT, all subjects were presumed to have node-negative, estrogen receptor-positive disease. Tumors were less than 1 cm, and all were prescribed 5 years of hormonal treatment and no chemotherapy. Results were virtually identical whether subjects took anastrozole or tamoxifen, so results are shown for anastrozole. The model simulated 15 years of follow-up, a time horizon that is considered clinically relevant [35]. We conservatively assumed that the rate of same-quadrant failures was equivalent across RT modalities, which ignored potential benefits of more efficient targeting of RT. A second conservative model assumption was that the rates of recurrence in other quadrants [i.e., “elsewhere-quadrant” failure (EF)] were three times higher for both balloon brachytherapy and 3-D CRT than for whole breast RT, because partial breast RT targets and treats primarily the same quadrant of the breast as the initial tumor, and because trial evidence indicates that compared to subjects receiving whole breast RT, women receiving no RT have a 3-times higher recurrence rate in the breast [36, 37]. Although these recurrences tend to be in the same quadrant, our model assumed a possible elsewhere-quadrant recurrence risk to account for the unknown consequences of targeting radiation only to the tumor bed. Outcomes were quality-adjusted life years (QALYs), costs (2008 US$), and incremental cost-effectiveness ratios (ICERs in terms of $/QALY gained). A 3 % discount rate was applied to all costs and benefits [38], and the willingness-to-pay (WTP) threshold was $100,000/QALY [39]. All analyses were conducted in TreeAge Pro 2009, v.1.0.2 (Williamstown, MA). This study was approved as exempt from the authors’ institutional review board.

Data sources

Natural history information included the probability of moving from the NED state to either recurrence [35] or metastatic disease [40, 41], from recurrence to metastatic disease [42], or from metastatic disease to death [43] (Table 1). Elsewhere-quadrant failures were assumed to be three times higher in years 0–5 for accelerated partial breast RT, and identical to same-quadrant failures in years 6–15 [33, 37, 44]. There is limited evidence concerning hazard rates of recurrence following delayed whole breast RT. Therefore, we assumed that the hazard ratio for recurrence following whole breast RT delayed by 8+ weeks was 1.14, and if delayed by 12+ weeks the risk was almost threefold (HR = 2.77) [11]. Rates of death from other causes were taken from US Life Tables [45]. Quality-of-life (i.e., utility) values ranged from 0-1 for each health state and came from previous studies [4648]. Utility values were multiplied by life expectancy to yield QALYs.
Table 1

Model inputs and assumptions

 

Base Case

References/notes

Natural history probabilities

  

 NED—recur, 1–5 year

6.7 %

[35]

 NED—recur, 6–15 year

3.3 %

[35]

 Recur—metastatic disease, 10 year

20 %

[42]

 NED—metastatic disease, 1–10 year

11 %

[59]

 NED—metastatic disease, 11–15 year

6.7 %

[59]

 Metastatic disease—death

27.89 %

[43]

 % of failures that are local, 1–5 year

93 %

[44]

 % of failures that are local, 6–10 year

62 %

[44]

Hazard ratios

 AI benefit, IBTR, 1–5 year

0.4559

[40, 41]

 AI benefit, IBTR, 6–10 year

0.6555

[40, 41]

 AI benefit, IBTR, 11–15 year

1

[40, 41]

 AI benefit, metastatic disease, 1–5 year

0.63168

[40, 41]

 AI benefit, metastatic disease, 6–10 year

0.67344

[40, 41]

 AI benefit, metastatic disease, 11–15 year

1

[40, 41]

 Elsewhere-failure rate, APBI*

3

[55]

 Recurrence hazard ratio for 12+ week delayed RT

2.77

[11]

 Recurrence hazard ratio for 8+ week delayed RT

1.14

[11]

Utilities (preferences/rating for health states)

 Recurrence

0.82

[46]

 NED after RT, year 1–2

0.92

[46]

 NED after RT, year 3–15

1

Estimate

 Metastatic disease

0.62

[47, 48]

 Brachytherapy balloon RT (e.g., MammoSite™)

0.92

Estimate

 3-D CRT

0.92

Estimate

 Whole breast RT

0.92

Estimate

Costs, 2008

 Whole breast RT

$8,000

[33], adjusted from 2003$

 MammoSite™

$19,360

[33], adjusted from 2003$

 3-D CRT

$7,800

[33], adjusted from 2003$

 Mastectomy

$17,000

[50], adjusted from 1998$

 Breast reconstruction

$12,700

[50], adjusted from 1998$

 Metastatic disease, year 1, survive

$21,000

[51], adjusted from 1995$

 Metastatic disease, year 1, die

$35,000

[51], adjusted from 1995$

 Metastatic disease, after year 1, survive

$8,400

[51], adjusted from 1995$

 Metastatic disease, after year 1, die within year

$14,000

[51], adjusted from 1995$

Patient time costs, 2008$**

 Whole breast RT

$1,200

[33, 60]

 Brachytherapy balloon RT (e.g., MammoSite™)

$550

[33]

 3-D CRT

$550

[33]

NED no evidence of local disease, AI aromatase inhibitor (anastrozole), IBTR ipsilateral breast tumor recurrence, APBI accelerated partial breast irradiation

* Elsewhere-failure rates for both accelerated partial breast irradiation modalities are 3 times the rate for whole breast radiotherapy

** Patient time costs include time and transportation, national average hourly wage for females aged 55–64, transportation time, cost/mile to drive, and cost of parking

Treatment costs for RT modalities were based on Medicare reimbursement rates for each and included the technical and professional components reimbursed by Medicare for each part of treatment, including treatment planning, physics and dosimetry, and treatment delivery [33]. Cost estimates for mastectomy for ipsilateral recurrences and subsequent breast reconstruction (for 50 % of subjects [49]) came from stage-specific estimates [50]. Estimates of the costs associated with metastatic disease were taken from the literature; when a subject experienced metastatic disease, she incurred an estimated initial $12,000 of health care costs, followed by $8,400 for each year she survived with metastatic disease. In the year of death, we assumed she incurred $14,000. These estimates reflected higher costs in both the initial year of metastasis and the final year of life [51]. All costs were updated to 2008 US$ using the consumer price index calculator [52]. Patient time costs were based on the national average hourly wage for females aged 55-64, and transportation costs were drawn from previous literature [33] including transportation time, waiting time, cost-per-mile traveled, and cost of parking. The societal perspective was used [38].

Model validation

We ran the model for 75,000 hypothetical patients and evaluated its validity by comparing the estimated mean survival with that predicted by Adjuvant!Online [53, 54], which provided expected mortality rates for breast cancer patients with specific characteristics. Adjuvant!Online is an on-line tool that helps oncologists estimate risks of mortality and side effects, given different clinical and treatment scenarios. The Markov cohort model only slightly underestimated overall survival of 85.2 % at 10 years, compared to the Adjuvant!Online estimate of 86.9 % for the same population.

Sensitivity analyses

One-way and probabilistic sensitivity analyses were conducted to identify how robust the model was to changes in the data and assumptions. One-way sensitivity analyses varied factors a priori expected to affect model results including the elsewhere-failure rate, the hazard rate for delayed RT, probability of breast reconstruction following mastectomy, and quality-of-life (utility) associated with treatment. We also explored the impact of removing patient time and transportation costs. Probabilistic sensitivity analysis was conducted using the societal perspective and varied patient time and transportation costs (uniform distribution ranging 30 % above to 30 % below the base case value), the utility of whole breast RT (uniform distribution from 90 to 100 % of base case partial breast RT value), the EF rate for the accelerated partial breast RT strategies (normal distribution with mean of 3 and standard deviation of 0.3), and the hazard ratios for recurrence for delayed RT (normal distributions based on base case and standard deviations calibrated separately for 8-week delay (0.07) and for 12-week delay (0.08). Results were based on a $100,000/QALY threshold for 5,000 samples.

Role of funding source

The funding source played no role in the study.

Results

The preferred strategy was 3-D CRT compared to whole breast RT completed on time, which costs over $367,000/QALY gained (Table 2). Brachytherapy (e.g., MammoSite™ or Contura™) was more expensive and equally as effective as 3D-CRT in all scenarios. Whole breast RT delayed by 8 weeks was not preferred compared to 3D-CRT, which in turn dominated a 12-week delay in whole breast RT.
Table 2

Costs, effectiveness, and cost effectiveness of radiotherapy modalities

 

Cost ($)

Incremental cost ($)

Effectiveness (QALYs)

Incremental effectiveness (QALYs)

ICER ($/QALY)

3-D CRT

10,826

 

11.208

  

On-time WBRT

11,534

708

11.209

0.002 = 17.5 QALDs

367,740

Brachytherapy

23,226

11,692

11.208

0

Dominated

3-D CRT

10,826

 

11.208

  

8+ week-delay WBRT

11,738

204

11.208

0

Dominated

12+ week-delay WBRT

11,815

281

11.207

0

Dominated

QALDs quality-adjusted life days, WBRT whole breast radiotherapy

All incremental costs and QALYs are calculated compared to 3-D CRT

Through reasonable variable ranges in our one-way sensitivity analyses, the model results were sensitive to the EF rate for the accelerated RT options, the main indicator of treatment effectiveness (Fig. 2), and the probability of breast reconstruction following mastectomy. With 3-D CRT, the preferred base-case strategy, whole breast RT was not cost effective until the rate of EFs following 3-D CRT was more than 5.4 times higher than whole breast RT (Fig. 2), well beyond the expected EF rate [24, 37, 55] and base-case assumption of 3.0. Even when the EF rate for brachytherapy is equivalent to 3-D CRT or whole breast RT, brachytherapy is always dominated because of its higher costs. In order for brachytherapy-based RT to become more attractive than 3-D CRT, it would have to be slightly more effective at reducing same-quadrant recurrences and slightly improve quality-of-life, or cost $11,560 less than current Medicare reimbursement amounts. Our base case assumption for the probability of breast reconstruction following mastectomy was 50 %; only when this probability exceeded 94 % did whole breast RT dominate all other strategies. Exploratory analyses using the payer perspective indicated that whole breast RT was very cost effective at $30K/QALY gained compared to 3-D CRT when patient time and transportation costs were excluded.
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-012-2242-y/MediaObjects/10549_2012_2242_Fig2_HTML.gif
Fig. 2

Sensitivity analysis of incremental cost-effectiveness ratio of whole breast radiotherapy relative to 3-D conformal radiotherapy when varying elsewhere-failure rate of 3-D CRT as ratio of whole breast RT. ICER incremental cost effectiveness ratio. Elsewhere-failure rate is different in years 0–5 only. Threshold point where ICER of elsewhere-failure rate crosses $100,000/QALY line is 5.4

In probabilistic sensitivity analyses with $100,000/QALY as a threshold, 3-D CRT dominated or was preferred in only 30 % of simulations compared to whole breast RT completed on time, whereas whole breast RT was more costly but below the WTP threshold in 70 % of simulations. When comparing 3-D CRT to delayed RT, however, 3-D CRT became slightly more favorable, dominating or preferred in 37 % and 40 % of simulations versus RT delayed by 8+ and 12+ weeks, respectively.

Discussion

For patients who may be likely to delay whole breast RT, our analyses showed that 3-D CRT is a cost-effective option to help improve health and reduce disparities in breast cancer treatment outcomes due to suboptimal care. In the base case, 3-D CRT is preferred, and the next best strategy, whole breast RT, costs $367K/QALY gained, well above the $100K/QALY WTP threshold. Brachytherapy-based RT was never a cost-effective strategy.

Our study considered different analytic perspectives to assess the effects of incorporating patient time and transportation costs and patients’ adherence to timely whole breast RT, unlike a previous study comparing 3-D CRT to whole breast RT [34]. That study also found that 3-D CRT was the most cost-effective treatment strategy in a similar population [34]. Our analyses discerned who might benefit more from accelerated partial breast RT modalities depending on their likelihood of receiving timely whole breast RT. Patient time and transportation costs and ability to obtain timely care were not only important in the decision about which RT modality to undergo, but also were meaningful in exploratory analyses. Preferences for a shorter course of RT and a patient’s likelihood of delaying whole breast RT appeared important for the decision of whether to adopt 3-D CRT. 3-D CRT is an example of an intervention that appears to have nearly similar effectiveness to standard care for certain select cancers, and costs less overall; it could, therefore, save the health system money for selected patients depending on their preferences [56]. The analysis showed that it is worthwhile to offer 3-D CRT to patients who might delay whole breast RT by more than 8 weeks, including women of lower socioeconomic status [11] or those who indicate that the time and transportation burden of whole breast RT would be a barrier to initiation or completion. It should be noted, however, that some patients may not be able to access twice-daily accelerated treatment regimens without reducing other responsibilities including work. Other shorter yet similarly effective interventions to traditional whole breast RT, such as Canadian fractionation of RT that takes 16 fractions over 22 days (instead of 25 fractions over 35 days) [5], also would cost less and be less burdensome than traditional whole breast RT and may be more effective with cost only slightly more than 3-D CRT.

Estimating the cost effectiveness of any intervention requires incorporating pre-existing data and making assumptions. We used the best available data on partial breast RT from case series and early-phase trials to evaluate the effectiveness of accelerated partial breast RT modalities. Although the results varied depending on the elsewhere-failure rate for 3-D CRT, an estimate with limited evidence from the literature, model conclusions were relatively stable for a wide range of EF values even up to 5.4 times that of whole breast RT. It is unlikely that the EF rate for 3-D CRT is that high, given short-term clinical trial evidence [17].

In addition, we incorporated information from a retrospective cohort study of Medicare beneficiaries to assess the cost effectiveness of delayed RT compared to the other treatment strategies. Both data types represent “all comers” to RT, but we know that eligibility criteria for accelerated modalities are much narrower [24, 57], and therefore patients’ risks for recurrence likely are lower than the general population who may seek RT. We did not incorporate health or quality-of-life impacts or costs of adverse effects or complications from RT, although it is known that the brachytherapy balloon can burst [58] requiring re-insertion or change in treatment strategy [17]. On the other hand, balloon brachytherapy may be the only RT modality that significantly reduces heart and lung toxicity and possibly allows for a second course of RT should there be a recurrence, because it is more focused than 3-D CRT or whole breast RT; these benefits were neither included in the model, nor were treatment-delivery complications such as machine breakdowns that would affect either treatment modality. Furthermore, we did not model the interventions for an older population that might have lower time costs due to being a predominantly retired group. Finally, probabilistic sensitivity analysis indicated that parameter uncertainty was a factor in stability of the results. With only about a third of simulations indicating that 3-D CRT would be cost effective, further exploratory analysis showed that there clearly is a need for higher quality data on treatment utilities and elsewhere-failure rates, which seemed to drive the variability in the probabilistic sensitivity analysis.

Use of 3-D CRT, even in the absence of long-term clinical trial data, is an RT strategy that likely would benefit most eligible patients, especially those who may have difficulty accessing a traditional whole breast RT protocol due to transportation issues or for other reasons causing them to delay RT by 8 or more weeks [11, 12]. It should be noted that this study does not attempt to reduce disparities directly but the modality itself may provide a novel means to reduce the hurdle for patients obtaining RT by making RT treatment easier; therefore, this modality should be studied in further disparities research. This work emphasizes that additional information about quality-of-life effects and elsewhere-failure rates of the newer partial breast RT modalities is needed; if brachytherapy-based partial breast RT is shown to yield better quality-of-life than 3-D CRT due to fewer side effects up front or long term, it may become more favorable. Completion of the on-going randomized trials will provide stronger clinical evidence regarding the modalities being compared, but in the meantime, decision analytic studies provide insights into the optimal treatment options for diverse populations.

Acknowledgments

This work was supported in part by the Center for Education and Research on Therapeutics (Agency of Health Research and Quality [Grant Number 1U18-HS016075]). The study was deemed exempt by the Institutional Review Board at Weill Cornell Medical College. Alvin I. Mushlin, MD, ScM, Bruce R. Schackman, PhD, and Natasha K. Stout, PhD, suggested critical revisions but any mistakes are the authors’ alone.

Conflict of interest

The authors have no conflicts of interest to disclose. This study was not supported by a pharmaceutical company.

Copyright information

© Springer Science+Business Media, LLC. 2012