Cost-Effectiveness Analysis of Trastuzumab in the Adjuvant Setting for Treatment of HER2-Positive Breast Cancer

489 Cost-Effectiveness Analysis of Trastuzumab in the Adjuvant Setting for Treatment of HER2-Positive Breast Cancer Louis P. Garrison Jr, PhD 1 Deborah Lubeck, PhD 2 Deepa Lalla, BPharm, PhD 2 Virginia Paton, PharmD 2 Amylou Dueck, PhD 3 Edith A. Perez, MD 4 1 Department of Pharmacy, University of Washington, Seattle, Washington. 2 Genentech, Inc., South San Francisco, California. 3 Department of Health Sciences Research, Mayo Clinic, Rochester, Minnesota. 4 Department of Hematology Oncology, Mayo Clinic, Jacksonville, Florida. Funded in part by a contract between Genentech and the University of Washington with unrestricted publication rights. Genentech and Mayo Clinic coauthors provided raw data, reviewed the model and assumptions, and commented on article drafts. University of Washington carried out all analyses and interpreted data and results independently. D. P. Lubeck is an employee of and owns stock in Genentech. E. A. Perez receives research grant support from Genentech, Bristol, Sanofi-Aventis, Novartis, Glaxo, Elli Lilly, Astra Zeneca, Pfizer, and Onyx, and has received honoraria for consulting from Bristol, Genentech, and Glaxo in the last year. L. P. Garrison Jr has been a consultant to Genentech and Roche. D. Lalla is an employee of and owns stock in Genentech. V. Paton is an employee of Genentech. The authors thank Nina Oestreicher and David Veenstra for their comments on an early version of the model. Support for third-party writing assistance for this article was provided by Genentech. Address for reprints: Louis P. Garrison Jr, PhD, Department of Pharmacy, University of Washington, Health Sciences Building, Seattle, WA 98195-7630; Fax: (206) 543-3835; E-mail: lgarrisn@ u.washington.edu Received December 22, 2006; revision received March 16, 2007; accepted March 21, 2007. BACKGROUND. Adding trastuzumab to adjuvant chemotherapy provides significant clinical benefit in patients with human epidermal growth factor receptor 2 (HER2)-positive breast cancer. A cost-effectiveness analysis was performed to assess clinical and economic implications of adding trastuzumab to adjuvant chemotherapy, based upon joint analysis of NSABP B-31 and NCCTG N9831 trials. METHODS. A Markov model with 4 health states was used to estimate the cost utility for a 50-year-old woman on the basis of trial results through 4 years and estimates of long-term recurrence and death based on a meta-analysis of trials. From 6 years onward, rates of recurrence and death were assumed to be the same in both trastuzumab and chemotherapy-only arms. Incremental costs were estimated for diagnostic and treatment-related costs. Analyses were from payer and societal perspectives, and these analyses were projected to lifetime and 20- year horizons. RESULTS. Over a lifetime, the projected cost of trastuzumab per quality-adjusted life year (QALY; discount rate 3%) gained was $26,417 (range, $9104-$69,340 under multiway sensitivity analysis). Discounted incremental lifetime cost was $44,923, and projected life expectancy was 3 years longer for patients who received trastuzumab (19.4 years vs 16.4 years). During a 20-year horizon, the projected cost of adding trastuzumab to chemotherapy was $34,201 per QALY gained. Key costeffectiveness drivers were discount rate, trastuzumab price, and probability of metastasis. The cost-effectiveness result was robust to sensitivity analysis. CONCLUSIONS. Trastuzumab for adjuvant treatment of early stage breast cancer was projected to be cost effective over a lifetime horizon, achieving a cost-effectiveness ratio below that of many widely accepted oncology treatments. Cancer 2007;110:489 98. Ó 2007 American Cancer Society. KEYWORDS: cost effectiveness, pharmacoeconomics, economics, HER2, trastuzumab, Herceptin, adjuvant, early breast cancer, cost utility. Breast cancer, the second leading cause of cancer death in women, is responsible for approximately 15% of cancer deaths in the United States (US). 1 It was estimated that in 2006 there would be 214,640 new cases of invasive breast cancer in the US and that 40,970 women would die from this disease. 1 Approximately 20% to 30% of patients with breast cancer have human epidermal growth factor receptor 2 (HER2)-positive disease, which is associated with a poor prognosis. 2 4 Trastuzumab (Herceptin; Genentech, South San Francisco, Calif), a monoclonal antibody that targets HER2, is approved for the treatment of HER2-positive metastatic breast cancer, either as first-line therapy in ª 2007 American Cancer Society DOI 10.1002/cncr.22806 Published online 25 June 2007 in Wiley InterScience (www.interscience.wiley.com).

490 CANCER August 1, 2007 / Volume 110 / Number 3 combination with paclitaxel or as monotherapy in patients who have received 1 or more chemotherapy regimens for their metastatic disease. Trastuzumab was also recently approved for the adjuvant treatment of HER2-positive early breast cancer, in combination with chemotherapy, after assessment in several trials. 5 8 The joint analysis of the National Cancer Institutesponsored National Surgical Adjuvant Breast and Bowel Project (NSABP) B-31 and the North Central Cancer Treatment Group (NCCTG) N9831 trials demonstrated that the addition of trastuzumab to the standard adjuvant regimen of doxorubicin and cyclophosphamide (AC) followed by paclitaxel reduced the risk of recurrence by 52% (P <.0001) and improved survival by 33% (P 5.015) in patients with HER2-positive, operable, early breast cancer. 5 The incidence of New York Heart Association (NYHA) class III or IV congestive heart failure or death from cardiac causes was higher in the trastuzumab-containing arms than in the control arms. 5,9 11 Five-year survival rates for early stage breast cancer have significantly improved during the past 3 decades (88% in 1995 2001 vs 75% in 1974 1976; P <.05) 1 ; the reduction in 5-year mortality reflects in part the cumulative progress made in adjuvant therapy for the disease during the past 20 years. However, individual incremental improvements in the efficacy of new adjuvant therapies are seldom large in either relative or absolute terms. Hence, the 52% reduction in disease recurrence observed with the addition of trastuzumab to standard adjuvant chemotherapy in the subgroup with aggressive HER2-positive cancers is a substantial development compared with recent advances in other cancers. The objective of this analysis was to develop a model to estimate the incremental cost effectiveness of adjuvant trastuzumab for early stage, HER2-positive, breast cancer in the US, based upon the joint analysis of the NSABP B-31 and NCCTG N9831 trials. Two analyses of adjuvant trastuzumab in the US setting were published recently. 12,13 Although their overall conclusions were similar to earlier meeting abstracts and European analyses, 14 21 the base-case cost-effectiveness estimates varied substantially from $18,970 to $39,982 per adjusted life-year (QALY) gained. This model and analysis, based upon a different modeling approach and assumptions, provide additional evidence on potential cost effectiveness. MATERIALS AND METHODS Model Overview This economic evaluation was a cost-utility assessment that compared projected costs and outcomes of 2 alternative treatment options on the basis of trial results from both a payer and a societal perspective. The cost-effectiveness analysis was based on the joint analysis comparison of the nontrastuzumabcontaining arm (Arm 1 in NSABP B-31 and Arm A in NCCTG N9831) with the trastuzumab-containing arm (Arm 2 in NSABP B-31 and Arm C in NCCTG N9831) (Fig. 1). 5 Disease-free survival was the prespecified primary endpoint of the combined efficacy analysis of these trials. Long-term survival and outcome data were projected from the 4-year joint analysis data 5 and extrapolated to 20-year and lifetime horizons based on a Markov model. A Markov model comprising 4 health states (treatment, disease-free, distant recurrence, and death) was developed to assess expected costs and life-years and QALYs gained over time (Fig. 2). The model assumed that all patients received treatment during the first year while they were still disease-free. After distant recurrence, the process of progression to metastasis and death was modeled as a simple aggregate cost and QALY outcome. In this aggregate, all patients were assumed to have received trastuzumab treatment after distant recurrence. To further simplify the model, and because no reliable estimates of differences between study arms were available, health states of contralateral disease or local recurrence were not explicitly defined, but their impact is reflected in the estimated time to distant recurrence based on trial data. Cardiac dysfunction associated with the addition of trastuzumab was assumed to be reversible (although more data are needed to confirm this) 10,22 and to have no direct mortality impact beyond that reflected in the 4-year trial period. Patient Population The base-case estimates were for a typical patient enrolled in the trials a 50-year-old woman. In the arms of the 2 trials, 50% to 52% of patients were aged younger than 50 years. 5 Estimates of actual disease recurrence and patient survival were based on the joint analysis. 5 An intent-to-treat basis was followed. Patients were eligible if they had invasive breast cancer resected by lumpectomy or mastectomy and axillary dissection with pathologically involved axillary nodes (amended in NCCTG N9831 to allow patients with high-risk, node-negative disease). The tumor had to be HER2-positive (immunohistochemistry score of 31 or fluorescence in situ hybridization [FISH]-positive) and had to be verified centrally [NCCTG N9831] or by an approved laboratory [NSABP B-31]). Patients had to have normal hematologic, hepatic, and renal function, and normal

Cost-Effectiveness of Adjuvant Trastuzumab/Garrison et al. 491 FIGURE 1. NSABP B-31 and NCCTG N9831 trial schemas. NCCTG N9831 s 3 regimen arms all begin with doxorubicin (60 mg/m 2 ) and cyclophosphamide (600 mg/m 2 ) given every 3 weeks for 4 cycles; this is followed by: weekly paclitaxel (80 mg/m 2 ) for 12 weeks (Arm A); 12 weeks of weekly paclitaxel followed by weekly trastuzumab (4 mg/kg loading dose, followed by 2 mg/kg) for 52 weeks (Arm B); or weekly paclitaxel plus trastuzumab for 12 weeks, followed by weekly trastuzumab alone for 40 weeks (Arm C). NSABP B-31 s 2 regimen arms begin with doxorubicin (60 mg/m 2 ) and cyclophosphamide (600 mg/m 2 ) given every three weeks for 4 cycles; this is followed by: every 3 weeks (175 mg/m 2 ) or weekly (80 mg/m 2 ) paclitaxel for 12 weeks (Arm 1); or every 3 weeks (175 mg/m 2 ) or weekly (80 mg/m 2 ) paclitaxel for 12 weeks plus trastuzumab for 12 weeks, followed by weekly trastuzumab alone for 40 weeks (Arm 2). FIGURE 2. Markov model comprising 4 health states and possible transitions between states for the trastuzumab-containing and control arms. AC indicates doxorubicin and cyclophosphamide; T, paclitaxel; H, trastuzumab. left ventricular ejection fraction, as assessed by multigated acquisition scan or echocardiogram. Patients were excluded when they had locally advanced or distant disease, had received prior anthracycline or taxane therapy, had significant sensory or motor neuropathy, or had past or active cardiac disease. Costs Healthcare resource costs were based on 2006 Medicare reimbursement rates and other published data (Table 1) 5,23 29 and are expressed in US dollars throughout. On the basis of the 20% to 30% incidence of HER2 positivity among breast tumors, 2 we assumed that 5 tests were performed on average for every patient identified for trastuzumab treatment, recognizing that some patients may have had more than 1 test. Both immunohistochemistry and FISH tests were available, and 30% of tests were assumed to be FISH in the base case. 29 The estimated drug cost per patient was based on the average wholesale price per vial 24 and the duration of treatment. 5 A total of 53 infusions were planned, but 19.9% of patients discontinued trastuzumab treatment 30 ; and based on trial data, the average total dose received was estimated to be 25% less than the protocol dose. In both trials, cardiac function was monitored before, during, and after therapy. On the basis of trial protocol and discontinuations, an estimated mean of 4.7 multigated acquisition scans or echocardiogram examinations per patient were performed to assess and monitor left ventricular ejection fraction. The mean cost per left ventricular ejection fraction examination was $367, yielding a total cost of $1726. 23 On the basis of results of the NCCTG N9831 trial, the rate of cardiac events (defined NYHA class III or IV congestive heart failure, or probable or definite cardiac death) was assumed to be 2.9% in the trastuzumab-containing arm and 0% in the control arm. 5 The mean cost of treating a cardiac event was assumed to be $1979. 23 On the basis of published data, the cost of metastatic breast cancer was modeled as a single cost of $40,000, which was adjusted for inflation, 25 and mean costs associated with death were assumed to be $10,000. As trastuzumab therapy is scheduled for 52 weeks, which is beyond the time period for standard chemotherapy alone, the patient incurs additional expense from a societal perspective. Incremental indirect costs related to trastuzumab treatment were estimated, reflecting time and expenses for travel and medical visits (Table 1). Time was valued at the average hourly compensation rate from the US Bureau of Labor Statistics, 31 and travel costs were

492 CANCER August 1, 2007 / Volume 110 / Number 3 TABLE 1 Input Parameters Input parameters Base case Range Source Direct Costs Diagnostic tests IHC $89 $75 $100 CMS 23 calculated FISH $482 $400 $600 CMS 23 calculated Proportion of HER2 tests by FISH method 30% 10% 50% Elkin et al. 2004 29 H infusion (per 440-mg vial) $2987 $2345 $3080 www.analysource.com 24 Administration of H infusion $172.81 $129 $216 CMS 23 calculated LVEF exam $367 $275 $459 CMS 23 calculated Mean cost per cardiotoxicity event $1979 $1500 $4000 CMS 23 calculated Recurrence cost (distant metastasis) $40,000 $30,000 $60,000 Rao et al. 2004 25 Inflation adjusted Cost associated with death $10,000 $5000 $20,000 Assumption Indirect costs Value of patient time per hr $25.57 $19.18 $31.96 BLS 31 Travel cost (auto) per infusion $14.55 $10.91 $18.19 IRS 32 Utilities Year 1 AC?T including treatment 0.770 0.750 0.950 Oestreicher et al. 2005 26 Year 1 AC?TH treatment 0.710 0.600 0.950 Oestreicher et al. 2005 26 Year 21 stable disease utility 0.900 0.850 0.950 Oestreicher et al. 2005 26 Year 21 recurrence utility 0.600 0.400 0.700 Hornberger et al. 2002 27 Cumulative probability, Year 4 Control arm (AC?T) Recurrence-free survival 0.719 Romond et al., 2005 5 Overall survival 0.870 Romond et al., 2005 5 H arm (AC?TH) Recurrence-free survival 0.856 0.820 0.890 Romond et al., 2005 5 Overall survival 0.911 0.875 0.945 Romond et al., 2005 5 Annual rate of distant recurrence after 5 y AC?T (Years 6 to 9) 0.048 0.037 0.061 EBCTCG, 2005 28 AC?TH (Years 6 to 9) 0.048 0.037 0.061 EBCTCG, 2005 28 AC?T (Years 101) 0.033 0.026 0.043 EBCTCG, 2005 28 AC?TH (Years 101) 0.033 0.026 0.043 EBCTCG, 2005 28 Annual probability of death following distant metastasis Control arm (AC?T) 0.323 0.242 0.404 Hornberger et al., 2002 27 H arm (AC?TH) 0.323 0.242 0.404 Hornberger et al., 2002 27 IHC indicates immunohistochemistry; FISH, fluorescence in situ hybridization; CMS, Centers for Medicare and Medicaid Services; H, trastuzumab; LVEF, left ventricular ejection fraction; AC, doxorubicin and cyclophosphamide; T, paclitaxel; BLS, US Bureau of Labor Statistics; IRS, Internal Revenue Service; EBCTCG, Early Breast Cancer Trialists Collaborative Group. valued at Internal Revenue Service business travel rates ($0.485 per mile for a 30-mile round trip). 32 This is only a partial societal perspective, as any gain from potential greater labor force attachment or productivity was not estimated. Quality-of-Life Utility Assumptions Adjustment was made for quality of life in the 4 different health states, but no differences were assumed across the arms with the exception of the treatment period. Utility during the first year of treatment was assumed to be 0.77 in the control arm and 0.71 in the trastuzumab arm (Table 1). The lower utility estimate for the trastuzumab arm was because of the longer duration of treatment and the assumption that utility during trastuzumab treatment was lower than utility without treatment (0.8 vs 0.9). In both arms, the utility level of 12 weeks of paclitaxel use was 0.5; the assumed values of 0.77 and 0.71 are weighted averages of these 3 respective values based on the chemotherapy protocol. During the stable-disease state before recurrence, utility was assumed to be 0.9. 26 After recurrence, utility with metastatic breast cancer was assumed to be 0.6. 27 These utility scores were used to adjust survival time in each health state to produce a composite measure of utility the quality-adjusted life year (QALY). The QALY is a widely used measure that adjusts the quantity of life gained by the quality of that life. The cost per QALY is a useful measure, as the ratio is not disease-

Cost-Effectiveness of Adjuvant Trastuzumab/Garrison et al. 493 specific and can be compared across a spectrum of therapeutic interventions. Cost-utility analyses are, thus, the preferred type of economic evaluation. 33,34 Both costs and outcomes were discounted at 3%, following the recommendation of Public Health Service Guidelines. 34 The discount adjusts for the finding that current dollars and health benefits are worth more than future dollars and health benefits. Projecting Recurrence-free and Overall Survival The first 5 years of distant-free recurrence and survival estimates were based on empirically fitting the observed distributions for the first 4 years of the reported trial data and extrapolating the trends to the fifth year. Thus, assumptions were made about the Markov transition probabilities during the first 4 years to attempt to reproduce the actual cumulative results at 3 and 4 years, recognizing that exact reproduction is not possible, as background mortality is an unobserved factor. From 6 years onward, annual transition probabilities to recurrence and death were assumed to be the same in both arms. Given the proportion of patients in each of 3 health states (disease-free, distant recurrence, and death) at the end of 5 years, transition rates to either recurrence or death were based on the literature. Patients in each arm were assumed to progress to recurrence based on the rates observed in the Early Breast Cancer Trialists Collaborative Group 2000 Report. 28 Before recurrence, the probability of dying ( background mortality ) was based on the standard rate of mortality for a 50-year-old woman in the US. 35 After recurrence, it was assumed that patients in both treatment arms would receive trastuzumab. The annual death rate for women with metastatic breast cancer was assumed to be 32.3%, a rate that produces the projected 2.1 years of overall survival for women with metastatic breast cancer treated with trastuzumab. 27 Sensitivity Analyses Robustness of results was assessed by using sensitivity analysis. Parameter values were varied across a range of plausible values to determine the impact of variation on the overall results. Sensitivities to changes in individual parameters and to simultaneous, multiway changes were assessed. TABLE 2 Cost-Effectiveness Results Payer Societal Perspective/Measures Base case Alternate base case Time Horizon Lifetime 20 Year Lifetime AC?T Costs $28,749 $25,055 $28,749 LYs 11.88 10.12 11.88 QALYs 10.08 8.56 10.08 AC?TH Costs $73,672 $71,847 $75,746 LYs 13.72 11.48 13.72 QALYs 11.78 9.85 11.78 Differences Costs $44,923 $43,913 $46,997 LYs 1.84 1.36 1.84 QALYs 1.70 1.28 1.70 Cost per LY $24,435 $32,361 $25,542 Cost per QALY $26,417 $34,201 $27,637 AC indicates doxorubicin and cyclophosphamide; T, paclitaxel; H, trastuzumab; LY, life year; QALY, quality-adjusted life year. RESULTS Cost Effectiveness Payer perspective In the base case (lifetime horizon), from a payer perspective, the addition of trastuzumab to chemotherapy is estimated to cost an additional $44,923, with an expected gain of 1.70 QALYs (Table 2). Thus, the incremental cost per QALY gained is $26,417. Without utility adjustment, the life-years gained would be 1.84 (13.72 years of discounted life expectancy for patients receiving trastuzumab vs 11.88 years for controls), and the cost per life-year gained would be $24,435. Over a 20-year horizon, treatment with trastuzumab is estimated to cost an additional $43,913 on average, with an expected gain of 1.28 QALYs. The incremental cost per QALY gained is $34,201, and the incremental cost per life-year gained is $32,361. Societal perspective From a societal perspective, the additional time and travel expenses for the patient, a result of trastuzumab therapy duration that extends beyond that for standard chemotherapy, is estimated to cost $2075 over a lifetime horizon. Accounting for this would only increase the cost-effectiveness ratio to $27,637 per QALY gained (Table 2). This does, however, ignore any indirect benefits in terms of improved productivity or labor force attachment over the remainder of the patient s working life. Projected Recurrence and Lifetime Survival The projected lifetime survival curves for the base case are shown in Figure 3. This reflects the first 4 years of trial data, extrapolated to a lifetime horizon

494 CANCER August 1, 2007 / Volume 110 / Number 3 extremes. The model estimates were most robust with respect to variations in the costs of detecting and treating cardiac events and the costs of diagnostic testing. This would suggest that the overall finding of a favorable cost-utility ratio, by usual standards, is robust. FIGURE 3. Projected overall survival based on 4-year trial data extrapolated to a lifetime horizon in the trastuzumab-containing and control arms. AC indicates doxorubicin and cyclophosphamide; T, paclitaxel; H, trastuzumab. based on background mortality, recurrence rates, and death after metastatic breast cancer assumptions. As described above, the first 5 years of transition probabilities in the Markov model were empirically derived to approximate the observed distribution through 4 years and extended 1 additional year. A conservative approach was taken by using a cumulative distant recurrence-free difference equal to 0.088 at 3 years versus 0.089, the actual difference, and by using a cumulative overall survival difference of 0.020 projected versus 0.026 actual. Similarly, at the end of 4 years, the differentials were 0.137 projected versus 0.160 actual for recurrence and 0.041 projected versus 0.048 actual for overall survival. Adding assumptions based on the literature from 6 years onward, the incremental life expectancy in patients receiving trastuzumab (area between the 2 curves) is 3.0 years (19.4 years vs 16.4 years). The absolute differential in survival increases from 4.1% at 4 years to a maximum of 13.6% at 10 years and then reduces thereafter, for example, to 9.7% at 20 years. Sensitivity Analyses One-way sensitivity analysis of key parameters is illustrated as a tornado diagram (Fig. 4). A multiway analysis changing all parameters to either upperbound or lower-bound values produced a best-case to worst-case range in the cost-utility ratio of $9104 to $69,340. The upper end of this range is reached only when all parameters are set to the worst case, a very unlikely scenario. The results were most sensitive to variations in the discount rate, the price of trastuzumab, and the probability of metastasis. The discount rate was varied from 0% to 5% at the DISCUSSION Compared with some other tumor types and latestage breast cancer, life expectancy for a 50-year-old woman with early stage breast cancer is favorable approximately 15 20 years before discounting (see Table 2). In this study, the projected lifetime cost per QALY for the addition of trastuzumab to adjuvant chemotherapy was estimated to be $26,417, a level that is below that of many treatments commonly used for oncology patients. This projection was driven by substantial absolute advantages in clinical outcomes that were observed with trastuzumab in the joint analysis trial results: a 16.0 percentage-point improvement in freedom from distant recurrence and a 4.8 percentage-point improvement in overall survival at 4 years. 5 Although the base-case assumptions (13.7 and 4.1, respectively) were slightly more conservative than these differences, these translate to an average life expectancy gain of 3.0 years (19.4 years for chemotherapy plus trastuzumab vs 16.4 years for chemotherapy alone). It is clear that both the difference in freedom from distant recurrence and the difference in overall survival affect this result. If only the differential in overall survival was used to calculate average life expectancy gain, then the mean gain would be smaller (4.8% 3 16.4 years, approximately 0.8 years). Therefore, the extra 16 women per 100 who had not experienced a distant recurrence by 4 years accounted for most of the 2.96 gain in undiscounted life years in the model. Further support for the importance of preventing recurrence by 4 years was provided by 1 of the sensitivity analyses for the model. If the differential in distant recurrence is forced to be 10.1% (instead of 14.0%), and the differential in overall survival is reduced from 4.8% to 0.5%, the cost-utility ratio increases to only $34,012. Although this is a 29% increase, the overall cost-effectiveness finding is robust. Although many may regard incremental discounted costs of approximately $45,000 as high, if the projected long-term costs and outcomes are achieved in the real world, then most observers would view the resulting cost-utility ratio favorably. The cost utility of adding trastuzumab to standard chemotherapy is comparable to that of adjuvant

Cost-Effectiveness of Adjuvant Trastuzumab/Garrison et al. 495 FIGURE 4. Tornado diagram of one-way sensitivity analyses showing best-case to worst-case ranges for key parameters. precur indicates probability of recurrence; AC, doxorubicin and cyclophosphamide; T, paclitaxel; H, trastuzumab; LVEF, left ventricular ejection fraction; FISH, fluorescence in-situ hybridization; IHC, immunohistochemistry. chemotherapy for colorectal cancer using oxaliplatin in addition to 5-FU/leucovorin; a recent study projected a cost per QALY of $22,800. 36 The use of hormonal therapy for early stage breast cancer is also generally less than $100,000 per QALY gained. 37 39 In 2 recently published analyses based on Markov models, Kurian and colleagues projected the cost effectiveness of trastuzumab to be $39,982 per QALY, and Liberato and colleagues projected it to be $18,970. 12,13 In the present study, the base-case estimate of $26,417 falls in between these, but closer to the lower figure. A large part of the difference between the estimates from this model and those from Kurian et al. can be explained by the cost of trastuzumab $40,063 versus $50,767; their estimate assumed 52 weeks of use based on protocol, whereas the estimate in this study was based on the average actual dose received in the trial. At the same time, although both studies took a lifetime horizon for a 49 or 50-year old woman, Kurian et al s projected differential in discounted QALYs was 1.42 versus 1.70 in the present study. This difference between the 2 studies is despite the more conservative assumption made in this analysis that there is no additional benefit of trastuzumab after Year 5. Presumably, this is the result of differences in how the 2 models fit the first 5 years of data and the hazard rates assumed from Year 6 onward. Liberato et al. used a 15-year horizon in the base case, 52 weeks of trastuzumab treatment, and also assumed no additional trastuzumab benefit after 5 years. Their projected QALY differential was 1.18 QALYs, which was lower than the estimate calculated here, due, in part, to the shorter horizon. Liberato et al s estimated discounted cost differential over a 15-year horizon was $22,385 versus $44,923 over a lifetime horizon in this study. Kurian et al. concluded that the incremental costeffectiveness ratio was comparable to other interventions, and Liberato et al. also concluded that for HER2-positive, high risk, early breast cancer adjuvant trasutuzumab is a cost-effective therapy. Other health economics studies have also predicted that adding trastuzumab to adjuvant chemotherapy is cost effective relative to other accepted breast-cancer therapies. 15 21 The methodologies and assumptions used, and the geographic contexts, differ among the studies. Because healthcare programs and funding vary among countries, conclusions concerning what constitutes a cost-effective therapy may differ between reports. Hillner and colleagues estimated that, based on a 20-year horizon, the cost per life-year gained for adjuvant trastuzumab would

496 CANCER August 1, 2007 / Volume 110 / Number 3 be between $39,600 and $51,300. 16 A Canadian study reported an estimated additive cost of trastuzumab of $0.6 million to $2.0 million (Canadian dollars) per 1000 new breast cancer cases per year. 15 In the United Kingdom, the National Institute for Health and Clinical Excellence (NICE) estimates that adjuvant trastuzumab has an incremental cost per QALY gained of 18,000 (approximately US $34,000), which is less than the general 30,000 threshold that NICE seems to apply in practice. 21 A cost-effectiveness analysis in Denmark has shown that, if indirect costs are not included, the cost per life-year gained would be DKK78,475 (approximately $14,000 in US dollars). 20 A European study, based on therapy regimens and results from the Breast Cancer International Research Group 006 trial, predicted that trastuzumab plus docetaxel after AC, or trastuzumab plus docetaxel and carboplatin would be cost effective provided long-term survival outcomes improved or drug costs were reduced. 19 It is important to note, however, that the comparative regimen in the BCIRG 006 cost-effectiveness analysis was 5-fluorouracil, epirubicin, and cyclophosphamide (FEC) 60, a regimen that is not frequently used in the US and which is much cheaper than AC plus a taxane. In addition, a 5% discount rate was used, compared with the 3% rate used in this study. The model used in this study is conservative in that it assumes that both trastuzumab-containing and control arms have equal efficacy from 6 years onward. Also, the model assumes that patients will receive trastuzumab upon disease recurrence or metastasis, which would tend to increase costs and reduce cost effectiveness. In a recent review of previous modeling efforts, Norum argues that at least an 8% to 10% differential in overall survival at 10 years is needed to produce an acceptable cost-effectiveness ratio (E40,000 to E60,000 per life-year gained). 14 The base-case projection in our model is a difference in overall survival of 13.6% at 10 years. Thus, our results are consistent with this hypothesis, producing a cost-effectiveness ratio below the range cited. But, importantly, this projection is based on an assumption that rates of distant recurrence and death in both treatment arms are the same after Year 5. Hence, the projection is likely to be conservative, and it appears to be plausible. The cost-effectiveness results from this study were robust to sensitivity analysis. Multiway analysis that examined the best-case and worse-case scenarios for utilities yielded a cost-utility ratio range of $9104 to $69,340. This range would generally be considered as cost effective in a US setting because many technologies in common use have higher ratios, including the use of trastuzumab in HER2- positive metastatic breast cancer. 28 As the long-term benefits of adjuvant trastuzumab are not yet known, the source of our estimates for clinical outcomes (disease-free survival and overall survival) were data from the first 4 years of the NSABP B-31 and NCCTG N9831 trials. These data were used in the Markov model to predict the 20-year and lifetime survival horizons. The Markov model used in this study comprised 4 health states: treatment, disease-free, distant recurrence, and death. The disease states of contralateral disease and local recurrence were not included in the model, as we did not have reliable estimates of differences between the control and trastuzumabcontaining treatment arms in the joint analysis of NSABP B-31 and NCCTG N9831. Adding these states to the model would have increased lifetime costs and disutility in both arms. However, it is highly unlikely that including these states would have had a significant impact on overall survival projections because their effect is reflected in the time to distant recurrence estimates. Cost-utility estimates from both a payer and a partial societal perspective were included in this analysis. The societal perspective accounted for the indirect time and financial costs associated with treatment. As a large proportion of patients with early stage breast cancer are of working age, however, the indirect costs and productivity gains are possibly substantial and have not been estimated in this study or in the literature to date. Studies are needed to address this gap. Anthracyclines (such as doxorubicin) and trastuzumab are associated with an increased risk of cardiac dysfunction, particularly when administered concomitantly. 40,41 For this reason, the trastuzumab portion of therapy was given after the AC-containing regimen in the NSABP B-31 and NCCTG N9831 trials. To estimate the rate of cardiac dysfunction in our analysis, we used the average incidences of cardiac events (NYHA class III or IV congestive heart failure or probable or definite cardiac death) reported in the NCCTG N9831 trial: 2.9% in the trastuzumab-containing arm versus 0% in the control arm. In both NSABP B-31 and NCCTG N9831, the cardiac function of most patients who experienced congestive heart failure improved after they received standard medical treatment. 10,11 This finding also correlated with recent reports by Ewer and colleagues, who reported that cardiotoxicity associated with trastuzumab is reversible and not dose-related, and they showed that trastuzumab treatment can often be continued or

Cost-Effectiveness of Adjuvant Trastuzumab/Garrison et al. 497 restarted in patients who develop cardiac dysfunction with no subsequent cardiac events. 22,42 As a result, an assumption in our analysis was that cardiac dysfunction associated with the addition of trastuzumab was at least partly reversible and would have no direct impact on mortality beyond that reflected in the 4-year trial period. Determining HER2 status in all cases of breast cancer is recommended practice. 43 HER2 test results (by immunohistochemistry and/or FISH) were verified centrally in the NCCTG N9831 trial and by an approved laboratory in the NSABP B-31 trial. There can be a high level of discordance between results from low-throughput local versus high-throughput reference laboratories, 44,45 as observed in the NCCTG N9831 trial. 44 The methods used to test HER2 status, and whether results need to be confirmed by reference laboratories, will have an effect on overall diagnostic costs. In conclusion, the addition of trastuzumab to standard adjuvant therapy reduces risk of recurrence and improves overall survival in patients with early stage breast cancer. Over a lifetime, the use of trastuzumab in the adjuvant setting is projected to cost $26,417 per QALY. This cost-effectiveness result is robust to sensitivity analysis. Trastuzumab for the adjuvant treatment of early stage breast cancer is projected to be cost effective over a lifetime horizon, achieving a cost-effectiveness ratio below that of many widely accepted oncology treatments. REFERENCES 1. Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2006. CA Cancer J Clin. 2006;56:106 130. 2. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science. 1987;235:177 182. 3. Slamon DJ, Godolphin W, Jones LA, et al. Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science. 1989;244:707 712. 4. Paik S, Hazan R, Fisher ER, et al. Pathologic findings from the National Surgical Adjuvant Breast and Bowel Project: Prognostic significance of erbb-2 protein overexpression in primary breast cancer. J Clin Oncol. 1990;8:103 112. 5. Romond EH, Perez EA, Bryant J, et al. Trastuzumab plus adjuvant chemotherapy for operable HER2-positive breast cancer. N Engl J Med. 2005;353:1673 1684. 6. Piccart-Gebhart MJ, Procter M, Leyland-Jones B, et al. Trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer. N Engl J Med. 2005;353:1659 1672. 7. Joensuu H, Kellokumpu-Lehtinen PL, Bono P, et al. Adjuvant docetaxel or vinorelbine with or without trastuzumab for breast cancer. N Engl J Med. 2006;354:809 820. 8. Slamon D, Eiermann W, Robert N, et al. Phase III randomized trial comparing doxorubicin and cyclophosphamide followed by docetaxel with doxorubicin and cyclophosphamide followed by docetaxel and trastuzumab with docetaxel, carboplatin and trastuzumab in HER2-positive early breast cancer patients: BCIRG 006 study [abstract]. Breast Cancer Res Treat. 2005;94:S5. 9. Perez EA, Suman VJ, Davidson NE, et al. Interim cardiac safety analysis of NCCTG N9831 Intergroup adjuvant trastuzumab trial [abstract]. Proc Am Soc Clin Oncol. 2005;23: 17s. 10. Tan-Chiu E, Yothers G, Romond E, et al. Assessment of cardiac dysfunction in a randomized trial comparing doxorubicin and cyclophosphamide followed by paclitaxel, with or without trastuzumab as adjuvant therapy in node-positive, human epidermal growth factor receptor 2-overexpressing breast cancer: NSABP B-31. J Clin Oncol. 2005; 23:7811 7819. 11. Perez EA, Suman VJ, Davidson NE, et al. Exploratory analysis from NCCTG N9831: Do clinical and laboratory characteristics predict cardiac toxicity of trastuzumab when administered as a component of adjuvant therapy [abstract]? Breast Cancer Res Treat. 2005;94:S96. 12. Kurian AW, Newton Thompson R, Gaw AF, Arai S, Ortiz R, Garber AM. A cost-effectiveness analysis of adjuvant trastuzumab regimens in early HER2/neu-positive breast cancer. J Clin Oncol. 2007;25:634 641. 13. Liberato NL, Marchetti M, Barosi G. Cost effectiveness of adjuvant trastuzumab in human epidermal growth factor receptor 2-positive breast cancer. J Clin Oncol. 2007;25: 625 633. 14. Norum J. The cost-effectiveness issue of adjuvant trastuzumab in early breast cancer. Expert Opin Pharmacother. 2006; 7:1617 1625. 15. Ragaz J, Spinelli JJ. Cost-benefit estimates of adjuvant trastuzumab for early breast cancer [abstract]. Breast Cancer Res Treat. 2005;94:S93. 16. Hillner B. Clinical and cost-effectiveness implications of adjuvant trastuzumab in HER2-positive breast cancer trials [abstract]. Breast Cancer Res Treat. 2005;94:S219. 17. Norum J, Olsen JA. Trastuzumab in adjuvant breast cancer: A model based cost-effectiveness analysis [abstract]. Proc Am Soc Clin Oncol. 2006;24:34s. 18. Wilson E, Crown J, Ballot J, McDonnell D, Shahan E, Healy J. Estimating the real cost of adjuvant trastuzumab in patients with HER2-positive early stage breast cancer [abstract]. Proc Am Soc Clin Oncol. 2006;24:320s. 19. Neyt M, Albrecht J, Cocquyt V. An economic evaluation of Herceptin in adjuvant setting: The Breast Cancer International Research Group 006 trial. Ann Oncol. 2006;17:381 390. 20. Danish National Board of Health. Trastuzumab (Herceptin) as adjuvant treatment of early breast cancer after surgical treatment. Preliminary report. Health Technology Assessment of Cancer Drugs 1: Available at: http://www. sundhedsstyrelsen.dk/upload/planlaegning_og_behandling/ cemtv/mtv_af_nye_kr%c3%a6ftl%c3%a6gemidler/herceptin_ en/herceptin_en.pdf, 2005. 21. National Institute for Health and Clinical Excellence. Trastuzumab for the adjuvant treatment of early-stage HER2- positive breast cancer. Available at: http://www.nice. org.uk/ta107, 2006. 22. Ewer MS, Vooletich MT, Durand JB, et al. Reversibility of trastuzumab-related cardiotoxicity: New insights based on clinical course and response to medical treatment. J Clin Oncol. 2005;23:7820 7826. 23. Centers for Medicare and Medicaid. Available at: http:// www.cms.hhs.gov/, 2006. 24. Analysource Online. Available at: http://www.analysource. com, 2006.

498 CANCER August 1, 2007 / Volume 110 / Number 3 25. Rao S, Kubisiak J, Gilden D: Cost of illness associated with metastatic breast cancer. Breast Cancer Res Treat. 2004;83: 25 32. 26. Oestreicher N, Ramsey SD, Linden HM, et al. Gene expression profiling and breast cancer care: What are the potential benefits and policy implications? Genet Med. 2005;7: 380 389. 27. Hornberger J, Kerrigan M, Foutel V. Cost-effectiveness of trastuzumab (Herceptin) for treatment of metastatic breast cancer [abstract]. Ann Oncol. 2002;13:52. 28. Early Breast Cancer Trialists Collaborative Group. Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: an overview of the randomised trials. Lancet. 2005;365:1687 1717. 29. Elkin EB, Weinstein MC, Winer EP, Kuntz KM, Schnitt SJ, Weeks JC. HER-2 testing and trastuzumab therapy for metastatic breast cancer: a cost-effectiveness analysis. J Clin Oncol. 2004;22:854 863. 30. Romond EH, Perez EA, Bryant J, et al. Doxorubicin and cyclophosphamide followed by paclitaxel with or without trastuzumab as adjuvant therapy for patients with HER2- positive operable breast cancer: Combined analysis of NSABP B-31 and NCCTG N9831. Presented at the 41st Annual Meeting of the American Society of Clinical Oncology, Orlando, FL, 2005. 31. US Department of Labor: US Bureau of Labor Statistics. Available at: http://www.bls.gov/ [accessed February 15, 2006]. 32. United States Department of the Treasury: Internal Revenue Service. Available at: http://www.irs.gov/, [accessed December 15, 2005]. 33. Siegel JE, Weinstein MC, Russell LB, Gold MR. Recommendations for reporting cost-effectiveness analyses. Panel on Cost-Effectiveness in Health and Medicine. JAMA. 1996;276: 1339 1341. 34. Gold M. Panel on cost-effectiveness in health and medicine. Med Care. 1996;34:DS197 DS199. 35. Arias E. United States life tables, 2002. Natl Vital Stat Rep. 2004;53:1 38. 36. Aballea S, Chancellor JV, Raikou M, et al. Cost-effectiveness analysis of oxaliplatin compared with 5-fluorouracil/leucovorin in adjuvant treatment of stage III colon cancer in the US. Cancer. 2007;109:1082 1089. 37. Noe LL, Becker RV, III, Gradishar WJ, Gore M, Trotter JP. The cost effectiveness of tamoxifen in the prevention of breast cancer. Am J Manag Care. 1999;5:S389 S406. 38. Delea TE, Karnon J, Smith RE. Cost-effectiveness of extended adjuvant letrozole therapy after 5 years of adjuvant tamoxifen therapy in postmenopausal women with early-stage breast cancer. Am J Manag Care. 2006;12:374 386. 39. Hillner BE. Benefit and projected cost-effectiveness of anastrozole versus tamoxifen as initial adjuvant therapy for patients with early-stage estrogen receptor-positive breast cancer. Cancer. 2004;101:1311 1322. 40. Slamon DJ, Leyland-Jones B, Shak S, et al. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med. 2001;344:783 792. 41. Seidman A, Hudis C, Pierri MK, et al. Cardiac dysfunction in the trastuzumab clinical trials experience. J Clin Oncol. 2002;20:1215 1221. 42. Ewer MS, Lippman SM. Type II chemotherapy-related cardiac dysfunction: time to recognize a new entity. J Clin Oncol. 2005;23:2900 2902. 43. Carlson RW, Moench SJ, Hammond ME, et al. HER2 testing in breast cancer: NCCN Task Force report and recommendations. J Natl Compr Canc Netw. 2006;4(suppl 3):S1 S22. 44. Perez EA, Suman VJ, Davidson NE, et al. HER2 testing by local, central, and reference laboratories in specimens from the North Central Cancer Treatment Group N9831 Intergroup adjuvant trial. J Clin Oncol. 2006;24:3032 3038. 45. Reddy JC, Reimann JD, Anderson SM, Klein PM. Concordance between central and local laboratory HER2 testing from a community-based clinical study. Clin Breast Cancer. 2006;7:153 157.