Presentation in oral format at the 36th Annual Meeting of the Society of Gynecologic Oncologists, March 18–23, 2005, Miami, Florida.
Current chemotherapy in platinum-resistant ovarian cancer patients has demonstrated minimal to no improvements in survival. Despite the lack of benefit, significant resources are utilized with such therapies. Therefore, the objective in the current study was to assess the cost-effectiveness of salvage chemotherapy for patients with platinum-resistant epithelial ovarian cancer (EOC).
A decision analysis model evaluated a hypothetical cohort of 4000 platinum-resistant patients with recurrent EOC. Several chemotherapy strategies were analyzed: 1) best supportive care (BSC); 2) second-line chemotherapy-monotherapy; 3) second-line chemotherapy-combination therapy; 4) third-line chemotherapy after disease progression on second-line monotherapy; and 5) third-line chemotherapy after disease progression on second-line combination therapy. Sensitivity analyses were performed on all pertinent uncertainties.
Using costs alone, BSC was the only definitive cost-effective treatment for platinum-resistant recurrent ovarian cancer patients, and second-line monotherapy was a reasonable cost-effective strategy with an incremental cost-effectiveness ratio (ICER) of $64,104. The cost-effectiveness ranged from $4,065 per month of overall survival (OS) for BSC to $12,927 for third-line previous combination therapy. Compared with BSC, second-line monotherapy gained an additional 3 months of OS, with a cost-effectiveness of $4,703 per month of OS. Second-line combination therapy and third-line therapies exhibited unfavorable ICER.
In the year 2005, ovarian cancer was estimated to affect over 20,000 women in the U.S. alone.1 Although the majority of patients diagnosed with ovarian cancer have advanced disease, 60% to 80% of patients will respond to cytoreductive surgery and first-line platinum-based combination chemotherapy. Unfortunately, many patients with advanced disease experience a recurrence and eventually succumb to progressive disease because effective curative therapy for recurrent ovarian cancer has yet to be determined.2–6
A well-established dichotomy of response exists based on the patient's response to first-line platinum-based chemotherapy. Patients who respond to initial platinum-based therapy with a progression-free interval (PFI) >6 months (platinum-sensitive) experience response rates generally >50% and overall survival (OS) of >1 year. Markman et al.7 demonstrated that this survival advantage extends as the platinum-free interval increases. Comparatively, patients who have arecurrence of disease within 6 months of first-line platinum-based therapy (platinum-resistant) have demonstrated response rates of usually <20% and an OS of <12 months.8–10
Despite numerous chemotherapeutic strategies, the poor prognosis has remained unchanged for patients with platinum-resistant recurrent ovarian cancer. Many factors must be considered before initiating additional chemotherapy in this patient population, including the patient's wishes, performance status, comorbidities, prior treatments, and emotional, social, and financial status. Although most therapies have demonstrated minimal to noimprovement in survival, substantial resources are utilized with each additional chemotherapeutic line of therapy. It is estimated that >10% of all medical resources areutilized within the last 6 months of life, which translates into >$150 billion in the U.S. annually. In addition, the use of best supportive care (hospice/palliative care) instead of chemotherapy could decrease end-of-life spending by 25% to 40% ($72 billion per year).11
Therefore, the goal of this study was to assess the medical costs and potential effectiveness of chemotherapy for patients with recurrent platinum-resistant advanced epithelial ovarian cancer (EOC).
MATERIALS AND METHODS
We constructed a decision analysis model consisting of a hypothetical cohort of 4000 patients to evaluate potential strategies for recurrent platinum-resistant EOC. The strategies analyzed included: 1) best supportive care (BSC); 2) second-line chemotherapy-monotherapy; 3) second-line chemotherapy-combination therapy; 4) third-line chemotherapy after disease progression on second-line monotherapy; and 5) third-line chemotherapy after disease progression on second-line combination therapy. The hypothetical cohort consisted of patients who were previously diagnosed with advanced EOC and subsequently underwent primary cytoreductive surgery followed by combination platinum-taxane based chemotherapy for 6 cycles. The cohort experienced a recurrence of disease within 6 months of the completion of primary chemotherapy (platinum-resistant).
BSC consisted of a multidiscipline approach consisting of physicians, nurses, healthcare aides, and social workers to address palliative care symptoms. All aspects of outpatient office visits, emergency department visits, hospitalizations, and home health care was included in the costs of BSC. This included the evaluation and treatment of symptomatic problems. In addition, all medications including intravenous analgesic and antibiotics were included. Other palliative supportive interventions such as blood transfusions, home oxygen, and intravenous hydration were included in the costs of BSC.12, 13 Although chemotherapy and/or radiation therapy may be utilized for palliation, our model did not incorporate these in the BSC strategy. However, other palliative procedures such as paracentesis, thoracentesis, and percutaneous gastrostomy tube placements were included.
To estimate OS for each strategy, we utilized the progression-free survival (PFS) from Phase II/III studies for the respective chemotherapy for each strategy. Additionally, we estimated a period of survival during which time the patient received hospice care after failing the respective chemotherapeutic strategy.
In the second-line monotherapy strategy, patients received monthly liposomal doxorubicin (Doxil; Ortho Biotech, Bridgewater, NJ) (at a dose of 40 mg/m2) for a total of 4 months. In the second-line combination therapy strategy, patients received combination chemotherapy with gemcitabine (Gemzar; Lilly, Indianapolis, IN) (at a dose of 750 mg/m2) and cisplatin (Platinol; Bristol-Myers Squibb, Princeton, NJ) (at a dose of 30 mg/m2) on Days 1 and 8 every 21 days for 4 months.
Third-line chemotherapy consisted of 3 cycles of topotecan (Hycamtin; GlaxoSmithKline, Philadelphia, PA) (at a dose of 1.5 mg/m2/day for 5 days every 21 days). Two separate third-line strategies were evaluated in our model, depending on the therapy previously received in the second-line setting (monotherapy vs. combination therapy).
Model Estimates: Clinical
Clinical estimates were obtained from a review of published literature (Table 1). An attempt was made to utilize data from Phase III trials; however, many estimates were collected from Phase II data. The interpretation of recurrent ovarian cancer survival data is difficult because many patients will receive further therapy after study protocol. Thus, it becomes problematic when assessing the effectiveness of a single agent if additional “off-protocol” therapy is received.
Several assumptions were made in the development of our model:
1Based on a review of 54 studies, OS for recurrent platinum-resistant patients is <1 year from diagnosis with subsequent therapy.
2Chemotherapy offers a small survival benefit over no further therapy.
3Each successive chemotherapy regimen is assumed to be less effective than the previous treatment, with shorter PFI in comparison to the previous treatment.14
The number of patients in each cohort was estimated to be 4000 based on the incidence of ovarian cancer in 2005 and recurrence figures (Fig. 1). We defined effectiveness as the months of OS from the time of initial recurrence. In addition, all patients received end-of-life home hospice care after their respective strategy treatment.
Liposomal doxorubicin was the chemotherapy agent used in the second-line monotherapy strategy. The reported response rate (RR) for liposomal doxorubicin ranges from 8% to 27%, with a mean RR of 16%. Based on 8 studies evaluating 575 patients with platinum-resistant EOC, the median PFS was 4.1 months15–22 (Table 2). Patients in this strategy were estimated to survive an additional 2 months from the time of progression on liposomal doxorubicin for an estimated OS of 6 months for this strategy.
Table 2. Liposomal Doxorubicin Literature for Platinum-Resistant Ovarian Cancer
No. of patients
RR indicates relative risk; PFS, progression-free survival; NA, not applicable.
Historically, combination chemotherapy for patients with platinum-resistant EOC has been disappointing. However, emerging evidence utilizing the combination of gemcitabine and cisplatin has been encouraging. A study by Rose et al.23 (Table 3) evaluated the combination of gemcitabine and cisplatin in platinum-resistant ovarian cancer. In the 35 patients evaluable for response, they demonstrated an overall RR of nearly 43%, with a PFS of 6 months. Similar results were also reported in a Gynecologic Oncology Group (GOG) study24 and a retrospective study by Tewari et al.25 The GOG study by Brewer et al.24 utilizing the same combination in 57 platinum-resistant patients showed a 15.8% overall RR and a PFS of 6 months. Therefore, we estimated a PFS of 6 months, with an additional 2 months of hospice care, for an OS of 8 months for second-line combination therapy.
Table 3. Gemcitibine Plus Cisplatin Literature for Platinum-Resistant Ovarian Cancer
Gemcitibine, 600–750 mg/m2 plus cisplatin, 30 mg/m2
6 mos (median)
Both third-line chemotherapy strategies utilized topotecan. In reviewing 9 studies, incorporating >600 patients who were treated with topotecan for platinum-resistant recurrent EOC, the median PFS was 3 months20, 26–33 (Table 4). In our model, each cohort in a third-line chemotherapy strategy gained an additional 3 months of PFS after progression on second-line chemotherapy and an estimated 1 month of hospice care at the time of third-line failure (the second recurrence). This translated to an OS of 8 months for third-line previous monotherapy and 10 months for third-line previous combination therapy.
Table 4. Topotecan Literature for Platinum-Resistant Ovarian Cancer
No. of patients
RR indicates relative risk; PFS, progression-free survival; NA, not applicable.
The true median life-expectancy of platinum-resistant ovarian cancer patients without treatment is unknown. Using clinical experience we estimated the OS for the BSC strategy to be 3 months.
Model Estimates: Costs
The analysis was conducted from the perspective of a third-party payer using 2004 USD. Direct costs were calculated for each strategy (Table 5). We did not use indirect costs or reimbursements because of the striking differences in reimbursement across various regions of the country. We estimated costs by adjusting local charges using a cost-to-charge ratio of 60%. For laboratory and procedure estimates, we consulted the University of Alabama at Birmingham. Pharmacy costs were calculated using average wholesale drug costs. Baseline costs were varied across clinically reasonable ranges in sensitivity analyses. To keep a conservative model, the costs of treating chemotherapy-related toxicity and complications (i.e., chemotherapy-induced anemia, neutropenia, thrombocytopenia) were not included in our estimates. In order to be congruent, the use of erythropoietic agents for palliative intent in the BSC strategy was not included.
Table 5. Cost Estimates
BSC indicates best supportive care.
BSC ($ 135.50 per day)
Associated infusion fees
Nonmedical associated costs
Second-line chemotherapy (monotherapy)
Cost of doxorubicin per treatment
Total (4 cycles)
Second-line chemotherapy (combination)
Cost of gemcitibine per treatment
Cost of cisplatin per treatment
Total (2 courses per cycle, 6 cycles)
Cost of topotecan per treatment
Total (5 courses per cycle, 3 cycles)
The costs of BSC were estimated to be $135.50/day.13 This resulted in $12,195 per patient over a 3-month time period. The costs of chemotherapy strategies included chemotherapy agents, infusion costs, laboratory tests, intravenous fluids, and associated support medications such as antiemetics and steroids. Chemotherapy agent costs were based on the recommended dosage for a 5′6″ woman weighing 65 kg (body surface area of 1.75).
The cost of the second-line monotherapy strategy utilizing liposomal doxorubicin over a 4-month period was $20,090 plus the costs of an additional 2 months of palliative care ($8130) for a total cost of $28,220 per patient. Similarly, the costs of second-line combination therapy was calculated using gemcitabine and cisplatin with a 2-month period of palliative care for a total cost of $78,606. Costs of using 3 months of topotecan for third-line previous monotherapy and third-line previous combination therapy with a 1-month period of palliative care for each was $78,884 and $129,270, respectively.
Model Estimates: Statistics
The cost-effectiveness was calculated as the overall cost of implementing the strategy for a hypothetical cohort of 4000 patients per life-year-saved (LYS). The incremental cost-effectiveness was calculated as the additional costs per LYS compared with the next-best strategy. All modeling and calculations were performed using a commercially available decision analysis program (DATA version 3.5; TreeAge Software, Williamston, MA).
Under baseline assumptions, a cohort of 4000 patients was applied to each of the 5 strategies (Table 6). BSC was the least expensive strategy at a total cost of $49 million (M) per 4000 patients, whereas third-line previous combination therapy was the most costly at $517 M. The effectiveness of the strategies ranged from 3 months to 10 months of OS, with BSC being the least effective and third-line previous combination therapy being the most effective. Cost-effectiveness ranged from $4065 to $12,927 per month of OS for BSC and third-line previous combination therapy, respectively.
USD indicates U.S. dollars; OS, overall survival; C-E ratio, cost-effectiveness ratio; BSC, best supportive care; M, million.
Third-line previous monotherapy
Third-line previous combination
The model demonstrates a marginal improvement in OS with increasing number of agents used and/or increasing number of lines of chemotherapy. However, these small improvements in survival came with significant costs. Compared with BSC, second-line monotherapy gained an additional 3 months of OS; however, it exhibited an unfavorable incrementalcost-effectiveness ratio (ICER) of $64,104 per LYS (Table 7). Second-line combination therapy demonstrated an additional 2 months of OS compared with second-line monotherapy; however, an unfavorable ICER of $302,316 per LYS limits its cost-effectiveness.
Table 7. Results Incremental Cost-Effectiveness Ratio
USD indicates U.S. dollars; OS, overall survival; ICER, incremental cost-effectiveness ratio; LYS, life-year saved; BSC, best supportive care; NA, not applicable; M, million.
Third-line previous combination
The third-line chemotherapy strategies showed small improvements in OS in comparison to second-line chemotherapy strategies, but were either dominated by other strategies or had unfavorable ICER. For example, third-line previous monotherapy demonstrated an OS of 8 months at a cost of $316 M, for a cost-effectiveness ratio of $9,861 per month of OS. This strategy was dominated by second-line combination therapy because it exhibited a lower total cost of $314 M with an equal efficacy of 8 months of OS.
Third-line chemotherapy had a small improvement in OS (2 months) compared with second-line chemotherapy. This marginal benefit of OS in third-line strategies was associated with substantial costs of approximately $200 M greater than the respective second-line strategies. This translates into an additional $101 M per additional month of OS gained with third-line chemotherapy.
One-Way Sensitivity Analyses
A series of 1-way sensitivity analyses were performed in which the value for a single estimate was varied across a reasonable range, whereas all other estimates were maintained at their baselines. This allowed us to determine the impact of the altered variable on the modeled outcome. Some examples of these analyses are discussed below.
The effectiveness was varied among strategies to determine how many months of OS would be required for a chemotherapy strategy to become cost-effective. By varying the OS of second-line monotherapy from 2 months to 12 months, we determined that an OS of 7 months would be considered cost-effective because the ICER fell below $50,000 per LYS at this point. Similarly, the OS of second-line combination chemotherapy was varied from 4 months to 24 months. Due to the high costs relative to combination chemotherapy, the OS would have to exceed 20 months to allow this strategy to be cost-effective (ICER < $50,000 per LYS).
We also wanted to determine at what cost would a chemotherapy strategy have to attain for it to become cost-effective. For second-line monotherapy, the costs were varied from $10,000 to $25,000; this analysis determined that the cost would have to fall below $16,500 to be cost-effective (ICER < $50,000 per LYS). In addition, varying the costs of second-line combination therapy from $15,000 to $80,000demonstrated that the cost would have to fall below $25,000 to become cost-effective.
Despite current advances in the treatment of advanced ovarian cancer, disease eventually recurs in the majority of patients,34 with an OS of <1 year for patients with platinum-resistant disease.35–37 The effectiveness of additional chemotherapy and the subsequent PFI appears to depend on a patient's response to first-line therapy.7, 38–41 Although numerous studies in patients with platinum-resistant recurrent ovarian cancer have failed to demonstrate significant improvements in survival, additional chemotherapy still may hold some utility. Opposed to initial treatment therapy with curative intent, subsequent chemotherapy goals encompass prolongation of survival and improvement in quality of life. The question remains: should costs and available resources be considered when initiating chemotherapy for recurrent platinum-resistant EOC? Therefore, this cost-effectiveness analysis was intended to be thought-provoking and bring attention to the significant amount of resources utilized for minimal gain in survival. The decision to give additional chemotherapy is ultimately complex and includes many factors beyond the scope of this project, which was costs alone. We concede and support the decision to give further therapy past what was evaluated in this analysis based on those other factors considered. However, this does not alleviate the need of more cost-effective therapeutic options for patients with recurrent ovarian cancer as depicted from this cost-effectiveness analysis.
Using a hypothetical cohort our decision analysis model revealed that utilizing BSC was the only definitive cost-effective treatment for platinum-resistant recurrent ovarian cancer patients, and second-line monotherapy was a reasonable cost-effective strategy,with an ICER of $64,104. Traditionally, an ICER <$50,000 per LYS has been established as a reasonable threshold of the cost-effectiveness of an intervention.38 Therefore, strategies costing >$50,000 per LYS would not be considered to be cost-effective and should not be utilized based on economic considerations. Although second-line monotherapy was not cost-effective by our baseline estimates, the ICER did approach the $50,000 threshold and therefore was evaluated closely in sensitivity analyses. The ICER of second-line monotherapy fell below $50,000 if OS achieved 7 months and, therefore, second-line monotherapy would be cost-effective. This trend, however, would be lost with second-line combination therapy or third-line chemotherapy due to the prohibitive costs of additional chemotherapy agents. In fact, the OS would have to be >20 months for second-line combination therapy to be cost-effective. In this patient population, this survival cannot be achieved with any currently available combination therapies.
Furthermore, the $50,000/LYS threshold for an ICER to be cost-effective is not absolute. Others have suggested the benchmark could be variable andgraduated up to $100,000/LYS.42 Using this ICER threshold, second-line monotherapy becomes cost-effective at baseline. However, second-line combination therapy and third-line previous combination would remain too costly a strategy, with ICER's eclipsing $300,000.
With all decision models there are potential limitations, especially with clinical estimates. Although one could criticize the choice of chemotherapy agents used in our model, we chose agents commonly utilized in the recurrent setting. Fortunately, a sensitivity analysis will encompass a wide range of costs and effectiveness, thereby including most alternative chemotherapy agents and their associated costs. Our sensitivity analysis revealed that second-line monotherapy must reach an OS of 7 months or cost <$16,500 to be cost-effective. Considering that larger institutions often purchase chemotherapy agents in consortia, they could potentially offer less expensive regimens, thereby making additional treatments more cost-effective. In addition, second-line combination therapy must exceed an OS of 20 months or cost <$25,000 to gain a cost-effectiveness advantage. However, there are limited data to suggest that we can accomplish this with our currently available combination regimens.
Furthermore, we believe that our model is quite conservative, as we did not include other costly factors associated with chemotherapy such ascolony-stimulating factors and chemotherapy-related complications. One could hypothesize that toxicities and complications would occur more frequently with subsequent therapies, and thus strategies would become less cost-effective with increasing lines of therapy.
The question of whether additional treatment improves survival over best supportive care in recurrent ovarian cancer is difficult to answer. One would have to perform a randomized controlled trial of further chemotherapy versus no further treatment in this patient population to answer this question. This type of study would be difficult to conduct due to patient and physician biases. However, even more important in these patients is how does additional chemotherapy affect the quality of life. This endpoint has not routinely been evaluated until recently and was not included in our analysis.
This model indicates that BSC is a cost-effective strategy for patients with platinum-resistant ovarian cancer. Second-line monotherapy is also a reasonable option if OS is predicted to be 7 months, the cost of therapy falls below $16,500, or if one accepts the ICER threshold of $100,000/LYS. However, further improvements in survival are needed in order for second-line combination therapy or third-line chemotherapy to be cost-effective strategies in this patient population from the perspective of costs alone.
In conclusion, we attempted to construct ahypothetical model to demonstrate the significant resources utilized with minimal improvements in survival in patients with platinum-resistant recurrent ovarian cancer. However, cost-effectiveness analyses are from the perspective of costs alone, and thus should never supersede sound clinical judgment and patient preference. Given that a hypothetical cohort is unable to reflect all relevant factors that are used in treatment decisions for actual patients, it should only be used as an adjunct to policy makers and hopefully direct attention toward the development of more effective treatment strategies for recurrent ovarian cancer.