|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2002 American Society for Clinical Oncology Outcomes and Cost-Effectiveness of Alternative Staging Strategies for Non–Small-Cell Lung CancerByFrom the Department of Surgery, Brigham and Women’s Hospital, CHASE Management Systems, Partners HealthCare, Inc, and the Department of Health Policy and Management, Harvard School of Public Health, Boston, MA. Address reprint requests to Nestor F. Esnaola, MD, MPH, University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Blvd, Box 444, Houston, TX 77030-4009; email: nesnaola{at}mdanderson.org
PURPOSE: To identify the optimal strategy for staging the mediastinum of patients with known non–small-cell lung cancer (NSCLC), stratified by tumor (T) classification. METHODS: We used a decision-analytic model to compare the health outcomes and cost-effectiveness of three staging strategies: (1) chest computed tomography alone, (2) selective mediastinoscopy, and (3) routine mediastinoscopy. The overall effectiveness and cost of each strategy was a function of the proportion of patients accurately staged and the risks, benefits, and costs of the diagnostic tests and treatments used. Probability estimates and costs were derived from primary data and the literature. We adopted a societal perspective and calculated incremental cost-effectiveness ratios (ICERs) as cost per quality-adjusted life year (QALY) gained. RESULTS: Both mediastinoscopy strategies correctly identified more patients with mediastinal involvement (N2/N3 disease) and assigned them to multimodal regimens. Routine mediastinoscopy maximized quality-adjusted life expectancy in all patients, irrespective of T classification, and this result was robust to varying the model estimates over their reported ranges. In T1 patients, selective mediastinoscopy cost $24,500 per QALY gained, compared with $78,800 per QALY gained for routine mediastinoscopy. In T2 and T3 patients, the ICER of routine mediastinoscopy was more favorable ($42,800 and $53,400 per QALY gained, respectively). CONCLUSION: Routine mediastinoscopy maximizes quality-adjusted life expectancy in patients with known NSCLC, and its ICER compares favorably with other currently accepted medical technologies. The survival benefit and cost-effectiveness of this strategy are greater in patients with T2 and T3 tumors and are likely to improve with advances in multimodal therapy.
LUNG CANCER remains the leading cause of cancer deaths among men and women in the United States.1 The majority of patients with lung cancer have non–small-cell lung cancer (NSCLC), and many of these patients present with advanced, locoregional disease or metastatic disease. Patients with tumor involvement of their ipsilateral mediastinal nodes have N2 disease and are staged as IIIA, whereas patients with involvement of their contralateral mediastinal nodes have N3 disease and are staged as IIIB.2,3 It is important to accurately stage the mediastinum of patients with NSCLC because the extent of mediastinal disease determines both treatment and prognosis. Patients with N2 disease are candidates for multimodal therapy and resection, whereas N3 disease is usually considered unresectable. Chest computed tomography (CT) can be used to assess the size of the tumor and the extent of mediastinal involvement (N classification). Although safe and relatively inexpensive, CT cannot differentiate between benign and malignant adenopathy. Normal-sized nodes can harbor metastases,4-6 while enlarged nodes may be the result of benign processes (eg, sarcoid, tuberculosis) or postobstructive pneumonia. Mediastinoscopy is a more invasive staging modality that involves direct access to the anterior mediastinum where nodes suggestive of abnormality can be directly visualized and biopsied. Although mediastinoscopy can usually be performed as an outpatient procedure, it requires general anesthesia and is associated with a low risk of morbidity and mortality. Chest CT and mediastinoscopy can be used alone or in combination to stage the mediastinum. In this study, we compared the health outcomes and cost-effectiveness of three strategies commonly used to stage patients with NSCLC.
Model to Compare Staging Strategies We developed a decision-analytic model7,8 to evaluate the health outcomes and costs associated with three staging strategies (Fig 1): (1) chest CT alone, (2) selective mediastinoscopy, and (3) routine mediastinoscopy. Our model was applicable to patients with pathologic evidence of NSCLC, a chest CT, and a metastatic work-up that revealed no evidence of locally unresectable (T4) or metastatic (M1) disease. In the CT staging strategy, patients with mediastinal adenopathy (ie, nodes > 1 cm) on chest CT were considered to have mediastinal disease. In the selective mediastinoscopy strategy, patients with mediastinal adenopathy underwent follow-up mediastinoscopy; in the routine mediastinoscopy strategy, all patients were subjected to up-front mediastinoscopy, irrespective of the results of their chest CT. On the basis of the final diagnostic result of each staging strategy, patients with normal-appearing or biopsy-negative mediastinal nodes were taken directly to resection. Because the only randomized, controlled trials comparing surgery with multimodal therapy in patients with stage IIIA NSCLC involved neoadjuvant chemotherapy (NAC) followed by resection and radical radiotherapy, we assumed that patients with N2 disease in our model were treated with similar trimodality regimens.9,10 Thus patients with enlarged or biopsy-positive N2 nodes in our model were treated with NAC, resection, and radiotherapy, while patients with enlarged or biopsy-positive N3 nodes were treated with chemoradiation alone.
The overall effectiveness and total cost of each staging strategy was a function of the proportion of patients accurately staged and the risks, benefits, and costs of the diagnostic tests and downstream treatments. Within each strategy, a certain proportion of patients with N2/N3 disease assumed to have localized disease were taken directly to surgery, while another group of patients with N0/N1 disease assumed to have mediastinal disease (ie, false-positive results) were treated with neoadjuvant therapy or chemoradiation alone. We stratified the results of our model by T classification. We used a state-transition (Markov) model11,12 to simulate the ongoing risk of death according to T classification, N classification, and treatment received and to calculate the time spent by patients in the various posttreatment health states. The effectiveness (ie, life expectancy) of each strategy was calculated by running the model in 1-year cycles until the entire cohort entered the final (absorbing) health state, death. To incorporate the effect of the various downstream treatments and health states on quality of life, we assigned health-related quality-of-life weights, or utilities, to these outcome states.13 By multiplying the time patients spent in the these states by their assigned utilities, we calculated the quality-adjusted life expectancy of each staging strategy, reported in quality-adjusted life years (QALYs). The subsequent cost-effectiveness analysis was performed according to the recommendations for a reference-case analysis of the Panel on Cost-Effectiveness in Health and Medicine.14,15 A societal perspective was used, and the direct medical costs, direct nonmedical costs, indirect (time) costs, and incurred (downstream) costs of the three staging strategies were considered. Before performing the cost-effectiveness analysis, the total cost and effectiveness (ie, quality-adjusted life expectancy) of the three staging strategies were calculated and ranked with respect to increasing effectiveness.16,17 The reference strategy used to calculate the incremental cost-effectiveness ratio (ICER) of each strategy was the next least-expensive strategy (ie, CT in the case of selective mediastinoscopy and selective mediastinoscopy in the case of routine mediastinoscopy). Strategies that are more costly but less effective than the next least-expensive strategy are eliminated based on simple dominance. The ICER of each strategy was calculated by dividing the additional cost by the additional benefit of each strategy (compared with the next least-expensive strategy). Strategies with ICERs greater than those of the next most-expensive strategy were eliminated by extended dominance. Both costs and benefits were discounted at an annual rate of 3% in the base-case analysis.
Data and Assumptions
Sensitivity and Specificity of Tests Base-case estimates of the sensitivity and specificity of CT with respect to N2/N3 disease were obtained from a published meta-analysis of 10 studies that used modern-generation scanners, a CT criterion of > 1 cm in the short-axis diameter to represent mediastinal lymph node involvement, and pathologic confirmation by mediastinal lymph node sampling.18 The sensitivity and specificity of mediastinoscopy were obtained from several prospective studies.19-25 Because of anatomic constraints, some nodal stations cannot be adequately sampled with either anterior or cervical mediastinoscopy. Although we assumed that the sensitivity of mediastinoscopy was 92% in the base case, we varied this value from 67% to 94% to account for the fact that the sensitivity may be lower at centers with more limited experience using with this procedure. Because biopsy-confirmation is required to classify a node as positive by mediastinoscopy (making the false-positive rate of mediastinoscopy essentially negligible), we assumed that the specificity of mediastinoscopy was 100%. Mediastinoscopy was assumed to be equally sensitive at identifying mediastinal involvement whether it was ipsilateral or contralateral.
Morbidity and Mortality of Tests and Treatments Cervical mediastinoscopy is usually performed as an outpatient procedure and is associated with minimal morbidity (< 2.5%)26 and mortality (0% to 0.1%).26-30 To avoid bias, we used the higher end of the reported range of mortality (0.1%) as our base-case estimate. Because the literature on anterior mediastinoscopy is limited, three series were combined to arrive at an estimate of the mortality of this procedure (0.5%).31-33 It was assumed that one third of patients would present with left upper-lobe tumors34 and require both anterior and cervical mediastinoscopy, exposing them to the cumulative risk of both procedures. Estimates for the mortality of surgical resection for lung cancer vary widely, from 0% to 20%.35-44 The 4% estimate used in the base-case analysis was based on a modern postoperative mortality of 3.8% reported by the Lung Cancer Study Group in 199444 and on several recent trials involving induction chemotherapy and surgical resection.37-41 The morbidity and mortality associated with neoadjuvant therapy for stage IIIA disease and sequential chemoradiation for stage IIIB disease were obtained from several published trials.9,10,39-41,45-47 Because these estimates were similar, they were combined to arrive at an average morbidity and mortality of 10% and 3%, respectively, for multimodal therapy for N2 and N3 disease. To account for the fact that multimodal regimens vary from institution to institution, we varied the morbidity, mortality, treatment duration, treatment effects, and costs of multimodal therapy in the sensitivity analysis.
Stage of Disease at Presentation, Treatment Effects, and Expected Survival Survival estimates of patients treated with surgery alone were obtained from the Mountain3 and Naruke et al48 series and from controls in published trials of multimodal therapy for stage IIIA/B disease.9,10,39-41,45-47 Survival was assumed to be independent of T classification for patients with N2/N3 disease. The treatment effects of neoadjuvant therapy for N2 disease and sequential chemoradiation for N3 disease were estimated from the aforementioned trials9,10 assuming a proportional hazards (multiplicative) treatment effect. Patients in the model with N0/N1 disease who were miscategorized as having N2 disease and treated with induction chemotherapy and resection were assumed to derive no additional treatment benefit. Patients with N0 to N2 disease who were miscategorized as having N3 disease and treated with chemoradiation alone were assumed to derive no treatment benefit and to behave like patients with N3 disease treated with chemoradiation. Finally, patients with N3 disease who were miscategorized as having N0/N1 disease and treated with surgery alone were assumed to derive no benefit from resection, whereas patients with N3 disease who were treated with neoadjuvant chemotherapy, resection, and postoperative chemoradiation were assumed to behave like patients with N3 disease treated with chemoradiation alone.
Health-Related Quality of Life
Costs Estimates of the costs of mediastinoscopy and lung resection were derived from primary cost data from the Brigham and Women’s Hospital cost accounting system. The average cost incurred by 100 consecutive patients with lung cancer who underwent mediastinoscopy between 1995 and 1996 was $2,100. Ninety-four patients were treated as outpatients and six patients were observed overnight, and there were no major complications or deaths in this group. The average cost incurred by 100 consecutive patients with documented lung cancer who underwent resection during the same time period was $22,000, and there were no deaths in this group. Both of these cost figures were well within the range reported in the literature.56-61 The costs of outpatient chemotherapy for stage IIIA and IIIB disease and radical radiotherapy were extrapolated from a published cost-effectiveness analysis of multimodal therapy for stage III NSCLC.62 This study identified the costs associated with three inpatient chemotherapy/radiotherapy regimens, including their complications. The costs of multimodal therapy for stage IIIA and IIIB disease vary widely in the literature,47,60,62-64 and our estimates were tested over a broad range in the sensitivity analysis. In our model, patients who died as a result of surgery, chemotherapy, or chemoradiation were assumed to incur the full cost of that treatment at the time of death. The long-term, direct medical costs patients with local or regional lung cancer were obtained from a study of costs incurred by health maintenance organization patients at Kaiser-Permanente over a 15-year period.65 In that study, cancer-attributable long-term costs were calculated by subtracting the annual continuing care and end-of-life costs of age and sex-matched controls. For the purpose of our analysis, those costs were assumed to be representative of the cancer-attributable, long-term, direct medical costs incurred by non–health maintenance organization patients with lung cancer. In our model, patients who died during the first year incurred the costs of any procedures or treatments during that year plus an end-of-life cost. For patients who died during subsequent years, the continuing care cost of that year was adjusted to account for the end-of-life period and its associated cost. The direct nonmedical costs associated with surgery, outpatient chemotherapy, and radiotherapy were estimated from a study of the non–health care costs (food, transportation, parking costs, and so on) incurred by patients undergoing outpatient chemotherapy during both treatment and nontreatment weeks.66 In our model, treatment and nontreatment week costs were assumed to be constant regardless of whether patients were recovering from lung resection, receiving chemotherapy, or undergoing radiotherapy. Patients were assumed to incur nontreatment costs during the 1-week postmediastinoscopy period and the 1-month postresection period and treatment costs during the chemotherapy and radiotherapy periods. An average wage rate of $15 per hour for individuals 55 years of age and older was used to estimate the time cost (ie, opportunity cost) of patients and their relatives/caregivers. This figure was based on a national, median weekly earning of $590 for this age group, assuming a 40-hour work week.67 The indirect costs (time costs of patients and caregivers) of the various staging strategies and treatments were estimated by multiplying this wage rate by the number of hours associated with mediastinoscopy (8 hours), chemotherapy (4 hours), and radiotherapy (4 hours). These time estimates were assumed to include transportation (both ways), parking, and waiting room time. Patients were assumed to be accompanied by one relative (an age-comparable spouse) at their mediastinoscopy and chemotherapy sessions and were assumed to be unaccompanied at their radiotherapy sessions.
Sensitivity Analysis
Outcomes and Cost-Effectiveness of Staging Strategies The outcomes and treatments received by a hypothetical cohort of NSCLC patients staged with each of the three staging strategies are shown in Table 2. Using the CT alone strategy, 11% of N0/N1 patients were miscategorized and treated with multimodal therapy. In addition, a large number of patients with N2 or N3 disease underwent resection alone or were treated with chemoradiation alone or neoadjuvant therapy/resection, respectively. The selective mediastinoscopy strategy resulted in a small risk of mediastinoscopy-related deaths, but none of the patients with N0/N1 disease received multimodal therapy. Because of the false-negative rate of mediastinoscopy, however, a slightly higher proportion of patients with N2/N3 disease were inadvertently treated with resection alone. In comparison, routine mediastinoscopy resulted in a slight increase in mediastinoscopy-related deaths. However, none of the N0/N1 patients received multimodal therapy, fewer patients with N2/N3 disease underwent resection alone, and more patients with N2 and N3 disease received neoadjuvant therapy and chemoradiation, respectively.
The expected cost, quality-adjusted life expectancy, and ICERs of the three staging strategies according to T classification are shown in Table 3. In the reference-case analysis, quality-of-life adjustments were used, and both costs and life expectancy were discounted at a rate of 3% per year. Routine mediastinoscopy maximized quality-adjusted life expectancy in all patients, irrespective of T classification, and was associated with the highest total costs.
The incremental cost associated with adopting routine rather than selective mediastinoscopy ranged from $1,200 to $1,600. For patients with T1 tumors, the ICER for selective mediastinoscopy was $24,500 per QALY gained compared with CT alone, whereas routine mediastinoscopy cost $78,800 per QALY gained compared with selective mediastinoscopy. For patients with T2 and T3 tumors, the ICER of routine mediastinoscopy was $42,800 and $53,400 per QALY gained, respectively. The ICERs of the mediastinoscopy strategies became less favorable with discounting due to the fact that the incremental costs associated with these strategies were incurred largely in the short term (during treatment), while their incremental gains were often realized over the long term (and were thus discounted to a greater degree). In T1 patients, routine mediastinoscopy cost more than $50,000 per QALY gained, irrespective of the time frame or perspective used in the analysis (Table 4). In contrast, the ICER of routine mediastinoscopy was quite favorable in both T2 and T3 patients, even when long-term costs were also considered and a societal perspective was adopted. In fact, among T3 patients, selective mediastinoscopy was eliminated by extended dominance in all four scenarios, leaving routine mediastinoscopy with an ICER of $30,000 to $53,400 per QALY gained.
Sensitivity Analysis One-way and two-way threshold analyses focusing on quality-adjusted life expectancy were used to determine the effect of the variables in the model on the rank ordering of the three staging strategies.68 For patients with T1, T2, and T3 tumors, routine mediastinoscopy resulted in the highest quality-adjusted life expectancy as long as the sensitivity of CT was less than 90%, 95%, and 99%, respectively. Varying the specificity of CT or the sensitivity of mediastinoscopy over broad ranges had no effect on the rank ordering of the strategies, regardless of T classification. In fact, routine mediastinoscopy maximized quality-adjusted life expectancy as long as the sensitivity of mediastinoscopy was greater than 10%. Estimates of the mortality of mediastinoscopy, lung resection, and chemoradiation were also tested. For patients with T1, T2, and T3 tumors, routine mediastinoscopy resulted in the highest quality-adjusted life expectancy until the mortality of mediastinoscopy reached 0.6%, 1%, and 8%, respectively, at which time selective mediastinoscopy became the most effective strategy. These threshold values exceeded even the highest reported mortalities quoted in the literature for cervical and anterior mediastinoscopy (0.1% and 0.5%, respectively). Two-way sensitivity analyses were performed in which the sensitivities of CT and mediastinoscopy, the mortality of mediastinoscopy, and the mortality of multimodal therapy were simultaneously varied. In all cases, routine mediastinoscopy remained the most effective strategy when the variables were varied over the reported range. Varying the baseline 5-year survival rates of patients with N2/N3 disease and the efficacy (hazard ratio) of multimodal therapy had no effect on the rank ordering of the staging strategies except in one extreme scenario. In patients with T3 tumors, CT alone resulted in the greatest health benefit when the baseline 5-year survival probability of patients with N2 disease was lower than 1.2%. In this case, the baseline survival of patients with N2 disease was so low that the benefit gained by identifying these patients at mediastinoscopy and treating them with NAC was negligible. The effect of varying the efficacy of trimodality therapy for N2 disease and chemoradiation for N3 disease was also analyzed. In T1 to T3 patients, routine mediastinoscopy maximized quality-adjusted life expectancy irrespective of the assumed efficacy of these regimens. Stratified sensitivity analyses of the cost-effectiveness of the three staging strategies were performed. For patients with T1 tumors, the ICER of routine mediastinoscopy exceeded $50,000 per QALY gained under most circumstances. In contrast, the ICER of adopting the selective mediastinoscopy strategy (over the CT alone strategy) consistently remained well below $50,000 per QALY gained under all circumstances. For patients with T2 tumors, routine mediastinoscopy resulted in a reasonable cost-effectiveness ratio as long as the sensitivity and specificity of CT were less than 82% and 93%, respectively, and the sensitivity of mediastinoscopy was greater than 70% (Table 5). Increasing the age at model entry resulted in less favorable ICERs because of the increased rate of death from competing causes, which decreased the incremental quality-adjusted life expectancies of the mediastinoscopy strategies. Varying the estimates for the mortality of mediastinoscopy and lung resection and the morbidity and mortality of NAC and chemoradiation over their reported ranges had minimal effect on the results. Routine mediastinoscopy became less favorable once the baseline 5-year survival of patients with N2 disease was assumed to be less than 10% and the survival after NAC was assumed to be less than 15%. The ICER of this strategy also exceeded $50,000 per QALY gained once the cost of mediastinoscopy exceeded $2,900, the cost of lung resection dropped below $15,000, or the cost of radiotherapy exceeded $15,000.
When the analysis was limited to patients with T3 tumors, results were less sensitive to changes in the sensitivity and specificity of CT but were more responsive to changes in the sensitivity of mediastinoscopy (Table 6). Once again, increasing the age at model entry increased the ICERs of the mediastinoscopy strategies because of higher death rates from competing causes. Varying the risks of mediastinoscopy, lung resection, NAC, and chemoradiation had minimal effect on the model results. CT alone dominated the two mediastinoscopy strategies if the baseline 5-year survival of N2 patients was less than 100%. Although the model was less sensitive to changes in the costs of mediastinoscopy and lung resection in these patients, routine mediastinoscopy became less cost-effective as the costs of radiotherapy increased (because of the large proportion of patients with mediastinal disease who received radiotherapy).
We also analyzed the effect of treating patients with N2 disease using dual-modality therapy (NAC/resection without radiotherapy) rather than trimodality therapy in the model, assuming that their efficacy was equal. In T1 patients, the ICER of selective mediastinoscopy was $23,600, compared with $71,700 for routine mediastinoscopy. In T2 and T3 patients, the ICERs of routine mediastinoscopy became more favorable once radiotherapy was eliminated. The ICERs of routine mediastinoscopy for T2 and T3 patients were $38,400 and $49,200, respectively. Our model was insensitive to the utility values used to quality-adjust life expectancy. In addition, halving and doubling the times spent in the various resulting health states had no effect on the results. This was probably due to the fact that total time spent in these treatment states was short compared with the time spent in the unadjusted remission state, which made up the bulk of patients’ life expectancies. Our results were robust to varying the costs of NAC and chemotherapy for stage IIIB disease, as well as the long-term medical costs, direct nonmedical costs, and indirect costs associated with the three staging strategies, irrespective of T classification.
The optimal strategy for staging patients with NSCLC remains controversial. Although chest CT can be used to assess patients’ T and M status, the ability of CT to accurately assess mediastinal involvement is suboptimal. Although mediastinoscopy provides a more accurate way to stage the mediastinum, it is associated with additional risks and costs. Several studies have attempted to identify the best strategy for staging the mediastinum of patients with NSCLC. Malenka et al69 used decision analysis to compare various staging strategies combining bronchoscopy/transbronchial needle aspiration, CT, and mediastinoscopy. In their model, patients without mediastinal disease underwent resection, whereas patients with evidence of mediastinal involvement were treated with radical radiotherapy alone. Life expectancy was estimated using the DEALE, but was not quality-adjusted. The various staging strategies resulted in similar life expectancies, likely because identifying and treating patients with mediastinal disease with radiotherapy provided little survival benefit. The subsequent cost-effectiveness analysis used short-term charges, rather than costs, and was limited to a health care system perspective. The authors concluded that strategies that used bronchoscopy and transbronchial needle aspiration to stage the mediastinum were preferable because of their lower cost. Eddy58 performed a cost-minimization analysis that compared selective mediastinoscopy with routine mediastinoscopy for staging NSCLC. Because the two staging strategies were assumed to have equivalent health outcomes, the analysis focused mainly on the cost consequences of the staging strategies with respect to total procedural costs and hospital lengths of stay. The analysis was also performed from a health care system perspective and considered only short-term costs. The total costs and hospital days associated with selective mediastinoscopy were consistently lower than those of routine mediastinoscopy. However, as the prior probability of mediastinal disease (and the proportion of positive CT scans) increased, more mediastinoscopies were performed in the selective strategy, and the differences between the two strategies diminished. More recently, the Canadian Lung Oncology Group performed a randomized, controlled trial comparing selective versus routine mediastinoscopy.70 Patients with both suspected and proven NSCLC were enrolled onto the study. Patients without evidence of mediastinal disease underwent resection, whereas patients with mediastinal involvement were considered to have unresectable disease and were not considered for multimodal therapy. The outcomes of interest were the number of thoracotomies without cure (defined as resection with recurrence within 3 years), the number of unnecessary thoracotomies (for benign disease), and the costs associated with the two strategies. The economic analysis was limited to hospital costs and professional fees within 3 months of randomization. After a follow-up period of 3 years, the two staging strategies seemed to have similar rates of thoracotomy without cure and costs. The inability of this study to show a difference may be due to the fact that approximately 10% of patients had either benign disease or SCLC. In addition, patients who were identified as node-positive by mediastinoscopy were not treated with multimodal therapy. In our study, we compared the outcomes and cost-effectiveness of three commonly used staging strategies for NSCLC. The risks, subsequent treatments, and long-term outcomes associated with each strategy were included to capture the full impact of each strategy. The analysis was performed from both a health care system and societal perspective, in an effort to address the interests of patients, physicians, and payors. To enhance the comparability and generalizability of our results, we used a reference-case analysis that involved quality adjustment of life expectancy, consideration of direct nonmedical and indirect (time) costs, and discounting of both costs and outcomes. In our model, routine mediastinoscopy was the optimal staging strategy with respect to both health outcomes (number of patients accurately staged/treated) and quality-adjusted life expectancy. Our results were stable to varying the estimates used in the model, except in one extreme scenario in which the baseline survival of patients with N2 disease was so low that the risks and incremental costs of routine mediastinoscopy (and its downstream treatments) outweighed its potential benefits. Otherwise, the rank ordering of the three staging strategies did not change unless estimates well beyond the reported ranges were used. In particular, the mortality of mediastinoscopy needed to exceed 0.6% before selective mediastinoscopy became the most effective strategy, illustrating the fact that the routine use of mediastinoscopy in patients with known NSCLC need not be limited to major referral centers. From an economic standpoint, the ICER of routine mediastinoscopy in patients with T2 or T3 tumors compared favorably with values cited for other currently accepted medical technologies,71 irrespective of whether the analysis was carried out from a health care system or societal perspective. This result was stable to variations of the test performance of CT and mediastinoscopy and the mortality of mediastinoscopy within their reported ranges. The ICER of routine mediastinoscopy was sensitive to variations in the baseline survival of patients with N2 disease and the efficacy of multimodal therapy, but only at the extremes. The ICER of routine mediastinoscopy became less favorable as the cost of mediastinoscopy approached $4,500. In a previous series of outpatient mediastinoscopies, the total hospital charges for outpatients was $1,400, compared with $4,110 for patients observed overnight after their procedure.72 The fact that the average cost incurred by six patients who were observed overnight in our series was $5,600 further highlights the need to perform mediastinoscopy on an outpatient basis to enhance the cost-effectiveness of this procedure. When the analysis was limited to patients with T1 tumors, the ICER of routine mediastinoscopy routinely exceeded $50,000 per QALY gained. In this subgroup, approximately 76% of patients had N0/N1 disease, whereas only 24% of patients had N2/N3 disease. Adopting the selective mediastinoscopy strategy (over the CT alone strategy) eliminated the number of N0/N1 patients subjected to multimodal therapy and its attendant risks. In contrast, the incremental change in life expectancy associated with adopting routine mediastinoscopy (over selective mediastinoscopy) was limited to the small number of patients with N2/N3 disease and thus did not justify the incremental cost of adopting this strategy. Our analysis had several potential limitations. Our model was limited to patients with known NSCLC and excluded patients with locally unresectable or widely metastatic disease. Because a large proportion of patients with NSCLC are referred for surgical evaluation after a biopsy and limited metastatic work-up have already been performed, we do not feel that our entry criteria limited the generalizability of our results. Alternative staging strategies, such as endoscopic ultrasonography–guided fine-needle aspiration and video-assisted thoracoscopy, were not considered in this analysis because they are not yet widely available and their performance characteristics have not been fully defined. Although recent reports suggest that the accuracy of positron emission tomography in detecting mediastinal and distant metastases exceeds that of CT,73,74 its availability remains limited to major referral centers. Because our aim was to design a model that would be applicable to all centers, we limited our analysis to widely available, commonly used staging strategies and did not incorporate positron emission tomography into our analysis. Because the only randomized, controlled trials that have compared surgery with multimodal therapy in patients with stage IIIA NSCLC involved NAC followed by resection and radiotherapy,9,10 we assumed that patients with N2 disease in our model were treated with similar trimodality regimens. Although these trials were small and have been criticized on the basis of their inclusion criteria and poor survival in the control arms, phase III trials comparing alternative dual-modality regimens have not been performed, and thus we could not test the impact of these regimens in our model. Another limitation of our study is that the costs of NAC and chemotherapy for stage IIIB disease were based on Canadian cost data. Given the differences in our health systems, these costs may not be valid when applied to patients treated in the United States, even after conversion to US dollars. Although varying the costs of chemotherapy had minimal effect of the model results in the sensitivity analysis, the ICERs of the mediastinoscopy strategies were sensitive to the cost of radiotherapy, which were estimated from American and Canadian studies. The fact that the cost of radical radiotherapy was highest in an American study47 highlights the need to collect cost data from patients with localized/locoregional NSCLC treated in the United States, particularly with respect to costs of radiotherapy. Finally, our model does not address the importance of accurately staging patients enrolled in clinical trials to ensure the interpretability and generalizability of the results. In summary, routine mediastinoscopy in patients with known NSCLC and T2/T3 tumors resulted in the greatest health benefit and was associated with an ICER comparable to other currently accepted medical technologies. Routine mediastinoscopy protected patients with localized (N0/N1) disease from unnecessary multimodal therapy and steered more patients with N2 or N3 disease toward more appropriate multimodal regimens. In patients with T1 tumors, the incremental cost of routine mediastinoscopy partly overshadowed its survival benefit, making selective mediastinoscopy potentially more attractive from an economic standpoint (although routine mediastinoscopy had a survival advantage over selected mediastinoscopy). Our results suggest that routine mediastinoscopy is a safe staging strategy and that its ability to accurately stage patients and optimize their treatment outweighs its risks and costs, particularly in patients with T2 and T3 tumors.
1. Greenlee RT, Murray T, Bolden S, et al: Cancer statistics, 2000. CA Cancer J Clin 50: 7-33, 2000[Abstract]
2.
Mountain CF: A new international staging system for lung cancer. Chest 89: 225S-233S, 1986 (suppl 4)
3.
Mountain CF: Revisions in the international system for staging for lung cancer. Chest 111: 1710-1723, 1997 4. Walliers E, Waters PF: Incidence of mediastinal node involvement in clinical T1 bronchogenic carcinomas. Can J Surg 30: 341-342, 1987[Medline]
5.
Gross DH, Glazer GM, Orringer MB, et al: Bronchogenic carcinoma metastatic to normal-sized lymph nodes: Frequency and significance. Radiology 166: 71-74, 1988
6.
Arita T, Kuratmisu T, Kawamura M, et al: Bronchogenic carcinoma: Incidence of metastases to normal sized lymph nodes. Thorax 50: 1267-1269, 1995 7. Pauker S, Kassirer JP: Decision analysis. N Engl J Med 316: 250-258, 1987[Medline] 8. Weinstein MC, Fineberg HV (eds): Clinical Decision Analysis. Philadelphia, PA, Saunders, 1980
9.
Rosell R, Gomez-Codina J, Camps C, et al: A randomized trial comparing pre-operative chemotherapy plus surgery with surgery alone in patients with non-small cell lung cancer. N Engl J Med 330: 153-158, 1994
10.
Roth J, Fossella F, Komaki R, et al: A randomized trial comparing pre-operative chemotherapy and surgery with surgery alone in resectable stage IIIA non-small cell lung cancer. J Natl Cancer Inst 86: 673-680, 1994 11. Beck JR, Pauker SG: The Markov process in medical prognosis. Med Decis Making 3: 419-458, 1983 12. Sonnenberg FA, Beck JR: Markov models in medical decision making: A practical guide. Med Decis Making 13: 322-338, 1993 13. Torrance GW: Measurement of health state utilities for economic appraisal: A review. J Health Econ 5: 1-30, 1986[CrossRef][Medline]
14.
Weinstein MC, Siegel JE, Gold MR, et al: Recommendation of the panel on cost-effectiveness in health and medicine. JAMA 276: 1253-1258, 1996 15. Gold MR, Siegel JE, Russell LB, et al (eds): Cost-effectiveness in Health and Medicine. New York, NY, Oxford University Press, 1996 16. Weinstein MC, Stason WB: Foundations of cost-effectiveness analysis for health and medical practices. N Engl J Med 296: 716-721, 1977[Abstract]
17.
Eisenberg JM: Clinical economics: A guide to the economic analysis of clinical practices. JAMA 262: 2879-2886, 1989 18. Dales RE, Start RM, Sankaranarayanan R: Computed tomography to stage lung cancer: Approaching a controversy using meta-analysis. Am Rev Respir Dis 141: 1096-1101, 1990[Medline] 19. Patterson GA, Ginsberg RJ, Poon PY, et al: A prospective evaluation of magnetic resonance imaging, computed tomography, and mediastinoscopy in the preoperative assessment of mediastinal node status in bronchogenic carcinoma. J Thorac Cardiovasc Surg 94: 679-684, 1987[Abstract] 20. Van Schil P, Van Hee R, Schoofs E: The value of mediastinoscopy in the preoperative staging of bronchogenic carcinoma. J Thorac Cardiovasc Surg 97: 240-244, 1989[Abstract] 21. Jolly PC, Hutchinson CH, Detterback F, et al: Routine computed tomographic scans, selective mediastinoscopy, and other factors in evaluation of lung cancer. J Thorac Cardiovasc Surg 102: 266-271, 1991[Abstract] 22. Dillemans B, Deneffe G, Verschakelen J, et al: Value of computed tomography and mediastinoscopy in preoperative evaluation of mediastinal nodes status in non-small cell lung cancer. Eur J Cardiothorac Surg 8: 37-42, 1994[Abstract] 23. Grover FL: The role of CT and MRI in staging of the mediastinum. Chest 106: 391S-396S, 1994 (suppl 6) 24. Gdeedo A, Van Schil P, Corthouts B, et al: Prospective evaluation of computed tomography in mediastinal lymph node staging. Eur Respir J 10: 1547-1551, 1997[Abstract]
25.
Porte H, Roumilhac D, Eraldi L, et al: The role of mediastinoscopy in the diagnosis of mediastinal lymphadenopathy. Eur J Cardiothorac Surg 13: 196-199, 1998 26. Kirschner PA: Cervical mediastinoscopy. Chest Surg Clin North Am 6: 1-20, 1996[Medline] 27. Ginsberg RJ: Evaluation of the mediastinum by invasive techniques. Surg Clin North Am 67: 1025-1035, 1987[Medline] 28. Ashbaugh DG: Mediastinoscopy. Arch Surg 100: 569-573, 1970 29. Luke W, Pearson F, Todd T, et al: Prospective evaluation of mediastinoscopy for assessment of carcinoma of the lung. J Thorac Cardiovasc Surg 91: 53-56, 1986[Abstract]
30.
Hammoud Z, Anderson R, Meyers B, et al: The current role mediastinoscopy in the evaluation of thoracic disease. J Thorac Cardiovasc Surg 118: 894-899, 1999 31. McNeill TM, Chamberlain JM: Diagnostic anterior mediastinoscopy. Ann Thorac Surg 2: 532-539, 1966[Medline] 32. Best L, Munichor M, Ben-Shakhar M, et al: The contribution of anterior mediastinoscopy in the diagnosis and evaluation of diseases of the mediastinum and lung. Ann Thorac Surg 43: 78-81, 1987[Abstract] 33. Jolly PC, Hill L, Lawless P, et al: Parasternal mediastinotomy and mediastinoscopy. J Thorac Cardiovasc Surg 66: 549-555, 1987[Medline] 34. Byers T, Vena J, Rzepka T: Predilection of lung cancer for the upper lobes: An epidemiologic inquiry. J Natl Cancer Inst 72: 1271-1275, 1984 35. Martini N, Flehinger B, Zaman M, et al: Results of resection in non-oat cell carcinoma with mediastinal lymph node metastases. Ann Surg 198: 386-397, 1983[Medline] 36. Watanabe Y, Shimizu J, Ota M, et al: Aggressive surgical intervention in N2 non-small cell lung cancer. Ann Thorac Surg 51: 253-261, 1991[Abstract] 37. Daly B, Mueller J, Faling L, et al: N2 lung cancer: Outcome in patients with false-negative computed tomography scans of the chest. J Thorac Cardiovasc Surg 105: 904-911, 1993[Abstract]
38.
Goldstraw P, Mannam G, Kaplan D, et al: Surgical management of non-small cell lung cancer with ipsilateral mediastinal lymph node metastases (N2 disease). J Thorac Cardiovasc Surg 107: 19-28, 1994 39. Martini N, Kris MG, Flehinger BJ, et al: Preoperative chemotherapy for stage IIIa (N2) lung cancer: The Sloan-Kettering experience with 136 patients. Ann Thorac Surg 55: 1365-1373, 1993[Abstract]
40.
Sugarbaker D, Herndon J, Kohman L, et al: Results of CALGB 8935: A multi-institutional phase II trimodality trial for stage IIIA non-small cell lung cancer. J Thorac Cardiovasc Surg 109: 473-485, 1995 41. Elias A, Skarin A, Leong T, et al: Neoadjuvant therapy for surgically staged stage IIIA N2 non-small cell lung cancer. Lung Cancer 17: 147-161, 1997[CrossRef][Medline]
42.
Evans E: Resection of bronchial carcinoma in the elderly. Thorax 28: 86-88, 1973 43. Ginsberg R, Hill L, Eagan R, et al: Modern thirty-day operative mortality for surgical resection in lung cancer. J Thorac Cardiovasc Surg 86: 654-658, 1983[Abstract]
44.
Deslauriers J, Ginsberg R, Piantadosi S, et al: Prospective assessment of 30-day operative morbidity for surgical resections in lung cancer. Chest 106: 329S-330S, 1994 (suppl 6)
45.
LeChevalier T, Arriagada R, Quoix E, et al: Radiotherapy alone versus combined chemotherapy and radiotherapy in non-resectable non-small cell lung cancer: First analysis of a randomized controlled trial in 353 patients. J Natl Cancer Inst 83: 417-423, 1991
46.
Sause W, Scott C, Taylor S, et al: Radiation Therapy Oncology Group (RTOG) 88-08 and Eastern Co-operative Oncology Group (ECOG) 4588: Preliminary results of a phase III trial in regionally advanced unresectable non-small cell lung cancer. J Natl Cancer Inst 87: 198-205, 1995 47. Dillman R, Herndon S, Propert K, et al: A randomized trial of induction chemotherapy plus high-dose radiation versus radiation alone in stage III non-small cell lung cancer. N Engl J Med 323: 940-945, 1990[Abstract] 48. Naruke T, Tomoyuki G, Ryosuke T, et al: Prognosis and survival in resected lung carcinoma based on the new staging system. J Thorac Cardiovasc Surg 96: 440-447, 1988[Abstract] 49. Beck J, Kassirer J, Pauker S: A convenient approximation of life expectancy (the "DEALE"): I. Validation of the method. Am J Med 73: 883-888, 1982[CrossRef][Medline] 50. Beck J, Pauker S, Gottlieb J: A convenient approximation of life expectancy (the "DEALE"): II. Use in medical decision making. Am J Med 73: 889-897, 1982[CrossRef][Medline]
51.
Goodwin P, Feld R, Evans W, et al: Cost-effectiveness of cancer chemotherapy: An economic evaluation of a randomized trial in small-cell lung cancer. J Clin Oncol 6: 1537-1547, 1988
52.
Ng A, Weeks J, Maunch P, et al: Decision analysis on alternative treatment strategies for favorable-prognosis, early-stage Hodgkin’s lymphoma. J Clin Oncol 17: 3577-3585, 1999 53. Ness R, Holmes A, Klein R, et al: Utility valuations for outcome states of colorectal cancer. Am J Gastroenterol 94: 1650-1657, 1999[CrossRef][Medline] 54. Hillner B, Smith T: Efficacy and cost-effectiveness of adjuvant chemotherapy in women with node-negative breast cancer. N Engl J Med 324: 160-168, 1991[Abstract] 55. Bureau of Labor Statistics: Consumer Price Index: Medical Component. Downloaded November 2, 1999. Http://146.142.4.24/cgi-bin/surveymost
56.
Black W, Armstrong P, Daniel T: Cost-effectiveness of chest CT in T1N0M0 lung cancer. Radiology 167: 373-378, 1988
57.
Scott W, Sheperd J, Gambhir S: Cost-effectiveness of FDG-PET for staging non-small cell lung cancer: A decision analysis. Ann Thor Surg 66: 1876-1885, 1998 58. Eddy R: Cost-effectiveness of CT compared to mediastinoscopy in the preoperative staging of lung cancer. J Can Assoc Radiol 40: 189-193, 1989 59. Aabakken L, Silvestri G, Hawes R, et al: Cost-efficacy of endoscopic ultrasonography with fine-needle aspiration vs. mediastinoscopy in patients with lung cancer and suspected mediastinal adenopathy. Endoscopy 31: 707-711, 1999[CrossRef][Medline]
60.
Evans W: Management of metastatic non-small cell lung cancer and a consideration of cost. Chest 103: 68S-71S, 1993 (suppl 1) 61. Evans W, Will B, Berthelot JM, et al: The cost of managing lung cancer in Canada. Oncology 9: 147-153, 1995 (suppl 11)[Medline] 62. Evans W, Will B, Berthelot JM, et al: Cost of combined modality interventions for stage III non-small cell lung cancer. J Clin Oncol 15: 3038-3048, 1997[Abstract] 63. Coy P, Schaafsma J, Schofield J, et al: Comparative costs of lung cancer management. Clin Invest Med 17: 577-587, 1994[Medline] 64. Earle C, Evans W: Cost-effectiveness of paclitaxel plus cisplatin in advanced non-small cell lung cancer. Br J Cancer 80: 815-820, 1999[CrossRef][Medline] 65. Fireman B, Quesenberry C, Somkin C, et al: Cost of care for cancer in a health care maintenance organization. Health Care Financing Rev 18: 51-76, 1997[Medline] 66. Houts P, Lipton A, Harvey H, et al: Nonmedical costs to patients and their families associated with outpatient chemotherapy. Cancer 53: 2388-2392, 1984[CrossRef][Medline] 67. U.S. Census Bureau: Money Income in the U.S.: Median Income of People by Selected Characteristics. Downloaded November 2, 1999. Http://www.census.gov/prod/2000pubs/p60-209.pdf 68. Pauker S, Kassirer J: The threshold approach to clinical decision making. N Engl J Med 302: 1109-1117, 1980[Abstract] 69. Malenka D, Colice G, Jacobs C, et al: Mediastinal staging in non-small cell lung cancer. Med Decis Making 9: 231-242, 1989
70.
The Canadian Lung Oncology Group: Investigation for mediastinal disease in patients with apparently operable lung cancer. Ann Thorac Surg 60: 1382-1389, 1995
71.
Smith T, Hillner B, Desch C: Efficacy and cost-effectiveness in oncology: Rational allocation of cancer care resources. J Natl Cancer Inst 85: 1460-1474, 1993 72. Bonadies J, D’Agostino R, Ruskis A, et al: Outpatient mediastinoscopy. J Thorac Cardiovasc Surg 106: 686-688, 1993[Abstract]
73.
Dwamena B, Sonnad S, Angobaldo J, Wahl R: Metastases from non-small cell lung cancer: Mediastinal staging in the 1990s—Meta-analytic comparison of PET and CT. Radiology 213: 530-536, 1999
74.
Pieterman R, van Putten J, Meuzelaar J, et al: Preoperative staging of non-small cell lung cancer with positron-emission tomography. N Engl J Med 343: 254-261, 2000 Submitted February 1, 2001; accepted August 6, 2001.
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2002 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
, decision nodes; 
, chance nodes. Abbreviations: Med, mediastinoscopy; CT- and CT+, patients with normal-appearing or enlarged mediastinal nodes on CT; "N2" and "N3", patients assumed to have N2 or N3 disease; Rsxn, lung resection; NAC, neoadjuvant chemotherapy; CR, chemoradiation therapy.
) is assumed to be constant, and survival is described by the formula S(t) = e -
