|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 22, No 16 (August 15), 2004: pp. 3261-3268 © 2004 American Society of Clinical Oncology. DOI: 10.1200/JCO.2004.02.051 Population Variations in the Initial Treatment of Non–Small-Cell Lung CancerFrom the Division of Cancer Control and Population Sciences, Division of Cancer Treatment and Diagnosis, National Cancer Institute, Bethesda, MD; and Department of Epidemiology, The University of Iowa, Iowa City, IA Address reprint requests to Arnold L. Potosky, PhD, Applied Research Program, National Cancer Institute, 6130 Executive Blvd, EPN Room 4005, Bethesda, MD 20892; e-mail: potosky{at}nih.gov
PURPOSE: Dissemination of recommended therapies for non–small-cell lung cancer (NSCLC) have not been described comprehensively. We report the patterns of initial therapy focusing on the investigation of differences in receipt of recommended therapies according to multiple clinical and nonclinical patient characteristics. METHODS: A population-based random sample of newly diagnosed NSCLC patients diagnosed in 10 separate geographic areas was collected in 1996 (n = 898). Data were obtained from medical records. Multiple logistic regression was used to assess the use of recommended therapies. RESULTS: Overall, 52% of NSCLC patients received recommended therapy. Approximately 69%, 48%, and 41% of patients with stages I and II, III, or IV NSCLC received recommended therapy, respectively. For all stages combined, the use of recommended therapy was significantly inversely associated with age and stage at diagnosis. Recommended therapy also was more common in white versus black patients, and in married versus single patients. Stage-specific analyses revealed a significant decline in the use of recommended surgery with increasing age at diagnosis for early-stage NSCLC only, and a significantly lower use of recommended therapy (primarily chemoradiotherapy) for stage III black and Hispanic patients compared with white patients. CONCLUSION: The overall use of recommended therapies for NSCLC is low. Large variations exist in the use of such therapies according to age, race or ethnicity, and marital status. Research combining medical record reviews with other sources of data is needed to better understand the contributions of both patient preferences and physician judgment to these treatment variations.
Lung cancer is the number one cause of cancer deaths in the United States, with an estimated 157,000 deaths in 2003.1 Non–small-cell lung cancer (NSCLC) comprises more than 80% of all new lung cancer cases. Although 2-year survival from the disease is only 25% and 5-year survival is only 17%,2 there are specific treatment strategies that are recommended in clinical guidelines or statements of evidence by the American Society for Clinical Oncology, the National Cancer Institute, the American College of Chest Surgeons, and others for the initial management and treatment of NSCLC patients.3-5 However, several population-based studies of clinical practice patterns in US patients have documented variations in the initial management of NSCLC patients according to age, race or ethnicity, education, comorbidity, and hospital type, suggesting that there are gaps in receipt of quality cancer care among certain population groups.6-10 In this study, we describe the clinical practice patterns in the United States by assessing a randomly selected cohort of nearly 900 NSCLC patients diagnosed in 1996 who were treated in diverse health care settings in 10 distinct geographic regions. We focus on receipt of specific therapies that we define as "recommended" to permit an assessment of community variations in receipt of best clinical practices, basing our definitions on the accumulation of clinical evidence and major consensus guidelines or evidence summaries.3-5 For American Joint Committee on Cancer tumor-node-metastasis system stage I and II NSCLC patients, primary therapy currently consists of surgical removal of the primary tumor, although radiotherapy is often used if the patient has comorbid conditions that do not allow removal of the primary tumor.11,12 On the basis of information demonstrating that lesser surgery results in a higher rate of cancer recurrence and death, we define recommended surgery for stages I and II NSCLC as pneumonectomy or lobectomy.13 For unresectable stage III NSCLC, randomized clinical trial evidence has accumulated in support of the use of combination chemoradiotherapy. A published meta-analysis demonstrated a 13% reduction in the risk of death for patients treated with radiation combined with cisplatin-based chemotherapy.14 For patients with stage IIIB disease on the basis of the presence of malignant pleural effusion, chemotherapy alone generally is recommended.15 For patients with technically resectable stage IIIA NSCLC, controversy exists about the most effective therapy. Depending on the extent of nodal involvement, definitive chemoradiotherapy, or surgery with or without chemotherapy or radiation therapy, can be recommended. Chemotherapy has been shown to impart modest survival and quality-of-life improvements for selected patients with stage IV NSCLC, increasing median survival by approximately 2 to 3 months.16 For stage IV patients we define primary chemotherapy as the recommended therapy. Using a population-based random sample of newly diagnosed NSCLC patients in 1996, we describe the patterns of practice by stage of disease, focusing on nonclinical factors associated with the use of recommended initial therapies.
From 1998 to 1999, National Cancer Institute's Surveillance, Epidemiology and End Results (SEER) registries in Connecticut; Iowa; New Mexico; Atlanta, GA; Detroit, MI; San Francisco-Oakland, CA; Seattle-Puget Sound, WA; Utah; San Jose-Monterey, CA; and Los Angeles County, CA, drew random samples of all NSCLC patients reported in their geographic catchment areas diagnosed during 1996. This special study was designed to assess the extent to which recommended therapies were received by NSCLC patients in community medical practices and to abstract additional variables such as comorbid conditions and smoking history. Patients with a previous diagnosis of cancer, except for nonmelanoma skin cancer, or with simultaneous cancer diagnoses, were not eligible for participation. There was oversampling of non-Hispanic whites and blacks by specific registries to obtain more stable estimates of practice patterns in these subgroups. After a patient case was sampled, the treating physician was identified from pathology reports and other registry sources and asked to directly verify the use of chemotherapy and radiotherapy for that patient. This method augments standard SEER registry data collection procedures that rely mostly on hospital and other facility records. Physician verification substantially improves the completeness of chemotherapy ascertainment because chemotherapy is frequently given outside the hospital setting, and a note of any planned chemotherapy might not be recorded in the hospital record. The participating registries monitored the percentage of patients in non–health maintenance organizations or Veterans Affairs settings (where medical records including ambulatory visits typically are consolidated in one record) whose physicians were requested to verify the receipt of outpatient therapies. This verification method covered 80% to 90% of sampled patient cases in this study. Comorbidity was assessed using the Charlson index, a weighted index measure shown to predict short- and long-term all-cause mortality in a breast cancer cohort.17 The index is composed of 19 conditions, each of which is assigned a weight according to its potential for influencing mortality. A comorbidity score was calculated as the sum of the weights. Smoking status also was obtained from the medical record. Because of difficulties in precisely ascertaining smoking history, former and current smokers were combined into categories of 1, 2, or unknown packs per day. SEER variables were obtained by trained abstractors from medical record reviews. The variables included in the analysis were stage and year of diagnosis, age at diagnosis, race or ethnicity, and marital status. Tumor-node-metastasis system stage at diagnosis was assigned using the extent-of-disease system employed by SEER using all available pathologic and clinical data from medical records. The SEER extent-of-disease system has been placed in a Comparative Staging Guide that is based on the American Joint Committee on Cancer tumor-node-metastasis system classification scheme.18 Given that the SEER program does not collect individual measures of socioeconomic status for patients, race-specific values for group-level median household income data from the US 1990 decennial census were merged with sampled patients using the census tract of the residence at the time of initial diagnosis. Geographic region was not included in final regression models because of small sample sizes in several region-stage strata. Treatment was measured using routinely collected SEER information on surgery and radiotherapy combined with new information obtained from physician verification, particularly for chemotherapy. Radiation therapy includes only radiation to the primary site, not radiation to metastatic sites. No details on dosage or duration were collected, but specific chemotherapeutic agents were identified. Approximately 90% of all chemotherapy reported was platinum-based regimens. To examine patient characteristics associated with variations in best treatment practice patterns, we defined, a priori, minimally recommended initial therapies for each stage at diagnosis (Table 1). For purposes of this analysis, the determination of recommended therapies was based on pertinent information from the published literature before 1996, including both randomized trials and meta-analyses of randomized trials, as well as definitions of accepted therapy reflected in the American Society of Clinical Oncology clinical practice guidelines for the treatment of unresectable NSCLC published in 1997.3 For tumor-node-metastasis system stages I and II, surgical resection, excluding more limited surgeries, was considered the recommended initial therapy. For stage IIIA patients, combination chemoradiotherapy, or surgery (pneumonectomy or lobectomy) with or without chemotherapy or radiation, was defined as recommended therapy. For some stage IIIB patients, chemotherapy combined with radiation therapy was considered recommended initial therapy. For selected patients with stage IIIB disease and pleural effusion, or patients with stage IV disease, chemotherapy alone was considered the recommended therapy.
 2 tests were used to evaluate bivariate relationships between patient characteristics and receipt of standard care, within each cancer site-stage combination. A two-tailed test was used. A logistic regression model was constructed to assess the association between independent variables and the receipt of therapy. We constructed one model with all stage groups combined, and then three models for each separate stage group. Age at diagnosis was initially included in all models as a linear independent variable, but is reported in the tables as a categoric variable for descriptive purposes. We report the adjusted percentages of patients receiving recommended therapy rather than odds ratios because the latter are not valid estimates of the risk ratio when the outcome is not a rare event. To obtain the adjusted percentages, the logistic regression models were used to generate estimates of the probability for each individual (predicted values) receiving therapy. The percentages in each group were then directly standardized to the distribution of the other model covariates among the entire weighted sample.19 These adjusted percentages are included in the tables to describe the distribution of model covariates with respect to therapies received. Sample weights were used in all analyses, calculated as the inverse of the percentage of all eligible SEER patient cases within sampling strata defined by registry and stage at diagnosis. All frequency estimates and regression coefficients, and associated variances, were computed using the SUDAAN statistical package.20 Wald-type F statistics with the robust variance estimator were used to assess the statistical significance of modeled associations adjusting for all other covariates. All P values were two-sided.
A total of 1,006 NSCLC patients diagnosed in 1996 were sampled. Of these, 13 were excluded because they were found to have benign or in situ disease, or did not have evidence of microscopic confirmation of malignancy. An additional 95 patients with unknown stage at diagnosis also were excluded. These exclusions left a final study cohort of 898 NSCLC patients. We combined stages I and II because of small sample sizes, and because curative resection is recommended for both. Table 2 shows the sample size according to stage groups, and the distribution of patient sociodemographic characteristics, comorbidity, and smoking history within each stage.
Table 3 shows results for all stages of NSCLC combined. Overall, 52% of NSCLC patients received recommended therapy. After adjustment for all other variables shown plus stage at diagnosis in a logistic regression model, age, race, and marital status remained statistically significantly associated with the receipt of recommended therapies. Age was inversely associated with receipt of therapy, with 75% of patients younger than age 50 years receiving recommended therapy, declining to just 32% in patients older than age 75 years. Non-Hispanic whites had an adjusted value of 55% receiving recommended therapy compared with 42% in non-Hispanic blacks and 46% in Hispanics. Married patients had the highest adjusted percentage receiving recommended therapy (59%), whereas single patients had the lowest (38%). Stage was the variable most strongly associated with differences in this combined model. Because of this finding, and the different types of therapies defined as recommended across stage at diagnosis, we next performed similar modeling analyses within stage.
Among 240 patients with stage I or II NSCLC, approximately 69% received recommended surgery (pneumonectomy or lobectomy), 10% received more limited local surgical resection, 16% received radiotherapy alone, and 5% received no reported therapy. Among those receiving resection, nearly 80% received surgery alone, and the remaining 20% received adjuvant radiotherapy and/or chemotherapy. Table 4 lists the percentage of patients who received recommended surgery according to various patient sociodemographic factors. After adjustment for all other variables in a multiple logistic regression model predicting receipt of recommended surgery, there were differences in the receipt of recommended surgery by age, with older patients less likely than younger patients to receive recommended surgery. The adjusted percentages in column 3 indicate that there was a steep decline in the use of surgery in patients older than 70 years. Although there were no other statistically significant model covariates, non-Hispanic whites had a higher adjusted prevalence of recommended surgery (71%) than did non-Hispanic blacks (53%), and surgery also was more frequent among currently married persons, those without any recorded comorbidity, and current and former smokers.
Among 297 stage III patients, 48% received recommended therapy, defined as either chemoradiotherapy or chemotherapy alone if pleural effusion was present. In addition, pneumonectomy or lobectomy was included as a recommended therapy if the patient was diagnosed with stage IIIA disease. Approximately 28% received radiotherapy alone, and 21% received no reported therapy. Table 5 lists the crude and adjusted percentages of stage III patients who received recommended therapy. After adjustment for all other variables, there remained a statistically significant difference in recommended therapy according to race or ethnicity, marital status, and group-level median income. Age at diagnosis was highly inversely associated with recommended therapy when included as a continuous model term (P < .01), but did not attain statistical significance when included as a categoric covariate. Fifty-six percent of non-Hispanic whites had recommended therapy compared with 33% of both non-Hispanic blacks and Hispanics. Currently married patients had the highest use of recommended therapy, whereas those living in areas with the highest median income had the lowest adjusted use of recommended therapy at 37%. In addition, stage III patients living in urban areas versus rural areas, and those without any comorbidity, displayed somewhat higher levels of recommended therapy according to those variables.
Among 361 stage IV NSCLC patients, 41% received chemotherapy as part of their initial therapy, whereas 31% received radiotherapy and 28% received no reported therapy. Among the 40% of patients who received chemotherapy, just more than two thirds also had radiotherapy as part of their initial treatment. Table 6 shows the crude and adjusted percentage of stage IV patients receiving chemotherapy according to various patient factors. Age at diagnosis was highly negatively associated with recommended therapy as a continuous, but not categoric, independent variable. Rural residence was statistically significantly associated with receipt of chemotherapy after adjustment for all other covariates. There are only small differences in chemotherapy according to race or ethnicity or sex, but the adjusted prevalence of chemotherapy was highest among widowed patients and lowest in single patients. Stage IV patients living in higher income areas relative to lower income areas, patients without comorbidity, and current or former smokers versus never smokers all had somewhat higher rates of chemotherapy use.
Although lung cancer is associated with a poor prognosis, there exist several recommended treatment strategies that can result in cure or at least marginally prolong survival and improve symptoms for patients coping with their disease. Prior studies of treatment patterns have found important differences in receipt of recommended therapies according to nonclinical factors, particularly age, race or ethnicity, comorbidity, and hospital type.6-10 Other multiregional studies have reported similar rates of curative resection for NSCLC during the same time period.6,10 A study using Medicare linked with SEER data reported that 22% of stage IV patients received chemotherapy during 1991 to 1993, which is generally consistent with our findings.7 Another national study reported that approximately 25% of stage IIIB patients received combined chemoradiotherapy in 1995, which is somewhat lower than our estimates.10 However, these earlier studies were limited by focusing on limited age groups, studying insured populations rather than population-based samples, or not verifying the administration of outpatient-based therapies in physicians' offices when relying on medical record–based data collection methods. In this study, we examined variables associated with the use of recommended initial treatments for NSCLC patients to assess systematic differences between US population subgroups, using a randomly selected population-based sample for which outpatient therapies were verified by contacting treating physicians. We found a consistent decline in the use of recommended therapy with increasing age at diagnosis, after adjustment for other factors including comorbidity and socioeconomic status. This pattern confirms findings from other studies.7-9 The weight of evidence from many patterns-of-care studies for other sites of cancer indicates that some physicians do not endorse aggressive therapy for elderly cancer patients, perhaps because they do not expect the gains in survival to outweigh the potential harms to patient morbidity and quality of life. The age association observed could be due partly to patient or family preferences, or unmeasured confounding of age with performance status and lung function, which were not included in our study. The summary records-based measure of comorbidity we used may also have been insufficiently comprehensive to capture the full range and severity of comorbidity in lung cancer patients. We observed a pattern of higher use of recommended therapies in whites compared with blacks or Hispanics after adjustment for other sociodemographic factors. The racial difference in recommended surgery for early-stage NSCLC is consistent with earlier findings by Bach et al6 in a much larger sample of early-stage patients age 65 years and older. The large racial difference we observed in stage III recommended therapies has not been reported previously to our knowledge, but suggests the existence of a persistent pattern of racial disparity in the treatment of NSCLC. Whether this disparity is due to unmeasured differences in disease severity (within-stage variations), comorbidity, patient preferences, or physician prescribing behavior remains to be confirmed with more detailed data collection efforts. We found a pattern of more use of recommended therapy among currently married persons after adjustment for other characteristics. Other patterns of cancer care studies have also found that marital status is an important predictor of receipt of more aggressive therapy, suggesting the importance of social support in cancer patients' attitudes toward and acceptance of recommended therapies.21,22 However, the specific factors underlying this association have yet to be clarified. We also found an inverse relationship between comorbidity and the use of recommended therapy in all stages. This finding, although not statistically significant, may be driven partly by the judgment of physicians that their patients with comorbidities cannot withstand the rigors of a more aggressive treatment strategy, or may be due to an age bias in physician referrals or treatment assignments. Comorbidity also may be associated with patients' and their families' decisions to avoid aggressive treatments. NSCLC patients who had no smoking history had lower receipt of recommended therapy than did current and former smokers for all stages combined, with the largest difference observed in receipt of surgery among stage I and II patients. However, these findings were limited by high variability in our estimates of recommended therapies received by never smokers. Exploring the potential reasons for this finding, such as differences in either the biology of tumors or in care-seeking behavior according to smoking status, would require a much larger sample size of never smokers, which in our sample was only 6% of the total. Other than patient preferences or physician judgment in prescribing behavior, there may be other reasons for observing that only 52% of all NSCLC patients received recommended therapies. Because we are unable to verify nonsurgical therapies in approximately 10% of sampled patients, it is possible that the overall use of recommended therapy is somewhat higher than reported, particularly for chemotherapy, which can be administered in a physician's office. However, we do not think this would have a large impact on our estimates because of the high rate of verification of chemotherapy and radiotherapy in this study. We included some patients as receiving recommended therapy who had received additional unproven therapies, such as adjuvant chemotherapy after recommended surgery for early-stage lung cancer. Therefore, our estimates of the proportion receiving recommended therapy might have been biased slightly upward. It also is possible that our definition of recommended therapy might be too narrow, or not yet universally accepted as standard by the oncology community by 1996, and that some patients not receiving such therapy did, in fact, receive appropriate treatment. For example, some guidelines note that radiotherapy is appropriate for early-stage NSCLC if removal of the primary tumor is not possible without undue morbidity. For some advanced NSCLC patients, there persists scientific uncertainty about the cost-benefit tradeoffs of aggressive chemotherapy versus other strategies such as best supportive care.16 Our use of recommended therapy was primarily to provide a conceptual basis for statistically analyzing community practice patterns, and we fully recognize that what we define as recommended therapy might not always be the most appropriate management strategy because of the heterogeneity of patients even within stage at diagnosis. Although our results are based on practice patterns from 1996, they can serve as an important benchmark for more updated studies that are necessary to track progress in the diffusion of recommended therapies for NSCLC. These conceptual challenges, combined with the practical difficulty of measuring the contribution of patient preferences versus physician behaviors on practice patterns using medical records alone, warrants more detailed data collection efforts at the patient, physician, and health care system level to help clarify the complex interplay of these components on dissemination of optimal therapies. More data also are needed on the dose and duration of chemotherapy and radiotherapy. Such efforts are currently being pursued by the American College of Surgeons, the American Society of Clinical Oncology, and the National Comprehensive Cancer Network. The National Cancer Institute is sponsoring the Cancer Care Outcomes Research and Surveillance Consortium, a large observational cohort study targeting accrual of 5,000 lung (and 5,000 colorectal) cancer patients in areas representing approximately 10% of the US population. These efforts aim to clarify the underlying reasons for observed practice differences according to nonclinical characteristics, with the ultimate objective of helping to improve patient outcomes by ensuring more equitable dissemination of recommended therapies to all segments of the population.
The authors indicated no potential conflicts of interest.
Supported by grant No. N01-PC-67005 to the Connecticut Department of Health, N01-PC-67006 to Emory University, N01-PC-67009 to Fred Hutchinson Cancer Research Center, N01-PC-65107 to Northern California Cancer Center, N01-PC-67008 to University of Iowa, N01-PC-67007 to Univeristy of New Mexico, N01-PC-67010 to University of Southern California, N01-PC-67000 to University of Utah, and N01-PC-65064 to Wayne State University from the National Cancer Institute, Division of Cancer Control and Population Sciences. This study was made possible through the efforts of the Principal Investigators and the cancer registry personnel at the participating SEER registries. Authors' disclosures of potential conflicts of interest are found at the end of this article.
1. Cancer Facts and Figures, 2003. http://www.cancer.org 2. Ries LAG, Eisner MP, Kosary CL, et al (eds): SEER Cancer Statistics Review, 1975-2000. Bethesda, MD, National Cancer Institute. http://seer.cancer.gov/csr/1975_2000 3. American Society of Clinical Oncology: Clinical practice guidelines for the treatment of unresectable non–small-cell lung cancer. J Clin Oncol 15:2996–3018, 1997[Abstract] 4. National Cancer Institute: Physician Data Query Cancer Information Summaries. http://www.nci.nih.gov/cancerinfo/pdq/treatment/non-small-cell-lung/healthprofessional/ 5. National Comprehensive Cancer Network and American Cancer Society: Lung Cancer: Treatment Guidelines for Patients. Version 1, December 2001. http://www.nccn.org
6. Bach PB, Cramer LD, Warren JL, et al: Racial differences in the treatment of early stage lung cancer. N Engl J Med 341:1198–1205, 1999
7. Earle CC, Venditti LN, Neumann PJ, et al: Who gets chemotherapy for metastatic lung cancer? Chest 117:1239–1246, 2000 8. Smith TJ, Penberthy L, Desch CE, et al: Differences in initial treatment patterns and outcomes of lung cancer in the elderly. Lung Cancer 13:235–252, 1995[CrossRef][Medline] 9. Hillner BE, McDonald MK, Desch CE, et al: A comparison of patterns of care of nonsmall cell lung carcinoma patients in a younger and Medigap commercially insured cohort. Cancer 83:1930–1937, 1998[CrossRef][Medline] 10. Fry WA, Phillips JL, Menck HR: Ten-year survey of lung cancer treatment and survival in the United States: A National Cancer Data Base report. Cancer 86:1867–1876, 1999[CrossRef][Medline] 11. Scott WJ, Howington J, Movsas B: Treatment of stage II non-small cell lung cancer. Chest 123:188S–201S, 2003 (suppl) 12. Smythe WR: Treatment of stage I non-small cell lung carcinoma. Chest 123:181S–187S, 2003 (suppl)
13. Ginsberg RJ, Rubinstein LV: Randomized trial of lobectomy versus limited resection for T1N0 non-small cell lung cancer. Ann Thorac Surg 60:615–623, 1995
14. Non-Small Cell Lung Cancer Collaborative Group: Chemotherapy in non-small cell lung cancer: A meta-analysis using updated data on individual patients from 52 randomized clinical trials. BMJ 311:899–909, 1995 15. Jett JR, Scott WJ, Rivera MP, et al: Guidelines on treatment of stage IIIB non-small cell lung cancer. Chest 123:221S–225S, 2003 (suppl) 16. Socinski MA, Morris DE, Masters GA, et al: Chemotherapeutic management of stage IV non-small cell lung cancer. Chest 123:226S–243S, 2003 (suppl) 17. Charlson ME, Pompei P, Alex KL, et al: A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 40:373–383, 1987[CrossRef][Medline] 18. Surveillance, Epidemiology, and End Results (SEER) Program: Comparative Staging Guide for Cancer, version 1.1. Bethesda, MD, National Cancer Institute, NIH publication 93–3640, 1993 19. Graubard BI, Korn EL: Predictive margins with survey data. Biometrics 55:652–659, 1999[CrossRef][Medline] 20. Shah BV, Folsom RE, Harrell FE, et al: Survey Data Analysis Software for Logistic Regression. Research Triangle Park, NC, Research Triangle Institute, 1984
21. Potosky AL, Harlan LC, Kaplan RS, et al: Age, sex, and racial differences in the use of standard adjuvant therapy for colorectal cancer. J Clin Oncol 20:1192–1202, 2002
22. Harlan LC, Clegg LX, Trimble EL: Trends in surgery and chemotherapy for women diagnosed with ovarian cancer in the United States. J Clin Oncol 21:3488–3494, 2003 Submitted February 6, 2004; accepted May 17, 2004.
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2004 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
