|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2005.04.6110 on June 12 2006 © 2006 American Society of Clinical Oncology. Postoperative Radiotherapy for Stage II or III Non–Small-Cell Lung Cancer Using the Surveillance, Epidemiology, and End Results Database
From the Department of Radiation Oncology, Wake Forest University School of Medicine, Winston-Salem, NC; Division of Biostatistics, Yale University School of Public Health; Departments of Therapeutic Radiology and Surgery, Yale University School of Medicine, New Haven, CT; and the Department of Radiation Oncology, University of Medicine and Dentistry of New Jersey–Robert Wood Johnson Medical School, Piscataway, NJ Address reprint requests to Brian E. Lally, MD, Wake Forest University School of Medicine, Department of Radiation Oncology, Medical Center Blvd, Winston-Salem, NC 27157-1030; e-mail: blally{at}wfubmc.edu
PURPOSE: To investigate the association between survival and postoperative radiotherapy (PORT) in patients with resected non–small-cell lung cancer (NSCLC). PATIENTS AND METHODS: Within the Surveillance, Epidemiology, and End Results database, we selected patients with stage II or III NSCLC who underwent a lobectomy or pneumonectomy. Only those patients coded as receiving PORT or observation were included. To account for perioperative mortality, we excluded patients who survived less than 4 months. As a result of our inclusion criteria, we selected a total of 7,465 patients, with a median follow-up time of 3.5 years for patients still alive. RESULTS: Predictors for the use of PORT included age less than 50 years, higher American Joint Committee on Cancer stage, T3-4 tumor stage, larger tumor size, advanced node stage, greater number of lymph nodes involved, and a ratio of lymph nodes involved to lymph nodes sampled approaching 1.00. On multivariate analysis, older age, T3-4 tumor stage, N2 node stage, male sex, fewer sampled lymph nodes, and greater number of involved lymph nodes had a negative impact on survival. The use of PORT did not have a significant impact on survival. However, in subset analysis for patients with N2 nodal disease (hazard ratio [HR] = 0.855; 95% CI, 0.762 to 0.959; P = .0077), PORT was associated with a significant increase in survival. For patients with N0 (HR = 1.176; 95% CI, 1.005 to 1.376; P = .0435) and N1 (HR = 1.097; 95% CI, 1.015 to 1.186; P = .0196) nodal disease, PORT was associated with a significant decrease in survival. CONCLUSION: In a population-based cohort, PORT use is associated with an increase in survival in patients with N2 nodal disease but not in patients with N1 and N0 nodal disease.
For patients diagnosed with early-stage non–small-cell lung cancer (NSCLC), surgical resection remains the mainstay of therapy. The possibility of patients having undetected microscopic disease within hilar or mediastinal lymph nodes is significant, with increasing risk correlated with stage. As a result, some individuals have advocated postoperative radiotherapy (PORT) as a means to sterilize regions with increased risk of harboring microscopic disease and, thus, improve outcomes. This hypothesis was tested in a trial preformed by the Lung Cancer Study Group that included patients with either stage II or stage III squamous cell carcinoma of the lung.1 After surgical resection, patients were randomly assigned to either observation or 50 Gy of radiation therapy to the bronchial stump, ipsilateral hilum, and vascular shadows within the mediastinum bilaterally. Although radiotherapy did not improve overall survival, it did lead to a reduction in local recurrences that reached statistical significance for patients with N2 disease. In 1998, the PORT Meta-Analysis Trialists Group report was published and has been quoted extensively.2 This study analyzed individual patient data from prospective trials of patients with resected NSCLC. Nine trials (involving 2,128 patients) were included in the analysis. Some trials had been initiated as early as 1965, and three had not been published in the peer-reviewed literature. The results strongly suggested that PORT had a detrimental effect on survival, presumably through the increased incidence of intercurrent death. Two-year overall survival rate was 48% in patients receiving PORT compared with 55% for patients treated with surgery alone. The detriment of PORT was inversely related to nodal status, with significantly reduced survival noted for N0 and N1 disease. The results for stage III and N2 patients favored PORT, although the difference was not significant. An update of the PORT meta-analysis was published recently with a 10th randomized trial and an additional 104 patients. The results did not differ greatly.3 One trial not included in the meta-analysis randomly assigned 155 patients to either surgery alone or surgery followed by PORT.4 With a median follow-up time of 3.6 years, there was no statistically significant difference in survival. Subgroup analysis conducted separately for each pathologic stage did not result in any significant survival benefit for PORT when compared with surgery alone. In another trial that was included in the meta-analysis, Dautzenberg et al5 randomly assigned 728 patients with stage I to III NSCLC after surgery to either observation or PORT. The trial showed a dramatically significantly worse survival with PORT for patients with stage II disease, but it did not show significant detriment in survival for patients receiving PORT who had stage I or III disease. Notably, the higher mortality seen was attributed to an excess of intercurrent deaths (ie, cardiac, infectious, or respiratory causes; neurovascular disease; new primary malignancy; or other causes, but not recurrence of the primary lung cancer). The 5-year intercurrent death rate was 8% for the control group compared with 31% for the PORT group; there was no significant difference in the number of cancer-related deaths between the two arms. However, this is complicated by the fact that the majority of patients were treated with cobalt-60 equipment and inferior treatment delivery techniques compared with currently available methods. Although it has yet to be proven in randomized trials, many clinicians suspect that PORT, delivered via modern treatment methods, may impact favorably on survival through decreased local failure in the group of patients with N2 disease. In the absence of randomized data, we sought to answer this question in a retrospective review of high-quality, population-based data to provide insight into the relationship between survival and PORT during a time period in which the linear accelerator dominated clinical treatment. We evaluated the relationship between PORT and survival and how various characteristics may have impacted this end point.
The Surveillance, Epidemiology, and End Results (SEER) database is a national cancer surveillance program that collects information on all incident cancer cases from 13 cancer registries covering approximately 26% of the United States population and is representative of national demographics. The specific histologic types selected were those coded as large-cell carcinoma (large-cell carcinoma, large-cell neuroendocrine carcinoma, and giant cell carcinoma), squamous cell carcinoma (papillary squamous cell carcinoma, squamous cell carcinoma, squamous cell carcinoma [keratinizing], squamous cell carcinoma [large cell, nonkeratinizing], and squamous cell carcinoma [small cell, nonkeratinizing]), adenocarcinoma (adenocarcinoma and acinar cell carcinoma), bronchioalveolar adenocarcinoma (bronchioloalveolar adenocarcinoma, alveolar adenocarcinoma, bronchioloalveolar carcinoma [nonmucinous], and bronchioloalveolar carcinoma [mucinous]), and adenocarcinoma with mixed subtypes (adenocarcinoma with mixed subtypes, papillary adenocarcinoma, clear cell adenocarcinoma, mucinous adenocarcinoma, mucin-producing adenocarcinoma, signet ring cell carcinoma, and adenosquamous carcinoma). We identified (within SEER) adult patients, who were age 21 years and older, with pathologically confirmed NSCLC between 1988 and 2002 (n = 173,429). We then excluded patients with stage I disease, metastatic disease, or unknown stage (n = 14,084). Patients were only included if they underwent a surgical procedure of either a lobectomy or pneumonectomy. Use of radiotherapy within the initial treatment course, typically defined as treatments received within 6 months of initial diagnosis, was abstracted by local tumor registries and reported to SEER. Only those patients coded as receiving postoperative external-beam irradiation or no irradiation were included in this analysis. All other patients, including patients who received radiotherapy such as preoperative radiotherapy or brachytherapy, were excluded. In addition, we also excluded patients with N3 nodal status or undefined nodal status. We required that patients have complete information regarding tumor size, tumor location, extent of disease/lymph node involvement, and cause of death if applicable. Patients who did not meet these criteria were excluded. We excluded 80 patients because they died and did not have a cause of death listed, and this did not impact the results significantly. In an effort to account for perioperative mortality, we excluded patients who died within 4 months of diagnosis. Figure 1 details the selection process for inclusion of patients. As a result of the inclusion criteria, we selected a total of 7,465 patients in our cohort for analysis. The time of last follow-up was November 2004, and the median follow-up time for all patients still alive was 3.5 years.
Categoric variables included histology, patient age at diagnosis (< 50, 50 to 69, or  70 years), sex, SEER reporting site, race, year of diagnosis, laterality (right v left), location (upper lobe, middle lobe, lower lobe, or main bronchus), American Joint Committee on Cancer stage (stage II, IIIa, or IIIb), tumor stage (T1-2 v T3-4), nodal stage (N0, N1, or N2), surgical procedure (pneumonectomy v lobectomy), the number of lymph nodes involved, the number of lymph nodes sampled, the ratio of lymph nodes involved to lymph nodes sampled, and the use of PORT. Patient age in years, the number of involved lymph nodes, and the number of lymph nodes sampled were investigated as categoric variables during univariate analysis but as continuous variables on multivariate analysis. Tumor size in millimeters was investigated as a categoric variable on univariate analysis, and the natural log of size was investigated as a continuous variable during multivariate analysis. The use of the log scale was done to reduce the influence of a few large values. Information on margin status, performance status, use of adjuvant chemotherapy, and specific radiotherapy technique (dose, fractionation, beam energy, and so on) was not available with the SEER database and is not included in the analysis here. Overall survival was the primary study end point.
The Pearson Multivariable Cox proportional hazards models were used to calculate adjusted hazard ratios (HRs) and their 95% CIs relating to the variables as described. Four models were constructed to assess the importance of external-beam radiotherapy as an independent predictor of survival after adjusting for other factors with the exception of patient race and the laterality of the tumor. The first model included all patients. Three subset models were subsequently performed to investigate the variables of interest only in cohorts of patients for each nodal stage. In these subset analyses, the SEER site information was not included. In patients with N0 disease, the number of involved lymph nodes and the ratio of involved lymph nodes to sampled lymph nodes were not included. All data were analyzed using the SAS Version 8.02 (SAS Institute, Cary, NC) statistical software package. Results were considered to be statistically significant when P < .05. This study was approved by the Yale Human Investigation Committee and is in full compliance with federal, state, and institutional regulations and guidelines.
As a result of selection criteria, the analysis included a total of 7,465 patients. Forty-seven percent received PORT. The median patient age at diagnosis for the patients within the cohort was 66 years (range, 24 to 95 years). The median number of lymph nodes involved out of lymph nodes sampled for the N0, N1, and N2 nodal stages was zero of seven, one of seven, and three of nine nodes, respectively. Examination of the categoric variables for sex, tumor laterality, and tumor location failed to demonstrate a statistical correlation with PORT use. Strong predictors for PORT use (P < .0001) included age younger than 50 years, higher American Joint Committee on Cancer stage, T3-4 tumor stage, larger tumor size, advanced nodal stage, greater number of lymph nodes involved, and a ratio of lymph nodes involved to lymph nodes sampled approaching 1.00. Trends among the 13 reporting SEER sites and the five histologies varied in both the number of patients reported and the frequency of PORT use. Before 1997, PORT was used in 54% to 60% of patients per year. However, since 1998, less than one third of the patients received PORT. Complete analyses of predictors of PORT use are shown in Table 1.
On univariate analysis for all patients, PORT was associated with significantly worse survival. Other adverse factors for prognosis were older age, male sex, larger tumor size, stage III disease, N2 disease, greater number of lymph nodes involved, fewer sampled lymph nodes, greater percentage of involved lymph nodes, and pneumonectomy. Race and laterality of tumor were not prognostic. There was improved prognosis with increasing year of diagnosis, with the critical year being 1998. The histologic types examined were significant for bronchioalveolar histology having a more indolent course. The results of all the univariate analyses are listed in Table 2. Kaplan-Meier plots for all patients and the subsets of N0, N1, and N2 patients are presented in Figures 2, 3, 4, and 5, respectively.
On multivariate analysis of all patients, older age (HR = 1.025; 95% CI, 1.022 to 1.028; P < .0001), T3-4 disease (HR = 1.288; 95% CI, 1.117 to 1.484; P = .0005), N2 nodal disease (HR = 1.281; 95% CI, 1.101 to 1.490; P = .0014), and greater number of involved lymph nodes (HR = 1.043; 95% CI, 1.027 to 1.060; P < .0001) had a negative impact on survival. Female sex (HR = 0.877; 95% CI, 0.827 to 0.930; P < .0001) and greater number of lymph nodes sampled (HR = 0.989; 95% CI, 0.984 to 0.994; P < .0001) were associated with a statistically improved survival. The use of PORT (HR = 1.048; 95% CI, 0.987 to 1.113; P = .1269) did not have a significant impact on survival. Subset analysis was then performed for patients based on N0, N1, or N2 stage. When evaluating all patients and the individual subsets, it is notable that increasing age, increasing log of the primary tumor size, and increasing percentage of involved lymph nodes were consistently associated with worse prognosis. However, PORT was associated with a significant decrease in survival for patients with N0 (HR = 1.1176; 95% CI, 1.005 to 1.376; P = .0435) and N1 nodal disease (HR = 1.097; 95% CI, 1.015 to 1.186; P = .0196) and a significant increase in survival for patients with N2 nodal disease (HR = 0.855; 95% CI, 0.762 to 0.959; P = .0077). Interactions were tested within the N2 subset, and none were detected that reached significance with application of Bonferroni corrections. The results of all multivariate analyses are listed in Table 3.
Five-year univariate estimates and multivariate analysis of overall survival and disease-specific survival were examined within each nodal group when stratified by PORT use. These results are listed in Table 4.
Through a large population-based cohort, we investigated the following clinical question. Is the use of PORT associated with improved survival in patients with resected stage II or III NSCLC? Our results suggest that the use of PORT significantly improved survival for patients with N2 nodal disease. However, our results also reveal that PORT is associated with a decrease in survival for patients with N0 and N1 nodal disease. The SEER data are retrospectively collected, and thus, the potential for error or bias may exist. SEER places heavy responsibility on individual pathologists and on professional pathology organizations to ensure that the diagnostic information provided to SEER is accurate and based on meaningful classification. In return, SEER serves as a mechanism of quality assurance for patterns of practice.6 A direct comparison between the SEER registry and independently reviewed lung cancer histologic types found the results obtained from the registry were felt to be reliable.7 We recognize that confounding factors, such as margin status and performance status, may have influenced the treating physician’s decision to prescribe PORT. This information is not available for analysis. Details demonstrating the nature and extent of lymph dissection, although prognostic in other reports,8,9 were limited to the number of sampled and involved nodes and the greatest involved nodal stage. However, given the large cohort of almost 7,500 patients, we hope that the impact of these variables was minimized. No information on the radiotherapy technique (total dose, fraction size, beam energy, length of treatment time, intensity modulated radiation therapy use, and respiratory gating use) was available either. If intercurrent death was increased secondary to the use of PORT, we would have expected a change in disease-specific survival when compared with overall survival. In the cohort that we have examined, most patients died secondary to lung cancer. A previous examination of the SEER database (1988 to 1995), including more than 4,000 patients, found that PORT was associated with improved 5-year overall survival compared with no PORT (22% v 16%, respectively).10 In this analysis, the author included many patients whose extent of disease was not specified and, thus, coded as not otherwise specified. By using our rigorous inclusion and exclusion criteria, we hoped to properly address the question that faces the clinician. The SEER data, like any retrospective cohort, can be susceptible to some forms of systemic bias. We also performed intensive statistical analysis and investigated the possibility of interaction between variables. Thus, we have taken the appropriate steps to ensure that our results and conclusions are accurate. The PORT meta-analysis is a collection of several prospective randomized trials that, when analyzed together, demonstrate that PORT can be detrimental. Many in the radiotherapy community are quick to identify the weaknesses in this report.11-13 The initial staging evaluation used in many of the trials would be considered suboptimal. The details of the surgical resection and nodal staging were often not clearly defined. The total dose and fractionation schedules in some of the trials would be considered nonstandard by current clinical practice. Still, in an era when it is believed that the linear accelerator was the primary instrument for delivery for clinical care, our findings are similar to the PORT meta-analysis for N0 and N1 disease. More importantly, our results differed from the meta-analysis because we suggest a survival benefit for patients with N2 nodal disease. As shown in Figure 6, a dramatic shift in the patterns of practice have occurred, with PORT use decreasing by approximately 50%. We also found that patients diagnosed more recently had improved survival. It is impossible to say whether this change in practice patterns was related to the publication of the PORT meta-analysis and accompanying media coverage. The trend towards improved survival may be secondary to better patient selection as a result of the publication of the PORT meta-analysis. Still, in the N2 subgroup, we found that the use of PORT was associated with a survival benefit, and no interaction between the variables investigated was found in this subset.
Examination of the Kaplan-Meier curves showed differences that are notable. Within the N2 nodal disease subgroup, the survival curves for patients treated with PORT and with observation did not diverge until approximately 2.5 years had elapsed. This suggests that a late effect, such as sterilizing residual microscopic disease, may be the mechanism by which PORT is increasing survival. For patients with N0 and N1 nodal disease, the survival curves demonstrated divergence between the arms before 1 year. In previous reports, such as the PORT meta-analysis, it was believed that PORT was detrimental because it increased the rate of intercurrent deaths, especially when older technology was used. However, modern linear accelerators can still induce pulmonary morbidity that can occur as both an early (radiation pneumonitis) and late (pulmonary fibrosis) effect. These findings raise the following question. Why is PORT detrimental for early-stage NSCLC? To answer this question, we need to examine the therapeutic ratio (ie, the ratio of the benefit that can be obtained by using PORT over the risk that is incurred in terms of normal tissue complications that PORT will induce). We speculate that the survival benefit identified for N2 nodal disease may have been because the PORT-induced morbidity was outweighed by the presence of residual disease. Because disease recurrence often leads to death, treatment of the microscopic disease likely improved outcomes. In patients with N0 and N1 disease, the potential for having undetected microscopic/residual disease is less, and the benefit gained by treatment with PORT is diminished. Also, a possibility may be that physicians prescribed PORT for N0-1 patients secondary to the presence of a positive margin. If true, PORT would instead be a surrogate for a more aggressive phenotype. We cannot deny that our analysis has some limitations and may be influenced by bias. However, we feel we have taken the appropriate steps regarding statistical analysis in an attempt to best control for the potential impact of any bias. We do acknowledge that the lack of chemotherapy data in SEER is a potential source of bias. Three prospective trials have been published recently demonstrating an advantage with the use of postoperative adjuvant chemotherapy using a variety of regimens.14-16 As a result, for patients with stage IB and higher disease, adjuvant chemotherapy is now considered the standard of care. In consideration of confounding of the data relating to benefits of systemic therapy, the documented advantages of postoperative chemotherapy are a relatively recent finding, and most patients in the series were not treated in this era. More follow-up and analysis are needed to determine how the use of chemotherapy impacts the patterns of failure. If chemotherapy is able to control metastatic disease such that failures at the primary tumor site become the primary location for disease recurrence, PORT would be beneficial. The published literature demonstrating that PORT delivered by modern techniques improves survival in NSCLC patients is limited to two moderately large single-institution experiences (224 patients from a Mayo Clinic study17 and 200 patients from a University of Pennsylvania study18). These studies were not randomized but involved retrospective comparisons. Both showed favorable overall survival and local control, whereas the risk of death from intercurrent disease for patients receiving less than 54 Gy of radiation was only 2%. All patients in these two studies underwent simulation and were treated on a linear accelerator using energies of 6 MV or higher. A large phase III randomized trial headed by the Eastern Cooperative Oncology Group randomly assigned patients with resected stages II and III NSCLC to either PORT or PORT plus etoposide and cisplatin.19 No difference in survival or recurrence was identified. The Radiation Therapy Oncology Group recently reported a phase II trial investigating paclitaxel and carboplatin in combination with PORT for patients with resected stage II or IIIa NSCLC.20 Using the Eastern Cooperative Oncology Group trial as a historical control, the Radiation Therapy Oncology Group felt that the paclitaxel/carboplatin doublet improved survival. However, combination therapy of radiotherapy and chemotherapy can also increase toxicity. Thus, future attempts at using such combination therapy in patients who are not found to have N2 stage disease should be considered carefully before being attempted. In summary, analysis of the SEER database shows that PORT offers a significant survival benefit for patients with N2 nodal disease. There was also a detrimental effect when PORT was administered to patients with N0 or N1 nodal disease. The decrease in survival may be related to PORT, other confounding factors, or both.
The authors indicated no potential conflicts of interest.
Supported by National Institutes of Health Grant No. 1T32CA113267-01, TRADONC Fellowship (B.E.L.). Presented at the 47th Annual Meeting of the American Society for Therapeutic Radiology and Oncology, Denver, CO, October 16-20, 2005. Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. The Lung Cancer Study Group: Effects of postoperative mediastinal radiation on completely resected stage II and stage III epidermoid cancer of the lung. N Engl J Med 315:1377-1381, 1986[Abstract] 2. PORT Meta-analysis Trialists Group: Postoperative radiotherapy in non-small-cell lung cancer: Systematic review and meta-analysis of individual patient data from nine randomised controlled trials. Lancet 352:257-263, 1998[CrossRef][Medline] 3. Burdett S, Stewart L: Postoperative radiotherapy in non-small-cell lung cancer: Update of an individual patient data meta-analysis. Lung Cancer 47:81-83, 2005[CrossRef][Medline] 4. Mayer R, Smolle-Juettner FM, Szolar D, et al: Postoperative radiotherapy in radically resected non-small cell lung cancer. Chest 112:954-959, 1997 5. Dautzenberg B, Arriagada R, Chammard AB, et al: A controlled study of postoperative radiotherapy for patients with completely resected nonsmall cell lung carcinoma: Groupe d’Etude et de Traitement des Cancers Bronchiques. Cancer 86:265-273, 1999[CrossRef][Medline] 6. Henson DE, Bores-Saavedra J: Checking up on the surveillance, epidemiology, and end results program. J Natl Cancer Inst 96:1050-1051, 2004 7. Field RW, Smith BJ, Platz CE, et al: Lung cancer histologic type in the surveillance, epidemiology, and end results registry versus independent review. J Natl Cancer Inst 96:1105-1107, 2004 8. Yano T, Yokoyama H, Inoue T, et al: Surgical results and prognostic factors of pathologic N1 disease in non-small-cell carcinoma of the lung: Significance of N1 level—Lobar or hilar nodes. J Thorac Cardiovasc Surg 107:1398-1402, 1994 9. Okada M, Tsubota N, Yoshimura M, et al: Prognosis of completely resected pN2 non-small cell lung carcinomas: What is the significant node that affects survival? J Thorac Cardiovasc Surg 118:270-275, 1999 10. Rescigno J: Use of postoperative radiotherapy for node-positive non-small-cell lung cancer. Clin Lung Cancer 4:35-44, 2002[Medline] 11. Bogart JA, Aronowitz JN: Localized non-small cell lung cancer: Adjuvant radiotherapy in the era of effective systemic therapy. Clin Cancer Res 11:5004s-5110s, 2005 12. Bonner JA: The role of postoperative radiotherapy for patients with completely resected nonsmall cell lung carcinoma: Seeking to optimize local control and survival while minimizing toxicity. Cancer 86:195-196, 1999[CrossRef][Medline] 13. Munro AJ: What now for postoperative radiotherapy for lung cancer? Lancet 352:250-251, 1998[CrossRef][Medline] 14. Arriagada R, Bergman B, Dunant A, et al: Cisplatin-based adjuvant chemotherapy in patients with completely resected non-small-cell lung cancer. N Engl J Med 350:351-360, 2004 15. Strauss GM, Herndon J, Maddaus MA, et al: Randomized clinical trial of adjuvant chemotherapy with paclitaxel and carboplatin following resection in stage IB non-small cell lung cancer (NSCLC). J Clin Oncol 22:621s, 2004 (suppl; abstr 7019) 16. Winton T, Livingston R, Johnson D, et al: Vinorelbine plus cisplatin vs. observation in resected non-small-cell lung cancer. N Engl J Med 352:2589-2597, 2005 17. Sawyer TE, Bonner JA, Gould PM, et al: Effectiveness of postoperative irradiation in stage IIIA non-small cell lung cancer according to regression tree analyses of recurrence risks. Ann Thorac Surg 64:1402-1407, 1997 18. Machtay M, Lee JH, Shrager JB, et al: Risk of death from intercurrent disease is not excessively increased by modern postoperative radiotherapy for high-risk resected non-small-cell lung carcinoma. J Clin Oncol 19:3912-3917, 2001 19. Keller SM, Adak S, Wagner H, et al: A randomized trial of postoperative adjuvant therapy in patients with completely resected stage II or IIIA non-small-cell lung cancer: Eastern Cooperative Oncology Group. N Engl J Med 343:1217-1222, 2000 20. Bradley JD, Paulus R, Graham MV, et al: Phase II trial of postoperative adjuvant paclitaxel/carboplatin and thoracic radiotherapy in resected stage II and IIIA non-small-cell lung cancer: Promising long-term results of the Radiation Therapy Oncology Group—RTOG 9705. J Clin Oncol 23:3480-3487, 2005 Submitted October 13, 2005; accepted January 24, 2006. Related Editorial
Related Correspondence
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
2 test was used to determine unadjusted associations between PORT use and categoric variables of interest. Overall survival was defined as the time from diagnosis until death. Disease-specific survival was defined as the time from diagnosis until death if secondary to lung cancer. Nonparametric estimates of the survivor function were calculated using the Kaplan-Meier method. The survival estimates were computed within defined strata levels, and the rank tests were pooled over the strata and are, therefore, adjusted for strata differences. The log-rank test was used to assess the statistical significance of the differences between the survival curves. 
, 
, and • represent N2, N1, and N0 nodal stages, respectively.
