Division of Hematology/Oncology, Department of Medicine, Chao Family Comprehensive Cancer Center, Orange, California
Genetic Epidemiology Research Institute, School of Medicine, University of California-Irvine, Irvine, California
Department of Epidemiology, School of Medicine, University of California-Irvine, Irvine California
Division of Hematology/Oncology, Department of Medicine, Chao Family Comprehensive Cancer Center, University of California Irvine Medical Center, 101 The City Drive South, Building 56, Room 241, RT 81, Orange, CA 92868-3298
The collection of cancer incidence data used in this study under subcontract 050N-8707-S1527 with the Public Health Institute, State of California, was supported by the California Department of Health Services as part of the statewide cancer reporting program mandated by California Health and Safety Code Sections 103875 and 103885; the National Cancer Institute's Surveillance, Epidemiology, and End Results Program; the Centers for Disease Control and Prevention National Program of Cancer Registries.
The ideas and opinions expressed herein are those of the authors, and endorsement by the State of California, Department of Health Services, the National Cancer Institute, the Centers for Disease Control and Prevention, and/or the Genetic Epidemiology Research Institute of the University of California-Irvine is not intended nor should be inferred.
Presented in part at the 43rd annual meeting of the American Society of Clinical Oncology, Chicago, Illinois, June 1–6, 2007.
Platinum-based adjuvant chemotherapy in randomized trials has failed to provide a survival benefit in patients with resected stage I nonsmall cell lung cancer (NSCLC). Using data from the California Cancer Registry (CCR), we explored factors that had detrimental effects on survival in patients with stage I NSCLC to identify a subset of patients at high risk for disease recurrence and subsequent mortality.
Between 1989 and 2003, 19,702 incident cases of stage I NSCLC in the CCR were identified and subgrouped into stage IA and IB disease. Patient demographic factors, tumor characteristics, and treatment delivered were examined. Kaplan-Meier survival curves were calculated to estimate survival rates. Cox proportional-hazards ratios were used to identify independent prognostic factors for survival.
Advanced age at diagnosis, male sex, low socioeconomic status (SES), nonsurgical treatment, and poor histologic grade (stage IA NSCLC: hazards ratio [HR], 1.13; 95% confidence interval [95% CI], 1.08–1.19; stage IB NSCLC: HR, 1.11; 95% CI, 1.07–1.16) were associated with increased mortality risk on multivariate analysis. Nonupper lobe tumor location (right middle lobe, right and left lower lobes) and tumor size ≥4 cm (vs <4 cm: HR, 1.23; 95% CI, 1.15–1.30) were additional factors that increased the risk of mortality among patients with stage IB disease. Bronchioloalveolar carcinoma and Asian ethnicity were associated with decreased mortality risk in stage I NSCLC.
Lung cancer is the number 1 cause of death from cancer in the world for both men and women.1 The majority of lung cancers in the U.S. are diagnosed at an advanced stage and usually are not amenable to curative surgical resection. Only approximately 25% of patients are diagnosed with stage I or II disease for which surgical resection offers the possibility of a cure. Tobacco cessation,2 early detection in high-risk patients,3 and multimodality treatment for early-stage nonsmall cell lung cancer (NSCLC) have the potential to improve the survival of patients with lung cancer in a manner similar to what has been achieved for patients with breast cancer.4 Recent randomized trials in patients who received adjuvant chemotherapy for resected stage I through III NSCLC demonstrated improved overall survival (OS)5–8; however, to our knowledge, none of those trials reported a survival advantage for patients with stage I NSCLC in subgroup analyses except with the use of uracil-tegafur.8 A meta-analysis of recent adjuvant, cisplatin-based chemotherapy trials also did not report any survival benefit from the use of chemotherapy for patients with stage I NSCLC.9 The Cancer and Leukemia Group B (CALGB) 9633 study reported a 12% reduction in mortality at 4 years with adjuvant carboplatin plus paclitaxel in patients with stage IB NSCLC; however, after a longer follow-up, the improvement in survival no longer was significant.10 Subgroup analysis showed that patients with stage IB NSCLC who had tumors that measured ≥4 cm did derive a survival benefit from the use of adjuvant chemotherapy.10 Despite surgical resection, approximately 40% of patients with stage I NSCLC died within 5 years.11 Therefore, it is important to identify routine, readily available clinicopathologic features that carry an increased risk of mortality in patients with stage I NSCLC so that such patients may be considered for adjuvant treatment or may be included in future adjuvant chemotherapy trials. Therefore, we undertook this population-based analysis of the California Cancer Registry (CCR) database to identify clinicopathologic features that were associated with increased mortality risk in patients with stage I NSCLC.
MATERIALS AND METHODS
Study Cohort and Diagnostic Codes
Data were obtained regarding 101,844 incident NSCLC cases in the CCR from 1989 to 2003 with tumor-lymph node-metastases (TNM) staging data and complete follow-up data available. Data were abstracted from medical and laboratory records by trained tumor registrars according to Cancer Reporting in California Volume 1: Abstracting and Coding Procedures for Hospitals.12
Tumor site and histology were abstracted as described previously.13 Only patients with histologically or cytologically confirmed NSCLC were included in the analysis. Nonsmall cell histologies that were not coded as adenocarcinoma, squamous cell carcinoma, large cell carcinoma, bronchioloalveolar carcinoma (BAC), or as a metastatic lung lesion from a separate primary tumor were categorized as undifferentiated NSCLC, as described previously.14 Ethnicity, histologic grade, tumor lobar location, and tumor size were abstracted using Surveillance, Epidemiology, and End Results (SEER) codes.
Radiation and surgical techniques, including local treatment, wedge/segmentectomy, lobectomy, and pneumonectomy, were abstracted using SEER codes. Chemotherapy that was given during the first course of therapy was ascertained using CCR codes. The last date of follow-up was either the date of death or the last date the patient was contacted.
The measurement of socioeconomic status (SES) used in this analysis was a composite measure using CCR and census data. Standardized component scores for the SES index for the 20,919 census block groups were sorted and categorized into quintiles, with a value of 1 representing the lowest SES level and a value of 5 representing the highest SES level.15
The Student t test was used to compare continuous variables for 2-category comparisons. The Pearson chi-square test or the Fisher exact test was used to compare categorical and dichotomous variables. Analysis of variance with the Tukey post-hoc test was used for multiple comparisons of continuous variables. Life tables and Kaplan-Meier curves were generated for OS. Comparisons between groups were analyzed with the log-rank test. Multivariate Cox regression analyses were used to determine which factors were associated significantly with survival. Statistical significance was assumed for a 2-tailed P value <.05. All statistical analyses were performed using SAS software (version 9.1; SAS Institute, Cary, NC).
This research study involved the analysis of existing data from the CCR database with no patient intervention. No identifiers were linked to patients. Therefore, this study was approved by the University of California-Irvine Institutional Review Board (IRB) under the category “exempt” status (IRB no. 2004-3971).
Beginning in 1989, all patients with lung cancer in the CCR had complete follow-up records. From 1989 to 2003, in total, 19,702 patients (19.4%) with stage I NSCLC were identified. The mean age at diagnosis was 68 years for patients with stage IA disease and 69 years for patients with stage IB disease. The most common age range at diagnosis for patients with stage IA and IB NSCLC was between ages 70 years and 74 years. The median follow-up was 53 months, and 92.7% of patients were treated with surgery, chemotherapy, or radiation. The clinicopathologic characteristics of the patients with stage IA and IB NSCLC are presented in Table 1.
Table 1. Comparison of Patient and Clinicopathologic Characteristics of 19,702 Patients With Stage IA and IB Nonsmall Cell Lung Cancer
No. of patients (%)
Stage IA tumors
Stage IB tumors
NOS indicates not otherwise specified.
Period of diagnosis
Squamous cell carcinoma
Large cell carcinoma
Right upper lobe
Right middle lobe
Right lower lobe
Left upper lobe
Left lower lobe
Right and left main bronchus/carina/hilus
Size (T2 only), cm
Quintile 1 (low)
Quintile 5 (high)
Chemotherapy [of 19,527 patients]
Any treatment (chemotherapy or radiation or surgery)
Three centimeters was the most common tumor size recorded (10.1%) followed closely by 2.0 cm (9.8%), 2.5 cm (8.1%), 4 cm (7.0%), 3.5 cm (5.5%), and 5 cm (4.9%). The majority of the tumors were located in the 2 upper lobes (62%). Adenocarcinoma was the most common histology for both stage IA NSCLC(46.3%) and IB NSCLC (39.1%), whereas BAC accounted for 12.2% of all stage I NSCLC. Moderately differentiated histology was the most common histologic grade for stage IA NSCLC (40.9%), whereas poorly differentiated histology was the most common histologic grade for stage IB NSCLC (47.2%).
More patients with stage IA NSCLC (85.5%) than patients with stage IB NSCLC (79.2%) underwent surgical resection (Table 1). For patients who underwent surgical resection, more lobectomies were performed on patients with stage IB tumors (82%) than on patients with stage IA tumors (78.8%) tumors. Similarly, more pneumonectomies were performed on patients with stage IB tumors (8.5%) than on patients with stage IA tumors (2.0%) among those who underwent surgical resection. Conversely, more wedge/segmentectomies were performed on patients with stage IA tumors (18.7%) than on patients with stage IB tumors (9%) among those who underwent surgical resection.
Univariate Survival Analysis
In the group with stage IA tumors, the 5-year survival rate, the 10-year survival rate, and the median OS were 71.0%, 38.6%, and 96 months, respectively, for patients who had well-differentiated tumors; 61.3%, 32.3%, and 85 months, respectively, for patients who had moderately differentiated tumors; 53.4%, 28.9%, and 67 months, respectively, for patients who had poorly differentiated tumors; and 59.6%, 35.7%, and 83 months, respectively, for patients who had undifferentiated tumors. The difference in survival among the 4 histologic grades was statistically significant (P < .0001) (Fig. 1A). In the group with stage IB tumors, the 5-year survival rate, the 10-year survival rate, and the median OS were 57.6%, 34.2%, and 77 months, respectively, for patients who had well-differentiated tumors; 49.6%, 25.4%, and 60 months, respectively, for patients who had moderately differentiated tumors; 46.5%, 26.2%, and 53 months, respectively, for patients who had poorly differentiated tumors; and 45.2%, 24.1%, and 54 months, respectively, for patients who had undifferentiated tumors. Again, the difference in survival among the 4 histologic grades was statistically significant (P < .0001) (Fig. 1B).
For this analysis, the 5-year survival rate, the 10-year survival rate, and the median OS for patients who underwent surgical resection were stratified by lobar location of the tumor. For stage IA tumors, the 5-year, 10-year, and median OS were 61.4%, 35.9%, and 88 months, respectively, for patients who had tumors with a right upper lobe (RUL) location; 60.8%, 30.4%, and 80 months, respectively, for patients who had tumors with a left upper lobe (LUL) location; 53.1%, 30.5%, and 71 months, respectively, for patients who had tumors with a right middle lobe (RML) location; 59.5%, 33.5%, and 78 months, respectively, for patients who had tumors with a right lower lobe (RLL) location; and 59.4%, 32.8%, and 71 months, respectively, for patients who had tumors with a left lower lobe (LLL) location. The difference in survival among the 5 lobar locations for patients with stage IA tumors was not statistically significant (P = .10). In the group with stage IB tumors, the 5-year survival rate, the 10-year survival rate, and the median OS were 53.5%, 31.4%, and 68 months, respectively, for patients who had RUL tumors; 49.9%, 27.3%, and 60 months, respectively, for patients who had LUL tumors; 46.7%, 27.8%, and 57 months, respectively, for patients who had RML tumors; 46.5%, 24.8%, and 54 months, respectively, for patients who had RLL tumors; and 44.5%, 23%, and 52 months, respectively, for patients who had LLL tumors. The difference in survival among the 5 major lobar locations for patients with stage IB tumors was statistically significant (P < .0001).
We also stratified the lobar locations into upper (RUL+LUL) and nonupper (RML+RLL+LLL) lobar location. For stage IA NSCLC, the 5-year survival rate, the 10-year survival rate, and the median OS were 54.3%, 29.5%, and 70 months, respectively, for patients who had tumors located in the upper lobes; and they were 50.8%, 27.2%, and 62 months, respectively, for patients who had tumors located in the nonupper lobes. The difference in survival was statistically significant (P = .0072) (Fig. 2A). The corresponding survival values for patients with stage IB NSCLC were 43.3%, 24.4%, and 46 months, respectively, for patients who had upper lobar tumor locations and 36.1%, 19.0%, and 36 months, respectively, for patients who had nonupper lobar tumor locations. The difference in survival also was statistically significant (P < .0001) (Fig. 2B).
Among the patients with stage IB NSCLC, the 5-year survival rate, the 10-year survival rate, and the median OS were 48.8%, 26.6%, and 58 months, respectively, for patients who had tumors that measured <4 cm; and they were 35.5%, 19.6%, and 33 months, respectively, for patients who had tumors that measured ≥4 cm. The difference in survival was statistically significant (P < .0001) (Fig. 3).
Lobectomy Versus Wedge/Segmentectomy
The survival values for lobectomy and wedge/segmentectomy stratified by tumor lobar location for both stage IA and stage IB NSCLC are presented in Table 2. Patients who underwent lobectomy had uniformly better survival than patients who underwent wedge/segmentectomy for both stage IA and stage IB NSCLC and for tumors in all lobar locations, and these differences were statistically significant, except for RML and LLL locations for stage IB tumors. There was a 10.4% 5-year survivorship decrease and an 18-month decrease in median OS for patients who had stage IB LLL tumors and underwent lobectomy compared with patients who had stage IB RUL tumors and underwent lobectomy.
Table 2. Survival of Patients with Stage IA and IB Nonsmall Cell Lung Cancer Stratified by Tumor Lobar Location and Lobectomy Versus Wedge Resection/Segmentectomy
Increasing age at diagnosis, male sex, low SES, no surgical intervention, and less differentiated histologic grade (stage IA NSCLC: hazards ratio [HR], 1.13; 95% confidence interval [95% CI], 1.08–1.19; stage IB NSCLC: HR, 1.11; 95% CI, 1.07–1.16) were identified as independent factors that carried increased mortality risk for both stage IA and stage IB NSCLC (Table 3). Conversely, BAC and Asian ethnicity were identified as independent favorable prognostic factors for survival for both stage IA and stage IB NSCLC. For stage IB NSCLC, tumor size ≥4 cm (vs <4 cm: HR, 1.23; 95% CI, 1.15–1.30) and nonupper lobar location (RML+RLL+LLL) were identified as additional independent factors that carried increased mortality risk (Table 3). Analyzed separately, RUL tumor location (vs LLL location: HR, 0.85; 95% CI, 0.78–0.91) and LUL tumor location (vs LLL location: HR, 0.83; 95% CI, 0.77–0.90) were found to be associated independently with improved survival, whereas RML tumor location (P = .34) and RLL tumor location (P = .15) were not associated with improved survival compared with LLL tumor location in patients with stage IB NSCLC (Table 3).
Table 3. Cox Proportional Hazards Model for Stage IA and IB Nonsmall Cell Lung Cancer
In this report, we identified increased age at diagnosis, male sex, no surgical intervention, low SES, and poorly differentiated histology as independent factors that carry an increased risk of mortality in patients with stage I NSCLC. Nonupper lobar tumor location (RML+RLL+LLL) and tumor size ≥4 cm are 2 additional factors that independently carry an increased risk of mortality that is unique to patients with stage IB NSCLC.
It has been reported previously that poorly differentiated histologic grade is an independent prognostic factor for poor survival in patients with stage I NSCLC.16–18 In the current study, there was a 5-year survival decrement of 17.6% in patients with stage IA NSCLC and 11.1% in patients with stage IB NSCLC from well-differentiated tumors to poorly differentiated tumors. Thus, the less differentiated the tumor, the worse the survival, with a relative increase in the risk of death of 14% for patients with stage IA NSCLC and 12% for patients with stage IB NSCLC in our multivariate model. This finding has external validity and most likely has practical significance, because the histologic grade was interpreted by pathologists from throughout the state of California.
We also demonstrated that tumors located in the nonupper lobes (RML+RLL+LLL) independently carried an increased risk of mortality in patients with stage IB NSCLC compared with tumors located in the upper lobes (RUL+LUL). Ichinose et al. reported similar differences in survival according to lobar location in patients with resected stage IIIA NSCLC who had RUL>LUL>RML/RLL/LUL tumor locations in terms of survival.19 Iwasaki et al. also reported that patients with an LLL tumor location had significantly poor survival outcomes compared with patients who had tumors with other lobar locations in a multivariate analysis of patients with resected T2 NSCLC.17 Anatomically, tumors located in upper lobe are less likely to metastasize to subcranial lymph nodes,20 whereas LLL tumors usually involve paraesophageal (Level 8) and inferior pulmonary ligament lymph nodes (Level 9), which are hard to detect even on mediastinoscopy.20, 21 In this report, approximately 66% of patients with stage I NSCLC had tumors located in the upper lobes. Because there was no routine screening of lung cancer, it is likely that many more tumors located in the nonupper lobes had metastasized to regional lymph nodes and were no longer stage I at the time of diagnosis. Rocha et al. reported that a primary tumor located in a lower lobe, and not age, smoking history, weight loss, prior mediastinoscopy, tumor size, or histology, was the only statistically significant factor associated with upstaging after surgery.22 Although the majority of the patients with stage I tumors in our study underwent surgical resection, we could not ascertain how the mediastinal lymph nodes were sampled. In a retrospective subset analysis of the Eastern Cooperative Oncology Group (ECOG) 3590 study, which compared the role of adding chemotherapy to postoperative radiation in patients with resected, stage II through IIIA NSCLC, complete mediastinal lymphadenectomy identified significantly more N2 disease and was associated with improved survival in patients with right-sided NSCLC. Those authors reported that access to mediastinal lymph nodes, such as paraesophageal lymph nodes, may be limited from the left chest, thus, there was no difference in the survival of patients who had tumors located in the left lung. It is noteworthy that, even for patients who underwent complete mediastinal lymphadenectomy, the median survival of those who had an RUL tumor location (not reached) was better numerically than for those who had an RLL tumor location (50.7 months).23 Thus, better staging of the mediastinal lymph nodes, especially Level 8 and 9 lymph nodes, by endoscopic, ultrasound-guided fine-needle aspiration; integrated computed tomography-positron emission tomography fusion scan; video-assisted thoracoscopy24; or more complete mediastinal lymphadenectomy25–28 will provide better management for nonupper lobe tumors in patients with early-stage NSCLC.
In this report, for all lobar locations in both stage IA and stage IB NSCLC, patients who underwent lobectomy had uniformly better survival than patients who underwent wedge/segmenctomy (Table 2). Furthermore, in the multivariate analyses, lobectomy had the lowest HR among all surgical interventions for both stage IA and stage IB NSCLC (Table 3). The results from a randomized trial in patients with stage IA NSCLC have indicated that lobectomy results in lower tumor recurrence and cancer deaths compared with segmentectomy or wedge resection29 and is considered the standard curative surgical approach in early-stage NSCLC. However, patients who only undergo wedge/segmentectomy may have a poorer performance status, increased comorbidities, or a lack of expert surgical care. Some of these factors may have been accounted for by adjusting for SES in the multivariate analysis model. Nevertheless, the 5-year survivorship of patients with stage IA NSCLC who had an RUL tumor location and underwent lobectomy was only 64.7% in the current report and was much lower than the 94% long-term survivorship of patients with resected stage I NSCLC reported in the International Early Lung Cancer Action Project (I-ELCAP) study.3 This difference most likely is because of the very small size of the tumors detected in the I-ELCAP study. Unless and until early detection of lung cancer is adopted and practiced widely in the future, a large number of patients who currently have resected stage IA NSCLC will have tumors larger than those reported in I-ELCAP and still may benefit from additional adjuvant chemotherapy.
In the current study, patients who had stage IB NSCLC with tumors that measured ≥4 cm had an increased risk of mortality in both univariate and multivariate analyses. There was a 13.3% decrease in 5-year survivorship and a 25-month decrease in the median OS for patients who had stage IB disease with tumors that measured ≥4 cm compared with patients who had stage IB with tumors that measured <4 cm. This finding provides further evidence that patients who have stage IB tumors that measure ≥4 cm most likely will benefit from adjuvant chemotherapy10 and supports the inclusion of this criterion in the new intergroup trial (ECOG 1505) assessing the role of adding antiangiogenesis agents to adjuvant chemotherapy.30 Although the biologic basis for a 4-cm cut-off size is unknown, it has been demonstrated that, as the size of the primary NSCLC tumor increases, the curability of NSCLC decreases.31
In this report, BAC and Asian ethnicity independently carried a decreased risk of mortality in both stage IA and IB NSCLC. It was demonstrated previously that BAC conferred better survival in patients with advanced NSCLC,13, 14 whereas Asians had a survival similar to that of Caucasian with advanced NSCLC.13 Currently, we are investigating whether this survival advantage for Asians extends to stage II and III NSCLC. Potential explanations include a greater prevalence of activating epidermal growth factor receptor (EGFR) mutations in Asians,32 which generally confer a better prognosis, or the greater prevalence of intra-allelic EGFR polymorphisms in Asians,33 which may confer a survival advantage in patients with NSCLC.34
This report is retrospective in nature and shares many limitations of other population-based studies, including the lack of central pathology review. We could not ascertain the timeline or sequence of the treatment modality rendered. Thus, the prognostic significance of bimodality or trimodality treatment could not be ascertained accurately, although the prognostic significance of each treatment modality was analyzed separately in the Cox proportional-hazards model. We do not have available information regarding how many patients underwent staging by mediastinoscopy. The CCR does not routinely record smoking history or quantify the amount of tobacco usage. Current smoking status and the amount oftobacco use at the time of lung cancer diagnosis, even in patients with stage I disease, were poor prognostic factors in a multivariate analysis model.16
Currently, adjuvant chemotherapy is not recommended for patients with resected stage I NSCLC because of the lack of a clear survival benefit from randomized trials, even though close to 40% of these patients died within 5 years of resection.11 Efforts to identify patients with stage I NSCLC who have a high risk of recurrence and death are currently being performed, including microarray-based molecular profiling studies.35, 36 CALGB is planning an adjuvant trial in stage IA NSCLC by stratifying patients into low- and high-risk categories based on molecular tumor profiling.35 Another current, large-scale, randomized adjuvant trial (RADIANT) is comparing erlotinib, an oral EGFR inhibitor, with placebo in patients with resected stage IB to IIIA NSCLC who had tumors that expressed EGFR.
In this report, we identified routine, readily available clinicopathologic features, poorly differentiated histology, tumor ≥4 cm, and nonupper lobe location, in patients with stage IB NSCLC that independently carry an increased risk of mortality. These factors—especially poorly differentiated histology16—should be considered as inclusion criteria for ongoing and future adjuvant trials. Furthermore, subgroup analysis of these adverse clinicopathologic features in meta-analyses of adjuvant NSCLC trials may help identify high-risk patients with early-stage NSCLC who may be considered for and benefit from adjuvant chemotherapy.