| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2003;76:1023-1028
© 2003 The Society of Thoracic Surgeons
a Department of Cardiothoracic Surgery, Northern General Hospital, Sheffield, United Kingdom
b Department of Cardiothoracic Surgery, City Hospital, Nottingham, United Kingdom
Accepted for publication April 18, 2003.
* Address reprint requests to Dr Alexiou, Department of Cardiothoracic Surgery, City Hospital, Nottingham NG5 1PB, UK.
e-mail: alexiou486{at}aol.com
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and methodsResultsCommentReferences |
|---|
METHODS: Four hundred eighty-five patients with stage I nonsmall cell lung cancer undergoing lung resection between 1991 and 2000 were studied. Three hundred seventy-four patients underwent a smaller resection than a pneumonectomy and 111 had a pneumonectomy.
RESULTS: Patients undergoing less extensive resections were older (mean age, 65 vs 63 years) (p = 0.01); these patients were also more likely to have a history of chronic obstructive airway disease (9% vs 2%) (p = 0.01) or asthma (10% vs 3%) (p = 0.04), nonsquamous cell type (56% vs 27%) (p < 0.0001), and T1 tumor stage (66% vs 17%) (p = 0.002) than patients having a pneumonectomy. Operative mortality was 2.4% versus 8% (p = 0.01). Overall 1-, 3-, and 5-year Kaplan-Meier survival rates (95% confidence interval [CI]) after less extensive resections were 85% (CI, 82% to 90%), 63% (CI, 56% to 69%), and 50% (CI, 42% to 57%), respectively, and after pneumonectomy the survival rates were 66% (CI, 53% to 73%), 47% (CI, 35% to 57%), and 44% (CI, 32% to 55%), respectively (p = 0.0006). When the Cox proportional hazards model was applied to all study patients (n = 485), pneumonectomy (p = 0.001), T2 stage (p = 0.006), older age (p = 0.03), and male gender (p = 0.03) were independent adverse predictors of survival. When the analysis was limited to the patients having T1N0 disease (n = 145), pneumonectomy (p = 0.0008), older age (p = 0.05), and nonsquamous cell type (p = 0.02) were independent adverse determinants of survival. When only the patients with T2N0 disease were analyzed (n = 340), male gender (p = 0.0005) and pneumonectomy (p = 0.01) were independent negative predictors of survival.
CONCLUSIONS: In this study, the patients who underwent pneumonectomy for stage T1N0 or T2N0 nonsmall cell lung cancer had a significantly poorer survival than those patients who underwent smaller lung resections.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentReferences |
|---|
The outlook after lung resection depends on the tumor stage, but the role of other factors, such as the age and gender of the patient, the histologic cell type, and the extent of the resection is less clear.
Stage I NSCLC is preferably treated with lobectomy and has a relatively favorable prognosis. Lesser lung resections have a higher local recurrence rate than lobectomy and are reserved for patients with impaired cardiorespiratory status [1]. Pneumonectomy carries higher operative risk than smaller resections and well-documented adverse physiologic effects, but because of anatomical and technical considerations it may be unavoidable [2–4].
The evidence regarding the effect of pneumonectomy on the overall survival is conflicting. A poorer survival after pneumonectomy has been recently attributed to a host of confounding variables, including a higher tumor stage [3]. Other reports have suggested an independent negative effect of pneumonectomy undertaken in early disease stages [5, 6].
The purpose of this study was to examine whether performing pneumonectomy in patients having pathologic stage I (T1N0 and T2N0) NSCLC impacts on survival.
|
Patients and methods |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentReferences |
|---|
These patients (n = 485) were divided in two groups according to the extent of their lung resections. Group A consisted of 374 patients who underwent a resection smaller than a pneumonectomy (24 bi-lobectomies, 272 lobectomies, 11 sleeve resections, 4 segmentectomies, and 63 wedge resections), and group B consisted of 111 patients who had a pneumonectomy.
The two groups were compared with regard to the demographic, clinical and histopathologic features, operative mortality, and overall survival.
Preoperative evaluation
Preoperative evaluations included physical examinations, hematological and biochemical investigations, chest roentgenogram, electrocardiogram, computerized tomography of the chest and abdomen, and bronchoscopy. In addition, investigations such as liver ultrasound, bone scan, head computed tomography and so forth were performed if required on the basis of clinical findings and laboratory measurements (eg, abnormal liver enzymes or serum calcium, skeletal symptoms, hepatomegaly, splenomegaly, lymphadenopathy, abnormal neurologic examination and so forth), or on the basis of clinic findings or laboratory measurements.
Cervical mediastinoscopy and anterior parasternal mediastinotomy are important diagnostic and staging procedures in patients having mediastinal lymph nodes greater than 1 cm on a computed tomographic scan, but not all surgeons used them routinely during the study period. These procedures are currently being used only when indicated.
All patients had spirometry and arterial blood gasses. In the presence of borderline predicted postoperative lung volumes, a ventilation perfusion isotopic scan was also performed. More recently, diffusing lung capacity studies also have been used. Exercise tests were carried out if a patient was older than 70 years of age, had lung volumes less than 60% of predicted value for age and height, or had a previous history of cardiovascular disease, and ischemic changes on the electrocardiogram (Q waves, ST segment depression, left ventricular hypertrophy, complete right bundle branch block, premature ventricular contractions and so forth). Patients exhibiting an abnormal exercise test were referred to a cardiologist for further evaluation.
Patients were considered for lung resection if there was no evidence of mediastinal involvement by the tumor or distant metastatic disease and they were deemed as having adequate cardiac reserve and a predicted postoperative forced expiratory volume in 1second of at least 1 L, as assessed by preoperative spirometry and ventilation-perfusion isotopic scan, or as assessed by both.
Operative approach
At operation the aim was to achieve complete clearance of all visible tumor mass with the least possible loss of healthy lung tissues. The hilar lymph nodes were dissected en-block routinely, and sampling of the mediastinal lymph nodes was performed as appropriate. Over the last 3 years of the study period, formal mediastinal lymph nodal clearance was routinely performed by some of the operating surgeons. In patients having a pneumonectomy, the main bronchus was closed with a stapling device applied to the main bronchus flush with the carina and was left uncovered or covered with a pleural flap depending on the surgeon's preference. Histopathologic analysis was carried out and the tumors were staged using the revised international TNM staging system for lung cancer [7]. Chemotherapy or radiotherapy before or after the operation was not used.
Follow-up and data collection
Patients were seen in the outpatient clinic every 3 months for the first year, every 6 months for the next 2 years, and yearly by the operating surgeons, the referring respiratory physicians, or the general practitioners thereafter. Survival time was calculated from the time of the operation until death or until the end of the study period.
Data were obtained from the thoracic surgery audit database, entered in a prospective systematic fashion by a dedicated thoracic surgery data manager and from the hospital medical records. Survival information (up to death or date last seen) was available for all patients. The mean follow-up was 2.4 ± 1.7 years (range, 0 to 10.4 years) with a total of 1,160 patient-years.
Definitions and statistics
Operative mortality includes deaths within 30 days from the operation or during the same hospital admission irrespective of the length of time elapsed since the operation.
Continuous data are presented as means (± standard deviation) and categorical variables as percentages. Thirty-one variables (see Appendix) were tested with univariate analysis with the end-point being survival. The continuous variables were screened with linear regression, means were compared with the unpaired t test, and proportions were compared with the 
2 test or Fischer's exact test as appropriate. The prediction of survival with 95% confidence interval (CI) limits was estimated with the Kaplan-Meier product limit method, and the resulting curves were compared with the log-rank, Breslow, and Tarone–Ware tests. To identify independent predictors of survival, the variables attaining a p value of less than or equal to 1.0 on univariate analysis were entered into the statistical program using the Cox proportional hazards regression model. A p value less than 0.05 was considered statistically significant in univariate and multivariate analyses. Independent predictors of survival are presented with theirp values and hazard ratios along with the corresponding 95% CI limits. The statistical package SPSS PC, version 11.5 (SPSS Inc, Chicago, IL) was used for analysis.
|
Results |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentReferences |
|---|
|
Histopathologic examination of the resected specimens revealed adenocarcinoma as the most common histologic type in group A and squamous cell carcinoma in group B (p < 0.0001) (Table 2).
|
Operative mortality
Operative mortality was 2.4% (9 patients) in group A and 8% (9 patients) in group B (p = 0.01).
Causes of death in group A were pneumonia (3 patients), pulmonary embolisms (2 patients), myocardial infarctions (2 patients), ventricular arrhythmias (1 patient), and peritonitis after unsuspected rupture of gall bladder (1 patient).
Causes of death in group B were pneumonia (3 patients), bronchopleural fistula (2 patients), pulmonary embolism (1 patient), myocardial infarction (1 patient), cardiorespiratory failure (1 patient), and adult respiratory distress syndrome (1 patient).
Survival
Overall (n = 485), 1-, 3-, and 5-year Kaplan-Meier survival rates (with upper and lower 95% CI limits) for patients with less extensive resections were 85% (82% to 90%), 63% (56% to 69%), and 50% (42% to 57%), respectively, and after pneumonectomy the rates were 66% (53% to 73%), 47% (35% to 57%), and 44% (32% to 55%), respectively (Fig 1).
The median survival for group A was 5 years (95% CI, 4 to 7 years) and for group B it was 3 years (95% CI, 4 to 6 years) (log rank test, p = 0.0006; Breslow test, p < 0.0001; and Tarone–Ware test, p < 0.0001). On multivariate analysis, pneumonectomy (p = 0.001), T2 stage (p = 0.006), older age (p = 0.03), and male gender (p = 0.03) were independent adverse predictors of survival (hazard ratios with 95% CI limits are shown in Table 3).
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentReferences |
|---|
The procedure of choice for stage I lung cancer is anatomic lobectomy. Lesser resections (wedge resection or segmentectomy) have been linked with higher probability of locoregional recurrence and are reserved for patients with impaired cardiorespiratory status [1, 8]. Pneumonectomy carries a greater operative risk than lobectomy and provides no benefit in terms of local recurrence and survival, but due to anatomical and technical reasons this may be unavoidable [9]. The reasons for performing a pneumonectomy in this series were central location of the tumor, tumors crossing the fissures, the presence of large lymph nodes or cancerous mass in close proximity to the pulmonary artery branches preventing their safe dissection, and intraoperative problems in which a pneumonectomy was undertaken as a bail-out procedure.
The proportion of patients undergoing pneumonectomies in this series was somewhat high (22.8%), and the greater proportion of these patients had T2N0 disease. Pneumonectomy rates between 6% and 20% for stage I NSCLC have been previously quoted [3, 6, 9–13]. The operative mortality for smaller resections (2.4%) and pneumonectomy (8%) were acceptable in keeping with earlier reports from specialist centers [3, 5, 6, 9–13].
The long-term consequences of pneumonectomy on cardiorespiratory function and quality of life are well founded [2–4]. Predictors of the outcome after removal of one lung have also been described [14], but whether pneumonectomy per se influences the overall survival remains controversial.
In this study, pneumonectomy was an independent predictor of poorer survival in the entire group and in the T1N0 and T2N0 subgroups. The significance of this finding is underscored by the fact that the patients who had a pneumonectomy were younger and had a reduced prevalence of respiratory comorbidity than the patients undergoing smaller resections preoperatively (Table 1). Although it could be argued that this may simply reflect the higher operative mortality of pneumonectomy, the Kaplan-Meier figures (Figs 1–3) reveal that the survival difference continues to increase over the first 2 to 3 postoperative years, when a sufficient number of patients in both groups remain under observation. This suggests that pneumonectomy exerts an early and a late negative effect.
The exclusion of noncancer-related deaths from the statistical analysis may have helped to further clarify the effect of pneumonectomy on survival. Although accurate information regarding the pattern of disease recurrence and the precise cause of late death was not available, which we acknowledge as a probable weakness of this study, because in the United Kingdom a patient who has been treated for lung cancer would almost certainly have lung cancer as the cause of death listed on the death certificate even if the patient died from other causes. Therefore, attempting to exclude noncancer-related deaths could introduce bias, as it may be difficult to rule out the contribution of an underlying, uncovered malignancy in an apparently noncancer-related death.
Earlier and recent reports from the Mayo Clinic [6] and St Marguerite Hospital [5] have also implicated pneumonectomy as an independent predictor of a poorer survival. Thomas and colleagues [5] described 515 stage I NSCLC patients, 81 of whom had a pneumonectomy. Survival after pneumonectomy was significantly reduced, despite the presence of similar recurrence rates in the pneumonectomy and the smaller lung resection groups.
Ferguson and Karrison [3] found a significantly poorer survival after pneumonectomy across all disease stages, but pneumonectomy ceased to be a significant factor after they adjusted for confounding variables. Their study could not address the effects of pneumonectomy on patients with stage I NSCLC, as there were only 12 of them available [3]. Other series on stage I NSCLC with rather limited numbers of pneumonectomies identified several variables, but not the extent of lung resection, as predictors of survival [9, 12, 15]. It seems likely that our study, including 111 pneumonectomy patients, was more conveniently placed to examine the role of the extent of lung resection on the outcome.
Male gender was an adverse predictor of survival in our series. A superior survival for women was previously documented by us [16] and by other authors after pulmonary resection for NSCLC [17, 18]. The reasons for this are not completely understood, but there is evidence that the growth of lung cancer cells may be partly dependent on reproductive hormones. In vitro research on human tissue with NSCLC has documented the presence of an abundance of estrogen receptors and has demonstrated an antiproliferative effect mediated by the binding of tamoxifen to these receptors [19]. Moreover, the estrogen metabolite 2-methoxyoestradiol was shown to inhibit angiogenesis to suppress tumor growth [20] and induce apoptosis in human lung cancer cells in vitro [21].
Squamous cell type was as an independent favorable predictor of survival in the T1N0 subgroup. There is uncertainty regarding the impact of the tumor histology on survival, but when a difference was demonstrated it was usually in favor of the squamous cell type [22]. Others showed no difference on the basis of the cell type [23], whereas Ferguson and colleagues [17] found a better late survival for patients having adenocarcinomas.
To avoid the physiologic impairment caused by pneumonectomy, lung parenchyma sparing procedures are used. In a recent report, sleeve lobectomy had similar mortality, but reduced loss of respiratory function compared with pneumonectomy [24]. Nevertheless, sleeve lobectomy carries specific risks including local recurrence, benign strictures or stenoses, and bronchial fistula formation [3]. We believe that this should be used judiciously, balancing the oncologic and functional circumstances of each individual patient.
We would like to believe that stage I NSCLC is a localized disease process and that complete removal of the tumor mass would result in a high cure rate. However, 5-year survival after complete lung resection is around 60% in most series and one third of the patients have distant metastases develop, which suggests that the tumor cells may spread before surgery [15]. On this basis, the use of adjuvant treatment has been proposed and there are ongoing trials testing its efficacy. In view of the lack of definitive, beneficial evidence of chemotherapy or radiotherapy and the likelihood of causing adverse effects, we do not normally offer these treatments to our patients who have completely resected stage I NSCLC [25].
In conclusion, this study suggests that pneumonectomy for stage T1N0 or T2N0 NSCLC has a potent, adverse impact on survival and provided that complete tumor removal with smaller resections is possible, then a pneumonectomy should be avoided.
|
Appendix |
|---|
|
References |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
