Limited Resection Versus Lobectomy For Stage â…  Non-small Cell Lung Cancer:a Meta-analysis Of Survival And Recurrence

BackgroudLung cancer has been the most common causes of cancer death around the word. According to United States Cancer Statistics, The estimated numbers of new lung cancer was221130cases, and The estimated numbers of lung deaths was106070cases, which is the leading estimated deaths in2011. In2008, Our country Ministry of Health released the third national population survey of deaths which show cancer has become the first cause of deaths for urban residents, accounting for25.0%, and lung cancer has accounting for22.7%of all cancer deaths.In1933, Graham completed the first left pneumonectomy for lung cancer. Then, some scholars have tried lesser lung resection for lung cancer to improve postoperative mortality and complication rates. In an attempt to preserve pulmonary function, in1973Jensik and colleagues were the first to suggest that a lesser resection (segmentectomy) might be an adequate operation for stage I of lung cancer. In1995, the Lung Cancer Study Group (LCSG) published final results of a randomized controlled trial (RCT) comparing local recurrence and survival after limited resection with those after standard lobectomy for stage IA non-small cell lung cancer(NSCLC). This study demonstrated that in patients undergoing limited resection, there was an observed tripling of the local recurrence rate, a30%increase in the overall death rate, and a50%increase in the cancer-related death rate. So, lobectomy remains the standard treatment for patients with resectable non-small cell lung cancer.With the aging of the population and the dramatic upsurge in early detection of small peripheral lung cancer through the development of integrated PET/CT, which is rapidly changing clinical practice. Improved imaging technology resolution, detection of ground glass opacities (GGO) associated with favorable histology (Bronchioloalveolar Adenocarcinoma), and increased screening of high-risk individuals has created a growing cohort of patients with smaller tumors. As an alternative, the advantages of limited resection include preservation of pulmonary function, decreased perioperative mortality and morbidity, the chance for further resection if a second or even a third NSCLC should develop.Although the randomised trial concluded that limited resection should not be standard surgery even for small peripheral NSCLC. several other groups continued to pursue limited resections, particularly anatomic segmentectomy, for early lung cancer. This was particularly the case in Japan, where the Study Group of Extended Segmentectomy for Small Lung Tumors organized a prospective multi-institutional trIal, which produced several reports. In2005, a meta-analysis was performed examining the pooled results of many of studies which compared survival in patients with stagel lung cancer after limited resection and lobectomy. Combined survival differences (survival rate for lobectomy minus that for limited resection) at1,3, and5years after resection were0.7%,1.9%, and3.6%respectively, although none was statistically significant. Minimally invasive concept of making the removal of diseased lung tissue to retain more of the healthy lung tissue, in line with the principle of treatment of surgical oncology is bound to the direction. Considering recent implementation of minimally invasive surgical techniques such as video-assisted thoracoscopic surgery (VATS), efficacy of limited resection in small, node-negative NSCLC needs to be re-evaluated. We therefore conducted a meta-analysis of published studies to quantitatively evaluate overall survival and recurrence after limited resection and lobectomy for patients with stage I NSCLC, so as to provide a more comprehensive and reliable theoretical and clinical reference. Meta-analysis is a statistical procedure that integrates the results of several independent studies considered to be combinable. The Meta-analysis approach may overcome the deficiencies of individual studies by combing trials evaluating the same intervention in a number of smaller, more accurately reflect the results in mathematics and realize the combination of literature evaluation and statistical methods. Since its introduction, Meta-analysis has established itself as an influential blanch of clinical epidemiology and health service research.ObjectiveTo systematically evaluate overall survival and recurrence after limited resection and lobectomy for patients with stage I NSCLC, so as to provide a more comprehensive and reliable theoretical and clinical basis.Methods1、According to the requisition of systematic evaluation, we formulate detailed standardization of the inclusion and exclusion criteria including the type of research subjects, interventions and outcome of events.2、Trials were identified by searching Pubmed, EMBASE, The Cochrane Library from1990to June2011. The following Mesh search heading were used:"limited resection" OR "subloectomy" OR "segmentectomy" OR "wedge resection" AND "lobectomy" AND "lung cancer" AND ("survival" OR "recurrence").The search was limited to English language papers. The related-articles function was used to expand the search from each relevant study identified.3、To assess the overall strength and quality of evidence for the various outcome parameters, a quality assessment was carried out in the style of the GRADEprofiler. Study design, study limitations, risk for bias, study inconsistency, indirectness, and imprecision were rated according to the GRADEprofiler(Grades of Recommendation, Assessment, Development and Evalution)4、The statistical analysis were performed using Review Manager version5.1.2(Co-chrane Collaboration, Oxford, UK). Results for recurrence were estimated as the odds ratio (OR) for each trial. The overall survivals were estimated as the hazard ratio (HR) for each trial. If possible, the HR and associated variance were obtained directly from each trial publication. If HR of overall survival not reported, the Kaplan-Meier survival curves of these studies were read to extract the data and to calculate the HR of overall survival by the methods of Parmer and Tierney. OR and HR for individual trials were combined across all trials, using the random-and fixed-effect models. Statistical heterogeneity was assessed using the inconsistency statistic. Publication bias was evaluated through funnel plots and with the Egger’s tests.ResultsDescription of general characteristics of the enrolled studiesTwenty-five articles published between1990and June2011met our inclusion criteria comparing limited resection versus lobectomy for stage I NSCLC. This meta-analysis evaluated8968patients in1randomized controlled trials,2prospective studies and22retrospective studies, of whom2787patients were included into the limited resection arm,6181patients into lobectomy arm. For the lobectomy arm, there were13cases of pneumonectomy in Kwiatkowski’study,1case of pneumonectomy in Campione’study and2case of pneumonectomy in Okami’study. Upon reviewing the data extraction, there was agreement among the reviews.Meta-analysis of overall survivalTwenty-four studies compared overall survival after limited resection and lobectomy for patients with stage Ⅰ NSCLC. The test of heterogeneity showed that there was statistical heterogeneity among studies included in the Meta-analysis (P=0.001, I2=53%). Meta-analysis demonstrated that the combined HR of these studies shows statistically significant difference in the HR for overall survival between limited resection arm and lobectomy arm; the HR was1.34(95%CI1.15-1.56, P=0.0002).Twelve studies compared overall survival after segmentectomy and lobectomy for patients with stage Ⅰ NSCLC. The test of heterogeneity showed that there was no statistical heterogeneity among studies included in the Meta-analysis(P=0.47, I2=0%). Meta-analysis demonstrated that the combined HR of these studies shows no statistically significant difference in the HR for overall survival between segmentectomy arm and lobectomy arm; the HR was1.25(95%CI0.98-1.58, P=0.07).Three studies compared overall survival after wedge resection and lobectomy for patients with stage Ⅰ NSCLC. The test of heterogeneity showed that there was no statistical heterogeneity among studies included in the Meta-analysis(P=0.44, I2=0%). Meta-analysis demonstrated that the combined HR of these studies shows statistically significant difference in the HR for overall survival between wedge resection arm and lobectomy arm; the HR was1.35(95%CI1.04-1.76, P=0.03).Seven studies compared overall survival after limited resection and lobectomy for patients with stage Ⅰ NSCLC with tumor2cm or less in size. The test of heterogeneity showed that there was no statistical heterogeneity among studies included in the Meta-analysis(P=0.61, I2=0%). Meta-analysis demonstrated that the combined HR of these studies shows no statistically significant difference in the HR for overall survival between limited resection arm and lobectomy arm; the HR was1.07(95%CI0.93-1.23, P=0.34).Meta-analysis of overall recurrenceData from fourteen studies were available to calculate OR for overall recurrence after limited resection and lobectomy for patients with stage I NSCLC. The test of heterogeneity showed that there was statistical heterogeneity among studies included in the Meta-analysis(P=0.010,I2=53%). Meta-analysis demonstrated that on the basis of these data, the pooled OR was estimated to1.41(95%CI1.05-1.89, P=0.02), providing a significantly increased risk for overall recurrence after limited resection.Data from nine studies were available to calculate OR for overall recurrence after segmentectomy and lobectomy for patients with stage I NSCLC. The test of heterogeneity showed that there was statistical heterogeneity among studies included in the Meta-analysis(P=0.02, I2=56%). Meta-analysis demonstrated that on the basis of these data, the pooled OR was estimated to1.20(95%CI0.74-1.94, P=0.46), providing a similar risk for overall recurrence between segmentectomy arm and lobectomy arm.Data from three studies were available to calculate OR for overall recurrence after limited resection and lobectomy for patients with stage I NSCLC with tumor2cm or less in size. The test of heterogeneity showed that there was statistical heterogeneity among studies included in the Meta-analysis(P=0.005, I2=81%). Meta-analysis demonstrated that on the basis of these data, the pooled OR was estimated to0.55(95%CI0.09-3.43, P=0.52), providing a similar risk for overall recurrence between limited resection arm and lobectomy arm.Meta-analysis of local recurrence Data from sixteen studies were available to calculate OR for local recurrence after limited resection and lobectomy for patients with stage I NSCLC. The test of heterogeneity showed that there was statistical heterogeneity among studies included in the Meta-analysis(P=0.02, I2=46%). Meta-analysis demonstrated that on the basis of these data, the pooled OR was estimated to2.64(95%CI1.77-3.93, P<0.00001), providing a significantly increased risk for local recurrence after limited resection.Ten studies compared local recurrence after segmentectomy and lobectomy for patients with stage I NSCLC. The test of heterogeneity showed that there was statistical heterogeneity among studies included in the Meta-analysis(P=0.06, I2=45%). Meta-analysis demonstrated that the combined result of these studies shows statistically significant difference in the OR for overall recurrence between segmentectomy arm and lobectomy arm; the OR was2.08(95%CI1.13-3.83, P=0.02).Four studies compared the local recurrence after limited resection and lobectomy for patients with stage I NSCLC with tumor2cm or less in size. The test of heterogeneity showed that there was statistical heterogeneity among studies included in the Meta-analysis(P=0.04,12=65%). Meta-analysis demonstrated that the combined result of these studies shows no statistically significant difference in the OR for local recurrence between limited resection arm and lobectomy arm; the OR was2.28(95%CI0.34-15.41, P=0.40).Conclusions1Compared with lobectomy, limited resection had statistically significant lower overall survival, higher local recurrence and higher overall recurrence for patients with stage I NSCLC.2Compared with lobectomy, segmentectomy had no statistically significant different overall survival, overall recurrence, but statistically significant higher local recurrence and wedge resection had lower overall survival for patients with stage I NSCLC.3Compared with lobectomy, limited resection has no statistically significant different overall survival, overall recurrence and local recurrence for patients with stage I NSCLC with tumor2cm or less in size.Since most of included studies were retrospective studies and interstudy heterogeneity was detected, we should interpret the present results carefully.