- Email: [email protected]
Three-Field or Two-Field Resection for Thoracic Esophageal Cancer: A Meta-Analysis
Ting Ye, MD, Yihua Sun, MD, Yiliang Zhang, MD, Yang Zhang, MD, and Haiquan Chen, MD Department of Thoracic Surgery, Shanghai Cancer Center, and Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, China
Background. There have been many controversies about the optimal extent of lymphadenectomy for thoracic esophageal cancer, whether three-field lymphadenectomy is superior to two-field lymphadenectomy with respect to the 5-year survival rate and perioperative morbidities and mortality. Methods. A comprehensive search of PubMed and Embase for relevant studies comparing three-field and two-field lymphadenectomies for thoracic esophageal cancer was conducted using the Preferred Reporting Items for Systemic Reviews and Meta-Analyses standards. Hazard ratios (HRs) were extracted from these studies to give pooled estimates of the effect of the two surgical procedures on the 5-year survival rate and perioperative morbidities and mortality. Results. Thirteen studies were included for analysis. Compared with two-field lymphadenectomy, three-field
lymphadenectomy provided a higher 5-year survival rate (HR 0.64, 95% confidence interval [CI]: 0.56 to 0.73, p [ 0.000) and incidence of anastomotic leakage (HR 1.46, 95% CI: 1.19 to 1.79, p [ 0.000), with a comparative perioperative mortality (HR 0.64, 95% CI: 0.38 to 1.10, p [ 0.110) and incidence of vocal cord palsy (HR 1.12, 95% CI: 0.82 to 1.54, p [ 0.470) and pulmonary complications (HR 1.00, 95% CI: 0.89 to 1.12, p [ 0.760). Conclusions. Published evidence indicated that threefield lymphadenectomy could be a priority for thoracic esophageal cancer, especially for tumors with positive lymph nodes. Given the lack of large-sample randomized controlled studies, further evaluations are necessary.
F
chain lymph nodes from the superior mediastinum up to the neck through a transthoracic approach and achieve the same outcome, in addition to easier perioperative management [8, 13]. The purpose of this study was to review the differences between three-field and two-field lymphadenectomy for thoracic esophageal cancer with respect to the 5-year survival rate as well as perioperative morbidities and mortality based on a series of published articles.
or years the procedure of lymphadenectomy for esophageal cancer has been indispensible as the number of positive lymph nodes has been reported to be an important prognostic factor [1-3]. However, recently there has been much debate regarding the optimal extent of lymph node dissection to improve long-term survival while minimizing the perioperative morbidity and mortality [1]. In particular, the question as to the benefits and risks of three-field and two-field lymphadenectomy for esophageal cancer has not yet been conclusively answered [4]. Three-field lymphadenectomy is proposed as an acceptable theory of surgical treatment for esophageal carcinoma because it is known to be essential in acquiring improved postoperative outcomes of patients with esophageal carcinoma, and the overall prevalence of cervical lymph node metastases has been documented as approximately 20% to 40% regardless of the level of the primary tumor [5-7]. Conversely, some investigators argue that two-field lymphadenectomy is enough because modern two-field dissection can dissect recurrent nerve Accepted for publication June 6, 2013. Address correspondence to Dr Chen, 270 Dong’an Rd, Shanghai 200032, China; e-mail: [email protected].
Ó 2013 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2013;96:1933–42) Ó 2013 by The Society of Thoracic Surgeons
Patients and Methods Search Strategy The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement was used as the basis for reporting the materials and methods of this study [9]. All studies comparing three-field and two-field lymphadenectomy for thoracic esophageal carcinoma were identified by searching two electronic databases (PubMed and Embase), using the medical subject headings “esophageal neoplasms,” “lymphadenectomy,” and “lymph node dissection.” These MeSH terms were combined using the Boolean operator “and” or “or” to find articles that contained information on all these terms. In 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2013.06.050
GENERAL THORACIC
Three-Field or Two-Field Resection for Thoracic Esophageal Cancer: A Meta-Analysis
GENERAL THORACIC
1934
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
Ann Thorac Surg 2013;96:1933–42
Fig 1. Flow diagram of studies included in this meta-analysis.
addition, another comprehensive search was performed using the following search terms: logical combinations of “two-field,” “2-field,” “three-field,” and “3-field,” respectively, with “lymphadenectomy” or “lymph node dissection” for achieving the highest likelihood of retrieving all relevant articles. All titles and abstracts were scanned, and appropriate citations were reviewed. A manual search of the bibliographies of relevant papers was also performed to identify publications for possible inclusion. Extra effort was taken to avoid inclusion of repeated (covert) publications.
Eligibility Criteria We have attempted to identify all relevant prospective and retrospective controlled studies comparing threefield and two-field lymphadenectomy and reporting the outcome of the two surgical procedures with curative intent for thoracic esophageal cancer, as only two randomized trials have been conducted. Three-field lymphadenectomy was defined as lymph node resection encompassing bilateral cervical lymph nodes (including the deep internal, deep external, and deep lateral nodes), total mediastinal lymph nodes (including superior, middle, and inferior mediastinal nodes), and abdominal lymph nodes (including paracardiac, perigastric lesser curvature, left gastric artery, common hepatic artery, and splenic artery nodes), whereas two-field lymphadenectomy includes the same abdominal and thoracic compartments as the three-field dissection with the exception of removal of the cervical nodes [1]. Articles in which three-field and two-field lymphadenectomy could not be distinguished were excluded. To ensure that the reviewed surgical series
reflected the outcomes for patients treated with modern surgical, anesthetic, and diagnostic care techniques, we have restricted our qualitative analysis to surgical series published between 1991 and December 2012. Moreover, there should be at least 60 patients in each study. Studies were required to have follow-up on patients for at least 30 days after operations for inclusion of data for postoperative morbidity and mortality. And for survival data, studies should contain the 5-year survival data and survival curve. All eligible English articles in thoracic surgery recruiting adult patients were included in this review. When several articles reported the same patient material, only the most recent article was included.
Data Extraction We created a data extraction sheet to capture all data needed to assess the quality and eligibility of studies and perform the meta-analysis. Data from eligible studies were extracted by one author and then independently by another author. Disagreements were solved by discussion between the two authors. The following information was extracted from each study: first author, year of publication, study type, population characteristics (including tumor location, pathologic subtype, pathologic T stage, and adjuvant or neoadjuvant therapy), number of patients, and key outcomes. The 5-year survival rate was considered as the primary outcome. Postoperative mortality and morbidities including anastomotic leakage, vocal cord paralysis, and pulmonary complications were regarded as the secondary outcomes. Mortality included both the inhospital and 30-day mortality. Anastomotic leakage
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
1935
Table 1. Characteristics of the Included Clinical Studies Study
Year
Study Design
Tumor Location
H. Shiozaki et al [12] H. Miyata et al [13] H. Fujita et al [14] Young MS et al [15] T. Noguchi et al [16] H. Fujita et al [17] Y. Tabira et al [18] H. Kato et al [19] H. Igaki et al [20] H. Fujita et al [21] H. Kato et al [22] H. Akiyama et al [23] T. Nishihira et al [24]
2001 2006 1992 2010 2004 1995 1999 1991 2004 2003 1995 1994 1998
Retrospective Prospective Retrospective Retrospective Retrospective Retrospective Retrospective RCT Retrospective Retrospective Retrospective Retrospective RCT
All thoracic locations Middle and lower thoracic locations All thoracic locations Upper thoracic location All thoracic locations All thoracic locations All thoracic locations All thoracic locations Lower thoracic location All thoracic locations All thoracic locations All thoracic locations All thoracic locations
Neo-adjuvant Therapy Two-Field 30 None 11 None 10 None 7 Unmentioned Unmentioned Unmentioned 2 Unmentioned Unmentioned None
Adjuvant Therapy
Pathological Subtype
pT-Stage
SCC Undefined Undefined SCC SCC SCC Mixed Mixed SCC SCC Mixed SCC SCC
T1-4 T0-4 T0-4 T0-4 T1-4 T0-4 T1-4 T0-4 T0-4 T1-4 T0-4 T0-4 T1-3
Total Number of Patients
Three-Field
Two-Field
Three-Field
Two-Field
Three-Field
52 None 15 None 22 None 0 Unmentioned Unmentioned Unmentioned 15 Unmentioned Unmentioned None
110 Unmentioned None 15 12 Unmentioned 19 Unmentioned 15 Unmentioned 21 Unmentioned Unmentioned 28
235 Unmentioned None 38 19 Unmentioned 35 Unmentioned 26 Unmentioned 89 Unmentioned Unmentioned 28
1176 123 56 100 34 78 65 86 73 55 65 121 290 30
1203 129 33 27 57 68 63 66 77 101 176 100 274 32
RCT ¼ randomized controlled trial; Unmentioned ¼ no pathological subtype described in the article or no neo-adjuvant therapy or adjuvant therapy mentioned in the article; Mixed ¼ a lower percentage of adenocarcinoma and other types of carcinoma included; SCC ¼ squamous cell carcinoma; None ¼ no patients received neo-adjuvant therapy or adjuvant therapy.
included both clinical leakage and subclinical leakage (only seen radiologically). Vocal cord paralysis included both persistent hoarseness of voice and vocal cord paralysis determined by indirect laryngoscopy. Pulmonary complications mainly included pneumonia and respiratory failure. When possible, the exact number of a specific complication was identified, otherwise the article was not included in the analysis (for that specific complication). We did not review adenocarcinoma and squamous cell carcinoma separately because the ratio of adenocarcinoma was very low in this analysis.
Statistical Analysis Studies were divided into the following two groups: randomized controlled trials and retrospective comparative trials when analyzing the primary outcome. Overall hazard ratio (HR) and its 95% confidence interval (CI) for the 5-year survival rate were estimated using a weighted
mean of HRs from the studies with each weight proportional to the reciprocal of the corresponding variance estimated by the Mantel-Haenszel procedure. The HR and its variance were extracted and estimated using the methods reported by Parmar and colleagues [10]. The statistical tests were performed with the STATA 11.0 software (StataCorp, College Station, TX). The secondary outcomes including perioperative mortality and morbidities were entered into the Cochrane Review Manager (RevMan) 5.1 (Cochrane Collaboration, Oxford, UK). Relative risk with the corresponding CI was the principal measure of effect. A p value less than 0.05 was considered statistically significant. Statistical heterogeneity among studies was evaluated with Cochrane’s Q statistic. Data were pooled using the fixed effect model. If statistical heterogeneity was suspected, the random effect model was applied. An estimation of potential publication bias was executed by Egger’s test [11]. A p value less than 0.05 was
GENERAL THORACIC
Ann Thorac Surg 2013;96:1933–42
GENERAL THORACIC
1936
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
Ann Thorac Surg 2013;96:1933–42
Fig 2. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the 5-year survival outcome for all included studies. (CI ¼ confidence interval.)
considered representative of statistically publication bias.
significant
Results Our initial review identified 3,048 studies, 3,032 of which were discarded after title and abstract review because they did not meet inclusion criteria. The full texts of the remaining 16 studies were further examined. Three of these studies were eliminated after full-text eligibility: two studies containing patients duplicated with other studies, and one study lacking the survival curve in the article. Thus, 13 articles involving a total of 2,379 patients were included in this analysis. There were two randomized controlled studies, one prospective nonrandomized study, and 10 retrospective studies (Fig 1; Table 1) [12-24].
Five-Year Survival Analysis When the survival data of the 13 articles were combined, the HR for the 5-year survival rate (defined as the primary
Fig 3. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the 5-year survival outcome for the two randomized controlled trials. (CI ¼ confidence interval.)
outcome), expressed as three-field lymphadenectomy versus two-field lymphadenectomy, was 0.64 (95% CI: 0.56 to 0.73, p ¼ 0.000; Fig 2), which indicated that three-field lymphadenectomy provided a better 5-year survival rate. When the data of the two randomized trials were extracted and analyzed, the HR was 0.56 (95% CI: 0.34 to 0.78, p ¼ 0.000; Fig 3), which also showed that three-field lymphadenectomy supplied a favorable survival outcome. Test for homogeneity of the HR did not show statistically significant heterogeneity concerning 5-year survival rate of all articles and one of the two randomized controlled trials (p ¼ 0.153 and 0.501, respectively). And the Egger’s test showed there was no publication bias in this analysis (p ¼ 0.237).
Mortality and Morbidity Analysis As to the secondary outcome in this study, eight studies showed the data of perioperative mortality. The difference of perioperative mortality between threefield and two-field lymphadenectomies was statistically insignificant (risk ratio [RR] 0.64; 95% CI: 0.38 to
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
1937
Fig 4. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the perioperative mortality. (CI ¼ confidence interval; M-H ¼ Mantel-Haenszel.)
1.10, p ¼ 0.110; Fig 4). Nine studies respectively provided the specific number of anastomotic leakage and pulmonary complications after esophagectomy. Three-field lymphadenectomy had a higher prevalence of anastomotic leakage (RR 1.46; 95% CI: 1.19 to 1.79, p ¼ 0.000; Fig 5) and a similar prevalence of pulmonary complications (RR 1.00; 95% CI: 0.89 to 1.12, p ¼ 0.760; Fig 6). In addition, eight studies provided the specific data of postoperative vocal cord palsy, and the prevalence was comparative between the two groups (RR 1.12; 95% CI: 0.82 to 1.54, p ¼ 0.470; Fig 7; Table 2).
Subgroup Analysis We did the subgroup analysis for the effect of three-field and two-field lymphadenectomy for esophageal carcinoma with pathologically positive lymph nodes. Four studies assessed the outcome of the two surgical procedures for patients with lymph nodes metastases. The HR for 5-year survival rate was 0.39 (95% CI: 0.31 to 0.46, p ¼ 0.000; Fig 8A), which favored three-field lymphadenectomy. Especially, for tumors with positive intrathoracic recurrent nerve nodes, the HR was 0.32 (95% CI: 0.23 to 0.41, p ¼ 0.000; Fig 8B), which favored the
more invasive lymphadenectomy as well. Furthermore, we assessed the effect of the two lymphadenectomies for different tumor locations. For upper/middle thoracic esophageal cancer with pathologically positive lymph nodes, the HR for the 5-year survival rate was 0.56 (95% CI: 0.34 to 0.78, p ¼ 0.000; Fig 8C). And for lower thoracic esophageal cancer with positive lymph nodes, the HR was 0.53 (95% CI: 0.42 to 0.64, p ¼ 0.000; Fig 8D). Nevertheless, for upper/middle thoracic esophageal cancer without lymph nodes metastases, the HR was 0.53 (95% CI: 0.01 to 1.04, p ¼ 0.044; Fig 8E). These results indicated that three-field lymphadenectomy could be favorable for esophageal cancer with positive lymph nodes regardless of the tumor location. However, for lower thoracic esophageal cancer without lymph nodes metastases, two-field lymphadenectomy could provide a 5-year survival rate similar to that of three-field dissection.
Comment The optimal type of lymphadenectomy for thoracic esophageal cancer is controversial. Only two randomized controlled trials have been published, involving only 212
Fig 5. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the postoperative anastomotic leakage. (CI ¼ confidence interval; M-H ¼ Mantel-Haenszel.)
GENERAL THORACIC
Ann Thorac Surg 2013;96:1933–42
GENERAL THORACIC
1938
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
Ann Thorac Surg 2013;96:1933–42
Fig 6. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the postoperative pulmonary complications. (CI ¼ confidence interval; M-H ¼ Mantel-Haenszel.)
patients (of a total of 2,379 patients in the articles reviewed [Table]). Meanwhile, the other studies are mostly retrospective analyses that described the investigators’ experiences, maybe reflecting the personal preference of the surgeons. The paucity of large-sample prospective studies yet provides little evidence to answer this question. Therefore, we did this study that included all published articles comparing the two types of lymphadenectomies. When all data were combined, the 5-year survival rate showed a significant difference between three-field and two-field lymphadenectomies. And three-field lymphadenectomy provided a better 5-year survival rate (Fig 2), which suggested that the addition of cervical lymph node dissection improved the longterm outcome over two-field lymphadenectomy alone, especially for esophageal cancer with positive lymph nodes (Fig 8A). Moreover, three-field lymphadenectomy interestingly presented a similar perioperative mortality compared with two-field dissection (Fig 4), whereas it used to be regarded as a more invasive procedure that was always associated with a higher mortality.
Fig 7. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the postoperative vocal cord palsy. (CI ¼ confidence interval; M-H ¼ Mantel-Haenszel.)
Despite the favorable results that three-field lymphadenectomy presented a higher 5-year survival rate and comparable perioperative mortality as two-field lymphadenectomy for thoracic esophageal cancer, some limitations in this study were worth noting. Firstly, the heterogeneity of these included studies was noticeable, though neither the test for homogeneity nor the Egger’s test for publication bias showed a statistical significance. Patients in these studies showed diverse clinicopathologic characteristics including various locations of the tumor and TN staging. It has been well supported that upper/middle thoracic esophageal cancer benefits much from three-field lymphadenectomy compared with lower lesions [7, 25]. In addition, different pathologic T stages have different ratios of cervical lymph nodes metastases, which determine the necessity of three-field lymphadenectomy [26]. Therefore, we did the subgroup assessment for esophageal cancer with positive lymph nodes and different tumor locations. Although results of the subgroup analyses were reasonable, the number of included patients was limited. Moreover, most articles in this study were from Japanese institutions. Considering that most Japanese surgeons prefer
Ten studies in this analysis provided the data of perioperative complications. Unmentioned indicates no relevant data provided in the study.
1939
cervical lymph nodes dissection for thoracic esophageal carcinoma, experiences from other countries are greatly needed, especially from Western colleagues. Furthermore, most studies were retrospective analyses, and the two parallel controlled groups of several studies were not well balanced. However, when the two randomized trials were analyzed, the results showed that three-field lymphadenectomy provided a better 5year survival as well (Fig 3), although conclusions from the two trials might be controversial because of the small sample sizes [1]. Three primary complications—anastomotic leakage, vocal cord palsy, and pulmonary complications (mainly pneumonia and respiratory failure)—were assessed in this study, because these frustrating issues were mainly attributed to three-field dissection and the overall number of complications in those studies was not precisely described. Unsurprisingly, three-field lymphadenectomy had a higher anastomotic leakage rate than two-field lymph node dissection (Fig 5). Anastomotic leakage is always linked with cervical lymph node dissection, probably partly due to the location of the anastomosis. The cervical anastomosis may carry a higher risk of leakage than the intrathoracic anastomosis, although management of the cervical leakage is much easier than the intrathoracic counterpart [27, 28]. In this case, three-field lymphadenectomy with intrathoracic anastomosis could be a promising choice for middle and lower thoracic esophageal cancer. Conversely, vocal cord paralysis from injury of the recurrent laryngeal nerve is another frequent complication of cervical dissection, although it is also mentioned after transthoracic anastomoses. A high incidence of vocal cord paralysis indicates that the recurrent nerve is mainly at risk during the cervical dissection and the construction of the anastomosis [8, 29]. The difference of the incidence of vocal cord palsy was statistically insignificant between the two groups (Fig 7). Moreover, the incidence of postoperative pulmonary complications was, interestingly, not higher in the three-field dissection group as well (Fig 6), which conflicted with the general idea that the more invasive lymphadenectomy might cause more pulmonary complications [30]. Therefore, meticulous surgical procedures with careful perioperative management may be a potential key for decreasing the incidence of the recurrent nerve injury and thus reducing the pulmonary complications of three-field lymphadenectomy. In summary, although three-field lymphadenectomy showed a higher incidence of anastomotic leakage than two-field lymphadenectomy did, it produced a higher 5-year survival rate, and comparable perioperative mortality and incidence of vocal cord palsy and pulmonary complications. Based on what was found in our study, three-field lymphadenectomy could be a priority for thoracic esophageal cancer, especially for tumors with positive lymph nodes. Large-sample randomized controlled studies are urgently needed for further evaluation.
GENERAL THORACIC
31.25 Unmentioned 7.4 33.33 9.09 10 18.75 1.47 8.77 31.25 13.85 Unmentioned 5 38.46 16.44 9.09 16.67 2.56 0 10.9 69.32 Unmentioned Unmentioned 69.84 14.29 15 56.25 2.94 28.07 8.91 55.38 Unmentioned Unmentioned 47.69 20.55 23.14 30 0 14.71 18.18 27.84 Unmentioned 22.22 33.33 33.77 43 6.25 1.47 15.79 38.61 20 Unmentioned 15 10.77 23.29 25.62 20 6.41 11.76 29.09 2.84 0 0 1.59 11.68 3 3.13 3.13 Unmentioned Unmentioned 6.15 1.79 6 3.08 2.74 10.74 6.67 6.67 Unmentioned Unmentioned
Three-Field Two-Field Three-Field Two-Field Three-Field Two-Field Study [Ref]
Fujita et al [21] Miyata et al [13] Fujita et al [14] Fujita et al [17] Kato et al [19] Kato et al [22] Nishihira et al [24] Noguchi et al [16] Young et al [15] Igaki et al [20]
Three-Field
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
Two-Field
p ¼ 0.760 p ¼ 0.470 p ¼ 0.000 p ¼ 0.110
Vocal Cord Palsy (%) Anastomotic Leakage (%) Perioperative Mortality (%)
Table 2. Incidence of Perioperative Complications Between Three-Field and Two-Field Lymphadenectomies in This Analysis
Pulmonary Complications (%)
Ann Thorac Surg 2013;96:1933–42
GENERAL THORACIC
1940
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
Ann Thorac Surg 2013;96:1933–42
Fig 8. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the 5-year survival outcome for subgroups: (A) esophageal cancer with pathologically positive lymph nodes; (B) esophageal cancer with positive intrathoracic recurrent laryngeal nerve nodes; (C) upper/middle esophageal cancer with positive lymph nodes; (D) lower esophageal cancer with positive lymph nodes; and (E) upper/middle esophageal cancer without lymph nodes metastasis. (CI ¼ confidence interval.)
References 1. Gockel I, Sgourakis G, Lyros O, Hansen T, Lang H. Dissection of lymph node metastases in esophageal cancer. Expert Rev Anticancer Ther 2011;11:571–8. 2. Darling G. The role of lymphadenectomy in esophageal cancer. J Surg Oncol 2009;99:189–93. 3. Tachibana M, Kinugasa S, Hirahara N, Yoshimura H. Lymph node classification of esophageal squamous cell carcinoma and adenocarcinoma. Eur J Cardiothorac Surg 2008;34: 427–31. 4. Yajima S, Oshima Y, Shimada H. Neck dissection for thoracic esophageal squamous cell carcinoma. Int J Surg Oncol 2012;2012:750456. 2012 Feb 27 [E-Pub]. 5. Lerut T, Nafteux P, Moons J, et al. Three-field lymphadenectomy for carcinoma of the esophagus and gastroesophageal junction in 174 R0 resections: impact on staging, diseasefree survival, and outcome. A plea for adaptation of TNM classification in upper-half esophageal carcinoma. Ann Surg 2004;240:962–74. 6. Altorki N, Kent M, Ferrara C, Port J. Three-field lymph node dissection for squamous cell and adenocarcinoma of the esophagus. Ann Surg 2002;236:177–83.
7. Udagawa H, Ueno M, Shinohara H, et al. The importance of grouping of lymph node stations and rationale of three-field lymphadenectomy for thoracic esophageal cancer. J Surg Oncol 2012;106:742–7. 8. Mariette C, Piessen G. Oesophageal cancer: how radical should surgery be? Eur J Surg Oncol 2012;38:210–3. 9. Moher D, Liberati A, Tetzlaff J, Altman DGfor the PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009;6: e1000097. 10. Parmar MK, Torri V, Stewart L. Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints. Stat Med 1998;17:2815–34. 11. Egger M, Davey SG, Schneider M, Minder C. Bias in metaanalysis detected by a simple, graphical test. BMJ 1997;315: 629–34. 12. Shiozaki H, Yano M, Tsujinaka T, et al. Lymph node metastasis along the recurrent nerve chain is an indication for cervical lymph node dissection in thoracic esophageal cancer. Dis Esophagus 2001;14:191–6. 13. Miyata H, Yano M, Doki Y, et al. A prospective trial for avoiding cervical lymph node dissection for thoracic esophageal cancers, based on intra-operative genetic diagnosis of
14. 15.
16.
17.
18.
19. 20.
21.
micrometastasis in recurrent laryngeal nerve chain nodes. J Surg Oncol 2006;93:477–84. Fujita H, Kakegawa T, Yamana H, et al. Cervico-thoracoabdominal (3-field) lymph node dissection for carcinoma in the thoracic esophagus. Kurume Med J 1992;39:167–74. Shim YM, Kim HK, Kim K. Comparison of survival and recurrence pattern between two-field and three-field lymph node dissections for upper thoracic esophageal squamous cell carcinoma. J Thorac Oncol 2010;5:707–12. Noguchi T, Wada S, Takeno S, Hashimoto T, Moriyama H, Uchida Y. Two-step three-field lymph node dissection is beneficial for thoracic esophageal carcinoma. Dis Esophagus 2004;17:27–31. Fujita H, Kakegawa T, Yamana H, et al. Mortality and morbidity rates, postoperative course, quality of life, and prognosis after extended radical lymphadenectomy for esophageal cancer. Comparison of three-field lymphadenectomy with two-field lymphadenectomy. Ann Surg 1995;222:654–62. Tabira Y, Kitamura N, Yoshioka M, et al. Significance of three-field lymphadenectomy for carcinoma of the thoracic esophagus based on depth of tumor infiltration, lymph nodal involvement and survival rate. J Cardiovasc Surg (Torino) 1999;40:737–40. Kato H, Watanabe H, Tachimori Y, Lizuka T. Evaluation of neck lymph node dissection for thoracic esophageal carcinoma. Ann Thorac Surg 1991;51:931–5. Igaki H, Tachimori Y, Kato H. Improved survival for patients with upper and/or middle mediastinal lymph node metastasis of squamous cell carcinoma of the lower thoracic esophagus treated with 3-field dissection. Ann Surg 2004;239: 483–90. Fujita H, Sueyoshi S, Tanaka T, et al. Optimal lymphadenectomy for squamous cell carcinoma in the thoracic
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
22. 23. 24.
25. 26. 27.
28.
29. 30.
1941
esophagus: comparing the short- and long-term outcome among the four types of lymphadenectomy. World J Surg 2003;27:571–9. Kato H. Lymph node dissection for thoracic esophageal carcinoma. Two- and 3-field lymph node dissection. Ann Chir Gynaecol 1995;84:193–9. Akiyama H, Tsurumaru M, Udagawa H, Kajiyama Y. Radical lymph node dissection for cancer of the thoracic esophagus. Ann Surg 1994;220:364–73. Nishihira T, Hirayama K, Mori S. A prospective randomized trial of extended cervical and superior mediastinal lymphadenectomy for carcinoma of the thoracic esophagus. Am J Surg 1998;175:47–51. Jang HJ, Lee HS, Kim MS, Lee JM, Zo JL. Patterns of lymph node metastasis and survival for upper esophageal squamous cell carcinoma. Ann Thorac Surg 2011;92:1091–7. Stiles BM, Mirza F, Port JL, et al. Predictors of cervical and recurrent laryngeal lymph node metastasis from esophageal cancer. Ann Thorac Surg 2010;90:1805–11. Klink CD, Binneb€ osel M, Otto J, et al. Intrathoracic versus cervical anastomosis after resection of esophageal cancer: a matched pair analysis of 72 patients in a single center study. World J Surg Oncol 2012;10:159. Hu H, Ye T, Zhang Y, Zhang J, Luketich JD, Chen H. Modifications in retrosternal reconstruction after oesophagogastrectomy may reduce the incidence of anastomotic leakage. Eur J Cardiothorac Surg 2012;42:359–63. Hulscher JB, Tijssen JG, Obertop H, van Lanschot JJ. Transthoracic versus transhiatal resection for carcinoma of the esophagus: a meta-analysis. Ann Thorac Surg 2001;72:306–13. Fang W, Kato H, Tachimori Y, Igaki H, Sato H, Daiko H. Analysis of pulmonary complications after three-field lymph node dissection for esophageal cancer. Ann Thorac Surg 2003;76:903–8.
INVITED COMMENTARY The debate about the extent of lymph node dissection for cancer of the thoracic esophagus has occupied esophageal surgeons for decades. In the early 1980s, Japanese surgeons proposed that nodal dissection should be extended to a “third field” that includes the nodes along both recurrent nerves and a cervical node dissection. Despite promising results, wider implementation of the technique in the West was impeded by concerns about its seemingly higher morbidity and unproven survival benefit. The lack of adequately powered randomized trials or well-conceived metaanalyses sustained the controversy. In this issue of The Annals, Ye and colleagues [1] take an important step toward resolving the debate. They report the first metaanalysis of published data comparing the outcomes of esophagectomy with either two-field or three-field dissection. The results show that three-field dissection was associated with a significant improvement in survival and a significant reduction in hospital mortality. Surprisingly, there was no difference in the incidence of recurrent nerve palsy or pulmonary morbidity. So should this metaanalysis resolve the simmering controversy in favor of three-field dissection? A careful review of the 15 studies included in this metaanalysis would suggest that the debate is likely to Ó 2013 by The Society of Thoracic Surgeons Published by Elsevier Inc
continue for some time. The highest quality metaanalyses usually result from “best evidence synthesis” where only methodologically sound studies are included. This metaanalysis falls short of that ideal since 12 of the 15 cited studies are retrospective series whose inherent selection biases cannot be controlled for by the method. The investigators are obviously keenly aware of that shortcoming, which may have driven them to analyze the results of the two randomized trials separately. The fact that these results were concordant with the overall analysis is reassuring; however, neither of these studies is of sufficient power to justify changes in current practice. Another important question to consider is whether these results are generalizable to patients with adenocarcinoma of the esophagus, the most common type of esophageal cancer throughout most of Europe and North America. Finally, there is the intriguing finding by Ye and colleagues [1] that the rate of recurrent nerve palsy between the two techniques is comparable. That certainly has not been my experience or the experience of many other Western surgeons. The explanation may well be in what is precisely meant by the term “two-field” dissection. More current versions of two-field dissection include dissection of the recurrent nodes without dissection of 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2013.07.076
GENERAL THORACIC
Ann Thorac Surg 2013;96:1933–42
Background. There have been many controversies about the optimal extent of lymphadenectomy for thoracic esophageal cancer, whether three-field lymphadenectomy is superior to two-field lymphadenectomy with respect to the 5-year survival rate and perioperative morbidities and mortality. Methods. A comprehensive search of PubMed and Embase for relevant studies comparing three-field and two-field lymphadenectomies for thoracic esophageal cancer was conducted using the Preferred Reporting Items for Systemic Reviews and Meta-Analyses standards. Hazard ratios (HRs) were extracted from these studies to give pooled estimates of the effect of the two surgical procedures on the 5-year survival rate and perioperative morbidities and mortality. Results. Thirteen studies were included for analysis. Compared with two-field lymphadenectomy, three-field
lymphadenectomy provided a higher 5-year survival rate (HR 0.64, 95% confidence interval [CI]: 0.56 to 0.73, p [ 0.000) and incidence of anastomotic leakage (HR 1.46, 95% CI: 1.19 to 1.79, p [ 0.000), with a comparative perioperative mortality (HR 0.64, 95% CI: 0.38 to 1.10, p [ 0.110) and incidence of vocal cord palsy (HR 1.12, 95% CI: 0.82 to 1.54, p [ 0.470) and pulmonary complications (HR 1.00, 95% CI: 0.89 to 1.12, p [ 0.760). Conclusions. Published evidence indicated that threefield lymphadenectomy could be a priority for thoracic esophageal cancer, especially for tumors with positive lymph nodes. Given the lack of large-sample randomized controlled studies, further evaluations are necessary.
F
chain lymph nodes from the superior mediastinum up to the neck through a transthoracic approach and achieve the same outcome, in addition to easier perioperative management [8, 13]. The purpose of this study was to review the differences between three-field and two-field lymphadenectomy for thoracic esophageal cancer with respect to the 5-year survival rate as well as perioperative morbidities and mortality based on a series of published articles.
or years the procedure of lymphadenectomy for esophageal cancer has been indispensible as the number of positive lymph nodes has been reported to be an important prognostic factor [1-3]. However, recently there has been much debate regarding the optimal extent of lymph node dissection to improve long-term survival while minimizing the perioperative morbidity and mortality [1]. In particular, the question as to the benefits and risks of three-field and two-field lymphadenectomy for esophageal cancer has not yet been conclusively answered [4]. Three-field lymphadenectomy is proposed as an acceptable theory of surgical treatment for esophageal carcinoma because it is known to be essential in acquiring improved postoperative outcomes of patients with esophageal carcinoma, and the overall prevalence of cervical lymph node metastases has been documented as approximately 20% to 40% regardless of the level of the primary tumor [5-7]. Conversely, some investigators argue that two-field lymphadenectomy is enough because modern two-field dissection can dissect recurrent nerve Accepted for publication June 6, 2013. Address correspondence to Dr Chen, 270 Dong’an Rd, Shanghai 200032, China; e-mail: [email protected].
Ó 2013 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2013;96:1933–42) Ó 2013 by The Society of Thoracic Surgeons
Patients and Methods Search Strategy The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement was used as the basis for reporting the materials and methods of this study [9]. All studies comparing three-field and two-field lymphadenectomy for thoracic esophageal carcinoma were identified by searching two electronic databases (PubMed and Embase), using the medical subject headings “esophageal neoplasms,” “lymphadenectomy,” and “lymph node dissection.” These MeSH terms were combined using the Boolean operator “and” or “or” to find articles that contained information on all these terms. In 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2013.06.050
GENERAL THORACIC
Three-Field or Two-Field Resection for Thoracic Esophageal Cancer: A Meta-Analysis
GENERAL THORACIC
1934
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
Ann Thorac Surg 2013;96:1933–42
Fig 1. Flow diagram of studies included in this meta-analysis.
addition, another comprehensive search was performed using the following search terms: logical combinations of “two-field,” “2-field,” “three-field,” and “3-field,” respectively, with “lymphadenectomy” or “lymph node dissection” for achieving the highest likelihood of retrieving all relevant articles. All titles and abstracts were scanned, and appropriate citations were reviewed. A manual search of the bibliographies of relevant papers was also performed to identify publications for possible inclusion. Extra effort was taken to avoid inclusion of repeated (covert) publications.
Eligibility Criteria We have attempted to identify all relevant prospective and retrospective controlled studies comparing threefield and two-field lymphadenectomy and reporting the outcome of the two surgical procedures with curative intent for thoracic esophageal cancer, as only two randomized trials have been conducted. Three-field lymphadenectomy was defined as lymph node resection encompassing bilateral cervical lymph nodes (including the deep internal, deep external, and deep lateral nodes), total mediastinal lymph nodes (including superior, middle, and inferior mediastinal nodes), and abdominal lymph nodes (including paracardiac, perigastric lesser curvature, left gastric artery, common hepatic artery, and splenic artery nodes), whereas two-field lymphadenectomy includes the same abdominal and thoracic compartments as the three-field dissection with the exception of removal of the cervical nodes [1]. Articles in which three-field and two-field lymphadenectomy could not be distinguished were excluded. To ensure that the reviewed surgical series
reflected the outcomes for patients treated with modern surgical, anesthetic, and diagnostic care techniques, we have restricted our qualitative analysis to surgical series published between 1991 and December 2012. Moreover, there should be at least 60 patients in each study. Studies were required to have follow-up on patients for at least 30 days after operations for inclusion of data for postoperative morbidity and mortality. And for survival data, studies should contain the 5-year survival data and survival curve. All eligible English articles in thoracic surgery recruiting adult patients were included in this review. When several articles reported the same patient material, only the most recent article was included.
Data Extraction We created a data extraction sheet to capture all data needed to assess the quality and eligibility of studies and perform the meta-analysis. Data from eligible studies were extracted by one author and then independently by another author. Disagreements were solved by discussion between the two authors. The following information was extracted from each study: first author, year of publication, study type, population characteristics (including tumor location, pathologic subtype, pathologic T stage, and adjuvant or neoadjuvant therapy), number of patients, and key outcomes. The 5-year survival rate was considered as the primary outcome. Postoperative mortality and morbidities including anastomotic leakage, vocal cord paralysis, and pulmonary complications were regarded as the secondary outcomes. Mortality included both the inhospital and 30-day mortality. Anastomotic leakage
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
1935
Table 1. Characteristics of the Included Clinical Studies Study
Year
Study Design
Tumor Location
H. Shiozaki et al [12] H. Miyata et al [13] H. Fujita et al [14] Young MS et al [15] T. Noguchi et al [16] H. Fujita et al [17] Y. Tabira et al [18] H. Kato et al [19] H. Igaki et al [20] H. Fujita et al [21] H. Kato et al [22] H. Akiyama et al [23] T. Nishihira et al [24]
2001 2006 1992 2010 2004 1995 1999 1991 2004 2003 1995 1994 1998
Retrospective Prospective Retrospective Retrospective Retrospective Retrospective Retrospective RCT Retrospective Retrospective Retrospective Retrospective RCT
All thoracic locations Middle and lower thoracic locations All thoracic locations Upper thoracic location All thoracic locations All thoracic locations All thoracic locations All thoracic locations Lower thoracic location All thoracic locations All thoracic locations All thoracic locations All thoracic locations
Neo-adjuvant Therapy Two-Field 30 None 11 None 10 None 7 Unmentioned Unmentioned Unmentioned 2 Unmentioned Unmentioned None
Adjuvant Therapy
Pathological Subtype
pT-Stage
SCC Undefined Undefined SCC SCC SCC Mixed Mixed SCC SCC Mixed SCC SCC
T1-4 T0-4 T0-4 T0-4 T1-4 T0-4 T1-4 T0-4 T0-4 T1-4 T0-4 T0-4 T1-3
Total Number of Patients
Three-Field
Two-Field
Three-Field
Two-Field
Three-Field
52 None 15 None 22 None 0 Unmentioned Unmentioned Unmentioned 15 Unmentioned Unmentioned None
110 Unmentioned None 15 12 Unmentioned 19 Unmentioned 15 Unmentioned 21 Unmentioned Unmentioned 28
235 Unmentioned None 38 19 Unmentioned 35 Unmentioned 26 Unmentioned 89 Unmentioned Unmentioned 28
1176 123 56 100 34 78 65 86 73 55 65 121 290 30
1203 129 33 27 57 68 63 66 77 101 176 100 274 32
RCT ¼ randomized controlled trial; Unmentioned ¼ no pathological subtype described in the article or no neo-adjuvant therapy or adjuvant therapy mentioned in the article; Mixed ¼ a lower percentage of adenocarcinoma and other types of carcinoma included; SCC ¼ squamous cell carcinoma; None ¼ no patients received neo-adjuvant therapy or adjuvant therapy.
included both clinical leakage and subclinical leakage (only seen radiologically). Vocal cord paralysis included both persistent hoarseness of voice and vocal cord paralysis determined by indirect laryngoscopy. Pulmonary complications mainly included pneumonia and respiratory failure. When possible, the exact number of a specific complication was identified, otherwise the article was not included in the analysis (for that specific complication). We did not review adenocarcinoma and squamous cell carcinoma separately because the ratio of adenocarcinoma was very low in this analysis.
Statistical Analysis Studies were divided into the following two groups: randomized controlled trials and retrospective comparative trials when analyzing the primary outcome. Overall hazard ratio (HR) and its 95% confidence interval (CI) for the 5-year survival rate were estimated using a weighted
mean of HRs from the studies with each weight proportional to the reciprocal of the corresponding variance estimated by the Mantel-Haenszel procedure. The HR and its variance were extracted and estimated using the methods reported by Parmar and colleagues [10]. The statistical tests were performed with the STATA 11.0 software (StataCorp, College Station, TX). The secondary outcomes including perioperative mortality and morbidities were entered into the Cochrane Review Manager (RevMan) 5.1 (Cochrane Collaboration, Oxford, UK). Relative risk with the corresponding CI was the principal measure of effect. A p value less than 0.05 was considered statistically significant. Statistical heterogeneity among studies was evaluated with Cochrane’s Q statistic. Data were pooled using the fixed effect model. If statistical heterogeneity was suspected, the random effect model was applied. An estimation of potential publication bias was executed by Egger’s test [11]. A p value less than 0.05 was
GENERAL THORACIC
Ann Thorac Surg 2013;96:1933–42
GENERAL THORACIC
1936
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
Ann Thorac Surg 2013;96:1933–42
Fig 2. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the 5-year survival outcome for all included studies. (CI ¼ confidence interval.)
considered representative of statistically publication bias.
significant
Results Our initial review identified 3,048 studies, 3,032 of which were discarded after title and abstract review because they did not meet inclusion criteria. The full texts of the remaining 16 studies were further examined. Three of these studies were eliminated after full-text eligibility: two studies containing patients duplicated with other studies, and one study lacking the survival curve in the article. Thus, 13 articles involving a total of 2,379 patients were included in this analysis. There were two randomized controlled studies, one prospective nonrandomized study, and 10 retrospective studies (Fig 1; Table 1) [12-24].
Five-Year Survival Analysis When the survival data of the 13 articles were combined, the HR for the 5-year survival rate (defined as the primary
Fig 3. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the 5-year survival outcome for the two randomized controlled trials. (CI ¼ confidence interval.)
outcome), expressed as three-field lymphadenectomy versus two-field lymphadenectomy, was 0.64 (95% CI: 0.56 to 0.73, p ¼ 0.000; Fig 2), which indicated that three-field lymphadenectomy provided a better 5-year survival rate. When the data of the two randomized trials were extracted and analyzed, the HR was 0.56 (95% CI: 0.34 to 0.78, p ¼ 0.000; Fig 3), which also showed that three-field lymphadenectomy supplied a favorable survival outcome. Test for homogeneity of the HR did not show statistically significant heterogeneity concerning 5-year survival rate of all articles and one of the two randomized controlled trials (p ¼ 0.153 and 0.501, respectively). And the Egger’s test showed there was no publication bias in this analysis (p ¼ 0.237).
Mortality and Morbidity Analysis As to the secondary outcome in this study, eight studies showed the data of perioperative mortality. The difference of perioperative mortality between threefield and two-field lymphadenectomies was statistically insignificant (risk ratio [RR] 0.64; 95% CI: 0.38 to
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
1937
Fig 4. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the perioperative mortality. (CI ¼ confidence interval; M-H ¼ Mantel-Haenszel.)
1.10, p ¼ 0.110; Fig 4). Nine studies respectively provided the specific number of anastomotic leakage and pulmonary complications after esophagectomy. Three-field lymphadenectomy had a higher prevalence of anastomotic leakage (RR 1.46; 95% CI: 1.19 to 1.79, p ¼ 0.000; Fig 5) and a similar prevalence of pulmonary complications (RR 1.00; 95% CI: 0.89 to 1.12, p ¼ 0.760; Fig 6). In addition, eight studies provided the specific data of postoperative vocal cord palsy, and the prevalence was comparative between the two groups (RR 1.12; 95% CI: 0.82 to 1.54, p ¼ 0.470; Fig 7; Table 2).
Subgroup Analysis We did the subgroup analysis for the effect of three-field and two-field lymphadenectomy for esophageal carcinoma with pathologically positive lymph nodes. Four studies assessed the outcome of the two surgical procedures for patients with lymph nodes metastases. The HR for 5-year survival rate was 0.39 (95% CI: 0.31 to 0.46, p ¼ 0.000; Fig 8A), which favored three-field lymphadenectomy. Especially, for tumors with positive intrathoracic recurrent nerve nodes, the HR was 0.32 (95% CI: 0.23 to 0.41, p ¼ 0.000; Fig 8B), which favored the
more invasive lymphadenectomy as well. Furthermore, we assessed the effect of the two lymphadenectomies for different tumor locations. For upper/middle thoracic esophageal cancer with pathologically positive lymph nodes, the HR for the 5-year survival rate was 0.56 (95% CI: 0.34 to 0.78, p ¼ 0.000; Fig 8C). And for lower thoracic esophageal cancer with positive lymph nodes, the HR was 0.53 (95% CI: 0.42 to 0.64, p ¼ 0.000; Fig 8D). Nevertheless, for upper/middle thoracic esophageal cancer without lymph nodes metastases, the HR was 0.53 (95% CI: 0.01 to 1.04, p ¼ 0.044; Fig 8E). These results indicated that three-field lymphadenectomy could be favorable for esophageal cancer with positive lymph nodes regardless of the tumor location. However, for lower thoracic esophageal cancer without lymph nodes metastases, two-field lymphadenectomy could provide a 5-year survival rate similar to that of three-field dissection.
Comment The optimal type of lymphadenectomy for thoracic esophageal cancer is controversial. Only two randomized controlled trials have been published, involving only 212
Fig 5. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the postoperative anastomotic leakage. (CI ¼ confidence interval; M-H ¼ Mantel-Haenszel.)
GENERAL THORACIC
Ann Thorac Surg 2013;96:1933–42
GENERAL THORACIC
1938
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
Ann Thorac Surg 2013;96:1933–42
Fig 6. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the postoperative pulmonary complications. (CI ¼ confidence interval; M-H ¼ Mantel-Haenszel.)
patients (of a total of 2,379 patients in the articles reviewed [Table]). Meanwhile, the other studies are mostly retrospective analyses that described the investigators’ experiences, maybe reflecting the personal preference of the surgeons. The paucity of large-sample prospective studies yet provides little evidence to answer this question. Therefore, we did this study that included all published articles comparing the two types of lymphadenectomies. When all data were combined, the 5-year survival rate showed a significant difference between three-field and two-field lymphadenectomies. And three-field lymphadenectomy provided a better 5-year survival rate (Fig 2), which suggested that the addition of cervical lymph node dissection improved the longterm outcome over two-field lymphadenectomy alone, especially for esophageal cancer with positive lymph nodes (Fig 8A). Moreover, three-field lymphadenectomy interestingly presented a similar perioperative mortality compared with two-field dissection (Fig 4), whereas it used to be regarded as a more invasive procedure that was always associated with a higher mortality.
Fig 7. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the postoperative vocal cord palsy. (CI ¼ confidence interval; M-H ¼ Mantel-Haenszel.)
Despite the favorable results that three-field lymphadenectomy presented a higher 5-year survival rate and comparable perioperative mortality as two-field lymphadenectomy for thoracic esophageal cancer, some limitations in this study were worth noting. Firstly, the heterogeneity of these included studies was noticeable, though neither the test for homogeneity nor the Egger’s test for publication bias showed a statistical significance. Patients in these studies showed diverse clinicopathologic characteristics including various locations of the tumor and TN staging. It has been well supported that upper/middle thoracic esophageal cancer benefits much from three-field lymphadenectomy compared with lower lesions [7, 25]. In addition, different pathologic T stages have different ratios of cervical lymph nodes metastases, which determine the necessity of three-field lymphadenectomy [26]. Therefore, we did the subgroup assessment for esophageal cancer with positive lymph nodes and different tumor locations. Although results of the subgroup analyses were reasonable, the number of included patients was limited. Moreover, most articles in this study were from Japanese institutions. Considering that most Japanese surgeons prefer
Ten studies in this analysis provided the data of perioperative complications. Unmentioned indicates no relevant data provided in the study.
1939
cervical lymph nodes dissection for thoracic esophageal carcinoma, experiences from other countries are greatly needed, especially from Western colleagues. Furthermore, most studies were retrospective analyses, and the two parallel controlled groups of several studies were not well balanced. However, when the two randomized trials were analyzed, the results showed that three-field lymphadenectomy provided a better 5year survival as well (Fig 3), although conclusions from the two trials might be controversial because of the small sample sizes [1]. Three primary complications—anastomotic leakage, vocal cord palsy, and pulmonary complications (mainly pneumonia and respiratory failure)—were assessed in this study, because these frustrating issues were mainly attributed to three-field dissection and the overall number of complications in those studies was not precisely described. Unsurprisingly, three-field lymphadenectomy had a higher anastomotic leakage rate than two-field lymph node dissection (Fig 5). Anastomotic leakage is always linked with cervical lymph node dissection, probably partly due to the location of the anastomosis. The cervical anastomosis may carry a higher risk of leakage than the intrathoracic anastomosis, although management of the cervical leakage is much easier than the intrathoracic counterpart [27, 28]. In this case, three-field lymphadenectomy with intrathoracic anastomosis could be a promising choice for middle and lower thoracic esophageal cancer. Conversely, vocal cord paralysis from injury of the recurrent laryngeal nerve is another frequent complication of cervical dissection, although it is also mentioned after transthoracic anastomoses. A high incidence of vocal cord paralysis indicates that the recurrent nerve is mainly at risk during the cervical dissection and the construction of the anastomosis [8, 29]. The difference of the incidence of vocal cord palsy was statistically insignificant between the two groups (Fig 7). Moreover, the incidence of postoperative pulmonary complications was, interestingly, not higher in the three-field dissection group as well (Fig 6), which conflicted with the general idea that the more invasive lymphadenectomy might cause more pulmonary complications [30]. Therefore, meticulous surgical procedures with careful perioperative management may be a potential key for decreasing the incidence of the recurrent nerve injury and thus reducing the pulmonary complications of three-field lymphadenectomy. In summary, although three-field lymphadenectomy showed a higher incidence of anastomotic leakage than two-field lymphadenectomy did, it produced a higher 5-year survival rate, and comparable perioperative mortality and incidence of vocal cord palsy and pulmonary complications. Based on what was found in our study, three-field lymphadenectomy could be a priority for thoracic esophageal cancer, especially for tumors with positive lymph nodes. Large-sample randomized controlled studies are urgently needed for further evaluation.
GENERAL THORACIC
31.25 Unmentioned 7.4 33.33 9.09 10 18.75 1.47 8.77 31.25 13.85 Unmentioned 5 38.46 16.44 9.09 16.67 2.56 0 10.9 69.32 Unmentioned Unmentioned 69.84 14.29 15 56.25 2.94 28.07 8.91 55.38 Unmentioned Unmentioned 47.69 20.55 23.14 30 0 14.71 18.18 27.84 Unmentioned 22.22 33.33 33.77 43 6.25 1.47 15.79 38.61 20 Unmentioned 15 10.77 23.29 25.62 20 6.41 11.76 29.09 2.84 0 0 1.59 11.68 3 3.13 3.13 Unmentioned Unmentioned 6.15 1.79 6 3.08 2.74 10.74 6.67 6.67 Unmentioned Unmentioned
Three-Field Two-Field Three-Field Two-Field Three-Field Two-Field Study [Ref]
Fujita et al [21] Miyata et al [13] Fujita et al [14] Fujita et al [17] Kato et al [19] Kato et al [22] Nishihira et al [24] Noguchi et al [16] Young et al [15] Igaki et al [20]
Three-Field
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
Two-Field
p ¼ 0.760 p ¼ 0.470 p ¼ 0.000 p ¼ 0.110
Vocal Cord Palsy (%) Anastomotic Leakage (%) Perioperative Mortality (%)
Table 2. Incidence of Perioperative Complications Between Three-Field and Two-Field Lymphadenectomies in This Analysis
Pulmonary Complications (%)
Ann Thorac Surg 2013;96:1933–42
GENERAL THORACIC
1940
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
Ann Thorac Surg 2013;96:1933–42
Fig 8. Meta-analysis of the effect of three-field or two-field lymphadenectomy on the 5-year survival outcome for subgroups: (A) esophageal cancer with pathologically positive lymph nodes; (B) esophageal cancer with positive intrathoracic recurrent laryngeal nerve nodes; (C) upper/middle esophageal cancer with positive lymph nodes; (D) lower esophageal cancer with positive lymph nodes; and (E) upper/middle esophageal cancer without lymph nodes metastasis. (CI ¼ confidence interval.)
References 1. Gockel I, Sgourakis G, Lyros O, Hansen T, Lang H. Dissection of lymph node metastases in esophageal cancer. Expert Rev Anticancer Ther 2011;11:571–8. 2. Darling G. The role of lymphadenectomy in esophageal cancer. J Surg Oncol 2009;99:189–93. 3. Tachibana M, Kinugasa S, Hirahara N, Yoshimura H. Lymph node classification of esophageal squamous cell carcinoma and adenocarcinoma. Eur J Cardiothorac Surg 2008;34: 427–31. 4. Yajima S, Oshima Y, Shimada H. Neck dissection for thoracic esophageal squamous cell carcinoma. Int J Surg Oncol 2012;2012:750456. 2012 Feb 27 [E-Pub]. 5. Lerut T, Nafteux P, Moons J, et al. Three-field lymphadenectomy for carcinoma of the esophagus and gastroesophageal junction in 174 R0 resections: impact on staging, diseasefree survival, and outcome. A plea for adaptation of TNM classification in upper-half esophageal carcinoma. Ann Surg 2004;240:962–74. 6. Altorki N, Kent M, Ferrara C, Port J. Three-field lymph node dissection for squamous cell and adenocarcinoma of the esophagus. Ann Surg 2002;236:177–83.
7. Udagawa H, Ueno M, Shinohara H, et al. The importance of grouping of lymph node stations and rationale of three-field lymphadenectomy for thoracic esophageal cancer. J Surg Oncol 2012;106:742–7. 8. Mariette C, Piessen G. Oesophageal cancer: how radical should surgery be? Eur J Surg Oncol 2012;38:210–3. 9. Moher D, Liberati A, Tetzlaff J, Altman DGfor the PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009;6: e1000097. 10. Parmar MK, Torri V, Stewart L. Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints. Stat Med 1998;17:2815–34. 11. Egger M, Davey SG, Schneider M, Minder C. Bias in metaanalysis detected by a simple, graphical test. BMJ 1997;315: 629–34. 12. Shiozaki H, Yano M, Tsujinaka T, et al. Lymph node metastasis along the recurrent nerve chain is an indication for cervical lymph node dissection in thoracic esophageal cancer. Dis Esophagus 2001;14:191–6. 13. Miyata H, Yano M, Doki Y, et al. A prospective trial for avoiding cervical lymph node dissection for thoracic esophageal cancers, based on intra-operative genetic diagnosis of
14. 15.
16.
17.
18.
19. 20.
21.
micrometastasis in recurrent laryngeal nerve chain nodes. J Surg Oncol 2006;93:477–84. Fujita H, Kakegawa T, Yamana H, et al. Cervico-thoracoabdominal (3-field) lymph node dissection for carcinoma in the thoracic esophagus. Kurume Med J 1992;39:167–74. Shim YM, Kim HK, Kim K. Comparison of survival and recurrence pattern between two-field and three-field lymph node dissections for upper thoracic esophageal squamous cell carcinoma. J Thorac Oncol 2010;5:707–12. Noguchi T, Wada S, Takeno S, Hashimoto T, Moriyama H, Uchida Y. Two-step three-field lymph node dissection is beneficial for thoracic esophageal carcinoma. Dis Esophagus 2004;17:27–31. Fujita H, Kakegawa T, Yamana H, et al. Mortality and morbidity rates, postoperative course, quality of life, and prognosis after extended radical lymphadenectomy for esophageal cancer. Comparison of three-field lymphadenectomy with two-field lymphadenectomy. Ann Surg 1995;222:654–62. Tabira Y, Kitamura N, Yoshioka M, et al. Significance of three-field lymphadenectomy for carcinoma of the thoracic esophagus based on depth of tumor infiltration, lymph nodal involvement and survival rate. J Cardiovasc Surg (Torino) 1999;40:737–40. Kato H, Watanabe H, Tachimori Y, Lizuka T. Evaluation of neck lymph node dissection for thoracic esophageal carcinoma. Ann Thorac Surg 1991;51:931–5. Igaki H, Tachimori Y, Kato H. Improved survival for patients with upper and/or middle mediastinal lymph node metastasis of squamous cell carcinoma of the lower thoracic esophagus treated with 3-field dissection. Ann Surg 2004;239: 483–90. Fujita H, Sueyoshi S, Tanaka T, et al. Optimal lymphadenectomy for squamous cell carcinoma in the thoracic
YE ET AL LYMPHADENECTOMY FOR ESOPHAGEAL CANCER
22. 23. 24.
25. 26. 27.
28.
29. 30.
1941
esophagus: comparing the short- and long-term outcome among the four types of lymphadenectomy. World J Surg 2003;27:571–9. Kato H. Lymph node dissection for thoracic esophageal carcinoma. Two- and 3-field lymph node dissection. Ann Chir Gynaecol 1995;84:193–9. Akiyama H, Tsurumaru M, Udagawa H, Kajiyama Y. Radical lymph node dissection for cancer of the thoracic esophagus. Ann Surg 1994;220:364–73. Nishihira T, Hirayama K, Mori S. A prospective randomized trial of extended cervical and superior mediastinal lymphadenectomy for carcinoma of the thoracic esophagus. Am J Surg 1998;175:47–51. Jang HJ, Lee HS, Kim MS, Lee JM, Zo JL. Patterns of lymph node metastasis and survival for upper esophageal squamous cell carcinoma. Ann Thorac Surg 2011;92:1091–7. Stiles BM, Mirza F, Port JL, et al. Predictors of cervical and recurrent laryngeal lymph node metastasis from esophageal cancer. Ann Thorac Surg 2010;90:1805–11. Klink CD, Binneb€ osel M, Otto J, et al. Intrathoracic versus cervical anastomosis after resection of esophageal cancer: a matched pair analysis of 72 patients in a single center study. World J Surg Oncol 2012;10:159. Hu H, Ye T, Zhang Y, Zhang J, Luketich JD, Chen H. Modifications in retrosternal reconstruction after oesophagogastrectomy may reduce the incidence of anastomotic leakage. Eur J Cardiothorac Surg 2012;42:359–63. Hulscher JB, Tijssen JG, Obertop H, van Lanschot JJ. Transthoracic versus transhiatal resection for carcinoma of the esophagus: a meta-analysis. Ann Thorac Surg 2001;72:306–13. Fang W, Kato H, Tachimori Y, Igaki H, Sato H, Daiko H. Analysis of pulmonary complications after three-field lymph node dissection for esophageal cancer. Ann Thorac Surg 2003;76:903–8.
INVITED COMMENTARY The debate about the extent of lymph node dissection for cancer of the thoracic esophagus has occupied esophageal surgeons for decades. In the early 1980s, Japanese surgeons proposed that nodal dissection should be extended to a “third field” that includes the nodes along both recurrent nerves and a cervical node dissection. Despite promising results, wider implementation of the technique in the West was impeded by concerns about its seemingly higher morbidity and unproven survival benefit. The lack of adequately powered randomized trials or well-conceived metaanalyses sustained the controversy. In this issue of The Annals, Ye and colleagues [1] take an important step toward resolving the debate. They report the first metaanalysis of published data comparing the outcomes of esophagectomy with either two-field or three-field dissection. The results show that three-field dissection was associated with a significant improvement in survival and a significant reduction in hospital mortality. Surprisingly, there was no difference in the incidence of recurrent nerve palsy or pulmonary morbidity. So should this metaanalysis resolve the simmering controversy in favor of three-field dissection? A careful review of the 15 studies included in this metaanalysis would suggest that the debate is likely to Ó 2013 by The Society of Thoracic Surgeons Published by Elsevier Inc
continue for some time. The highest quality metaanalyses usually result from “best evidence synthesis” where only methodologically sound studies are included. This metaanalysis falls short of that ideal since 12 of the 15 cited studies are retrospective series whose inherent selection biases cannot be controlled for by the method. The investigators are obviously keenly aware of that shortcoming, which may have driven them to analyze the results of the two randomized trials separately. The fact that these results were concordant with the overall analysis is reassuring; however, neither of these studies is of sufficient power to justify changes in current practice. Another important question to consider is whether these results are generalizable to patients with adenocarcinoma of the esophagus, the most common type of esophageal cancer throughout most of Europe and North America. Finally, there is the intriguing finding by Ye and colleagues [1] that the rate of recurrent nerve palsy between the two techniques is comparable. That certainly has not been my experience or the experience of many other Western surgeons. The explanation may well be in what is precisely meant by the term “two-field” dissection. More current versions of two-field dissection include dissection of the recurrent nodes without dissection of 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2013.07.076
GENERAL THORACIC
Ann Thorac Surg 2013;96:1933–42