Invasion of neighboring tissues (that is, trachea, aorta,

Transcription

1 Factors Predictive of Complete Resection of Operable Esophageal Cancer: A Prospective Study Christophe Mariette, MD, Laetitia Finzi, MD, Sylvain Fabre, MD, Jean-Michel Balon, MD, Isabelle Van Seuningen, PhD, and Jean-Pierre J. Triboulet, MD Service de Chirurgie Digestive et Générale Hôpital Claude Huriez and Unité INSERM 560 CHRU, Lille, France Background. Esophagectomy remains a standard treatment for patients with resectable esophageal cancer, but the 5-year survival is only 20% to 25%. After complete resection survival is significantly longer than after incomplete resection with microscopic or macroscopic penetration. The purpose of this study was to prospectively identify the factors predictive of complete resection of operable esophageal cancers. Methods. Betwen January 1995 and January 2002, 372 patients with esophageal cancer underwent surgery with curative intent. Complete resection was performed in 304 patients (81.7%), incomplete resection with microscopic penetration in 28 (7.5%), and incomplete resection with macroscopic penetration in 40 (10.8%). Univariate and multivariate analysis included 16 preoperative and operative factors. Results. Factors predictive of complete resection were absence of any modification of the esophageal axis on the barium swallow (p 0.019) and a partial or complete Invasion of neighboring tissues (that is, trachea, aorta, lung, pericardium) is common in cancer of the esophagus and is caused by anatomic proximity and lack of a protective serosa. Complete surgical resection is one of the most important prognostic factors [1 4]. Therefore it is fundamental to establish a precise assessment of tumor extension to avoid incomplete surgical resection and establish an appropriate therapeutic strategy. Despite recent advances in imaging techniques, the rate of incomplete resection has varied from 19.2% to 64.0% depending on the series [5]. In a retrospective study [6] we showed that complete resection of esophageal cancer was predictable. The purpose of this study was to prospectively identify factors predictive of complete resection (R0), as defined in the Union Internacional Contra la Cancrum (UICC) 1993 classification of operable esophageal cancer, to determine the appropriate pre-therapeutic workup for defining a population likely to benefit from surgical resection in terms of survival. Patients and Methods Patients Between January 1995 and January 2002, 372 patients (mean age, 58.1 years; standard deviation [sd], 9.8; range, Accepted for publication Jan 14, Address reprint requests to Dr Triboulet, Service de Chirurgie Digestive et Générale Hôpital, Claude Huriez CHRU, Place de Verdun, Lille Cedex 59037, France; jp-triboulet@chru-lille.fr. response to preoperative radiochemotherapy (p 0.042). Three groups of patients were identified: group 1 had no deviation of the axis on the barium swallow (n 253); group 2 had deviation of the axis on the barium swallow and partial or complete response to radiochemotherapy (n 66); and group 3 had deviation of the axis on the barium swallow and no response to radiochemotherapy or no preoperative treatment (n 53). Rates of complete resection were 90.1%, 74.2%, and 50.9%, and 5-year actuarial survivals were 46%, 37%, and 0%, respectively (p < 0.001). Conclusions. Complete resection of esophageal cancer is predictable. Deviation axis on the barium swallow and morphologic response to neoadjuvant radiochemotherapy are variables available for all patients at onset of therapeutic management. (Ann Thorac Surg 2003;75:1720 6) 2003 by The Society of Thoracic Surgeons 31 to 78 years) with cancer of the cervical or thoracic esophagus underwent surgery with curative intent in our digestive surgical unit of the Claude Huriez university hospital (Lille, France). The hospital files of these patients were prospectively noted. Patients were considered to be operable with resectable esophageal cancer after a complete pre-therapeutic workup (including physical examination; standard laboratory tests; ear, nose, and throat examination; panendoscopy under general anesthesia for squamous cell tumors; digestive fibroscopy and esogastroduodenal barium study; bronchial fibroscopy with biopsies; ultrasound exploration of the cervical and abdominal areas; computed tomographic (CT) scan of the thorax, mediastinum and abdomen; and endoscopic ultrasound). Criteria for nonresecability were adherence to the aorta ( 90 ), invasion of the pericardium, diaphragm, pleura, trachea-bronchi, azygos vein, recurrent nerve, tumor diameter ( 4 cm), celiac or subclavian lymph node enlargement, and visceral metastasis. Criteria for nonoperability were cirrhosis (any stage) associated with portal hypertension, respiratory failure, forced expiratory volume ( 1,000 ml/s), weight loss more than 20%, heart failure (New York Heart Association functional class III to IV). Surgical Approach The detailed resection techniques have been described elsewhere [1]. Surgical resection consisted of a transthoracic esophagectomy for tumor of the middle-third or 2003 by The Society of Thoracic Surgeons /03/$30.00 Published by Elsevier Inc PII S (03)

2 Ann Thorac Surg MARIETTE ET AL 2003;75: ESOPHAGEAL CANCER RESECTION lower-third of the esophagus, completed with a cervical incision for anastomosis in case of tumor of the upperthird of the thoracic esophagus. The surgical approach included an abdominal lymphadenectomy and an extended en bloc mediastinal lymphadenectomy (two-field lymphadenectomy). No cervical lymphadenectomy was undertaken. Abdominal lymphadenectomy comprised en bloc removal of all lymphatic tissue in the lower posterior mediastinum, in the left and right pericardial regions, along the lesser curve, and along the left gastric artery. A meticulous lymphadenectomy of the peritracheal, carinal, and left and right bronchial nodes was performed followed by en bloc resection of the thoracic duct together with the periaortic nodes. The nodes in the aortopulmonary window were removed but routinely a full dissection of the left recurrent laryngeal nerve chain was not carried out. For patients with respiratory insufficiency and limited tumor of the esophagus, a subtotal esophagectomy without thoracotomy was realized, with only an abdominal and mediastinal inferior lymphadenectomy. The esophagus was replaced by the stomach in 96.8% of patients, excepting patients with a history of gastric surgery or who required composite plasty for cancer of the remaining esophagus. Histopathological Analysis Assessment of the Removed Specimen and Lymph Nodes All nodal material was dissected separately from the specimen at the end of the procedure by the surgeon, and the resection specimen was assessed by an experienced pathologist according to the ptnm classification [7]. Squamous cell carcinoma (SCC) was found in 283 patients, adenocarcinoma in 79, and other tumoral types in 10. Tumors were well or moderately differentiated in 251 patients and poorly or undifferentiated in 121 patients. On histopathologic assessment of the 372 resected specimens, the distribution of pt and pn categories was as follows: the lesion was restricted to mucosa in 58 patients, submucosa in 88 patients, and muscularis propria in 51 patients; the lesion had invaded the adventitia in 152 patients (57% of these were adenocarcinoma) and the neighboring structures in 23 patients; and lymph node metastases were found in 180 patients. Tumoral stages were as follows: esophageal cancer stage I (ptispt1n0m0), n 109; stage IIA (pt2-t3n0m0), n 76; stage IIB (pt1-t2n1m0), n 60; stage III (pt3n1 or pt4nxm0], n 127. A mean (sd) of 19.8 lymph nodes (10.9) (range, 2 to 56) was dissected from each specimen by the surgeon and pathologist. The mean (sd) number of histopathologically positive lymph nodes was 1.9 (3.2) (range, 0 to 25). Variables Studied The preoperative and operative variables presented in Table 1 were used to search for factors predictive of complete resection. Tumor localization was determined by endoscopy and expressed as the distance from the upper pole of the tumor to the mandibular arcade (distance T-MA). A threshold distance of 25 cm was retained because this stratification level has been identified as a prognostic factor in esophageal cancer [6]. Siewert type 1 cancers of the cardia were included in the analysis. Patients with tumors of the hypopharynx or of the esophagus involving the upper esophageal sphincter were excluded. Deviation of the esophageal axis corresponded to a deviation from a virtual axis drawn through the middle of the esophageal lumen over its entire height (Fig 1). Local tumor extension was assessed before treatment by measuring the tumor height on the barium swallow ( 10 cm or 10 cm) and the maximal diameter of the tumor visualized on the CT scan, in accordance with the Wurtz classification as modified [2]. Endoscopic ultrasonography was used to assess the integrity of the different layers of the esophageal wall. Involvement of coronal and lesser curvature stomachal, pericardial, lateral esophageal, lateral aortic, intertracheobronchial, intertracheocaval, and subclavian nodes was considered to be regional extension. Invasion of the celiac or supraclavian nodes was considered to be metastatic. Node size was measured on the CT scan; nodes were considered invaded when the greater diameter measured greater than or equal to 10 mm. Endoscopic ultrasonography was used to study the size, sphericity, echostructure, and contour of the nodes that were suspected to be metastatic when well-delimited and round, with a homogeneous echostructure and a greater diameter ( 10 mm). Neoadjuvant radiochemotherapy was proposed for locally advanced tumors or within the framework of therapeutic protocols. The following morphologic criteria were used to assess response to neoadjuvant radiochemotherapy: Complete response: absence of identifiable tumor on plain roentgenograms or CT scan. Endoscopy was not performed routinely in all patients as negative biopsies can not rule out residual tumor fragments. Partial response: greater than or equal to 50% reduction in tumor size determined as the product of the largest dimensions of formations with two measurable dimensions or less than or equal to 30% reduction of the dimension measured for formations with one measurable dimension, without development of a new formation during treatment. Stability: less than 50% reduction or less than 25% progression of tumors with two measurable dimensions. Progression: more than 25% progression of measurable tumors or development of a new tumor Among the 174 patients who had preoperative radiochemotherapy, 104 (59.8%) exhibited morphologic partial response and 45 (25.9%) exhibited complete response. Postoperative radiochemotherapy or chemotherapy, or both, were proposed in all patients with incomplete resection with microscopic penetration (R1) or incomplete resection with macroscopic penetration (R2). Statistical Analysis The survival status of patients was ascertained in July Follow-up was complete for all 372 patients. Data are shown as prevalence or mean (standard deviation). Comparison of continuous data between groups was determined by the Student s t test and for ordinal data by GENERAL THORACIC

3 1722 MARIETTE ET AL Ann Thorac Surg ESOPHAGEAL CANCER RESECTION 2003;75: Table 1. Preoperative and Operative Parameters Used to Search for Factors Predictive of Complete Resection (RO) Variables N 372 % Age 60 yrs yrs Gender Male Female Malnutrition a No Yes Dysphagia No Yes Distance T-MA 25 cm cm Esophageal deviation on barium swallow No Yes Tumor height on barium swallow 10 cm cm Tumor diameter on CT 30 cm mm Lymph node invasion CT No Yes T (endoscopic ltrasonography (n 249) us T1 T us T N (endoscopic ltrasonography) (n 249) us No us N Histological type Squamous cell Adenocarcinoma Other Tumor differentiation Moderate to well Poor or none Treatment Surgery alone Neoadjuvant radiochemotherapy Response to radiochemotherapy Complete or partial Stability or progression Type of resection With thoracotomy Without thoracotomy a 10% weight loss, distance T-MA: distance between the upper pole of the tumor and the mandibular arcades. the 2 test or Fischer s exact test when appropriate. In analyzing survival time, we used the Statistical Package for Social Sciences (SPSS, Chicago, IL). The survival function has been estimated by the actuarial method without excluding the postoperative deaths. The log rank test was used for comparison of survival curves. To determine which of many covariates was the most significant risk factor regarding quality of surgical resection, we used the stepwise logistic regression model, adjusting all the covariates simultaneously. The 0.1 level was defined for entry into the model. Groups with different risks were identified using a stepwise procedure starting with the two most significant variables from the multivariate analysis. The model retained included two variables known at the time of diagnosis that would be potentially useful for determining resectability. Survival curves were compared between the different subgroups thus defined to establish sets of groups with significantly different risk patterns. Differences in the rate of R0 resection between the defined groups were analyzed with the 2 test. Differences were considered to be significant at 5% alpha risk. Results Postoperative Mortality and Morbidity The in-hospital mortality rate was 4.3% (n 16). Significant complications occurred in 142 patients (38.2%).

4 Ann Thorac Surg MARIETTE ET AL 2003;75: ESOPHAGEAL CANCER RESECTION 1723 interval was 27.9 months (20.8) (range, 6 to 91 months), with a median follow-up of 22 months. Median survival for the overall population was 40 months. Overall survival at 1, 3, and 5 years was 84%, 56%, and 40%, respectively. Median survival was 54.9 months in the R0 group and 10.8 months in the R1 and R2 group. The 1, 3, and 5-year survival rates were 93%, 65%, and 47% in the R0 group and 42%, 12%, and 4% in the R1 and R2 group, respectively (p 0.001). GENERAL THORACIC Univariate Analysis The results of the univariate analysis are presented in Table 2. Eight variables were found to be statistically related to the R0 resection group: (1) absence of dysphagia, (2) absence of esophageal deviation on barium swallow, (3) tumor height ( 10 cm), (4) tumor diameter ( 30 mm on CT scan), (5) tumor restricted to the muscularis propria at endoscopic ultrasonography, (6) no evidence of lymph node invasion on CT scan, (7) no evidence of lymph node invasion on endoscopic ultrasonography, and (8) total or partial response to preoperative radiochemotherapy. Multivariate Analysis Multivariate analysis (Table 3) identified two preoperative variables predictive of R0 surgical resection: (1) esophageal deviation on barium swallow (p 0.019), and (2) response to preoperative radiochemotherapy (p 0.042). These two variables were equally allocated among histologic subtypes. Fig 1. Deviation of the esophagus on the barium swallow. Deviation of the virtual axis representing the middle of the esophageal canal. Groups Three groups were constructed with two variables which were identified as predictive of R0 resection by multivariate analysis: (1) esophageal deviation on barium swallow and (2) response to preoperative radiochemotherapy (Fig 3). The first group was composed of 253 patients without axial deviation of the esophagus on barium swallow with or without preoperative radiochemotherapy, irrespective of tumor response to treatment. The second group was composed of 66 patients with axial deviation of the Thirty-three patients (8.9%) had anastomotic leaks that were observed clinically or by radiology. Respiratory complications were observed in 95 patients (25.5%). Type of Resection According to the 1993 UICC criteria, macroscopically and microscopically R0 resection was achieved in 304 patients (81.7%), R1 resection (histologic evidence of invasion of the section margin or lateral clearance) was achieved in 28 patients (7.5%), and R2 resection (macroscopic residual tumor after surgery) was achieved in 40 patients (10.8%). We assigned patients to two groups for analysis of factors predictive of complete resection: R0, patients who had complete R0 resection (n 304; 81.7%); R1-R2, patients who had incomplete R1 or R2 resection (n 68; 18.3%). Actuarial Survival The actuarial survival is shown in Figure 2. Follow-up was complete for all patients. The mean (sd) follow-up Fig 2. Actuarial survival for the entire population and of group R0 and group R1 and R2 (p 0.001). The number of subjects at risk at each interval is shown in the table at the bottom of the graph.

5 1724 MARIETTE ET AL Ann Thorac Surg ESOPHAGEAL CANCER RESECTION 2003;75: Table 2. Univariate Analysis: Factors Influencing Type of Resection Variables Group R 0 n (%) Group R1 R2 n (%) 2 p Age 60 y y Gender Male Female 32 8 Malnutrition a No Yes Dysphagia No Yes Distance T-MA b 25 cm cm Esophageal deviation on barium swallow No Yes Tumor height 10 cm cm Tumor diameter on computed tomography 30 mm mm Lymph node invasion on computed tomography No Yes T (endoscopic ultrasonography [us]) Us T1 T Us T N (endoscopic ultrasonography) Us N Us N Histologic type Squamous cell Adenocarcinoma Other 8 2 Tumor differentiation Moderate to well Poor or none Treatment Surgery alone Neoadjuvant radiochemotherapy Response to radiochemotherapy Complete or partial Stability or progression 17 8 Type of resection With thoracotomy Without thoracotomy a Malnutrition: 10% weight loss; b distance T-MA is the distance between the upper pole of the tumor and the mandibular arcades. esophagus on barium swallow who responded totally or partially to preoperative radiochemotherapy. The third group was composed of 53 patients with esophageal deviation who either did not respond to preoperative radiochemotherapy or were not given preoperative adjuvant therapy. Rates of R0 resection (Table 4) and survival (Fig 4) were significantly different between groups (p 0.001) and two-by-two (p 0.013). Comment Pre-therapeutic exploration of patients with esophageal cancer is hampered by difficulty determining the potential for curative resection before histologic examination of the surgical specimen [8]. In fact, the survival rates of noncurative surgery are essentially identical to those achieved with nonsurgical therapy using combined che-

6 Ann Thorac Surg MARIETTE ET AL 2003;75: ESOPHAGEAL CANCER RESECTION Table 3. Multivariate Analysis: Preoperative Factors Predictive of Complete Resection (RO) Variables p 2 Ratio Odds- 95% Confidence Interval Esophageal deviation on barium swallow No Yes 1.0 Response to neoadjuvant radiochemotherapy Complete or partial Stability or progression GENERAL THORACIC Fig 3. Group constitution according to the two factors predictive of complete resection: esophageal axis deviation on barium swallow and morphologic response to preoperative radiochemotherapy (RCT). moradiation [9]. Consequently, surgery can be retained as a valid option for patients with esophageal cancer, with the goal of improved 5-year survival only if R0 resection can be achieved [1 4]. This requirement was confirmed in our series of 372 patients in which median survival was 54.9 months in the R0 group and 10.8 months in the R1 and R2 group. The 5-year survival rates were 47% and 4% (p 0.001), respectively. In a modern series [3, 4, 10 13], 5-year survival rates after curative surgery in esophageal cancer ranged from 23% to 55%, including series with three-field lymph node dissection. Long-term survival has clearly improved since 1980, because of advances in surgical and anesthetic techniques, together with improvements in perioperative management, tumoral staging, and adjuvant therapy. Nevertheless, 5-year survival remains low. In order to establish what elements of the pretherapeutic workup enable identifying a population of patients who can benefit from surgical resection, we identified two preoperative variables predictive of R0 resection: (1) complete or partial response to preoperative radiochemotherapy and (2) absence of esophageal deviation on the barium swallow. Objective response to neoadjuvant radiochemotherapy [14, 15] or chemotherapy [16, 17] has been associated with better survival for patients with resectable esophageal carcinoma. Inversely, others have not demonstrated better survival with neoadjuvant radiochemotherapy [18, 19] or chemotherapy [20, 21]. Phase II trials have demonstrated that 20% to 40% of patients given preoperative radiochemotherapy achieve complete histologic response [22, 23]. Patients treated with surgery alone were more likely to have an esophageal resection, but those treated with chemotherapy and surgery were more likely to have R0 resection [16, 20]. Work by the French Federation of Digestive Tract Cancerology [15] demonstrated that clinical response, assessed by the World Health Organization radioendoscopic criteria, is overestimated in 10% of patients and underestimated in 29% compared with histologic response. This discordance between clinical and histologic response is well known, the predictive value of negative endoscopic biopsy for diagnosis of complete response is 21% to 47% [24]. Thus, new tools are necessary to assess the efficacy of nonsurgical treatments for esophageal cancers. Deviation of the esophageal axis on the barium swallow is associated with noncurative resection [5, 6], and it is also predictive of extension to neighboring organs [5]. The deviation of the esophageal axis is usually a typical sign of advanced squamous cell carcinoma and is rare in Barrett cancer. In our study, axis deviation was equally allocated among both histologic subtypes and only a few specimens showed Barrett mucosa. An explanation could be the importance of pt3 stage patients (57%) with adenocarcinoma. Surgical resection of esophageal cancer should only be undertaken to attempt curative treatment. The limitations of preoperative explorations to predict tumor resectability require considering other factors predictive of R0 resection. Using two variables predictive of R0 resection (ie, deviation of the esophageal axis on barium swallow and response to neoadjuvant radiochemo- Table 4. Rate of Complete Resection and Survival in the 3 Groups Identified With Multivariate Analysis Group Patients Complete Resection (R0) Median Survival Survival Rate (%) N % N % Months 1 Year 3 Years 5 Years Group 1: no deviation of the esophagus on barium swallow; group 2: deviation of the esophagus on barium swallow and partial or complete response to preoperative radiochemotherapy; and Group 3: deviation of the esophagus on barium swallow and either no response to radiochemotherapy or no preoperative treatment.

7 1726 MARIETTE ET AL Ann Thorac Surg ESOPHAGEAL CANCER RESECTION 2003;75: Fig 4. Survival by the two factors predictive of complete resection. Group 1: no deviation of the esophagus on barium swallow; group 2: deviation of the esophagus on barium swallow and partial or complete response to preoperative radiochemotherapy; group 3: deviation of the esophagus on barium swallow and either no response to radiochemotherapy or no preoperative treatment (p less than between all groups and p less than when compared individually). The number of subjects at risk at each interval is shown in the table at the bottom of the graph. therapy), enabled us to identify three groups of patients with significantly different rates of R0 resection. These two variables are available for all patients at the onset of therapeutic management allowing easy classification of the patients into three groups. For patients with 50% risk of R1 or R2 resection (group 3), we recommend preoperative radiochemotherapy for those with axis deviation on barium swallow, and in patients with no response, only palliation. Further oncology research are necessary to refine the patient selection. References 1. Mriette C, Castel B, Toursel H, Fabre S, Balon JM, Triboulet JP. Surgical management of and long-term survival after adenocarcinoma of the cardia. Br J Surg 2002;89: Mariette C, Maurel A, Fabre S, Balon JM, Triboulet JP. Preoperative prognostic factors for squamous cell carcinomas of the thoracic esophagus. Gastroenterol Clin Biol 2001;25: Orringer MB, Marshall B, Iannettoni MD. Transhiatal esophagectomy: clinical experience and refinements. Ann Surg 1999;230: Karl RC, Schreiber R, Boulware D, Baker S, Coppola D. Factors affecting morbidity, mortality, and survival in patients undergoing Ivor Lewis esophagogastrectomy. Ann Surg 2000;231: Sugimachi K, Watanabe M, Sadanaga N, et al. Pre-operative estimation of complete resection for patients with oesophageal carcinoma. Surg Oncol 1994;3: Mariette C, Fabre S, Balon JM, Finzi L, Triboulet JP. Factors predictive of complete resection of operable esophageal cancer: review of 746 patients. Gastroenterol Clin Biol 2002; 26: Sobon LH, Wittekind C, eds. UICC TNM Classification of Malignant Tumors, 5th ed. New York: Wiley-Liss, Matsubara T, Ueda M, Kokudo N, Takahashi T, Muto T, Yanagisawa A. Role of esophagectomy in treatment of esophageal carcinoma with clinical evidence of adjacent organ invasion. World J Surg 2001;25: Herskovic A, Martz K, Al-Sarraf M, et al. Combined chemotherapy and radiotherapy compared with radiotherapy alone in patients with cancer of the esophagus. N Engl J Med 1992;326: Visbal AL, Allen MS, Miller DL, Deschamps C, Trastek VF, Pairolero PC. Ivor Lewis esophagogastrectomy for esophageal cancer. Ann Thorac Surg 2001;71: Swanson SJ, Batirel HF, Bueno R, et al. Transthoracic esophagectomy with radical mediastinal and abdominal lymph node dissection and cervical esophagogastrostomy for esophageal carcinoma. Ann Thorac Surg 2001;72: Elias D, Lasser P, Hatchouel JM, et al. A multivariate prospective study of prognostic factors in 200 operated epidermoid cancers of the esophagus. Definition of patients for whom surgical resection was of benefit. Gastroenterol Clin Biol 1993;17: Nozoe T, Saeki H, Ohga T, Sugimachi K. Clinicopathologic characteristics of esophageal squamous cell carcinoma in younger patients. Ann Thorac Surg 2001;72: Le Prise E, Etienne PL, Meunier B, et al. A randomized study of chemotherapy, radiation therapy, and surgery versus surgery for localized squamous cell carcinoma of the esophagus. Cancer 1994;73: Bedenne L, Seitz JF, Milan C, et al. Cisplatin, 5-FU, and preoperative radiotherapy in esophageal epidermoid cancer. Multicenter phase II FFCD 8804 study. Gastroenterol Clin Biol 1998;22: Medical Research Council Oesophageal Cancer Working Party. Surgical resection with or without preoperative chemotherapy in oesophageal cancer: a randomised controlled trial. Lancet 2002;359: Ancona E, Ruol A, Santi S, et al. Only pathologic complete response to neoadjuvant chemotherapy improves significantly the long term survival of patients with resectable esophageal squamous cell carcinoma: final report of a randomized, controlled trial of preoperative chemotherapy versus surgery alone. Cancer 200;91: Bosset JF, Gignoux M, Triboulet JP, et al. Chemoradiotherapy followed by surgery compared with surgery alone in squamous-cell cancer of the esophagus. N Engl J Med 1997;337: Entwistle JW 3rd, Goldberg M. Multimodality therapy for resectable cancer of the thoracic esophagus. Ann Thorac Surg 2002;73: Urschel JD, Vasan H, Blewett CJ. A meta-analysis of randomized controlled trials that compared neoadjuvant chemotherapy and surgery to surgery alone for resectable esophageal cancer. Am J Surg 2002;183: Kelsen DP, Ginsberg R, Pajak TF, et al. Chemotherapy followed by surgery compared with surgery alone for localized esophageal cancer. N Engl J Med 1998;339: Forastiere AA, Orringer MB, Perez-Tamayo C, Urba SG, Zahurak M. Preoperative chemoradiation followed by transhiatal esophagectomy for carcinoma of the esophagus: final report. J Clin Oncol 1993;11: Fink U, Stein HJ, Bochtler H, Roder JD, Wilke HJ, Siewert JR. Neoadjuvant therapy for squamous cell esophageal carcinoma. Ann Oncol 1994;5(Suppl 3): Bates BA, Detterbeck FC, Bernard SA, Qaqish BF, Tepper JE. Concurrent radiation therapy and chemotherapy followed by esophagectomy for localized esophageal carcinoma. J Clin Oncol 1996;14: