The incidence of adenocarcinoma of the esophagus

GENERAL THORACIC A Retrospective Analysis of Locally Advanced Esophageal Cancer Patients Treated With Neoadjuvant Chemoradiation Therapy Followed by or Alone Kenneth A. Kesler, MD, Paul R. Helft, MD, Elizabeth A. Werner, MS, Neel P. Jain, MS, Jo Ann Brooks, DNS, John M. DeWitt, MD, Julia K. Leblanc, MD, Naomi S. Fineberg, PhD, Lawrence H. Einhorn, MD, and John W. Brown, MD Departments of, Thoracic Division, and Medicine, Divisions of Oncology, Gastroenterology, and Biostatistics, Indiana University School of Medicine, Indianapolis, Indiana Accepted for publication Aug 23, 2004. Address reprint requests to Dr Kesler, Indiana University School of Medicine, Dept of, Thoracic Div, 545 Barnhill Dr, EH #215, Indianapolis, IN 46202; e-mail: kkesler@iupui.edu. Background. We conducted an institutional review of patients with locally advanced esophageal cancer who had complete pretreatment and surgical staging to identify variables predictive of outcome. Methods. From 1993 through 2002, 286 patients presented for surgical therapy of esophageal cancer. Of these, 176 patients met criteria for review including pretreatment endoscopic ultrasound stages IIA through IVA and a transthoracic surgical approach with twofield lymph node dissection. This cohort was primarily male (84.7%, n 149) with adenocarcinoma (88.6%, n 156), and 101 patients (57.3%) demonstrated endoscopic ultrasound stage III or IVA. Results. Eighty-five (48.3%) patients presented to surgery after receiving neoadjuvant chemoradiation therapy, and 91 (51.7%) underwent surgery alone. Both groups were well matched with respect to comorbidities and pretreatment stage. Patients receiving neoadjuvant chemoradiation demonstrated a nonsignificant trend toward increased operative mortality and nonfatal morbidity. The overall median survival was 16.8 months, and there was no survival difference comparing patients treated with neoadjuvant chemoradiation followed by surgery or surgery alone (p 0.82). The subset of 25 patients (29.4%) demonstrating a complete pathologic response after neoadjuvant chemoradiation therapy however had superior survival (median survival 57.6 months, p < 0.01) as compared with neoadjuvant chemoradiation patients demonstrating partial downstaging (n 36, 42.3%), no downstaging (n 24, 28.2%), and surgery alone patients. Multivariate analysis identified a complete pathologic response, endoscopic ultrasound stage, and number of pathologically positive lymph nodes as independent predictors of survival. Conclusions. These data support the use of neoadjuvant chemoradiation for locally advanced esophageal cancer as the subset of patients who demonstrate a complete pathologic response experienced significantly better survival. (Ann Thorac Surg 2005;79:1116 21) 2005 by The Society of Thoracic Surgeons The incidence of adenocarcinoma of the esophagus has been increasing during the last two decades, and esophageal cancer now represents the seventh leading cause of all cancer-related deaths in the US male population [1, 2]. It is a lethal disease, with less than 20% of all patients expected to survive beyond 5 years after diagnosis. Several factors contribute to this poor outcome, the most important of which is that the vast majority of patients demonstrate either locally advanced or metastatic disease at the time of diagnosis [3]. Neoadjuvant chemoradiation therapy followed by surgical extirpation of residual disease has become the most common treatment strategy for locally advanced disease despite the lack of a convincing phase III trial supporting this treatment approach. To further define the role and risks of neoadjuvant chemoradiation therapy, we conducted an institutional review of patients with locally advanced esophageal carcinoma who underwent complete pretreatment and surgical staging to identify variables predictive of outcome. Material and Methods From 1993 through 2002, 286 patients with adenocarcinoma or squamous cell esophageal cancer underwent surgical therapy at our institution. Of these patients, 110 patients were deemed not eligible for review owing to stage I disease (n 59), neoadjuvant chemotherapy without radiation therapy or non cisplatin/5-fluorouracil based neoadjuvant chemoradiation therapy (n 23), incomplete pretreatment or surgical staging (n 18), and surgery after failed definitive chemoradiation therapy (n 10). We identified a subset of 176 patients who met the criteria for review: pretreatment (clinical) stages IIA through IVA as determined by transesophageal ultrasound and chest and abdominal computed tomogra- 2005 by The Society of Thoracic Surgeons 0003-4975/05/$30.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2004.08.042

Ann Thorac Surg KESLER ET AL 2005;79:1116 21 LOCALLY ADVANCED ESOPHAGEAL CANCER Fig 1. Treatment of 176 patients reviewed during the study period. Solid bars represent patient numbers treated by neoadjuvant chemoradiation therapy followed by surgery. Shaded bars represent patient numbers treated with surgery alone. phy scans, a transthoracic surgical approach with en bloc two-field (celiac and periesophageal mediastinal) lymph node dissections, and treatment with either surgery alone (n 91, 51.7%) or surgery after standard neoadjuvant chemoradiation therapy as a planned strategy (n 85, 48.3%). Hospital records of eligible patients were reviewed, and follow-up information was obtained on an Institutional Review Board approved protocol. The tendency to treat locally advanced esophageal cancer patients with neoadjuvant chemoradiation therapy at our institution increased during the study period [4, 5]. Figure 1 demonstrates the trend in this series, with virtually all patients treated with surgery alone for the first 4 years, then a 2-year transition period where mainly better-risk patients with clinical stage III or IVA disease underwent neoadjuvant chemoradiation followed by surgery. Finally, during the last 4 years essentially all patients, including poorer-risk and clinical stage IIA or IIB disease patients were offered neoadjuvant chemoradiation therapy before surgery as a first treatment option. Standard cisplatin-based and 5-fluorouracil based chemotherapy regimens with concurrent radiation therapy (mean dose, 4,907 cgy) were used in 76 patients. Nine patients additionally received oral celecoxib as part of a prospective phase II protocol [6]. Eleven (12.9%) patients suffered grade III or IV toxicities from chemoradiation therapy, but only 2 patients failed to complete the entire course of planned chemoradiation. Twenty-one patients (24.7%) required enteric tube feeding for nutritional support during chemoradiation therapy. Repeat chest and abdominal computed tomography scans were obtained after chemoradiation to rule out interval evidence of systemic metastases. Patients who received neoadjuvant chemoradiation therapy underwent surgery after satisfactory hematologic and functional recovery at a median of 36 days (range, 26 to 237 days). Esophagogastric resections were planned to achieve a 6- to 8-cm proximal and 4- to 6-cm distal tumor-free margin when possible. The vast majority of patients (75.0%, n 132) underwent an Ivor Lewis approach (staged laparotomy then right thoracotomy), with 18.2% (n 32) undergoing a three-port approach (staged right thoracotomy then laparotomy and neck), and 6.8% (n 12) underwent an approach through a single left thoracoabdominal incision. Esophageal resections included en bloc removal of American Thoracic Society lymph node stations numbers 17, 20, 7, 8M, 8L, 9, 15, and 16 [7]. An average of 12 lymph nodes (range, 4 to 42) was harvested from the surgical specimens during pathologic evaluation. The stomach, with pyloroplasty, was used for esophageal replacement in 96.0% (n 169), with the remaining patients undergoing reconstruction with either large or small intestine. Patients were maintained on a nothingby-mouth basis for 7 days postoperatively, at which time a contrast esophagram was obtained before resuming oral nutrition. Multiple variables were recorded and shown in Table 1. For statistical analysis, clinical and pathologic stages III and IVA were combined, as the IVA designation was introduced in 1997 by the American Joint Committee on Cancer and therefore not discernable on earlier transesophageal endoscopic ultrasound or surgical pathology reports [8]. Statistical comparisons between neoadjuvant chemoradiation therapy followed by surgery and surgery alone groups were done with the Fisher exact test and independent Student s t tests for discrete and continuous variables, respectively. If continuous variables were not normally distributed, the Mann-Whitney U test was used Table 1. Preoperative Demographics and Clinical Staging Variable Alone (n 91) 1117 NCR and (n 85) p Value a Sex (M/F) 74/17 75/10 NS Age (y) 61.5 10.6 57.8 9.7 0.02 BMI (kg/m 2 ) 29.4 6.3 27.2 5.8 NS Weight loss (kg) 6.6 6.6 7.3 9.9 NS No. of comorbidities b 1.0 0.9 1.0 0.9 NS Pathologic diagnosis 80/11 76/9 NS (adenocarcinoma/squamous) EUS N stage NS N0 28 (30.8) 21 (24.7) N1 62 (69.2) 65 (75.3) N1 FNA confirmed 16 (18.8) 29 (34.1) EUS stage NS IIA/IIB 42 (46.1) 30 (35.3) III/IVA 49 (53.9) 55 (64.7) Preoperative Hemoglobin (g/dl) 13.6 2.0 12.3 1.4 0.01 White blood cell ( 10 3 ) 7.5 2.3 5.5 2.2 0.01 Creatinine (mg/dl) 1.0 0.3 1.0 0.2 NS a Not significant (NS) if p 0.10. b Comorbidities include chronic obstructive pulmonary disease, ethanol abuse, cardiac disease, peripheral vascular disease, cerebrovascular disease, and diabetes. Continuous variables expressed as mean standard deviation; discrete variables given as number of patients with percentage of subset in parentheses. BMI body mass index; EUS transesophageal endoscopic ultrasound; FNA fine-needle aspiration; NCR neoadjuvant chemoradiation therapy; NS not significant. GENERAL THORACIC

GENERAL THORACIC 1118 KESLER ET AL Ann Thorac Surg LOCALLY ADVANCED ESOPHAGEAL CANCER 2005;79:1116 21 and medians reported. Univariate assessments of discrete risk factors possibly predictive of survival were made using the Kaplan-Meier method with log-rank tests. For continuous risk factors, the Cox proportional hazard model was used. Multivariate Cox hazard modeling of patient survival was done using risk factors found to be significant (p 0.05) by univariate analysis. Results Table 2. Type of Resection, Final Pathologic Diagnosis, and Long-Term Outcome Variable Alone (n 91) NCR and (n 85) p Value a Type of resection NS R0 84 (92.3) 83 (97.6) R1 6 (6.6) 2 (2.4) R2 1 (1.1) 0 (0.0) Pathologic stage 0.01 0/I 2 (2.2) 34 (40.0) IIA/IIB 25 (27.5) 30 (35.3) III/IVA/IVB 64 (70.3) 21 (24.7) Pathologically positive 0.01 lymph nodes 0 19 (20.9) 52 (61.2) 1 3 29 (31.9) 21 (24.7) 4 43 (47.2) 12 (14.1) Site of first recurrence.06 Local/regional 27 (45.8) 7 (23.3) Distant 23 (38.9) 19 (63.3) Both 0 (0.0) 1 (3.3) Unknown 9 (15.2) 3 (10.0) Status at last follow-up Alive and well 19 (21.6) 38 (48.7) Alive with disease 3 (3.4) 2 (2.6) Died of disease 57 (64.7) 29 (37.1) Died nondisease or unknown 9 (10.2) 9 (11.5) a Not significant (NS) if p 0.10. Number of patients given with percentage of subset in parentheses. NCR neoadjuvant chemoradiation therapy; NS not significant. The cohort was primarily male (84.7%, n 149) at a mean age of 59.8 years with adenocarcinoma as the predominant histology type (88.6%, n 156). One hundred one patients (57.3%) demonstrated transesophageal endoscopic ultrasound stage III or IVA before treatment. Patients had an average of 1.0 comorbid risk factors, with 30.7% (n 54) having two or more comorbid risk factors before treatment. Patients undergoing neoadjuvant chemoradiation therapy had a younger mean age (p 0.02) and a slight trend toward higher percentage of clinical stage III or IVA disease as compared with patients undergoing surgery alone (p 0.29). Patients who received neoadjuvant chemoradiation therapy did present to surgery with significantly lower mean hemoglobin level and white blood cell count (Table 1). Both groups were otherwise well matched with respect to comorbid risk factors. The overall operative mortality was 5.1% (n 9). There was a trend toward increased operative mortality in patients receiving neoadjuvant chemoradiation therapy followed by surgery as compared with patients treated with surgery alone (n 6, 7.1% versus n 3, 3.3%; p 0.26). Seven patients died of complications from acute respiratory distress syndrome or pneumonia, including all 3 patients in the surgery alone cohort and 4 patients treated with neoadjuvant chemoradiation therapy followed by surgery. Three patients treated with neoadjuvant chemoradiation therapy died secondary to other postoperative complications (myocardial infarction, cerebrovascular accident, and anastomotic leak). Forty patients (24.0%) experienced nonfatal operative morbidity with respiratory complications predominating. There was also increased nonfatal postoperative morbidity in patients receiving neoadjuvant chemoradiation therapy, which was not statistically significant (p 0.58). Nineteen patients (22.7%) who underwent surgery alone experienced operative morbidity. Seventeen patients exhibited acute respiratory distress syndrome or pneumonia, 2 patients were treated for pulmonary embolism, and 1 patient had transient renal failure. Twenty-one patients (26.9%) who underwent neoadjuvant chemoradiation therapy (including 19 patients demonstrating acute respiratory distress syndrome or pneumonia) experienced nonfatal morbidity. Three of these patients exhibited transient renal failure, 2 patients had nonfatal myocardial infarction, and 1 patient was treated for a pulmonary embolus. Two patients, one in both groups, experienced anastomotic leak requiring reoperation and both survived. Finally, in comparing patients treated with surgery alone and patients treated with neoadjuvant chemoradiation therapy followed by surgery, there was no significant difference in mean blood loss (542 395 ml versus 498 338 ml), mean transfusion requirements (2.5 1.7 U versus 2.1 1.0 U), or median length of hospital stay in surviving patients (14 versus 13 days). One hundred sixty-seven patients (94.9%) demonstrated pathologic tumor-free margins (R0 resection; Table 2). Eight patients (4.5%) had microscopic positive margins on final pathologic examination (R1 resection), and 1 patient had residual gross disease adjacent to the thoracic aorta (R2 resection). Despite a higher percentage of clinical stage III or IVA disease, patients receiving neoadjuvant chemoradiation therapy demonstrated earlier stage by final pathologic analysis including the number of pathologically positive lymph nodes, representing a significant downstaging effect of preoperative treatment (Table 2). Overall, 61 patients (71.7%) who received chemoradiation therapy demonstrated lower pathologic T or N stages as compared with their respective pretreatment clinical stages. As a comparison, 15 patients (16.5%) treated with surgery alone did have clinical T or N overstaging. Conversely, 30 (32.9%) of the patients treated with surgery alone demonstrated higher pathologic T or N stages as compared with their respective

Ann Thorac Surg KESLER ET AL 2005;79:1116 21 LOCALLY ADVANCED ESOPHAGEAL CANCER Fig 2. Survival based on clinical (transesophageal endoscopic ultrasound) staging. Stage IIA is represented by a solid line (median survival, 26.4 months; 95% confidence interval, 14.4 to 38.4 months), IIB is represented by a dashed line (median survival, 22.8 months; 95% confidence interval, 4.8 to 42.0 months), and clinical III/IVA is represented by a dotted line (median survival, 13.2 months; 95% confidence interval, 8.4 to 18.0 months). Patient numbers at risk per year are given. 1119 response, or patients undergoing surgery alone (p 0.01; Fig 4). Despite statistically similar overall survival, there was a trend (p 0.06) for patients treated with surgery alone to demonstrate local or regional disease as the first identifiable site of recurrence compared with patients receiving chemoradiation therapy who more frequently demonstrated systemic metastases as the first site of recurrence (Table 2). Finally, there were three significant break points with respect to the number of lymph nodes pathologically containing metastatic disease and survival. Seventy-one patients (40.3%) demonstrated no pathologic lymph node metastases. These patients experienced a median survival of 33.6 months, which was significantly longer (p 0.01) than either the 50 patients (28.4%) who demonstrated from one to three pathologic lymph nodes or the 55 patients (31.3%) who demonstrated four or more lymph node metastases (p 0.01 versus one to three positive lymph nodes; Fig 5). Multivariate analysis demonstrated three independent predictors of survival including a pcr to chemoradiation therapy (pcr versus rest, hazard ratio 0.34; 95% confidence interval, 0.13 to 0.87; p 0.02), clinical stage category (stages IIA and IIB versus III/IVA, hazard ratio 1.34; 95% confidence interval, 1.05 to 1.71; p 0.01), and the number of lymph nodes containing metastatic disease on final pathologic examination (0 versus 1 to 3 versus 4, hazard ratio 1.53; 95% confidence interval, 1.18 to 1.98; p 0.01). GENERAL THORACIC preoperative clinical stages, whereas only 8 patients (9.4%) receiving neoadjuvant therapy were found to have higher pathologic T or N stages. Of the patients who underwent neoadjuvant chemoradiation therapy, 24 patients (28.2%) demonstrated no T or N downstaging, 36 patients (42.3%) demonstrated partial T or N downstaging, and 25 patients (29.4%) demonstrated a complete pathologic response (pcr). No pretreatment variables were significantly predictive of a pcr; however, trends were identified in the subset of pcr patients, including more patients with squamous cell carcinoma histology as compared with adenocarcinoma (55.6%, n 5 versus 26.7%, n 20; p 0.11) and clinical stage IIA or IIB as compared with clinical stage III or IVA (41.3%, n 12 versus 23.6%, n 13; p 0.09) [9]. The overall median survival was 16.8 months. Patients demonstrating clinical stages IIA and IIB had significantly better survival than patients with clinical stages III and IVA (p 0.01; Fig 2). There was no significant survival difference in patients treated with surgery alone as compared with patients treated with neoadjuvant chemoradiation therapy followed by surgery (p 0.82; Fig 3). The subset of patients who received preoperative chemoradiation therapy and demonstrated a pcr, however, had significantly improved survival compared with patients who demonstrated no response or a partial Fig 3. Survival based on treatment. Patients undergoing surgery alone are represented by a solid line (median survival, 16.8 months; 95% confidence interval, 12.0 to 21.6 months) and patients receiving neoadjuvant chemoradiation therapy (NCR) before surgery are represented by a dotted line (median survival, 15.6 months; 95% confidence interval, 9.6 to 20.4 months). Patient numbers at risk per year are given.

GENERAL THORACIC 1120 KESLER ET AL Ann Thorac Surg LOCALLY ADVANCED ESOPHAGEAL CANCER 2005;79:1116 21 Fig 4. Survival based on response to chemoradiation therapy. Patients achieving a complete pathologic response are represented by a thick solid line (median survival, 57.6 months; 95% confidence interval, 8.4 to 115.2 months), patients having no response are represented by a dotted line (median survival, 16.8 months; 95% confidence interval, 6.0 to 28.8 months), and patients demonstrating either a clinical T or N downstaging are represented by a dashed line (median survival, 13.2 months; 95% confidence interval, 7.2 to 20.4 months). Survival of patients treated with surgery alone is included for comparison and represented by a thin solid line. Patient numbers at risk per year are given. The survival advantage conferred by clinical to pathologic downstaging without achieving a pcr is more controversial. Rice and coworkers [13] reviewed the experience at the Cleveland Clinic including 69 patients with clinical N disease undergoing surgery after neoadjuvant chemoradiation and reported a survival advantage to 37 patients who demonstrated N downstaging. Unfortunately the majority of patients in our series who received neoadjuvant chemoradiation therapy demonstrated either no or partial clinical to pathologic downstaging with survival similar to patients undergoing surgery alone. This finding probably explains why only transesophageal endoscopic ultrasound stages were independently predictive of survival and final pathologic stage was not predictive in our series. Further study is warranted to determine whether patients who do not achieve a pcr but demonstrate a lower pathologic stage as compared with pretreatment stage retain a prognosis based on their clinical stage or have an upgraded prognosis based on the final pathologic stage. Definitive chemoradiation with surgery reserved only as a salvage treatment for patients who do not achieve pcr has also been proposed as a legitimate treatment option [17]. Two recent European phase III trials have demonstrated no survival difference in patients receiving definitive chemoradiation therapy as compared with patients undergoing preoperative chemoradiation therapy Comment Patients achieving a pcr after neoadjuvant chemoradiation therapy are a select subset reported to have a favorable prognosis after surgery in several retrospective studies [10 13]. Ancoma and coworkers [14] demonstrated superior survival in 6 of 47 patients with squamous cell cancer who achieved a pcr after treatment on the experimental arm of a phase III trial. Forastiere and associates [15] reported on a phase II multimodality trial from the University of Michigan involving 43 patients with nearly equal numbers of patients demonstrating adenocarcinoma and squamous cell histology. Patients with pcr had a median survival of 70 months as compared with only 26 months for patients with residual cancer in the surgical specimen. The percentage of patients who achieve a pcr with most chemoradiation regimens unfortunately remains low. A recent metaanalysis of nine randomized phase III trials comparing the results of surgery alone with neoadjuvant chemoradiation therapy followed by surgery reported an average pcr rate of only 21% [16]. Despite the retrospective nature of these data, on the basis of the outcome of the pcr subset alone, we are in agreement with Rice and colleagues [13] that a randomized phase III trial, if appropriately powered, would likely demonstrate significant survival advantage to neoadjuvant chemoradiation followed by surgery treatment as compared with surgery alone. Fig 5. Survival based on the number of pathologically positive lymph nodes (LN). Patients without any pathologic positive lymph nodes are represented by a solid line (median survival, 33.6 months; 95% confidence interval, 0.0 to 72.0 months), patients who demonstrated one to three positive lymph nodes are represented by a dashed line (mean survival, 21.6 months; 95% confidence interval, 14.4 to 27.6 months), and patients who demonstrated four or more lymph node metastases are represented by a dotted line (median survival, 9.6 months; 95% confidence interval, 8.4 to 12.0 months). Patient numbers at risk per year are given.

Ann Thorac Surg KESLER ET AL 2005;79:1116 21 LOCALLY ADVANCED ESOPHAGEAL CANCER followed by surgery [18, 19]. However, both of these studies enrolled patients with primarily squamous cell carcinoma histology, which by our data and other reported series produces a higher fraction of pcr patients after neoadjuvant therapy as compared with adenocarcinoma histology [13]. Moreover, there was relatively high mortality in the surgical arms of these studies, which perhaps reflects higher operative risks in patients with squamous cell cancer histology who would be anticipated to have more endorgan (eg, lung, liver) compromise in addition to poorer preoperative nutritional status as compared with patients presenting with adenocarcinoma. Until clinical staging methods can identify patients who achieve a pcr and, more importantly, until pcr patients are shown to demonstrate durable long-term survival without surgery, we believe surgery continues to remain an essential component of therapy for patients with good performance status and locally advanced adenocarcinoma. Several retrospective and prospective studies have cited additional operative risks, albeit minor, after neoadjuvant chemoradiation therapy for esophageal cancer [13, 16]. Patients receiving chemoradiation therapy would seem more likely to present to surgery with diminished nutritional, functional, and immunologic reserves. Our data demonstrate that patients receiving neoadjuvant chemoradiation therapy do present to surgery with significantly lower white blood cell and hemoglobin levels. The trend toward lower median survival in patients not achieving a pcr as compared with patients undergoing surgery alone may however not only be reflective of a higher fraction of clinical stage III or IVA patients in the neoadjuvant chemoradiation group, but a higher perioperative mortality rate after neoadjuvant therapy in our series. Selection bias inherent to any retrospective study is arguably best demonstrated by the younger mean age of our patients receiving neoadjuvant chemoradiation. The relatively higher operative mortality after neoadjuvant therapy underscores the need for patients to demonstrate good to excellent pretreatment performance status to minimize risks. Many issues remain controversial with respect to the optimal treatment strategy for locally advanced esophageal cancer. Significant subsets, such as the excellent survival in patients demonstrating a pcr in these data, would however support the use of neoadjuvant chemoradiation therapy as a planned approach in patients with good to excellent performance status. Risk models or biomarkers that can predict pathologic responses using current chemoradiation therapy to avoid the costs and potential added morbidity to patients who will not respond would appear to be valuable. Finally, it appears additionally imperative that novel neoadjuvant chemoradiation treatment strategies continue to be tested in phase II trials in an attempt to improve pcr rates. References 1. Blot WJ, McLaughlin JK. The changing epidemiology of esophageal cancer. Semin Oncol 1999;26(Suppl 15):2 8. 1121 2. Jemal A, Tiwari R, Murray T, et al. Cancer statistics, 2004. CA Cancer J Clin 2004;54:8 29. 3. Enzinger PC, Mayer RJ. Esophageal cancer. N Engl J Med 2003;349:2241 52. 4. Daly JM, Fry WA, Little AG, et al. Esophageal cancer: results of an American College of Surgeons patient care evaluation study. J Am Coll Surg 2000;190:562 72. 5. Visbal AL, Allen MS, Miller DL, Deschamps C, Trastek VF, Pairolero PC. Ivor Lewis esophagogastrectomy for esophageal cancer. Ann Thorac Surg 2001;71:1803 8. 6. Mantravadi RP, Fineberg N, Ansari R, et al. A phase II study of cisplatin, 5-fluorouracil, radiation therapy, and celecoxib in patients with respectable esophageal cancer: preliminary results from the Hoosier Oncology Group (abstract 1168). Proceedings of the American Society of Clinical Oncology 39th Annual Meeting 2003;22:291. 7. Casson AG, Rusch VW, Ginsberg RJ, Zankowicz N, Finley RW. Lymph node mapping of esophageal cancer. Ann Thorac Surg 1994;58:1569 70. 8. American Joint Committee on Cancer. AJCC cancer staging manual, 5th ed. Philadelphia: Lippincott-Raven, 1997:65 8. 9. Brooks JA, Werner EA, Jain NP, et al. A retrospective analysis to identify variables predictive of achieving a pcr after neoadjuvant chemoradiation therapy for locally advanced esophageal cancer (abstract 4061). Proceedings of the American Society of Clinical Oncology 40th Annual Meeting 2004;23:327. 10. 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;72:1918 25. 11. Yano M, Tsujinaka T, Shiozaki H, et al. Concurrent chemotherapy (5-fluorouracil and cisplatin) and radiation therapy followed by surgery for T4 squamous cell carcinoma of the esophagus. J Surg Oncol 1999;70:25 32. 12. Mandard AM, Dalibard F, Mandard JC, et al. Pathologic assessment of tumor regression after preoperative chemoradiotherapy of esophageal carcinoma: clinicopathologic correlations. Cancer 1994;73:2680 6. 13. Rice TW, Blackstone EH, Adelstein DJ, et al. N1 esophageal carcinoma: the importance of staging and downstaging. J Thorac Cardiovasc Surg 2001;121:454 64. 14. Ancona E, Ruol A, Santi S, et al. Only pathologic complete response to neoadjuvant chemotherapy improves significantly the long term survival of patient with respectable esophageal squamous cell carcinoma: final report of a randomized, controlled trial of preoperative chemotherapy versus surgery alone. Cancer 2001;91:2165 74. 15. 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;11:1118 23. 16. Urschel JD, Vasan H. A meta-analysis of randomized controlled trials that compared neoadjuvant chemoradiation and surgery to surgery alone for resectable esophageal cancer. Am J Surg 2003;185:538 43. 17. Urschel JD, Ashiku S, Thurer R, Sellke FW. Salvage or planned esophagectomy after chemoradiation therapy for locally advanced esophageal cancer a review. Dis Esophagus 2003;16:60 5. 18. Bedenne L, Pierre M, Bouch O, et al. Randomized phase III trial in locally advanced esophageal cancer: radiochemotherapy followed by surgery versus radiochemotherapy alone (FFCD 9102) (abstract 519). Proceedings of the American Society of Clinical Oncology 38th Annual Meeting 2002;21(1 of 2):130a. 19. Stahl M, Wilke H, Walz MK, et al. Randomized phase III trial in locally advanced squamous cell carcinoma (SCC) of the esophagus chemoradiation with and without surgery (abstract 1001). Proceedings of the American Society of Clinical Oncology 39th Annual Meeting 2003;22:130a. GENERAL THORACIC