Barrett s esophagus is a recognized predisposing condition

Transcription

1 Differential Response to Preoperative Chemoradiation and Surgery in Esophageal Adenocarcinomas Based on Presence of Barrett s Esophagus and Symptomatic Gastroesophageal Reflux Banke Agarwal, MD, Stephen G. Swisher, MD, Jaffer Ajani, MD, Kaitlyn Kelly, Ritsuko R. Komaki, MD, Emad Abu-Hamda, MD, Arlene M. Correa, PhD, and Jack A. Roth, MD Departments of Gastrointestinal Medicine and Nutrition, Medical Oncology, Radiation Oncology, and Thoracic and Cardiovascular Surgery, M. D. Anderson Cancer Center, Houston, Texas Background. Barrett s esophagus and gastroesophageal reflux disease (GERD) are recognized to predispose to esophageal adenocarcinoma. Abdel-Latif and colleagues recently suggested that esophageal adenocarcinoma patients with GERD might be resistant to multimodality treatment. In this study, we investigated potential differences in clinical outcomes in esophageal adenocarcinoma patients based on the presence of identifiable Barrett s mucosa and/or history of symptomatic GERD. Methods. Eighty-four patients with resectable esophageal adenocarcinoma, who completed the planned preoperative chemoradiation and underwent a potentially curative esophageal resection were retrospectively evaluated. Postoperative survival was compared between patients with or without underlying Barrett s esophagus and history of symptomatic GERD. Patients with pathologic complete response (path CR) and those with partial or no response (path PR) were compared to determine if presence of Barrett s esophagus and history of symptomatic GERD influence the path CR rates. Results. We found significantly lower postoperative survival in patients with Barrett s associated adenocarcinoma (vs adenocarcinoma arising de novo, p 0.031) and patients with symptomatic GERD (vs patients without symptomatic GERD, p 0.019). Furthermore, the subset of patients with path PR (vs path CR) after chemoradiation have a significantly higher proportion of patients with Barrett s esophagus (HR 4.38, confidence interval [CI] 1.39 to 13.83, p 0.012) and patients with GERD (HR 2.71, CI 1.13 to 6.50, p 0.026). Conclusions. Patients with esophageal adenocarcinoma may have differences in response to preoperative chemoradiation based on the presence of Barrett s esophagus and history of symptomatic GERD. (Ann Thorac Surg 2005;79: ) 2005 by The Society of Thoracic Surgeons Barrett s esophagus is a recognized predisposing condition for esophageal adenocarcinoma [1, 2] and therefore endoscopic surveillance is currently recommended in these patients [3 5]. Although two earlier studies have suggested that Barrett s associated adenocarcinomas tend to have better overall survival as they are diagnosed earlier than esophageal adenocarcinomas arising de novo, any differences in clinical behavior between the two groups after matching for tumor stage have not been reported [6, 7]. Abdel-Latif and colleagues [8] recently found differences in pathologic complete response (path CR) rates and survival in esophageal adenocarcinomas based on the expression of NF- B in the tumors. They further Accepted for publication Oct 18, Address reprint requests to Dr Agarwal, Therapeutic Endoscopy Division of Gastroenterology and Hepatology, St. Louis University School of Medicine, 3635 Vista Avenue, St. Louis, MO 63110; agarwalb@ slu.edu. demonstrated that NF- B expression increases in esophageal cancer cells in culture systems when they are exposed to bile acid (deoxycholic acid [DCA]) or low ph due to hydrochloric acid (HCl). The NF- B is a known inhibitor of apoptosis. Thus, esophageal adenocarcinoma patients with gastroesophageal reflux might be resistant to multimodality treatment due to upregulation of NF- B in their tumors. Surgery has been the basis of most treatment regimens for esophageal cancer. In an effort to further improve the outcomes, combination regimens with chemotherapy and/or radiation with surgery have been studied. Preoperative chemoradiation regimens are currently under intense study [9 15] and have shown considerable initial promise [9, 15, 16]. It has also been suggested that patients who achieve a path CR after preoperative chemoradiation may have best postoperative survival [17 19]. In the present study, involving esophageal adenocar by The Society of Thoracic Surgeons /05/$30.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg AGARWAL ET AL 2005;79: ESOPHAGEAL ADENOCARCINOMA AND PREOPERATIVE CHEMORADIATION 1717 cinoma patients treated with preoperative chemoradiation and surgery, we investigated if there were any differences in postoperative survival between Barrett s associated adenocarcinoma (vs adenocarcinoma arising de novo) and patients with or without symptomatic GERD. We also compared the patients with path CR and partial or no response (path PR) to identify patient and tumor characteristics that might significantly correlate with path CR. Material and Methods Patient Enrollment A retrospective analysis of patients from sequential phase II studies at MD Anderson Cancer Center (MDACC), evaluating preoperative chemoradiation regimens, was performed. All studies were approved by the institutional review board at MDACC. Patients were included if they satisfied all of the following criteria: (1) had previously untreated, biopsyproven adenocarcinoma confined to either thoracic esophagus or involving the gastroesophageal (GE) junction; (2) tumors were considered resectable before and after preoperative chemoradiation; (3) successfully completed full planned course of preoperative chemoradiation; and (4) underwent potentially curative esophageal resection. Patients with tumors that were involving distal esophagus and proximal stomach were not included if they were centered below the GE junction. Patients treated between March 1998 and January 2001 were evaluated. Before chemoradiation was initiated, a thoracic surgeon, a medical oncologist, a radiation oncologist, and a gastroenterologist jointly evaluated each patient to determine if the primary cancer was resectable and if the patient was medically fit for surgery. To be eligible for preoperative chemoradiation and surgery, the patients needed to have Zubrod performance status of 2 or less, a serum creatinine of less than 1.5 mg/dl, serum bilirubin levels less than 1.6 mg/dl, absolute granulocyte count greater than 1,500 cells/ L, and platelet count of greater than 100,000/ L. Patients with a T1N0 (treated with surgery alone) or T4N0 1 (considered unresectable) as determined by endoscopic ultrasound (EUS) staging were excluded. A tumor was also considered unresectable if malignant supraclavicular adenopathy, effusions, or suspected distant lesions were histologically confirmed to be malignant. Any patient with malignant involvement of the tracheobronchial tree was defined as having unresectable esophageal cancer. All patients gave informed consent and surgery alone or definitive chemotherapy was offered as the alternate therapy. Preoperative Assessment and Staging Patients had a complete history and physical assessment at the time of presentation to MDACC. Pretreatment tests included a chest radiograph, computed tomographic (CT) scan of the chest and abdomen, upper gastrointestinal endoscopy with endoscopic ultrasound, and double contrast upper gastrointestinal radiographs. Tumor staging was primarily based on EUS findings. Mediastinal and celiac lymph nodes were considered malignant if EUS showed enlarged lymph node ( 10 mm in diameter) or if the lymph nodes were round and hypoechoic with sharp margins. In the oncology clinic at MDACC, patients are considered to have a history of symptomatic GERD if they have symptoms of heartburn (chest pain improved with antacids or proton pump inhibitors) and reflux of gastric contents (food with sour or bitter taste in mouth) for several months or years before diagnosis is esophageal cancer. Recent onset (within last 3 months) of dyspeptic symptoms in patients presenting for initial work-up of esophageal cancer are not considered indicative of GERD. Study Design This study is based on a separate analysis of data from patients treated on sequential phase II clinical trials at MDACC evaluating different preoperative chemoradiation regimens for treatment of esophageal cancer. Preoperative chemoradiation comprised of either five weeks of concurrent chemoradiation alone or induction chemotherapy followed by 5 weeks of concurrent chemoradiation. Chemoradiation Protocols 1. Induction chemotherapy and concurrent chemoradiation: induction chemotherapy consisting of 2 cycles of paclitaxel (200 mg/m 2 ) on day 1, 5-FU (750 mg/m 2 ) on day 1 to 5, and cisplatin (15 mg/m 2 )on days 1 to 5 with second cycle 28 days later. Concurrent chemoradiation was administered 4 weeks later and consisted of 5 weeks of 45 Gy (1.8 Gy/Fx) in association with 5-FU (300 mg/m 2 ) Monday through Friday, and cisplatin (15 mg/m 2 ) on days 1 to 5 of radiation therapy. 2. Induction chemotherapy and concurrent chemoradiation: induction chemotherapy consisting of 2 cycles of irinotecan (70 mg/m 2 ) and cisplatin (15 mg/m 2 ) on days 1, 7, 21, and 28 with the second cycle 42 days later. Concurrent chemotherapy was administered 4 weeks later and consisted of 5 weeks of 45 Gy (1.8 Gy/Fx) in association with 5-FU (300 mg/m 2 ) Monday through Friday, and paclitaxel (45 mg/m 2 ) on Monday of each week of radiation therapy. 3. Concurrent chemoradiation alone: 5 weeks of 45 Gy (1.8 Gy/Fx) in association with 5-FU (300 mg/m 2 ) Monday through Friday and cisplatin (15 mg/m 2 ) on days 1 to 5 of radiation therapy. 4. Concurrent chemoradiation alone: 5 weeks of 45 Gy (1.8 Gy/Fx) in association with 5-FU (300 mg/m 2 ) Monday through Friday and Taxol (45 mg/m 2 )on Monday of each week of radiation therapy. All patients underwent esophageal resection 4 to 6 weeks after the completion of radiation therapy.

3 1718 AGARWAL ET AL Ann Thorac Surg ESOPHAGEAL ADENOCARCINOMA AND PREOPERATIVE CHEMORADIATION 2005;79: Preoperative Evaluation and Surgery Upper gastrointestinal barium radiographs were performed after each course of induction chemotherapy and just before surgery. Esophagogastroduodenoscopy, CT of the chest and abdomen, and chest radiography were performed prior to surgery. Four to six weeks after completion of preoperative chemoradiation patients were treated with surgical resection of the esophagus through a transhiatal or transthoracic approach, followed by reconstruction of the esophagus with a gastric conduit according to the surgeon s preference. Regional lymph nodes from the abdomen and chest were included in the surgical resection. Response to Chemoradiation Pathologic complete response was defined as the complete absence of tumor cells in the resection specimens. Patients with macroscopic residual tumor or microscopic foci of residual tumors were considered to have path PR. Follow-Up and Survival Follow-up data were obtained from the hospital records, telephone communication, and the MDACC tumor registry. Dates of treatment, time to relapse and location of relapse (local and distant), date of death, and cause of death were ascertained. Survival was calculated from the date of operation. Last contact was based on actual date of death from hospital records or confirmation of death from Social Security Death Index or telephone conversation with relatives, last documented hospital visit, or last telephone communication with the patient. The mean follow-up in the survivors was 25.6 months and the longest follow-up was 54.4 months. Statistical Analysis Statistical analysis was performed by our biostatistician (AMC). The 2 test and logistic regression analysis were use to test significance of observed differences in tumor characteristics in patients with path CR and path PR. Kaplan-Meier estimates and Cox regression analysis were used to determine differences in postoperative survival. The SPSS statistical software was used for analysis (SPSS Inc, Chicago, IL). All analyses were twotailed and statistical significance was accepted as p less than Results Patient Characteristics The characteristics of the 84 patients included in our study are summarized in Table 1. The mean age of the patients was 58 years (range, 32 to 73 years). Eighty of the 84 patients were males. Forty-seven patients had reported a history of symptomatic GERD at the time of initial evaluation for esophageal cancer. Forty-eight patients had Barrett s-associated adenocarcinoma. In 38 of 48 patients, Barrett s mucosa was identified during esophagogastroduodenoscopy performed before therapy. In the remaining 10 patients, Barrett s mucosa was Table 1. Patient Characteristics (n 84) n (% of total) Gender Male 80 (95) Female 4 (5) Age 58 (range, yrs) Associated etiologic factors Tobacco use 61 (72) Alcohol use 39 (46) Family history of cancer 60 (71) Grade I 2 (2) II 32 (38) III 50 (60) Barrett s esophagus 48 (57) Symptomatic gastroesophageal reflux 47 (56) disease Pretreatment endoscopic ultrasound stage of study patients Tumor stage T1 2 (2) T2 10 (12) T3 72 (86) Lymph nodes stage N0 30 (36) N1 54 (64) identified postoperatively in esophageal resection specimens. Barrett s mucosa was histologically defined as the presence of specialized columnar epithelium, irrespective of the presence of dysplasia. Tumor grade (differentiation) was assessed from endoscopic biopsies taken before treatment. In tumors with mixed differentiation, tumor grade was designated based on the poorest level of differentiation. Two patients had well-differentiated tumors (grade I), 32 moderately differentiated tumors (grade II), and 50 patients had poorly differentiated tumors (grade III). Lymphadenopathy (suspicious for malignant infiltration) was identified in 54 patients (64%) by EUS. The tumor was staged T1N1 in 2 patients, T2N0 1 in 10 patients, and T3N0 1 in 72 patients. At surgery, two of these patients had T4 tumors (one of these two patients had Barrett s-associated adenocarcinoma and had history of GERD). Eighty-three of 84 patients underwent R0 esophageal resection and one patient was found to have residual tumor in the resection margins. Postoperative Survival Based on Barrett s Esophagus and GERD Figure 1 shows the Kaplan-Meier cumulative survival estimates in our study patients. Postoperative survival was significantly lower in patient s with Barrett s associated adenocarcinoma (vs adenocarcinoma arising de novo, p 0.031, log-rank test, Fig 1A) and in patients with a history of symptomatic GERD (vs patients without GERD, p 0.018, Fig 1B). Similar differences in postop-

4 Ann Thorac Surg AGARWAL ET AL 2005;79: ESOPHAGEAL ADENOCARCINOMA AND PREOPERATIVE CHEMORADIATION 1719 Fig 1. Kaplan-Meier estimates of cumulative postoperative survival in patients with esophageal adenocarcinoma treated with preoperative chemoradiation and potentially curative surgical resection. (A) Patients with or without identifiable Barrett s esophagus. (B) Patients with or without symptomatic GERD. (C) Patients with Barrett s or GERD or both versus patients with neither Barrett s nor GERD. (GERD gastroesophageal reflux disease.) erative survival were seen when patients with Barrett s esophagus and GERD were grouped together and compared with the patients without Barrett s or GERD (p 0.024, Fig 1C). Table 2 summarizes the distribution of Barrett s esophagus and history of symptomatic GERD in our study patients. Table 3 summarizes the Cox regression analysis to determine the influence of tumor grade and presence of significant lymphadenopathy by EUS, chemoradiation Table 2. Distribution and GERD and Barrett s Esophagus in Study Patients Barrett s Positive Barrett s Negative Total GERD positive GERD negative Total GERD gastroesophageal reflux disease.

5 1720 AGARWAL ET AL Ann Thorac Surg ESOPHAGEAL ADENOCARCINOMA AND PREOPERATIVE CHEMORADIATION 2005;79: Table 3. Multivariate Analysis in Study Patients Based on Barrett s Esophagus (A), GERD (B), or Combination of Barrett s and GERD (C) A HR (CI) p value Barrett s 4.38 ( ) Tumor grade 3.74 ( ) LN 2.13 ( ) 0.20 Protocol used 0.30 B GERD 2.71 ( ) Tumor grade 1.49 ( ) 0.34 LN 1.08 ( ) 0.84 Protocol used C Barrett s and/or GERD 3.82 ( ) Tumor grade 1.63 ( ) 0.24 LN 1.13 ( ) 0.76 Protocol used GERD gastroesophageal reflux disease; LN lymph node. protocol used on the differences in survival based on presence of Barrett s and/or GERD. Among our study patients, there were no significant differences in patients with or without Barrett s esophagus (Table 4) or patients with or without symptomatic GERD (Table 5), or patients with Barrett s esophagus and/or GERD (vs patients without Barrett s or GERD [Table 6]) in the proportion of grade I or II tumors, tumor staging (T stage), presence of significant lymphadenopathy by EUS, or the proportion of patients from the four chemoradiation regimens. Barrett s Esophagus and Pathological Complete Response Rates Table 7 summarizes the tumor characteristics of the patients in our study group who achieved path CR rates after chemoradiation. The patient group who achieved path CR had a significantly higher number of patients without Barrett s esophagus (p 0.011), patients without history of symptomatic GERD (p 0.016), and had higher proportion of well and moderately differentiated tumors than patients with path PR (p 0.023). However, patients with path CR and path PR did not differ significantly with regard to tumor location (p 1.00), presence of significant lymphadenopathy (p 0.43), or tumor stage by EUS (p 1.00). Patients with path CR also had a significantly lower number of patients with Barrett s and/or symptomatic GERD versus patients without Barrett s or GERD (p 0.005). Cox regression analysis adjusting for tumor grade, presence of significant lymphadenopathy, and the protocol used revealed significant differences between Barrett s associated adenocarcinoma (vs adenocarcinoma arising de novo, p 0.013), GERD positive (vs GERD negative, p 0.010), and Barrett s and/or GERD positive (vs Barrett e negative and GERD negative, p 0.003). Comment In the present study, we observed that among esophageal adenocarcinoma patients who underwent potentially curative esophageal resection after preoperative chemoradiation, patients with identifiable Barrett s mucosa and/or history of GERD had significantly lower cumulative postoperative survival. Furthermore, patients who achieve path CR after chemoradiation had a significantly higher proportion of patients without identifiable Barrett s esophagus and/or GERD, even after adjusting for other covariables. Based on the observations by Abdel-Latif and colleagues [8], we sought and found significant survival differences based on the history of symptomatic GERD in esophageal adenocarcinoma patients who underwent potentially curative esophageal resection after completing planned preoperative chemoradiation. We also found significant survival differences between the patients with or without underlying Barrett s esophagus. These differences held up even after multivariate analysis with other potentially confounding factors including tumor stage, tumor grade, presence of significant lymphadenopathy, and chemoradiation protocol used. The present study was retrospective with a relatively small number of patients, but involved sequential similarly staged patients. A potential and valid criticism of this study is the accuracy of diagnoses of GERD and Barrett s mucosa in our patients. We agree that Barrett s mucosa might have been significantly underdiagnosed since most of the endoscopists do not rigorously look for Barrett s mucosa in patients with esophageal cancer. We believe that the 10 patients in this study in whom Barrett s mucosa was found only by surgical pathology account for all or most of them. This is supported by data from Theisen and colleagues [20], who identified Barrett s mucosa in 18 of 20 patients in whom it was not identified endoscopically before chemoradiation, showing thereby that chemoradiation does not obliterate Barrett s mucosa but unmasks it. Even if we assume that some patients with Barrett s esophagus were missed, it is likely that contamination of patient group without Barrett s esophagus with Barrett s patients would diminish the survival differences between the two groups. We tested this assumption by recalculating the differences in survival and path CR rates after moving the 10 patients in whom the Barrett s esophagus was identified only in surgical pathology to the patient group without Barrett s esophagus. As expected, there was reduction in degree of statistical significance in survival (p 0.172, cw p 0.03) and path CR rates (p 0.03, compared with p 0.01). The same explanation may also apply for a well-justified criticism that the history of symptomatic GERD might have been underdiagnosed because a detailed and standardized questionnaire was not used for diagnosis in our study patients. In support of our data, similar differences in

6 Ann Thorac Surg AGARWAL ET AL 2005;79: ESOPHAGEAL ADENOCARCINOMA AND PREOPERATIVE CHEMORADIATION 1721 Table 4. Characteristics of Esophageal Adenocarcinoma Patients With or Without Barrett s Esophagus Barrett s ve (n 48) Barrett s ve (n 36) p Value Tumor grade Grade I or II 20 (42) 14 (39) Grade III 28 (58) 22 (61) 0.79 Tumor stage T1 1 (2) 1 (3) T2 6 (13) 4 (11) T3 41 (85) 31 (86) 1.00 Lymphadenopathy LN ve 32 (67) 22 (61) LN ve 16 (33) 14 (39) 0.59 Protocol used Protocol 1 16 (34) 13 (36) Protocol 2 15 (31) 11 (31) Protocol 3 14 (29) 8 (22) Protocol 4 3 (6) 4 (11) 0.79 LN lymph node. path CR rates and survival were observed based on Barrett mucosa and history of symptomatic GERD. The GERD and Barrett s esophagus were noted in the study patients at different time points by different observers (GERD by physician assistants or oncologists in oncology clinic, and Barrett s mucosa by gastroenterologists at the time of endoscopy and pathologists in surgical pathology specimens). Moreover, as would be expected based on an etiologic relationship between GERD and Barrett s, the survival differences persisted when patients with Barrett s and/or GERD were compared with patients with neither Barrett s nor GERD. It seems highly unlikely Table 5. Characteristics of Esophageal Adenocarcinoma Patients With or Without Symptomatic GERD GERD ve (n 47) GERD ve (n 37) p Value Tumor grade Grade I or II 20 (43) 14 (38) Grade III 27 (57) 23 (62) 0.79 Tumor stage T1 1 (2) 1 (3) T2 7 (15) 3 (8) T3 39 (83) 33 (89) 0.75 Lymphadenopathy LN ve 27 (57) 27 (73) LN ve 20 (43) 10 (27) 0.14 Protocol used Protocol 1 12 (26) 17 (46) Protocol 2 19 (40) 7 (19) Protocol 3 13 (28) 9 (24) Protocol 4 3 (6) 4 (11) 0.10 GERD gastroesophageal reflux disease; LN lymph node. Table 6. Characteristics of Esophageal Adenocarcinoma Patients With and Without Barrett s Esophagus and/or Symptomatic GERD Barrett s and/or GERD ve (n 60) Barrett s and GERD ve (n 24) p Value Tumor grade Grade I or II 26 (43) 8 (33) Grade III 34 (57) 16 (67) 0.40 Tumor stage T1 1 (2) 1 (4) T2 8 (13) 2 (8) T3 51 (85) 21 (88) 0.61 Lymphadenopathy LN ve 36 (60) 18 (75) LN ve 24 (40) 6 (25) 0.20 Protocol used protocol 1 19 (32) 10 (42) protocol 2 20 (33) 6 (25) protocol 3 18 (30) 4 (17) protocol 4 3 (5) 4 (17) 0.19 GERD gastroesophageal reflux disease; LN lymph node. that the observed differences in this study are artifactual. The subset of patients with Barrett s esophagus without GERD likely included patients with asymptomatic GERD or those in which GERD might have been missed. Two studies have suggested that, as a group, patients with Barrett s associated esophageal adenocarcinoma tend to have better survival largely because these tumors are diagnosed at an earlier stage. Sabel and colleagues [6] had found that overall survival in patients with Barrett s associated adenocarcinoma was 22 months versus 16.4 months in esophageal adenocarcinoma arising de novo (p 0.02). They also noted that the Barrett s group had smaller tumors, which were detected at an earlier stage (stage 0, I, or II) (36% of patient with identifiable Barrett s mucosa vs 21% without it, p 0.04). In another study [7], 1-year and 5-year survival in esophageal adenocarcinoma patients with Barrett s esophagus was 71% and 64%, respectively, and was 70% and 32%, respectively, in patients without Barrett s mucosa. Even in this study, higher 5-year survival in Barrett s associated adenocarcinoma was due to smaller size and lower preoperative stage of the tumors. In the present study, patient groups with or without identifiable Barrett s mucosa were matched for tumor stage, thereby eliminating the survival advantage in the Barrett s group in previous studies [6, 7] due to more patients with early stage tumors. The observed differences in path CR rates and subsequent survival between esophageal adenocarcinomas with or without Barrett s mucosa in this study likely reflect differences in biology of these tumors. What is the reason for differences in response to preoperative chemoradiation and subsequent postoperative survival in esophageal adenocarcinoma patients

7 1722 AGARWAL ET AL Ann Thorac Surg ESOPHAGEAL ADENOCARCINOMA AND PREOPERATIVE CHEMORADIATION 2005;79: Table 7. Tumor Characteristics of Patients With Path CR and Path PR After Preoperative Chemoradiation Path CR (n 21) Path PR (n 63) p Value Tumor location: midesophagus 2 (10) 5 (8) distal esophagus 4 (19) 14 (22) distal esophagus GE 9 (42) 27 (43) junction distal esophagus GE 6 (29) 17 (27) 1.00 junction proximal stomach Differentiation Grade I (well) 2 (9) 0 (0) Grade II (moderate) 10 (48) 22 (35) Grade III (poor) 9 (43) 41 (65) 0.02 Tumor stage: T1 0 (0) 2 (3) T2 2 (10) 8 (13) T3 19 (90) 53 (84) 1.00 Lymphadenopathy LN ve 9 (43) 21 (33) LN ve 12 (57) 42 (66) 0.43 Barrett s esophagus Barrett s ve 7 (33) 41 (65) Barrett s ve 14 (67) 22 (35) Symptomatic GERD GERD positive 7 (33) 40 (63) GERD negative 14 (67) 23 (37) Barrett s and/or GERD Barrett s and/or GERD 10 (48) 50 (79) positive No Barrett s no GERD 11 (52) 13 (21) CR complete response; GE gastroesophageal; GERD gastroesophageal reflux disease; LN lymph node; PR partial response. based on underlying Barrett s esophagus and history of symptomatic GERD. In the report by El-Rifai and colleagues [21], in the 19 patients with distal esophageal and gastroesophageal junction (GEJ) tumors among the 24 patients grouped as proximal tumors, there were significant genetic differences between tumors with or without identifiable Barrett s mucosa (losses in 5q and 8p and gains at 2q, 6p, 12p, and 20q were more frequent in Barrett s associated tumors). Abdel-Latif and colleagues [8] found significant differences in response to multimodality treatment and survival in patients with or without increased expression of NF- B in their tumors. They further showed that esophageal cancer cells in culture upregulate NF- B activity on exposure to bile acids (DCA) or low ph due to hydrochloric acid. The NF- B is well known to inhibit apoptosis [22, 23] and thus it would seem reasonable that upregulation of NF- in tumor cells by acid and bile reflux should make these tumors resistant to apoptosis induced by chemotherapeutic agents and radiation. We are planning to study these and other potential mechanisms in our future studies. It has been suggested that all esophageal adenocarcinomas arise from Barrett s mucosa [24]. Theisen and colleagues [20] reported that in patients with esophageal adenocarcinoma, they endoscopically identified Barrett s epithelium in 75% (59 of 79) of patients before, and in 97% (77 of 79) of patients after, chemoradiation by endoscopic biopsies or in resection specimens. A vast majority of published studies have reported lower prevalence of Barrett s esophagus in patients with esophageal adenocarcinomas [8, 21, 20, 25 30]. Lagergren and colleagues [25] reported that the risk of esophageal adenocarcinoma was identical in GERD patients with or without Barrett s esophagus, suggesting thereby that Barrett s esophagus is important, but not essential, for development of esophageal cancer. Cameron and colleagues [31] and Ruol and colleagues [32] reported higher prevalence of identifiable Barrett s mucosa in early tumors. However, early esophageal cancers are likely to be detected during Barrett s surveillance or incidentally in patients undergoing upper gastrointestinal endoscopy for unrelated symptoms, and Barrett s esophagus is more frequently found in patients with upper abdominal symptoms who undergo endoscopy [33]. It is still debatable whether more advanced tumors overgrow and obscure the Barrett s mucosa or Barrett s mucosa is usually found with early esophageal cancers as these early cancers are more likely to be diagnosed in patients with Barrett s esophagus. The study has limitations inherent to its retrospective design. As a result, the observed differences in survival and response to preoperative chemoradiation are not definitive. A potential limitation of the study is that patients received one of four chemoradiation protocols. However, the distribution of patients from each protocol in groups with or without Barrett s mucosa or symptomatic GERD was similar and survival differences based on presence of Barrett s mucosa or GERD persisted in a multivariate analysis that included the protocols used in the study. One of the major limitations of the study is that history of GERD was made without the use of detailed questionnaires or specific testing. However, though accurate diagnosis based on objective tests is desirable in a study like this, no correlation between GERD and treatment outcomes has previously been reported and the questionnaires and specific tests for GERD are not routinely used in pretreatment evaluation of patients with esophageal adenocarcinoma. The findings of this study are potentially of considerable interest, but need to be confirmed by future studies. Future studies may also evaluate if survival differences between these two groups exist irrespective of the treatment used (single modality vs multimodality) or these are unique to patients treated with preoperative chemoradiation. To conclude, there seem to be differences in survival and path CR rates among patients with esophageal adenocarcinoma treated with preoperative chemoradiation and surgery based on the presence of identifiable Barrett s esophagus and history of symptomatic GERD.

8 Ann Thorac Surg AGARWAL ET AL 2005;79: ESOPHAGEAL ADENOCARCINOMA AND PREOPERATIVE CHEMORADIATION 1723 Further studies may be justified to confirm these observations, to determine their basis and their significance in planning treatment for patients with esophageal adenocarcinoma. References 1. Cameron AJ, Ott BJ, Payne WS. The incidence of adenocarcinoma in columnar lined (Barrett s) esophagus. N Engl J Med 1985;313: Drewitz DJ, Sampliner RE, Garewal HS. The incidence of adenocarcinoma in Barrett s esophagus: a prospective study of 170 patients followed 4.8 years. Am J Gastroenterol 1997;92: Morales TG, Sampliner RE. Barrett s esophagus: update on screening, surveillance and treatment. Ann Intern Med 1999;159: Wright TA, Gray MR, Morris AI, et al. Cost effectiveness of detecting Barrett s esophagus. Gut 1996;39: Streitz JM Jr, Andrews CW Jr, Ellis FH Jr. Endoscopic surveillance of Barrett s esophagus. Does it help? J Thorac Cardiovasc Surg 1993;105: Sabel MS, Pastore K, Toon H, Smith JL. Adenocarcinoma of the esophagus with or without Barrett mucosa. Arch Surg 2000;135: Hoff SJ, Sawyers JL, Blanke CD, Choy H, Stewart JR. Prognosis of adenocarcinoma arising in Barrett s esophagus. Ann Thorac Surg 1998;65: Abdel-Latif MMM, O Riordan J, Windle HJ, et al. NF- B activation in esophageal adenocarcinoma: relationship to Barrett s metaplasia, survival, and response to neoadjuvant chemoradiotherapy. Ann Surg 2004;239: Kok TC, Lanschot J, Siersema PD, van Overhagen H, Talanus HV. Neoadjuvant chemotherapy in operable esophageal squamous cell cancer: final report of a phase III multicenter randomized controlled trial. Proc Am Soc Clin Oncol 1997; 16: Ruol A. Multimodality treatment for non-metastatic cancer of the thoracic esophagus. Results of a consensus conference held at the 6th World Congress of the International Society for Diseases of the Esophagus. Dis Esophagus 1996;9: Bates BA, Detterbeck FC, Bernard SA, et al. Concurrent radiation therapy and chemotherapy followed by esophagectomy for localized esophageal carcinoma. J Clin Oncol 1996;14: Jones DR, Detterbeck FC, Egan TM, et al. Induction chemotherapy followed by esophagectomy in patients with carcinoma of the esophagus. Ann Thorac Surg 1997;64: Naunheim KS, Petruska PJ, Roy TS, et al. Multimodality therapy for adenocarcinoma of the esophagus. Ann Thorac Surg 1995;59: Bosset JF, Gignoux M, Triboulet JP, et al. Chemoradiotherapy followed by surgery compared with surgery alone in squamous cell carcinoma of esophagus. N Engl J Med 1997;337: Walsh TN, Noonan N, Hollywood D, et al. A comparison of multi-modality therapy and surgery for esophageal adenocarcinoma. N Engl J Med 1996;335: Medical Research Council oesophageal cancer working party. Surgical resection with or without pre-operative chemotherapy in esophageal cancer: a randomized 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. Cancer 2001;91: Vogel SB, Mendenhall WM, Sombeck MD, Marsh R, Woodward ER. Downstaging of esophageal cancer after preoperative radiation and chemotherapy. Ann Surg 1995;221: Urba SG, Orringer MB, Turrisi A, Lannettoni M, Forastiere A, Strawderman M. Randomized trial of preoperative chemoradiation versus surgery alone in patients with locoregional esophageal carcinoma. J Clin Oncol 2001;19: Thiesen J, Stein HJ, Dittler HJ, et al. Preoperative chemoradiation unmasks underlying Barrett s mucosa in patients with adenocarcinoma of the distal esophagus. Surg Endosc 2002;16: El-Rifai W, Frierson HF, Moskaluk CA, et al. Genetic differences between adenocarcinomas arising in Barrett s esophagus and gastric mucosa. Gastroenterology 2001;121: Uetsuka H, Haisa M, Kimura M, et al. Inhibition of inducible NF- B activity reduces chemoresistance to 5-fluorouracil in human stomach cancer cell line. Exp Cell Res 2003;289: Huang Y, Fang Y, Dziadyk JM, Norris JS, Fan W. The possible correlation between activation of NF- B/I B pathway and the susceptibility of tumor cells to paclitaxel induced apoptosis. Oncol Res 2002;13: Mendes de Almeida JC, Chaves P, Periera AD, Altorki NK. Is Barrett s esophagus the precursor of most adenocarcinomas of the esophagus and cardia? A biochemical study. Ann Surg 1997;226: Lagergren J, Bergstrom R, Lindgren A, Nyren O. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 1999;340: Thomas P, Daddoli C, Lienne P, et al. Changing patterns and surgical results in adenocarcinoma of the oesophagus. Br J Surg 1997;84: Van Sandick JW, Van Lanschot JJB, Kuiken BW, Tytgat GNJ, Offerhaus GJA, Obertop H. Impact of endoscopic biopsy surveillance of Barrett s esophagus on pathological stage and clinical outcome of Barrett s carcinoma. Gut 1998;43: Hamilton SR, Smith RR, Cameron JL. Prevalence and characteristics of Barrett s esophagus in patients with adenocarcinoma of the esophagus or esophagogastric junction. Hum Pathol 1988;19: Cameron AJ, Lomboy CT, Pera M, Carpenter HA. Adenocarcinoma of the esophagogastric junction and Barrett s esophagus. Gastroenterology 1995;109: Corley DA, Levin TR, Habel LA, Weiss NS, Buffler PA. Surveillance and survival in Barrett s adenocarcinomas: a population based study. Gastroenterology 2002;122: Cameron AJ, Souto EO, Smyrk TC. Small adenocarcinomas of the esophagogastric junction: association with intestinal metaplasia and dysplasia. Am J Gastroenterol 2002;97: Ruol A, Parenti A, Zaninotto G, et al. Intestinal metaplasia is the probable common precursor of adenocarcinoma in Barrett esophagus and adenocarcinoma of the gastric cardia. Cancer 2000;88: Kamberoglou DK, Savva SC,. Kalapothakos PN, et al. Prevalence and risk factors associated with specialized intestinal metaplasia at the esophagogastric junction Hepatogastroenterology 2002;49:995 8.