Surgical Treatment of Esophageal High-Grade Dysplasia

Surgical Treatment of Esophageal High-Grade Dysplasia

GENERAL THORACIC ORIGINAL ARTICLES: GENERAL THORACIC Surgical Treatment of Esophageal High-Grade Dysplasia Michael F. Reed, MD, George Tolis, Jr, MD...

147KB Sizes 0 Downloads 74 Views

GENERAL THORACIC

ORIGINAL ARTICLES: GENERAL THORACIC

Surgical Treatment of Esophageal High-Grade Dysplasia Michael F. Reed, MD, George Tolis, Jr, MD, Barish H. Edil, MD, James S. Allan, MD, Dean M. Donahue, MD, Henning A. Gaissert, MD, Ashby C. Moncure, MD, John C. Wain, MD, Cameron D. Wright, MD, and Douglas J. Mathisen, MD Division of Thoracic Surgery, Department of Surgery, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts

Background. Barrett’s esophagus, high-grade dysplasia (HGD), and invasive cancer are steps in the progression of esophageal adenocarcinoma. While surgery is recommended for resectable invasive adenocarcinoma, a number of treatment modalities are advocated for HGD. The purpose of this study is to determine the outcomes after surgery for HGD. Methods. We identified cases of HGD based on endoscopic biopsy in a single institution’s databases from 1980 through 2001. Records were reviewed for patient characteristics, treatments, staging, and outcomes. Results. In a 22-year period, 869 cases of esophageal adenocarcinoma and 1,614 cases of Barrett’s esophagus were diagnosed. Of these, 115 had HGD without pretreatment evidence of invasion. Forty-nine patients with HGD underwent resection (mean age, 59 years) as initial treatment. Forty-seven had endoscopic treatment (mean age, 70 years) by photodynamic therapy or endoscopic

mucosal resection. Seven of the endoscopically treated patients failed, with three undergoing surgery and four observation. Nineteen patients were initially observed, with six eventually having surgery. For the 49 initially treated surgically, one (2%) operative mortality occurred. Invasive adenocarcinoma was present in 18 (37%). The five-year survival was 83% for all resected HGD patients (91% for those without invasion, 68% with invasion). Three of the eight deaths in those with invasion were from recurrent adenocarcinoma. Conclusions. Surgical resection of esophageal HGD can be performed with low mortality and allows longterm survival. A significant percentage with an initial diagnosis of HGD will have invasive disease at resection. Surgery is the optimal treatment for HGD unless contraindicated by severe comorbidities. (Ann Thorac Surg 2005;79:1110 –5) © 2005 by The Society of Thoracic Surgeons

T

frequent endoscopic follow-up of those patients with Barrett’s. This has led to the identification of more individuals with HGD. High-grade dysplasia is characterized by the cytologic features of malignancy. The only histologic difference between HGD and invasive adenocarcinoma is penetration through the basement membrane. Thus HGD, the esophageal equivalent to carcinoma in situ, is more than a mere risk factor for adenocarcinoma. Rather, it is the point before invasion when therapeutic intervention should be curative. Moreover, with suboptimal techniques for definitively differentiating between HGD and early invasive disease, a subset of HGD patients is expected to have occult invasive adenocarcinoma. Indeed, this has been supported by a number of studies indicating a rate of approximately 40% of invasive adenocarcinoma in patients with the preoperative diagnosis of HGD [5–13]. Yet controversy remains about the optimal treatment of HGD [14]. In light of the common perception that all esophageal resections carry high risk of mortality, some have advocated alternative, less invasive therapies including endoscopic mucosal resection (EMR), photodynamic therapy (PDT), and endoscopic laser ablation. We therefore employed our databases to identify all patients with HGD in order to determine what therapies were provided, the outcome of those treatments, and the long-term survival of these individuals.

he incidence of adenocarcinoma of the esophagus continues to increase in the United States and Western Europe [1, 2]. Yet the only documented risk for adenocarcinoma is gastroesophageal reflux disease (GERD) [3]. The progression to neoplasia, along a continuum from Barrett’s esophagus, through low-grade dysplasia and high-grade dysplasia (HGD), and eventually into invasive carcinoma, is well-established. Adenocarcinoma of the esophagus carries a dismal prognosis. Five-year survival for all newly diagnosed patients is less than 10% [4]. Ongoing studies into the optimal treatment for specific stages, incorporating improved surgical technique, novel combinations and timing of chemotherapy and radiation therapy, and better palliative modalities, contribute to incremental improvements. Diagnosis and intervention at an earlier stage should therefore lead to improved survival. Thus, the foremost strategy for achieving substantial progress toward enhanced survival remains intervention early in the development of disease. In line with this approach has been the increased use of endoscopy for screening patients with GERD who are a risk for Barrett’s esophagus, as well as the Accepted for publication Sept 3, 2004. Presented at the Thirty-ninth Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 31–Feb 2, 2003. Address reprint requests to Dr Mathisen, General Thoracic Surgery, Massachusetts General Hospital, 55 Fruit St, Blake 1570, Boston, MA 02114; e-mail: [email protected].

© 2005 by The Society of Thoracic Surgeons Published by Elsevier Inc

0003-4975/05/$30.00 doi:10.1016/j.athoracsur.2004.09.006

REED ET AL SURGERY FOR HIGH-GRADE DYSPLASIA

1111

Patients and Methods

Table 1. Surgical Management of High-Grade Dysplasia

To identify all new diagnoses of HGD, the Department of Pathology database at the Massachusetts General Hospital (MGH) was searched for all cases of Barrett’s esophagus, esophageal HGD, and esophageal adenocarcinoma for the 22-year period from January 1, 1980 through December 31, 2001. The Division of Thoracic Surgery database was searched for all esophagectomies performed during the same period. Operative, endoscopy, radiology, and pathology reports were reviewed to verify adenocarcinoma of the esophagus, including adenocarcinomas of the gastroesophageal junction, but excluding gastric tumors from analysis. Pathology reports were reviewed to identify those with carcinoma limited to only HGD. Any patient with pretreatment pathologic evidence of invasive adenocarcinoma, including intramucosal (T1a) disease, was excluded from the HGD group. The original reports from MGH pathologists were utilized. Typically, the slides had been reviewed in a consensus conference before finalization. The records of these patients were reviewed for age, gender, endoscopic findings, radiologic results, endoscopic ultrasound (EUS) findings, prior surveillance studies, details of the operative or endoscopic intervention, mortality, subsequent surveillance studies, and survival. Endoscopy was performed by gastroenterologists and surgeons. Endoscopic treatment (EMR, PDT) was performed by gastroenterologists. Patients undergoing resection typically had endoscopy repeated at the time of surgery. Patients with pretreatment radiologic or EUS-based evidence of invasion were excluded from the HGD group. All patients included in this study of HGD were therefore considered stage 0 before intervention. The final pathologic staging, based on resection specimens, followed the tumor, mode, metastasis (TNM) classification guidelines of the American Joint Committee for Cancer Staging [15]. Follow-up data on each patient were obtained by review of the hospital record, Department of Pathology database, Division of Thoracic Surgery records, and/or phone contact with the patient or patient’s family. The presence of recurrent carcinoma, time of diagnosis of any recurrence, and site of recurrence were determined. The cause of death was recorded. Operative mortality was defined as death within 30 days of the operation or during the same hospitalization. Nonparametric estimates of survival were obtained using the method of Kaplan and Meier [16] with the date of operation at the starting point. Analysis of variables relating to survival was performed using the log-rank test. A p value less than 0.05 was considered statistically significant. During the 22-year period, we identified 1,614 new cases of Barrett’s esophagus in 1,047 men (64.9%) and 567 women (35.1%). In the same time frame there were 869 new cases of adenocarcinoma in 722 men (83.1%) and 147 women (16.9%). The mean age of diagnosis for Barrett’s was 59 years (range, 0 to 96 years) and the mean age for diagnosis of esophageal adenocarcinoma was 65 years (range, 23 to 95 years). Among this large cohort of patients was a subset of 115 cases (95 men, 20 women) of

Characteristic

Number

Number of patients Age (range) Gender Male Female Surgical approach Left thoracoabdominal Ivor Lewis Transhiatal Other Stage 0 I (T1a) I (T1b) IIA IIB III

49 59 (32–79)

Percentage 100%

40 9

81.6% 18.4%

20 18 7 4

40.8% 36.7% 14.3% 8.2%

31 7 3 3 3 2

63.3% 14.3% 6.1% 6.1% 6.1% 4.1%

HGD without evidence of invasive adenocarcinoma, based upon histologic evaluation of biopsy specimens, computed tomography (CT), or EUS. The mean age of HGD patients was 65 years (range, 30 to 89 years). Five patients with HGD had prior fundoplications for GERD, one patient had two previous Heller myotomies for achalasia, and one patient had undergone gastric bypass with a Rous-en-Y gastrojejunostomy for morbid obesity.

Results Treatment by Immediate Esophageal Resection Among the patients with HGD by preoperative evaluation, 49 (40 men, 9 women) underwent surgical resection within 60 days after the diagnosis was made (Table 1). The mean age for those undergoing surgery was 59 years (range, 32 to 79 years). A left thoracoabdominal esophagogastrectomy was performed in 20 patients (41%). An Ivor Lewis technique was chosen for 18 cases (37%). A transhiatal approach was performed in seven patients (14%). In the other four patients (8%), a variety of techniques were required, primarily from an abdominal approach. A gastric conduit was used for esophageal replacement unless prior gastric surgery precluded its use. En bloc lymph node dissection was utilized. The choice of operative strategy was the individual surgeon’s preference. One operative mortality occurred (2.0%). This was secondary to a large cerebrovascular accident in the postoperative period. Two anastomotic leaks were documented. One was in the neck after a transhiatal esophagogastrectomy in a patient with scleroderma that was managed with drainage. Another was an esophagogastric dehiscence in the chest after an Ivor Lewis procedure. The gastric conduit was immediately removed and a jejunal interposition was performed six months later. The patient is alive, tolerating a regular diet, five years after the initial surgery.

GENERAL THORACIC

Ann Thorac Surg 2005;79:1110 –5

GENERAL THORACIC

1112

REED ET AL SURGERY FOR HIGH-GRADE DYSPLASIA

Ann Thorac Surg 2005;79:1110 –5

Table 2. Endoscopic Management of High-Grade Dysplasia Characteristic

Number

Number of patients Age (range) Gender Male Female Reason for no surgery Significant comorbidities Patient preference Technique PDT EMR Recurrent HGD or invasion PDT EMR Recurrent HGD or invasion resected Stage I Stage IIA Recurrent HGD or invasion not resected Stage 0 Stage I Stage IV

47 70 (30–89)

EMR ⫽ endoscopic mucosal resection;

Percentage 100%

38 9

80.9% 19.1%

37 10

78.7% 21.3%

42 5

89.4% 10.6%

5 2 2 1

1 2 1 PDT ⫽ photodynamic therapy.

Treatment by an Endoscopic Approach Forty-seven patients (38 men, 9 women) were treated using endoscopic techniques for HGD without evidence of invasive adenocarcinoma (Table 2). The mean age was 70 years (range, 30 to 89 years). Among the patients managed initially by endoscopy, 37 were considered high risk for surgery due to advance age and/or medical comorbidities. The other ten patients refused surgery. Endoscopic mucosal resection was performed in five patients and the other 42 had PDT. Five of the PDT-treated patients and two of the EMR-treated patients failed, with recurrent HGD or invasive adenocarcinoma. The average time from first endoscopic treatment to recurrence or invasion was 13.5 months. Three of these seven patients, treated initially by endoscopy, had progression of disease and then underwent surgery: two had stage I and one had stage IIA disease (Table 2). Of the four who were followed without surgical intervention, one had continued HGD without evidence of invasion, two had stage I adenocarcinoma, and one, who was medically unfit for surgery, progressed to invasive adenocarcinoma and eventually developed metastatic spread.

Management by Observation First Nineteen patients had biopsy-proven HGD without evidence of invasive adenocarcinoma and elected to defer all available options for intervention at the time of diagnosis (Table 3). Six of the 19 patients had persistence of HGD or progression to invasion and later elected surgery. The average time from initial diagnosis to subsequent surgery was 4.0 months. Based on histologic evaluation of the surgical specimen, as well as intraoperative findings, one of the six was stage 0 (only HGD),

Table 3. Outcome of Patients Managed Initially by Observation Number of patients Delayed surgery Stage 0 Stage I Stage IIA Stage IV Total No surgery Stage 0 Stage IV Total No follow-up Total

19 1 3 1 1 6 5 2 7 6

three were stage I, one was stage IIA, and one underwent palliative gastric bypass for the intraoperative evidence of stage IV disease. Seven of the 19 patients who were initially observed had no subsequent intervention and two of this group eventually developed invasive carcinoma with metastatic spread, one of whom underwent palliative esophageal stenting. Six patients were lost to follow-up. Thus, seven of the 13 patients who were followed after an initial period of observation subsequently developed invasive adenocarcinoma.

Pathologic Staging For all cases of HGD treated by resection, the pathologic staging was determined based on the American Joint Committee for Cancer Staging criteria [15]. For the 49 patients who underwent surgery as the primary treatment, as opposed to salvage therapy after failed endoscopic therapy or after a period of observation, 31 (63%) were stage 0 with no evidence of invasive adenocarcinoma (Table 1). The other 18 patients (37%) had tumor invasion through the basement membrane, indicating invasive adenocarcinoma. Of those patients with invasive disease, 10 were stage I, three were stage IIA, three were stage IIB (T1N1), and two were stage III (T3N1). Within the group of ten with stage I adenocarcinoma, seven had intramucosal (T1a) disease and three exhibited submucosal (T1b) invasion. Five of the eighteen with invasive disease had positive lymph nodes. Thus, of the 49 treated surgically, 10% had nodal involvement not identified by preoperative studies.

Survival The median follow-up for the 48 patients who were operative survivors after resection for HGD was 56 months (range, 1 to 250 months). The overall five-year survival for all 49 patients who underwent esophagectomy was 83% (Fig 1). The ten-year survival was 64%. The causes of the twelve deaths in the operative survivors were recurrent esophageal adenocarcinoma (three patients), non-small cell lung cancer (two), renal cell carcinoma (one), lymphoma (one), and unknown (five). The three deaths from recurrent disease were among the 18 patients who had surgery for HGD but had invasive

Fig 1. Estimated survival of 49 patients who underwent esophagectomy for the preoperative diagnosis of high-grade dysplasia. The five-year survival is 83%.

adenocarcinoma on pathologic examination, one with stage IIA and two with stage IIB disease. The diseasespecific five-year survival for all 49 patients was 94%. None of the patients who underwent esophagectomy and had only HGD died of esophageal adenocarcinoma. The presence of invasive adenocarcinoma in the resected specimen significantly lowered survival (Fig 2). Despite the absence of preoperative evidence of invasion, when penetration of the basement membrane was identified, five-year survival was 68%. But with invasive disease limited to the mucosa or submucosa, without spread to the lymph nodes (stage I), survival remained high at 90% (Fig 3). The diseasespecific five-year survival was 100%, with no mortality from esophageal adenocarcinoma among stage I patients. Deeper invasion or lymph node involvement lowered the five-year survival to 50% or less. However, if there was no

Fig 2. Estimated survival of 31 patients with high-grade dysplasia (HGD) only, compared to 18 patients with the invasive adenocarcinoma on pathologic evaluation of the specimen after esophagectomy. The five-year survival is 91% for HGD (- - -) and 68% for invasive adenocarcinoma (—). p less than 0.03.

REED ET AL SURGERY FOR HIGH-GRADE DYSPLASIA

1113

Fig 3. Estimated survival of 49 patients who underwent esophagectomy for high-grade dysplasia, with separate curves for each stage: (—), stage 0 (31 patients); ( · - - - · · ), stage I (10); (–), stage IIA (3); ( · - · · - · ·), stage IIB (3); ( - – - –), stage III (2). The five-year survivals are 91% (stage 0), 90% (stage I), 33% (stage IIA), and 50% (stages IIB and III).

evidence of invasive adenocarcinoma, five-year survival was 91% and disease-specific survival was 100%.

Comment Gastroesophageal reflux disease is a common diagnosis, affecting 40% of adults in the United States. Yet the majority of cases of adenocarcinoma are diagnosed at advanced stages when five-year mortality exceeds 90% [4], so early identification is essential for improving survival rates. The higher number of patients now being diagnosed with HGD has made its optimal management an important question for which considerable controversy exists. In this study, of the 115 patients with HGD, 49 underwent surgical resection while 47 were treated endoscopically and 19 were initially followed. The efficacy of esophagectomy for HGD has been demonstrated by a number of groups [5–7, 9 –11, 17]. Furthermore, when patients have undergone esophageal resection for HGD without preoperative evidence of invasion, histologic demonstration of invasive adenocarcinoma is seen in up to 50% of cases. In this series 37% had invasive disease. Many caregivers in both surgical and medical specialties have therefore promoted esophagectomy as the appropriate treatment for HGD in patients medically fit for major surgery. Yet a substantial number of physicians do not support surgical resection of HGD, unless progression to invasion is documented. A common misperception is that esophagectomy is a highrisk operation associated with a substantial morbidity rate, often up to 50%, as well as mortality rates of up to 10%. While some of the quoted high morbidity and mortality rates associated with esophagectomy are from historical series that do not represent current practice, recent large multiinstitutional studies of outcomes after esophagectomy maintain the notion of high risk [18].

GENERAL THORACIC

Ann Thorac Surg 2005;79:1110 –5

GENERAL THORACIC

1114

REED ET AL SURGERY FOR HIGH-GRADE DYSPLASIA

However, these studies include low-volume hospitals, surgeons with limited esophageal experience, and esophagectomy for various indications, high-risk patients, advanced disease, and after neoadjuvant therapy. In a recent report, hospitals in which a high volume of esophagectomies are performed had better results than lowvolume institutions with a 3.7-fold decrease in hospital mortality [19]. And esophagectomy by experienced surgeons is similarly associated with low mortality [20]. A number of publications from groups performing esophagectomy specifically for HGD have reported operative mortality rates less than 3% [6, 7]. In this series we had one operative death, from a postoperative cerebrovascular accident, for a mortality rate of 2.0%. Endoscopic mucosal ablative techniques, including PDT, laser ablation, and EMR, have been evaluated for treatment of HGD, especially among patients unfit for surgery. In this series, those undergoing resection had an average age of 59 years compared to 70 years in those treated endoscopically, and the majority of those receiving mucosal ablative therapy had significant medical risk factors. Unfortunately, PDT can preclude pathologic evaluation for submucosal invasion and incomplete ablation may result in submerged Barrett’s epithelium that is not amenable to endoscopic surveillance. Further disadvantages include the cutaneous photosensitivity of 4 to 8 weeks in duration and a 34% to 58% incidence of stricture formation. Using EMR permits the analysis of depth of penetration to identify those patients with HGD who also have invasive adenocarcinoma. But EMR is associated with a 17% risk of cancer recurrence in 12 months [21]. In this study, we identified persistent HGD or progression to invasive adenocarcinoma in five patients who had PDT and two who had EMR. Three underwent esophagectomy, of which two were stage I and one was stage IIA. Of the four who were observed, one continued with HGD, two had stage I adenocarcinoma, and one developed metastatic adenocarcinoma. In total, 60 patients were followed after endoscopic treatment or initial observation, with 14 developing recurrent HGD or progression to invasive adenocarcinoma. Thus, the risk of subsequent invasion is not insignificant. The purpose of early surgical intervention, ideally when only HGD is present, is prevention of death from esophageal cancer. For all patients undergoing esophagectomy for HGD, we demonstrate a five-year estimated survival rate of 83%, with a disease-specific survival of 94%. This includes the 37% of patients who had invasive disease. For those with only HGD, the five-year survival is 91%, and diseasespecific survival is 100%. Moreover, if invasion is present, but limited to the mucosa or submucosa (T1), the five-year survival remains high at 90% with disease-specific survival of 100%. Not surprisingly, deeper invasion or lymph node spread resulted in five-year survival less than 50%, despite the absence of preoperative evidence of invasive carcinoma. The three individuals who died from recurrent adenocarcinoma were all patients who had invasive disease in the resected specimen. Notably, the rate of invasive adenocarcinoma in patients undergoing resection for HGD changed over the 22-year period. During the first 11 years,

Ann Thorac Surg 2005;79:1110 –5

seven of 15 patients had invasive disease (47%). Over the second 11-year period, 11 of 34 had invasion (32%) with eight stage I, one stage IIA, and only two with nodal spread (one stage IIB and one stage III). The lower rate of invasion likely reflects advancements in preoperative staging, including improved CT, the addition of EUS, and the recent use of positron emission tomography. While endoscopic methods employed for the management of esophageal HGD carry the allure of lower mortality, diminished morbidity, and equal efficacy when compared to esophagectomy, this should not currently be accepted. Esophagectomy for HGD, without any evidence of invasion, provides the opportunity for longterm survival that is not achieved by any other method if the disease has progressed to invasion. In experienced hands, it carries a low operative risk. And, despite improved diagnostic techniques, about one-third of patients will have invasion at the time of resection. Mucosal ablative techniques should continue to be evaluated, in a setting of prospective trials among patients who are at high risk for major surgery. For patients with HGD without significant comorbid disease, esophagectomy remains the standard treatment.

Assistance with statistical analysis and figure preparation was generously provided by the Center for Biostatistical Services, University of Cincinnati Medical Center.

References 1. Devesa SS, Blot WJ, Fraumeni JF Jr. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 1998;83:2049 –53. 2. Blot WJ, McLaughlin JK. The changing epidemiology of esophageal cancer. Semin Oncol 1999;26:2– 8. 3. Lagergren J, Bergstrom R, Lindgren A, Nyren O. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 1999;340:825–31. 4. Farrow DC, Vaughan TL. Determinants of survival following the diagnosis of esophageal adenocarcinoma (United States). Cancer Causes Control 1996;7:322–7. 5. Pera M, Trastek VF, Carpenter HA, Allen MS, Deschamps C, Pairolero PC. Barrett’s esophagus with high-grade dysplasia: an indication for esophagectomy? Ann Thorac Surg 1992;54: 199 –204. 6. Headrick JR, Nichols FC III, Miller DL, et al. High-grade esophageal dysplasia: long-term survival and quality of life after esophagectomy. Ann Thorac Surg 2002;73:1697–702; discussion 1702–3. 7. Rice TW, Blackstone EH, Goldblum JR, et al. Superficial adenocarcinoma of the esophagus. J Thorac Cardiovasc Surg 2001;122:1077–90. 8. Altorki NK, Sunagawa M, Little AG, Skinner DB. High-grade dysplasia in the columnar-lined esophagus. Am J Surg 1991;161:97–9; discussion 99 –100. 9. Edwards MJ, Gable DR, Lentsch AB, Richardson JD. The rationale for esophagectomy as the optimal therapy for Barrett’s esophagus with high-grade dysplasia. Ann Surg 1996;223:585–9; discussion 589 –91. 10. Heitmiller RF, Redmond M, Hamilton SR. Barrett’s esophagus with high-grade dysplasia. An indication for prophylactic esophagectomy. Ann Surg 1996;224:66 –71. 11. Rice TW, Falk GW, Achkar E, Petras RE. Surgical management of high-grade dysplasia in Barrett’s esophagus. Am J Gastroenterol 1993;88:1832– 6.

12. Peters JH, Clark GW, Ireland AP, Chandrasoma P, Smyrk TC, DeMeester TR. Outcome of adenocarcinoma arising in Barrett’s esophagus in endoscopically surveyed and nonsurveyed patients. J Thorac Cardiovasc Surg 1994;108:813–21; discussion 821–2. 13. Falk GW, Rice TW, Goldblum JR, Richter JE. Jumbo biopsy forceps protocol still misses unsuspected cancer in Barrett’s esophagus with high-grade dysplasia. Gastrointest Endosc 1999;49:170 – 6. 14. Schnell TG, Sontag SJ, Chejfec G, et al. Long-term nonsurgical management of Barrett’s esophagus with high-grade dysplasia. Gastroenterology 2001;120:1607–19. 15. AJCC Cancer Staging Manual. Sixth ed. New York, NY: Springer-Verlag; 2002. 16. Kaplan EL, Meier P. Non-parametric estimation from incomplete observation. J Am Stat Assoc 1958;53:457– 81.

REED ET AL SURGERY FOR HIGH-GRADE DYSPLASIA

1115

17. Ferguson MK, Naunheim KS. Resection for Barrett’s mucosa with high-grade dysplasia: implications for prophylactic photodynamic therapy. J Thorac Cardiovasc Surg 1997;114:824 –9. 18. Bailey SH, Bull DA, Harpole DH, et al. Outcomes after esophagectomy: a ten-year prospective cohort. Ann Thorac Surg 2003;75:217–22. 19. Kuo EY, Chang Y, Wright CD. Impact of hospital volume on clinical and economic outcomes for esophagectomy. Ann Thorac Surg 2001;72:1118 –24. 20. Mathisen DJ, Grillo HC, Wilkins EW Jr, Moncure AC, Hilgenberg AD. Transthoracic esophagectomy: a safe approach to carcinoma of the esophagus. Ann Thorac Surg 1988;45:137– 43. 21. Ell C, May A, Gossner L, et al. Endoscopic mucosal resection of early cancer and high-grade dysplasia in Barrett’s esophagus. Gastroenterology 2000;118:670 –7.

DISCUSSION DR JEFFREY A. HAGEN (Los Angeles, CA): I have two questions. If I understood you correctly there were seven patients treated nonsurgically that developed recurrent high-grade dysplasia or progressed to cancer. While this seems a small number it is critical that this be viewed in the context of the length of follow-up when nonsurgical therapy was chosen. Could you provide us with this information? Second, you seem to have a much higher frequency of locally advanced tumors than is generally reported when people are reporting their experience with high-grade dysplasia as the highest lesion on preoperative biopsies. The question is, what type of biopsy protocol was used and was it done by the operating surgeon? Were there lesions identified preoperatively on endoscopy in these patients? DR REED: First, in terms of the follow-up of that subgroup who had recurrence or progression, of the three patients who underwent salvage surgery and had early-stage disease, their recurrences were at 4, 8, and 31 months after endoscopic therapy and there were no deaths from esophageal cancer. Of those who were not treated operatively, there was one patient who continued to progress who was inoperable and was stented for palliation and did die of esophageal cancer. The other three who were managed nonoperatively did not die of cancer. For all seven patients, the time to recurrence ranged from one to 32 months with an average of 11 months. As to the second question, we did notice that there were a few with advanced disease. The nodal disease was primarily microscopic. These patients did not have adenopathy picked up by endoscopic ultrasound or CT. This was also a 22-year period and technology has changed over this time with improved detection of nodal spread. The hope is that we are not going to see as many of those patients with these more recent imaging advancements including EUS. DR DOUGLAS E. WOOD (Seattle, WA): Nice presentation and obviously excellent results, but in spite of the good results, I am not sure that your last conclusion is supported by your data. Your data is compelling that one can do an esophagectomy in these patients with low risk of mortality. However, it does not automatically follow that these patients with high-grade dysplasia have a high incidence of more advanced disease. That goes back to the last question that you answered regarding the thoroughness and the extent of the protocol for biopsies, and over this period of time there are probably a variety of biopsy protocols that may have been in place. I will say that at the University of Washington, where we have a large Barrett’s program, we have followed the opposite strategy that you have concluded; that actually the morbidity of

esophagectomy, in spite of good results in experienced hands, is still substantial. Since we do not have a good understanding of the natural history of these patients, if one can enroll them in a good biopsy protocol that follows them closely, that may better select patients who then will progress and should have an esophagectomy. How are you managing these patients in terms of their biopsy protocol and are you discriminating between patients with different types of high-grade dysplasia or patients that progress to show nodules or invasive disease? DR REED: We biopsy patients with significant reflux disease. For patients with Barrett’s esophagus, we advocate at least annual biopsies. These are performed in the standard four quadrants every 1 cm until normal esophageal tissue is reached. For patients who have low-grade dysplasia we usually are following these patients with at least quarterly biopsies in the early period after diagnosis. For high-grade dysplasia we don’t differentiate necessarily between the endoscopic appearance of nodularity versus high-grade dysplasia that is only picked up based on pathologic findings. As to the conclusion, I think the way we feel is that esophagectomy is curative. No patient who underwent esophagectomy for high-grade dysplasia or stage I disease died of esophageal cancer. There were patients who were managed nonoperatively who progressed. DR THOMAS J. WATSON (Rochester, NY): I just want to concur with the comments that Dr Wood made. I think the real issue is whether there is a group of patients we can safely follow. Your last recommendation is that esophagectomy should be the treatment of choice; it sounds like, for all patients with highgrade dysplasia. Do you still believe that is true for, say, an 80-year old with a single focus of high-grade dysplasia versus a 50-year old with multiple foci of high-grade dysplasia? Were you able to get that from your data, the extent of high grade dysplasia as a predictor of invasive carcinoma or as an indication for surgery? DR REED: We didn’t look specifically at the extent of disease. A patient needs to be an appropriate candidate for surgery for us to advocate a surgical approach. For the 50-year old, absolutely, we advocate esophagectomy. I think right now that the nonoperative management of high-grade dysplasia has to be considered experimental, regardless of the extent of high-grade dysplasia, and should be done in a protocol setting, and we would support that for a patient with significant comorbidities.

GENERAL THORACIC

Ann Thorac Surg 2005;79:1110 –5