ORIGINAL SUBMISSIONS

Eliminating a Need for Esophagectomy: Endoscopic Treatment of Barrett Esophagus With Early Esophageal Neoplasia Michal J. Lada, MD,* Thomas J. Watson, MD, FACS,1,* Aqsa Shakoor, MD,* Dylan R. Nieman, MD, PhD,* Michelle Han, MD,* Andreas Tschoner, MD,* Christian G. Peyre, MD,* Carolyn E. Jones, MD, FACS,* and Jeffrey H. Peters, MD, FACS† Over the past several years, endoscopic ablation and resection have become a new standard of care in the management of Barrett esophagus (BE) with high-grade dysplasia (HGD) or intramucosal adenocarcinoma (IMC). Risk factors for failure of endoscopic therapy and the need for subsequent esophagectomy have not been well elucidated. The aims of this study were to determine the efficacy of radiofrequency ablation (RFA) with or without endoscopic mucosal resection (EMR) in the management of BE with HGD or IMC, to discern factors predictive of endoscopic treatment failure, and to assess the effect of endoscopic therapies on esophagectomy volume at our institution. Data were obtained retrospectively for all patients who underwent endoscopic therapies or esophagectomy for a diagnosis of BE with HGD or IMC in our department between January 1, 2004, and December 31, 2012. Complete remission (CR) of BE or HGD or IMC was defined as 2 consecutive biopsy sessions without BE or HGD or IMC and no subsequent recurrence. Recurrence was defined by the return of BE or HGD or IMC after initial remission. Progression was defined as worsening of HGD to IMC or worsening of IMC to submucosal neoplasia or beyond. Overall, 57 patients underwent RFA with or without EMR for BE with HGD (n ¼ 45) or IMC (n ¼ 12) between 2007 and 2012, with a median follow-up duration of 35.4 months (range: 18.5-52.0 months). The 57 patients underwent 181 ablation sessions and more than half (61%) of patients underwent EMR as a component of treatment. There were no major procedural complications or deaths, with only 2 minor complications including 1 symptomatic stricture requiring dilation. Multifocal HGD or IMC was present in 43% (25/57) of patients. CR of IMC was achieved in 100% (12/12) at a median of 6.1 months, CR of dysplasia was achieved in 79% (45/57) at a median of 11.5 months, and CR of BE was achieved in 49% (28/57) at a median of 18.4 months. Following initial remission, 28% of patients (16/57) had recurrence of dysplasia (n ¼ 12) or BE (n ¼ 4). Progression to IMC occurred in 7% (4/57). All patients without CR continue endoscopic treatment. No patient required esophagectomy or developed metastatic disease. Overall, 6 patients died during the follow-up interval, none from esophageal cancer. Factors associated with failure to achieve CR of BE included increasing length of BE (6.0 ⫾ 0.6 vs 4.0 ⫾ 0.6 cm, P ¼ 0.03) and shorter duration of follow-up (28.5 ⫾ 3.8 months vs 49.0 ⫾ 5.8 months, P ¼ 0.004). Shorter surveillance duration (17.8 ⫾ 7.6 months vs 63.9 ⫾ 14.4 months, P ¼ 0.009) and shorter follow-up (21.1 ⫾ 6.1 months vs 43.2 ⫾ 4.1 months) were the only significant factors associated with failure to eradicate dysplasia. Our use of esophagectomy as primary therapy for BE with HGD or IMC has diminished since we began using endoscopic therapies in 2007. From a maximum of 16 esophagectomies per year for early Barrett neoplasia in 2006, we performed only 3 esophageal *Department †

of Surgery, University of Rochester Medical Center, Rochester, New York. Chief Operating Officer, University Hospitals, Cleveland, Case Western Reserve University.

Dr. Watson has been a paid consultant to Covidien, Inc for projects unrelated to the current study. Dr Peters reports receiving consulting fees from Torax Medical and Given Imaging. Drs Lada, Shakoor, Nieman, Han, Tschoner, Peyre, and Jones have no commercial interests to disclose. Address reprint requests to Thomas J. Watson, MD, FACS, Division of Thoracic and Foregut Surgery, University of Rochester Medical Center, 601 Elmwood Ave, Box Surgery, Rochester, NY 14642. E-mail: Thomas_watson@urmc. rochester.edu 1 AATS member.

274

1043-0679/$-see front matter ª 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1053/j.semtcvs.2014.12.004

ENDOSCOPIC TREATMENT OF BE WITH EARLY ESOPHAGEAL NEOPLASIA resections for such early disease in 2012, all for IMC, and we have not performed an esophagectomy for HGD since 2008. Although recurrence of BE or dysplasia/IMC was not uncommon, RFA with or without EMR ultimately resulted in CR of IMC in all patients, CR of HGD in the majority (79%), and CR of BE in nearly half (49%). No patient treated endoscopically for HGD or IMC subsequently required esophagectomy. In patients with BE with HGD or IMC, RFA and EMR are safe and highly effective. The use of endoscopic therapies appears justified as the new standard of care in most cases of BE with early esophageal neoplasia. Semin Thoracic Surg 26:274–284 I 2014 Elsevier Inc. All rights reserved. Keywords: Barrett Esophagus, dysplasia, adenocarcinoma, ablation, esophagectomy

INTRODUCTION The link between gastroesophageal reflux disease (GERD), Barrett esophagus (BE), esophageal dysplasia, and esophageal adenocarcinoma (EAC) has been well established. The prevalence of BE has been estimated to be between 1.6% and 6.8% of the general population, which translates into approximately 5-20 million people within the United States alone.1,2 The 0.12%-0.4% annual risk of progression of nondysplastic BE to EAC has fueled surveillance of known BE.3 Along with the increasing rates of obesity and GERD, the incidence of EAC in the United States has been rising over the past several decades. Data from the American Cancer Society for 2014 predict 18,100 esophageal cancer diagnoses with 15,450 deaths and an overall 5-year survival of 17%.4-6 Although patients commonly present with the manifestations of advanced, incurable esophageal malignancy, early-stage esophageal neoplasia can be treated with the expectation of cure. Esophagectomy has been the standard of care for BE with highgrade dysplasia (HGD) or intramucosal adenocarcinoma (IMC) for decades. Multiple surgical series have shown that esophagectomy, when performed by expert surgeons in specialty centers for early neoplasia, is associated with low perioperative mortality and acceptable morbidity.7 More recently, endoscopic ablative and resective technologies and techniques have been introduced into clinical practice, altering the landscape of therapies available for eradication of early esophageal neoplasia in the setting of BE. Current guidelines proposed by specialty medical societies in the United States, as well as the National Comprehensive Cancer Network, recommend endoscopic therapies as the preferred treatment for BE with HGD, relegating esophagectomy to the minority of cases not suitable for endoscopic approaches.8-11 These same societies, however, have yet to take a stance on the optimal management of IMC. In addition, the risk factors for failure of endoscopic

therapy and the need for subsequent esophagectomy in the setting of early neoplasia have not been well elucidated. The aims of this study were to determine the efficacy of radiofrequency ablation (RFA) with or without endoscopic mucosal resection (EMR) in the management of BE with HGD or IMC, to discern factors predictive of endoscopic treatment failure, and to assess the effect of endoscopic therapies on esophagectomy volume at our institution when undertaken for early esophageal neoplasia. METHODS Study Design Data were obtained from a prospectively maintained clinical database and medical record review of all patients who underwent endoscopic treatment or esophagectomy for a diagnosis of BE with HGD or IMC in the Division of Thoracic and Foregut Surgery, Department of Surgery, at the University of Rochester Medical Center between January 1, 2004, and December 31, 2012. Endoscopic therapies were initiated in 2007. The clinical records of 102 consecutive patients undergoing endoscopic therapies were available for review. After exclusion of 42 patients with only low-grade dysplasia and 3 patients who underwent EMR alone with detection of submucosal disease, 57 patients treated with RFA with or without EMR for HGD or IMC remained in the final study population. Patient demographics and treatment characteristics, including age, sex, body mass index, comorbidities, clinical factors, treatment efficacy, complications, and survival were collected and compared. The data were recorded in a Microsoft Excel database file (Microsoft Corp, Redmond, WA). Complete remission (CR) of BE or HGD of IMC was the primary end point. The secondary end point was survival from the time of initial treatment. Survival was assessed via medical record review and the Social Security Death Index (accessed on October 2, 2013) with patient status classified as

Seminars in Thoracic and Cardiovascular Surgery  Volume 26, Number 4

275

ENDOSCOPIC TREATMENT OF BE WITH EARLY ESOPHAGEAL NEOPLASIA alive or deceased. Total survival time was calculated by subtracting the date of the first ablation or EMR from the date of last follow-up or death. The study protocol was approved by the University of Rochester Medical Center Research Subjects Review Board. Definitions BE was defined by histologic identification of columnar epithelium with goblet cells in the esophageal biopsy specimens. HGD and IMC were defined by a histologic diagnosis made via consensus of at least 2 experienced gastrointestinal pathologists. Almost all cases (56/57, 98%) were reviewed at a gastrointestinal pathology review conference. Surveillance was defined as pretreatment endoscopic monitoring of known BE up to the time of referral. Multifocal dysplasia or carcinoma was defined by the presence of more than 1 area of dysplasia or carcinoma identified on histologic assessment. CR of BE or HGD or IMC was defined as 2 or more consecutive biopsy sessions at least 4 weeks apart without BE or HGD of IMC, and with no subsequent recurrence. Recurrence was defined by a biopsy result indicating BE or HGD/IMC after an initial remission. Progression was defined as the development of IMC after initial treatment for HGD or worsening of IMC to submucosal neoplasia or beyond. Major complications were defined as death, perforation, or bleeding requiring transfusion, whereas minor complications included strictures, minor bleeding, or admission to the hospital for observation or pain control. Follow-up was defined by the duration of time from initial treatment until the last known clinic visit. Endoscopic Therapies Procedural Descriptions Before any intervention, all patients underwent flexible esophagogastroduodenoscopy by a thoracic or foregut surgeon at our institution for the purpose of mapping and staging of disease. Endoscopic interventions were performed on an outpatient basis. A standard or high-definition flexible adult Olympus upper endoscope was used for visual inspection of the entire esophagus, stomach, pylorus, and duodenum up to at least the duodenal bulb. Areas of BE were carefully examined using combinations of white-light endoscopy and narrow-band imaging at the discretion of the surgeon. A standard biopsy protocol with 4-quadrant biopsies every 1 cm throughout the length of BE was used in all patients. Biopsies included the esophagus and gastric cardia. In addition to routine mapping biopsies, additional biopsies of any regions of mucosal nodularity or

276

irregularity were also obtained. EMR was used liberally on any such lesions with the intent of obtaining a generous biopsy for histologic analysis. All patients with any mucosal nodularity underwent EMR before ablative therapy. Some of the nodular areas had small ulcerations, which were endoscopically resected as well. Any patient with nonnodular or smooth BE underwent primary RFA rather than EMR. Similarly, no patients with mucosal abnormalities amenable to EMR underwent RFA before EMR. Radiofrequency Ablation RFA was performed with either the circumferential Barrx 360 catheter or any of a number of focal ablation catheters, most commonly the Barrx 90 (Covidien Medical, Minneapolis, MN). A sizing balloon was used to select a circumferential catheter of appropriate diameter. Radiofrequency current was applied at 12 J/cm2 along the entire extent of the Barrett mucosa, with slight overlap of the treatment zones in cases of long segments of BE. After application of the ablation balloon, the endoscope was reinserted with a fitted cap, which was used to debride the coagulum from the esophageal mucosa. Subsequently, the entire Barrett segment was reablated. The scope was then reinserted for final inspection. The Barrx90 procedure involved fitting the tip of the endoscope with the ablation device and focally ablating any areas of BE. Overall, 2 doses of 12 J/cm2 were applied twice at each site with debridement of the coagulum between applications. Endoscopic Mucosal Resection In cases of endoscopically visible mucosal irregularities or nodules, EMR was liberally used for pathologic staging before any ablative procedures. EMR was performed via the “suck-and-ligate” technique. An endoscope with a variceal banding device (Duette Multi-band Mucosectomy kit, Cook Medical, Inc, Bloomington, IN) attached to the tip was inserted. After opposing the cap to the mucosal lesion, suction was applied to draw the mucosa into the cap and a variceal band was fired, to create a pseudopolyp. An electrocautery snare was tightened just above or just below the rubber band. Resection was then performed at an energy level of 20-25 W. Resection typically occurred at the interface of the submucosa and the muscularis propria. The specimen was retrieved with a net or grasper. The EMR crater was inspected for bleeding or evidence of perforation. Histologic assessment was performed by experienced gastrointestinal pathologists. If the EMR

Seminars in Thoracic and Cardiovascular Surgery  Volume 26, Number 4

ENDOSCOPIC TREATMENT OF BE WITH EARLY ESOPHAGEAL NEOPLASIA specimen revealed tumor characteristics deemed low risk for nodal metastasis, including tumor size r20 mm, tumor depth no greater than the muscularis mucosa, good-moderate differentiation, the absence of lymphovascular invasion (LVI), and a negative deep resection margin, ablation of the remaining BE was performed at subsequent treatment sessions. If tumor characteristics were unfavorable, patients were considered for esophagectomy with regional lymphadenectomy unless they were medically inoperable. Initial Staging and Follow-up Protocols Patients initially found to have IMC underwent a computed tomography (CT) scan of the chest and abdomen at the discretion of the surgeon. CT imaging was not used in cases of HGD alone. Positron emission tomography was not performed in any patient. Endoscopic ultrasonography was used before EMR only for larger mucosal nodules where concern existed for penetration into the muscularis propria (T2). Neither endoscopic ultrasonography nor CT was used routinely for follow-up surveillance. Endoscopic treatment and follow-up protocols involved combinations of ablation, endoscopic resection, and surveillance mapping biopsies. Ablations and EMR sessions typically took place at 2-month intervals, and no more frequently than every 4 weeks. Between treatment sessions, interval mapping biopsies including any suspicious areas were performed approximately every 2 months at the discretion of the endoscopist. Patients were maintained on acid suppression consisting of a proton pump inhibitor at least twice daily throughout the treatment and surveillance periods. Persistent BE, dysplasia, or neoplasia proven on histologic assessment was treated within 4-12 weeks with repeat ablation or resection. Once all BE had been eradicated as proven on 2 successive biopsies spaced 3 months apart, the surveillance interval was increased to every 6 months for 1-2 years and then to every year thereafter, assuming the lack of recurrent BE or dysplasia. Statistical Analysis Statistical analysis was performed using SPSS statistical software (SPSS Inc, version 21.0, Chicago, IL). Continuous variables were analyzed via descriptive statistics such as means, medians, standard deviations, and interquartile ranges (IQRs) and were compared via the unpaired t test or Mann-Whitney U test as appropriate. Categorical variables were compared via Fischer exact or chi-square test. Survival

analysis was carried out via the Kaplan-Meier method, with the log-rank test to determine significant differences. P o 0.05 was considered statistically significant. RESULTS Patient Characteristics A total of 57 patients with HGD (n = 45) or IMC (n = 12) underwent RFA with (n = 35) or without (n = 22) EMR between January 1, 2007, and December 31, 2012. The mean age for the cohort was 66.2 ⫾ 1.4 years, measured at the time of the initial ablation or EMR. Most patients (50/57, 88%) were male. The median body mass index was 29.4 kg/m² (range: 19.9-46.3 kg/m2). The most common comorbidities included hypertension in 57% (33/57), hypercholesterolemia in 37% (21/57), and coronary artery disease in 33% (19/57). The most common primary symptoms were heartburn (40/57, 70%), regurgitation (4/57, 7%), and dysphagia (4/57, 7%), whereas the most common secondary symptom was regurgitation (19/57, 33%). A history of GERD and proton pump inhibitor use was present in most patients (51/ 57, 89%). Known BE before the first intervention was present in 32 of 57 patients (56%), all of whom were part of a surveillance program, with a mean surveillance duration of 55.6 ⫾ 12.3 months. Most of these patients (31/32, 97%) were undergoing surveillance by their referring gastroenterologists. Previous foregut surgery had been undertaken in 8 patients, including Nissen fundoplication in 6, resection of a gastrointestinal stromal tumor in 1, and a cricopharyngeal myotomy with Zenker diverticulectomy in 1 (Table 1). Treatment The pretreatment biopsies revealed HGD in 45 patients and IMC in 12 patients. The extent of circumferential BE (C) as per the Prague classification12 was 4.1 ⫾ 0.43 cm, and the maximal extent of any tongue of BE (M) was 5.1 ⫾ 0.46 cm. More than half of patients, 32 of 57 (56%), had an endoscopically visualized hiatal hernia with mean hernia size of 1.8 ⫾ 0.2 cm. Nodular mucosal changes were present in 22 of 57 (39%) patients and ulceration in 6 of 57 (11%) patients. Multifocal dysplasia was present in 25 of 57 (44%) patients. EMR was used at some point during the treatment course in 35 of 57 (61%) patients, whereas EMR was the initial procedure in 28 of 57 (49%) patients. A total of 181 ablation sessions (median ¼ 3 ablation sessions per patient, range: 1-8) and 58 EMR sessions (median ¼ 1 EMR session per patient, range: 0-4) were performed for a median of 4 therapeutic procedures per patient (range: 1-11).

Seminars in Thoracic and Cardiovascular Surgery  Volume 26, Number 4

277

ENDOSCOPIC TREATMENT OF BE WITH EARLY ESOPHAGEAL NEOPLASIA Table 1. Patient Characteristics Characteristics Demographics Age (y) Sex (male) BMI (kg/m2) Comorbidities CAD Hypertension Hypercholesterolemia Diabetes mellitus Smoking Alcohol Primary symptoms Heartburn Dysphagia Regurgitation Epigastric pain Melena Chest pain Cough Asymptomatic

Mean ⫾ SD or n (%)

Mean ⫾ SD or n (%)

Characteristics Secondary symptoms Heartburn Dysphagia Regurgitation Epigastric pain Bloating Chest pain Throat clearing None Reflux history GERD PPI Surveillance of BE Surveillance duration (mo) Previous foregut surgery Nissen fundoplication Zenker diverticulectomy GIST resection

66.2 ⫾ 1.4 50 (88) 29.7 ⫾ 0.6 19 (33) 33 (57) 21 (37) 9 (16) 13 (23) 11 (19) 40 (70) 4 (7) 4 (7) 1 (2) 2 (4) 2 (4) 2 (4) 2 (4)

4 (7) 5 (9) 19 (33) 1 (2) 1 (2) 1 (2) 1 (2) 25 (44) 51 (89) 51 (89) 32 (56) 55.6 ⫾ 12.3 8 (14) 6 (11) 1 (2) 1 (2)

BMI, body mass index; CAD, coronary artery disease; GIST, gastrointestinal stromal tumor; PPI, proton pump inhibitor; SD, standard deviation.

When surveillance endoscopies were included (514 procedures, median 9 endoscopies per patient, range: 1-23), the 57 patients underwent 753 procedures (median ¼ 13 procedures per patient, range: 3-33). A fundoplication after ablation was used in 2 patients. Alternatively, ablation after fundoplication was performed in 6 patients (Table 2). Remission, Recurrence, and Progression CR of cancer was achieved in all patients (12/12) at a median of 6.1 months (IQR: 1.4-7.5 months), CR of dysplasia was achieved in 49 of 57 patients (79%) at a median of 11.5 months (IQR: 5.3-26.8 months), and CR of BE occurred in 28 of 57 patients (49%) at a median of 18.4 months (IQR: 9.6-34.5 months). Recurrence of dysplasia occurred in 12 of 57 patients (21%) at a median of 29.1 months (IQR: 12.3-38.0), whereas recurrence of BE occurred in 4 of 57 patients (7%) at a median of 14.8 months (IQR: 5.3-34.6). Progression from HGD to IMC during treatment occurred in 4 of 57 patients (7%), all successfully retreated endoscopically with combinations of EMR and ablation without persistent cancer. The median follow-up of the 12 patients who had recurrence of HGD was 56.4 months (range: 20.277.7). The median follow-up of the 4 patients who had recurrence of BE was 40.9 months (range: 34.381.5). The median follow-up of the 4 patients who progressed to IMC was 42.6 months (range: 27.277.7). None of the 57 patients had disease

278

Table 2. Treatment Characteristics Characteristics

Mean ⫾ SD, n (%) or Median (Range)

Pretreatment pathology HGD 45 (79) IMC 12 (21) Extent of BE (cm) Circumferential BE Maximum BE

4.1 ⫾ 0.43 5.1 ⫾ 0.46

Endoscopic findings Hiatal hernia Hiatal hernia size (cm) Nodules Ulceration Multifocal

32 (56) 1.8 ⫾ 0.24 22 (39) 6 (11) 25 (43)

Procedures EMR ever EMR first Total ablation sessions Total EMR sessions Total surveillance endoscopies Total procedures Ablation after fundoplication Fundoplication after ablation Follow-up (mo)

35 (61) 28 (49) 181 (3.0, range: 1.0-8.0) 58 (1.0, range: 0-4.0) 514 (9.0, range: 1.0-23.0) 753 (13.0, range: 3.0-33.0) 6 (11) 2 (4) 35.4 (18.5-52.0)

SD, standard deviation.

Seminars in Thoracic and Cardiovascular Surgery  Volume 26, Number 4

ENDOSCOPIC TREATMENT OF BE WITH EARLY ESOPHAGEAL NEOPLASIA Table 3. Remission, Recurrence, and Progression Characteristics Remission CR IMC Time to CR (mo) CR of dysplasia Time to CR (mo) CR of BE Time to CR (mo) Recurrence Recurrence IMC Recurrence dysplasia Time to recurrence (mo) Recurrence IM Time to recurrence (mo) Progression HGD - IMC Time to progression (mo) IMC - submucosal disease Esophagectomy

Median (IQR) or n (%) 12 (100) 6.1 (1.4-7.5) 45 (79) 11.5 (5.3-26.8) 28 (49) 18.4 (9.6-34.5) 0 (0) 12 (21) 29.1 (12.3-38.0) 4 (7) 14.8 (5.3-34.6) 4 (7) 12.6 (6.2-22.5) 0 (0) 0 (0)

SD, standard deviation.

progression into the submucosa, required esophagectomy, or developed metastatic disease (Table 3). Shorter surveillance duration (17.8 ⫾ 7.6 months vs 63.9 ⫾ 14.4 months, P ¼ 0.009), fewer total procedures (12.1 ⫾ 0.8 months vs 17.4 ⫾ 2.1 months, P ¼ 0.009), and shorter posttreatment follow-up (21.1 ⫾ 6.1 months vs 43.2 ⫾ 4.1 months, P ¼ 0.007) were the only significant factors for failure to eradicate dysplasia (Table 4). On multivariate Cox regression analysis, none of these factors was significant as an independent predictor (shorter surveillance duration—hazard ratio (HR) ¼ 1.004, 95% CI: 0.998-1.011, P ¼ 0.205; fewer total procedures—HR ¼ 0.954, 95% CI: 0.867-1.050, P ¼ 0.335; and shorter follow-up time—HR ¼ 1.012, 95% CI: 0.984-1.041, P ¼ 0.407). Factors associated with failure to eradicate BE included increasing maximum length of BE (M) (6.0 ⫾ 0.6 cm vs 4.0 ⫾ 0.6 cm, P ¼ 0.030) and shorter duration of follow-up (28.5 ⫾ 3.8 months vs 49.0 ⫾ 5.8 months, P ¼ 0.004) (Table 5). None were significant as an independent predictor (increasing maximum length of BE—HR ¼ 0.894, 95% CI: 0.765-1.046, P ¼ 0.163; and shorter duration of follow-up— HR ¼ 0.999, 95% CI: 0.982-1.016, P ¼ 0.899). Patients developing recurrence of dysplasia were found to be older (73.8 ⫾ 10.5 vs 64.2 ⫾ 1.4 years, P ¼ 0.003) and to have undergone more total procedures, including therapeutic interventions and posttreatment surveillance endoscopies (17.4 ⫾ 2.1 vs 12.1 ⫾0.8, P ¼0.009), than those remaining dysplasia free. Likewise, a longer duration of posttreatment

follow-up was also associated with a higher incidence of recurrent dysplasia (52.1 ⫾ 5.8 months vs 34.9 ⫾ 4.3 months, P ¼ 0.025). Interestingly, patients with recurrence of dysplasia had a significantly longer survival (57.8 ⫾ 4.5 months vs 35.5 ⫾ 2.6 months, P o 0.001) than those without, a finding likely related to their longer period of posttreatment follow-up (Table 6). Again, none of these factors were significant as an independent predictor (increasing age— HR ¼ 1.026, 95% CI: 0.905-1.163, P ¼ 0.691; more total procedures—HR ¼ 0.798, 95% CI: 0.599-1.06, P ¼ 0.126; and longer follow-up time—HR ¼ 0.946, 95% CI: 0.868-1.031, P ¼ 0.206). Complications None of the 57 patients undergoing RFA with or without EMR died because of the procedure, experienced a perforation, or had major bleeding. However, 2 patients experienced minor complications, including 1 admission to the hospital for pain control and 1 symptomatic esophageal stricture requiring 3 dilations (Table 7). Postprocedural pain was common after RFA, though was not considered a complication unless hospital admission was required. Survival Patients achieving a CR of dysplasia had a median survival of 42.3 months (IQR: 27.4-53.9 months) vs 34.2 months (IQR: 12.5-53.1 months) for patients without CR of dysplasia (P ¼ 0.145). Patients achieving a CR of BE had a median survival of 42.5 months (IQR: 31.7-54.4 months) vs 37.3 months (IQR: 17.0-51.6 months) for patients without CR of BE (P ¼ 0.188). For the entire cohort, the median survival from the time of first ablation or EMR was 40.4 months (range: 8.878.3). Overall, 6 patients died during the follow-up interval, none from esophageal cancer (1 from hemorrhagic stroke, 1 from cardiac arrest, and 4 older than 80 years of age who died of natural causes). Based on Kaplan-Meier analysis, the actuarial 5-year survival was estimated at 77% and the 5-year disease-free survival was estimated at 100%. Esophagectomy Data Between 2004 and 2012, 84 patients underwent a primary esophagectomy for the diagnosis of HGD or IMC. The use of esophagectomy as primary therapy for BE with HGD or IMC has decreased since endoscopic therapies were initiated in 2007. From a maximum of 16 esophagectomies per year for early Barrett neoplasia in 2006, only 3 such resections were undertaken in 2012, all for IMC. No esophagectomies have been performed for HGD since

Seminars in Thoracic and Cardiovascular Surgery  Volume 26, Number 4

279

ENDOSCOPIC TREATMENT OF BE WITH EARLY ESOPHAGEAL NEOPLASIA Table 4. Comparison of Patients With and Without Complete Remission of Dysplasia CR (N ¼ 45) n (% of N) or Mean ⫾ SD

Non-CR (N ¼ 12) n (% of N) or Mean ⫾ SD

P Value

Demographics Age (y) Sex Male BMI CAD Hypertension Hyperlipidemia Diabetes mellitus Smoking Alcohol

65.7 ⫾ 1.5

68.1 ⫾ 2.8

0.459

38 (84) 29.6 ⫾ 0.6 16 (36) 25 (56) 15 (33) 7 (16) 21 (47) 8 (18)

12 (100) 30.1 ⫾ 1.6 3 (25) 8 (67) 6 (50) 2 (17) 1 (8) 3 (25)

0.172 0.715 0.374 0.393 0.232 0.614 0.161 0.421

Reflux history GERD PPI use Surveillance of known BE Surveillance duration (mo) Previous foregut surgery

41 (91) 42 (93) 25 (56) 63.9 ⫾ 14.4 6 (13)

10 (83) 9 (75) 6 (50) 17.8 ⫾ 7.6 2 (17)

0.372 0.101 0.362 0.009 0.541

Treatment Circumferential BE (cm) Maximum BE (cm) Hiatal hernia Hiatal hernia size (cm) Nodules Ulcerations Multifocal Total ablations EMR first EMR ever EMR total Total interventions Total procedures Fundoplication after ablation Ablation after fundoplication

4.0 ⫾ 0.5 4.8 ⫾ 0.5 25 (56) 1.8 ⫾ 0.3 17 (38) 4 (9) 21 (47) 3.3 ⫾ 0.3 24 (53) 30 (67) 1.1 ⫾ 0.2 4.4 ⫾ 0.4 14.2 ⫾ 0.9 2 (4) 6 (13)

4.7 ⫾ 6.2 ⫾ 7 (58) 2.1 ⫾ 5 (42) 2 (17) 4 (33) 2.6 ⫾ 4 (33) 5 (42) 0.8 ⫾ 3.3 ⫾ 9.5 ⫾ 0 (0) 0 (0)

0.495 0.214 0.564 0.587 0.529 0.381 0.312 0.228 0.183 0.107 0.352 0.163 0.040 0.620 0.224

Outcomes Progression to IMC Survival (mo) Follow-up (mo)

2 (4) 42.1 ⫾ 2.7 43.2 ⫾ 4.1

2 (17) 33.2 ⫾ 6.6 21.1 ⫾ 6.2

0.8 1.0 0.6

0.5

0.3 0.6 1.9

0.192 0.162 0.007

BMI, body mass index; CAD, coronary artery disease; PPI, proton pump inhibitor; SD, standard deviation.

2008. The reasons for primary esophagectomy in the 42 patients with HGD or IMC treated since 2007 were patient preference in 13 (31%), long-segment BE in addition to multifocal HGD in 10 (24%), poorly differentiated grade in 7 (17%), surgeon recommendation in 5 (12%), long-segment BE only in 3 (7%), multifocal HGD only in 3 (7%), and the presence of ulceration in 1 (2%) (Fig.). DISCUSSION The introduction of endoscopic ablation, most notably RFA, and resection into clinical practice has dramatically altered the treatment options for BE

280

with HGD or IMC over the past several years. Our experience showed that RFA in combination with EMR resulted in CR of IMC in all patients, CR of HGD in most (79%), and CR of BE in nearly half (49%). When encountered, cancer or dysplasia recurrences were managed successfully with endoscopic techniques. No patient started on a course of endoscopic treatment for BE with HGD or IMC subsequently required esophagectomy in our center. Further, no patients developed nodal or systemic metastases or died of their disease. An extensive literature regarding the safety and efficacy of endoscopic ablation and resection in the

Seminars in Thoracic and Cardiovascular Surgery  Volume 26, Number 4

ENDOSCOPIC TREATMENT OF BE WITH EARLY ESOPHAGEAL NEOPLASIA Table 5. Comparison of Patients With and Without Complete Remission of Barrett Esophagus CR (N ¼ 28) n (% of N) or Mean ⫾ SD

Non-Cr (N ¼ 29) n (% of N) or Mean ⫾ SD

P Value

Demographics Age (y) Sex Male BMI CAD Hypertension Hyperlipidemia Diabetes mellitus Smoking Alcohol

64.5 ⫾ 2.0

67.6 ⫾ 1.8

0.277

24 (86) 28.7 ⫾ 0.6 9 (32) 15 (54) 8 (29) 3 (11) 7 (25) 6 (21)

26 (89) 30.6 ⫾ 1.0 10 (34) 18 (62) 13 (45) 6 (21) 6 (21) 5 (17)

0.480 0.125 0.537 0.412 0.159 0.253 0.500 0.474

Reflux history GERD PPI use Surveillance of known BE Surveillance duration (mo) Previous foregut surgery

25 (89) 27 (96) 15 (54) 76.5 ⫾ 20.6 4 (14)

26 (90) 24 (83) 16 (55) 31.54 ⫾ 8.7 4 (14)

0.648 0.105 0.565 0.068 0.480

Treatment Circumferential BE (cm) Maximum BE (cm) Hiatal hernia Hiatal hernia size (cm) Nodules Ulcerations Multifocal Total ablations EMR first EMR ever EMR total Total interventions Total procedures Fundoplication after ablation Ablation after fundoplication

3.4 ⫾ 0.6 4.0 ⫾ 0.6 15 (54) 1.5 ⫾ 0.3 8 (29) 3 (11) 10 (36) 3.5 ⫾ 0.4 12 (43) 17 (61) 1.07 ⫾ 0.2 4.6 ⫾ 0.5 14.5 ⫾ 1.1 1 (4) 4 (14)

4.8 ⫾ 0.6 6.0 ⫾ 0.6 17 (59) 2.2 ⫾ 0.4 14 (28) 3 (10) 15 (52) 2.8 ⫾ 0.3 16 (55) 18 (62) 0.96 ⫾ 0.2 3.8 ⫾ 0.37 12.0 ⫾ 1.1 1 (3) 2 (7)

0.102 0.027 0.453 0.146 0.104 0.630 0.171 0.169 0.253 0.566 0.725 0.210 0.135 0.746 0.318

Outcomes Progression to IMC Survival (mo) Follow-up (mo)

1 (4) 43.2 ⫾ 3.0 49.0 ⫾ 5.8

3 (10) 37.3 ⫾ 4.1 28.5 ⫾ 3.8

0.319 0.244 0.004

BMI, body mass index; CAD, coronary artery disease; PPI, proton pump inhibitor; SD, standard deviation.

management of BE with HGD has evolved. As a result, gastroenterological and surgical societies in the United States, as well as the National Comprehensive Cancer Network, have recommended endoscopic therapies as the treatments of choice for HGD, relegating esophagectomy to a minority of cases.8-11 Our experience is consistent with such recommendations; esophagectomy has disappeared as a treatment for HGD in our institution since 2008. Specialty society guidelines have not yet taken a definitive stance on the optimal therapy for BEassociated IMC despite a considerable body of data supporting the role of endoscopic approaches. In the

largest experience to date with EMR specifically for IMC, 1000 consecutive patients were assessed at a mean follow-up of 56.6 ⫾ 33.4 months.13 The CR rate was 96.3%, with esophagectomy being necessary in 3.7% after failed endoscopic treatment. Metachronous or recurrent cancer developed in 14.5% of patients during their follow-up period, though endoscopic retreatment was successful in most. Importantly, only 2 of the 1000 patients (0.2%) died of esophageal cancer, a rate below the 1% perioperative mortality observed after esophagectomy in specialty centers for early-stage disease.7 Of the 2026 individuals referred to those authors

Seminars in Thoracic and Cardiovascular Surgery  Volume 26, Number 4

281

ENDOSCOPIC TREATMENT OF BE WITH EARLY ESOPHAGEAL NEOPLASIA Table 6. Comparison of Patients With and Without Recurrence of Dysplasia Recurrence (N ¼ 12) n (% of N) or Mean ⫾ SD

No Recurrence (N ¼ 47) n (% of N) or Mean ⫾ SD

P Value

Demographics Age (y) Sex Male BMI CAD Hypertension Hyperlipidemia Diabetes mellitus Smoking Alcohol

73.4 ⫾ 3.0

64.2 ⫾ 1.4

0.003

10 (83) 29.2 ⫾ 1.0 6 (50) 5 (42) 3 (25) 3 (25) 4 (33) 1 (8)

40 (85) 29.8 ⫾ 0.7 13 (28) 28 (60) 18 (38) 6 (13) 9 (19) 10 (21)

0.459 0.683 0.151 0.249 0.272 0.281 0.282 0.262

Reflux history GERD PPI use Surveillance of known BE Surveillance duration (mo) Previous foregut surgery

10 (83) 11 (92) 5 (42) 33.0 ⫾ 15.0 1 (8)

41 (87) 40 (85) 26 (54) 58.4 ⫾ 13.7 7 (15)

0.372 0.628 0.259 0.536 0.541

4.2 ⫾ 0.5 5.4 ⫾ 0.5 7 (58) 2.7 ⫾ 0.7 2 (17) 1 (8) 3 (25) 4.0 ⫾ 0.7 3 (25) 6 (50) 1.0 ⫾ 0.4 5.0 ⫾ 0.7 17.4 ⫾ 2.1

3.6 ⫾ 1.0 3.9 ⫾ 0.9 25 (53) 1.6 ⫾ 0.2 20 (43) 5 (11) 22 (57) 3.0 ⫾ 0.3 25 (53) 29 (62) 1.0 ⫾ 0.2 4.0 ⫾ 0.2 12.1 ⫾ 0.8

0.581 0.201 0.564 0.073 0.074 0.665 0.123 0.091 0.059 0.278 0.952 0.191 0.009

6 (50) 6 (50) 0 (0) 1 (8)

16 (73) 29 (83) 2 (4) 5 (11)

0.278 0.376 0.620 0.628

2 (17) 57.8 ⫾ 4.5 52.1 ⫾ 5.8

2 (4) 35.5 ⫾ 2.6 34.9 ⫾ 4.3

Treatment Circumferential BE (cm) Maximum BE (cm) Hiatal hernia Hiatal hernia size (cm) Nodules Ulcerations Multifocal Total ablations EMR first EMR ever EMR total Total interventions Total procedures Procedure combination RFA only RFA þ EMR Fundoplication after ablation Ablation after fundoplication Outcomes Progression to IMC Survival (mo) Follow-up (mo)

0.192 o0.001 0.025

BMI, body mass index; CAD, coronary artery disease; IMC, intramucosal adenocarcinoma; PPI, proton pump inhibitor; SD, standard deviation.

with suspected early esophageal neoplasia, 96 underwent esophagectomy as primary therapy after a diagnostic EMR, some presumably for IMC not deemed favorable for an endoscopic treatment paradigm, and were not included in the study cohort. Consistent with their experience, we have continued to use esophagectomy for IMC under select circumstances, though with decreasing frequency. Our current study was based on an assessment, treatment, and follow-up protocol that was rigorous

282

yet practical. The basic management tenets included the following: (1) thorough endoscopic visualization of the esophageal mucosa by an experienced surgical endoscopist to detect mucosal nodularity or irregularity that might harbor an invasive cancer; (2) liberal use of EMR to allow an accurate histologic assessment of any such abnormalities by experienced gastrointestinal pathologists; (3) strictly defined criteria for use of an endoscopic treatment pathway with curative intent, limiting such an approach to

Seminars in Thoracic and Cardiovascular Surgery  Volume 26, Number 4

ENDOSCOPIC TREATMENT OF BE WITH EARLY ESOPHAGEAL NEOPLASIA Table 7. Complications Characteristics

n (%)

Major Death Perforation Major bleeding

0 0 0 0

(0) (0) (0) (0)

Minor Symptomatic Stricture Admission for observation of pain Minor bleeding

2 1 1 0

(4) (2) (2) (0)

neoplasia deemed to be at low risk for nodal or systemic spread; (4) RFA of remaining BE at subsequent sessions, continuing every 2 months until all gross columnar-lined esophagus had been eradicated; and (5) detailed mapping biopsies of the esophagus and gastric cardia during posttreatment surveillance until a CR for both dysplasia and BE had been obtained. Patients were counseled about the need for multiple endoscopies, averaging 13 per patient in our experience, for the initial staging, subsequent treatment, and eventual surveillance of their disease, a process that required diligence by both the patient and the treating physician. The average number of endoscopies certainly will increase with ongoing surveillance, a potentially lifelong process. Our ability to achieve a CR for BE or HGD was time dependent, a factor explained by the need for multiple treatment sessions. Although esophagectomy is a one-time treatment that might be considered radical prophylaxis for microscopic neoplasia, less invasive endoscopic therapies are clearly a multiyear commitment. Given the usefulness of EMR as both a diagnostic and therapeutic tool, it must be emphasized that if the EMR specimen revealed factors suggesting a significant risk of nodal metastasis, such as a tumor

Figure. Esophagectomy for HGD or IMC from 2004-2012.

invading the submucosa (T1b cancer), a large tumor 420 mm, poor differentiation, or LVI, esophagectomy was considered rather than attempts at curative endoscopic therapy. An extensive surgical literature, derived from series of esophagectomies during which an extensive lymphadenectomy had been performed, has shown the significant risk of nodal spread for tumors invading into the submucosa or deeper.14-18 As a result, we have considered the presence of submucosal neoplasia a contraindication to a curative endoscopic treatment approach, except in select circumstances for patients deemed to be at high risk for undergoing esophagectomy. The role of endoscopic resection for “low-risk” invasion into the superficial submucosa (sm1) is a matter of controversy, with data both supporting and refuting its appropriateness.19,20 Analysis of other tumor-related factors, especially tumor size, differentiation, and the presence of LVI, also determines a spectrum of risk for nodal spread that must be considered before embarking on a course of endoscopic or surgical treatment.21 Another important consideration is that recurrence rates for cancer and dysplasia as high as 30% have been reported after EMR, though most recurrences can be managed by endoscopic means.22 Previously described factors associated with recurrence include long-segment BE, piecemeal resections, tumors 42 cm, multifocal disease, and the lack of ablation of remaining BE.23 Cancer and dysplasia recurrence rates should be low if all remaining intestinal metaplasia is ablated aggressively after EMR of focal neoplastic lesions.24 Our study had several limitations. The series was analyzed retrospectively from our esophageal cancer and BE database. The number of patients from our single institution experience was relatively small, especially for IMC, and our follow-up period was only approximately 3 years on average. Larger numbers of patients followed for longer periods of time, ideally 5-10 years, will be required to confirm the strength of our conclusions, particularly regarding the efficacy of endoscopic therapies at prevention of more advanced malignancy. Other forms of mucosal ablation besides RFA (such as cryotherapy or argon plasma coagulation) or endoscopic resection (such as submucosal dissection) were not evaluated, as they have not been used in our unit. Owing to the lack of an esophagectomy specimen in patients undergoing endoscopic therapies, it is not possible to determine with absolute confidence that CR was achieved. Our definition of CR does not account for sampling error that may occur at the time of biopsies. Our rigorous biopsy protocol, including meticulous 4-quadrant biopsies of the esophagus

Seminars in Thoracic and Cardiovascular Surgery  Volume 26, Number 4

283

ENDOSCOPIC TREATMENT OF BE WITH EARLY ESOPHAGEAL NEOPLASIA and cardia as well as biopsies of any suspicious areas of mucosal nodularity or irregularity, hopefully minimized the false-negative rate. Another study limitation is that variables identified as risk factors on univariate analysis, such as the length of the pretreatment surveillance period and posttreatment follow-up, may in fact be proxies of disease rather than true risk factors. No independent risk factors were identified on multivariable regression analysis. Finally, the comprehensiveness of the pathology reports for our EMR specimens initially was suboptimal, though improved after 2011. Before that time, LVI was not routinely reported, and the pathologic assessment was less descript. Thus, some of the earlier EMR specimens may have harbored unfavorable histologic characteristics that were not

1. Ronkainen J, Aro P, Johansson SE, et al: Prevalence of Barrett’s esophagus in the general population: An endoscopic study. Gastroenterology 129(6):1825-1831, 2005 2. Rex DK, Cummings OW, Shaw M, et al: Screening for Barrett’s esophagus in colonoscopy patients with and without heartburn. Gastroenterology 125(6):1670-1677, 2003 3. Hvid-Jensen F, Pedersen L, Drewes AM, et al: Incidence of adenocarcinoma among patients with Barrett’s esophagus. N Engl J Med 365 (15):1375-1383, 2011 4. Blot WJ, McLaughlin JK: The changing epidemiology of esophageal cancer. Semin Oncol 26:2-8, 1999 (5 suppl 15) 5. Devesa SS, Blot WJ, Fraumeni Jr JF: Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 83(10):2049-2053, 1998 6. Siegel R, Ma J, Zou Z, et al: Cancer statistics, 2014. CA Cancer J Clin 64(1):9-29, 2014 7. Williams VA, Watson TJ, Herbella FA, et al: Esophagectomy for high grade dysplasia is safe, curative, and results in good alimentary outcome. J Gastrointest Surg 11(12):1589-1597, 2007 8. Fernando HC, Murthy SC, Hofstetter W, et al: The Society of Thoracic Surgeons practice guidelines series: Guidelines for the management of Barrett’s esophagus with high-grade dysplasia. Ann Thorac Surg 87:1993-2002, 2009 9. Spechler SJ, Sharma P, Souza RF, et al: American Gastroenterological Association medical position statement on the management of Barrett’s esophagus. Gastroenterology 140: 1084-1091, 2011 10. Wang KK, Sampliner RE, Practice Parameters Committee of the American College of

284

11.

12.

13.

14.

15.

16.

17.

detected or reported. Fortunately, even if this were the case, no patients developed metastatic disease during follow-up. CONCLUSION Our results have shown that the combination of RFA and EMR was safe and efficacious in eradicating BE with HGD or IMC, with no patients requiring esophagectomy, developing metastases, or dying of their disease. The need for esophagectomy has been eliminated for HGD, and greatly reduced for IMC, over the past several years at our institution. In experienced centers, the use of endoscopic therapies for cure of BE-associated early esophageal neoplasia appears appropriate in most cases and is justified as the new standard of care.

Gastroenterology: Updated guidelines 2008 for the diagnosis, surveillance and therapy of Barrett’s esophagus. Am J Gastroenterol 103: 788-797, 2008 National Comprehensive Cancer Network (NCCN): Clinical Practice Guidelines in Oncology (NCCN Guidelines). Esophageal and esophagogastric junction cancers. Version 1. 2014. Available at: . Available at: www.NCCN. org. Accessed July 12, 2014. Sharma P, Dent J, Armstrong D, et al: The development and validation of an endoscopic grading system for Barrett’s esophagus: The Prague C & M criteria. Gastroenterology 131 (5):1392-1399, 2006 Pech O, May A, Manner H, et al: Long-term efficacy and safety of endoscopic resection for patients with mucosal adenocarcinoma of the esophagus. Gastroenterology 146:652-660, 2014 Rice TW, Zuccaro Jr. G, Adelstein DJ, et al: Esophageal carcinoma: Depth of tumor invasion is predictive of regional nodal status. Ann Thorac Surg 65:787-792, 1998 Buskens CJ, Westerterp M, Lagarde SM, et al: Prediction of appropriateness of local endoscopic treatment for high-grade dysplasia and early adenocarcinomas by EUS and histopathologic features. Gastrointest Endosc 60: 703-707, 2004 Liu L, Hofstetter WL, Rashid A, et al: Significance of the depth of tumor invasion and lymph node metastasis in superficially invasive (T1) esophageal adenocarcinoma. Am J Surg Pathol 29:1079-1085, 2005 Altorki NK, Lee PC, Liss Y, et al: Multifocal neoplasia and nodal metastases in T1 esophageal carcinoma: Implications for endoscopic treatment. Ann Surg 247:434-439, 2008

18. Leers JM, DeMeester SR, Oezcelik A, et al: The prevalence of lymph node metastases in patients with T1 esophageal adenocarcinoma: A retrospective review of esophagectomy specimens. Ann Surg 253(2):271-278, 2011 19. Manner H, Pech O, Heldmann Y, et al: Efficacy, safety, and long-term results of endoscopic treatment for early stage adenocarcinoma of the esophagus with low-risk sm1 invasion. Clin Gastroenterol Hepatol 11 (6):630-635, 2013 20. Sepesi B, Watson TJ, Zhou D, et al: Are endoscopic therapies appropriate for superficial submucosal esophageal adenocarcinoma? An analysis of esophagectomy specimens. J Am Coll Surg 210:418-427, 2010 21. Lee L, Ronellenfitsch U, Hofstetter WL, et al: Predicting lymph node metastases in early esophageal adenocarcinoma using a simple scoring system. J Am Coll Surg 217:191-199, 2013 22. Gupta M, Iyer PG, Lutzke L, et al: Recurrence of esophageal intestinal metaplasia after endoscopic mucosal resection and radiofrequency ablation of Barrett’s esophagus: Results from a US Multicenter Consortium. Gastroenterology 145(1):79-86, 2013 23. Pech O, Behrens A, May A, et al: Long-term results and risk factor analysis for recurrence after curative endoscopic therapy in 349 patients with high-grade intraepithelial neoplasia and mucosal adenocarcinoma in Barrett’s oesophagus. Gut 9:1200-1206, 2008 24. Manner H, Rabenstein T, Pech O, et al: Ablation of residual Barrett’s epithelium after endoscopic resection: A randomized long-term follow-up study of argon plasma coagulation vs. surveiilance (APE study). Endoscopy 46(1): 6-12, 2014

Seminars in Thoracic and Cardiovascular Surgery  Volume 26, Number 4