Endoscopic Mucosal Resection in Barrett’s Esophagus John M. Poneros, MD, FASGE Assistant Professor Assistant Director of Endoscopy Division of Digestive and Liver Diseases New York Presbyterian/Columbia College of Physicians and Surgeons

Early ADC in BE: Definitions HGD: Dysplasia limited to crypts Intramucosal carcinoma: Dysplasia limited to lamina (IMCA) propria Invasive carcinoma: Dysplasia has invaded through muscularis mucosa (ie submucosal invasion, T1b) Esophagus lacks serosa, therefore no anatomic barriers to local invasion Network of lymphatics in submucosa

Barrett’s with HGD 30% of HGD patients will progress to esophageal ADC Have your path re-read at a tertiary care center Cribiform branching glands High nuclear:cytoplasmic ratio Loss of polarity Numerous mitoses Wang et al Am J Gastro 2008

Risk of Invasive ADC in HGD Risk for occult invasive ADC in HGD previously estimated at 40%, thus esophagectomy recommended for surgical candidates Recent meta-analysis suggests that risk overestimated when defining invasion to be submucosal (Stage T1b) and beyond 14 studies with total of 213 HGD cases taken to surgery 27 pts (12.7%) had invasive submucosal cancer at esophagectomy Konda et al. Clin Gastro Hep 2008

Visible Lesions in BE with HGD Presence of endoscopically visible lesions (ulcer, nodule, stricture) associated with occult invasive esophageal ADC HGD meta-analysis: 28 pts with visible lesions: 11% had T1b cancers 29 pts with no visible lesions:1 patient (3%) had T1b (submucosal) invasive cancer Konda et al. Clin Gastro Hep 2008

20 year experience with HGD/IMCA Retrospective review of all esophagectomy specimens at BWH who underwent surgery for HGD or IMCA 41 with HGD, 19 with IMCA Rate of submucosally invasive carcinoma was 6.7% (5% for pre-op HGD and 11% for pre-op IMCA) All 4 patients with submucosal invasion had either nodular or ulcerated mucosa on pre-op EGD Wang, Poneros et al GIE 2009

Risk of Lymph Node Metastasis at Esophagectomy Final Pathology HGD IMC Submucosal invasion Lymph Node Met Rate 0% 0-5% 25% Nigro et al. J Thorac Cardiovasc Surg 1999 Stein et al. Ann Surg 2000 Rice et al. J Thorac Cardiovasc Surg 2001 Hulscher et al NEJM 2002

Barrett’s with HGD

Role of EUS in HGD Patients referred for endoscopic therapy must be accurately staged (extensive biopsies and EUS) 7.5 MHz in all cases plus 12.5 or 20 MHz for elevated and/or depressed lesions Pech et al Am J Gastro 2008 EUS superior to CT in differentiating between patients with carcinoma in whom endoscopic therapy is suitable vs those in whom surgical treatment is required

Accuracy of EUS Rampado et al Ann Thorac Surg 2008 55 patients with superficial carcinoma of the esophagus who underwent EUS Compared to esophagectomy or mucosectomy Positive predictive value for submucosal invasion 67%, NPV 86%, sensitivity 88%, specificity 63% and diagnostic accuracy 75% Accuracy of EUS in evaluating LN metastases 71%, NPV 84%

EUS in HGD & IMCA in BE Columbia: EUS proved accurate staging in 41/48 patients (85%) vs EMR or surgery U Chicago: By detecting unsuspected malignant LNs or submucosal invasion, EUS-FNA changed clinical course of 20% of patients with HGD/IMCA referred for endoscopic therapy All patients with HGD or IM carcinoma should undergo a EUS prior to therapy Larghi et al GIE 2005 Shami et al Endoscopy 2006

Endoscopic Mucosal Resection (EMR) Saline lift technique vs Band ligator technique

EMR: Band ligator technique Duette

Endoscopic Mucosal Resection (EMR) Provides en bloc pathology specimen, which has been demonstrated to useful in staging invasion level 5% risk of bleeding; 1% risk of perforation Strictures are most common late complication: more likely if EMR used over 3/4th of circumference Prasad et al. Amer J Gastroenterol 2007 Larghi et al. GIE 2005 Peters et al. Am J Gastroenterol 2006

EMR in HGD/IMCA in BE EMR is diagnostic and therapeutic

EMR in HGD/IMCA in BE Retrospective review of 150 EMRs: EMR specimen histology led to change in diagnosis in 49% and relevant change in treatment in 30% Mayo study of 25 EMRs: Peters et al GIE 2008 Prasad et al Am J Gastro 2007 Esophagectomy staging consistent with preoperative EMR staging in all patients No patient with negative EMR margins had residual tumor at esophagectomy Submucosal invasion (T1b), 50% had residual carcinoma and 31% had metastatic LNs

EMR: Which technique is best? “Inject, suck and cut” vs “Band and snare” Both cap assisted Study comparing 20 EMR specimens from each technique found equivalent mean depths (0.51cms vs 0.50cms, p= 0.76) All specimens contained submucosa, allowing accurate staging Muscularis propria in 65% “band and snare” vs 50% “lift and snare” Abrams et al Endoscopy 2008

Beware the Double MM Increasing recognition of double muscularis mucosa in BE making staging by biopsy difficult Consider EMR for flat lesions read as invasive ADC prior to esophagectomy 122 EMRs: MM duplication in 67% whether or not dysplasia Carcinomas that invade through superficial MM into deep LP should be considered “intramucosal” rather than “submucosal” Some beginning to argue for explicit statement regarding depth of ADC invasion Hahn et al Am J Surg Path 2008 Lewis et al Am J Surg Path 2008

100 consecutive pts with early ADC in BE < 2cm in diameter Suck and cut technique 1.47 resections per patient No major complications (11 episodes of bleeding) Complete local remission in 99/100 after 1.9 months and max of 3 resections During follow-up of 36.7 months 11% recurrent or metachronous carcinomas found All treated with successfully with repeat resection EMR in HGD/IMCA: Long term follow-up Ell et al GIE 2007

132 patients managed endoscopically (ENDO) vs 46 treated surgically (SURG) EMR alone 75 (57%); EMR + PDT 75 (43%) Mean follow-up 43 months in ENDO and 64 months in SURG Cumulative mortality comparable: 17% ENDO vs 20% SURG Recurrent carcinoma detected in 12% of ENDO group, all successfully re-treated without impact on survival EMR in T1a BE ADC: Mayo Experience Prasad et al Gastro 2009

“At the current time it appears as if surveillance with intensive biopsies, endoscopic ablative techniques (most likely a combination of techniques) or esophagectomy may produce similar outcomes in retrospective cohort studies from expert centers.” “Selection of which of these therapies must be individualized and will depend on the expertise available in the patient’s community, the patient’s preferences and the gastroenterologists own experience.” 2008 ACG BE Guidelines for HGD Wang and Sampliner Am J Gastro 2008