Barrett's Esophagus Related Neoplasia (BERN) Project

1. BACKGROUND AND SIGNIFICANCE Burden of disease Barrett's esophagus (BE) is an acquired condition resulting from chronic gastroesophageal reflux disease and is a well recognized pre-malignant condition for the development of esophageal adenocarcinoma (EAC) (1, 2). BE is defined as the displacement of the squamocolumnar junction proximal to the gastroesophageal junction with the presence of intestinal metaplasia (15). This condition entails a 30 to 50 fold greater risk for the development of EAC and has an incidence of development of EAC that approaches 0.5% annually (3, 4). EAC is a highly lethal cancer and is the most rapidly increasing cancer in the United States and Western Europe with an incremental increase of 4-10% per year. Although survival rates have improved during the recent years in some countries, the overall 5-year survival rate is still a dismal 10% in most Western populations (5-7). Clinical strategies for preventing deaths from this cancer focus on techniques for identification of esophageal neoplasms in an asymptomatic, early, and curable stage. Therefore, endoscopic screening of subjects with chronic reflux symptoms has been recommended as a method of detecting BE and early cancer (8). Patients with BE are then routinely enrolled in surveillance programs in an attempt to identify those who might benefit from treatment at a pre-invasive stage of EAC with an ultimate aim of reducing cancer related deaths (8).

Limitations of current Barrett's surveillance protocol Current guidelines for surveillance include taking biopsy specimens from endoscopically visible mucosal abnormality followed by random 4-quadrant biopsies every 2 cm throughout the entire length of BE (15). Unfortunately, the effectiveness of this strategy is hampered by numerous factors. Apart from being labor intensive and time consuming for the patient; the accuracy of this protocol is limited by sampling errors. Biopsy specimens from short segments or tongues of columnar mucosa generally reveal intestinal metaplasia in only 40-60% of patients (16). In a study of 570 patients undergoing upper endoscopy, BE was suspected in 146 patients; however, only 60 patients had diagnosis confirmed by biopsy. Short segment BE (SSBE) was more frequently suspected than long segment BE (LSBE) but was correctly diagnosed only 25% of the time compared with 55% for LSBE (17). Similar to the distribution of the metaplastic tissue, the presence of dysplasia or early adenocarcinoma within a Barrett's segment is patchy and focal. Standard endoscopy and random biopsies might fail to detect these lesions. Early neoplastic lesions are not visible to the eye of the endoscopist at standard endoscopy and random biopsies sample only a small proportion of the epithelium at risk. Also, this "hit and miss" nature of biopsy increases the cost of the procedure and limits the reliability of histologic interpretation of dysplasia. The increasing use of endoscopy has led to more and more patients being diagnosed with BE and offered surveillance. Also, the number of patients being followed up, and to whom potentially curative therapy can be offered, have increased. The advent of newer methods of endoscopic treatment of high-grade dysplasia (HGD) and early EAC in the form of ablative therapy and EMR makes it highly desirable to diagnose dysplasia and EAC early in the disease process. This may help in alleviating the morbidity and mortality associated with esophagectomy for EAC. These compelling reasons coupled with the limitations of current surveillance protocols make the development of new approaches aimed at improving efficacy of Barrett's surveillance mandatory.

Novel imaging techniques and endoscopic therapies Significant effort has been expended on development of new GI techniques in order to provide a precise and even a "real time" endoscopic diagnosis. Chromoendoscopy is a technique that involves the application of agents to improve characterization of the mucosa resulting in selective uptake or enhancement of mucosal surface pattern (11). HRE units are equipped with charge-coupled devices with up to a million pixels that allow clear visualization of fine mucosal details which may facilitate the detection of early neoplastic lesions (9). NBI is a novel endoscopic technique that is based on the optical phenomenon that the depth of light penetration into tissues is dependent on the wavelength; the shorter the wavelength, the more superficial the penetration. Use of blue light with narrow band filters enables detailed imaging of the mucosal and vascular surface patterns within the BE segment with a high level of resolution and contrast without the need for chromoendoscopy. The main chromophore in esophageal tissues in the visible wavelength is hemoglobin, which has a maximum absorptive wavelength near 415 nm. This is within the wavelength for NBI and responsible for revealing the superficial vasculature (10).

Esophagectomy was considered as the criterion standard for treatment of patients with early EAC. However, it is associated with significant morbidity and mortality, even in experienced surgical hands and high-volume surgical centers. EMR has been used increasingly to replace surgery as a curative treatment modality for early EAC in patients with BE (12, 13). It allows effective local treatment of early cancer, histological analysis of all resected specimens and thus confirmation of diagnosis and complete resection of the lesion. Ablative therapies in BE patients have provided promising results as well (18, 19).

1. PRELIMINARY DATA Our group recently assessed the potential of NBI for the prediction of histology during screening and surveillance endoscopy in BE patients (10). Images obtained by this system were classified according to the mucosal (ridge/villous, circular, irregular/distorted) and vascular (normal, abnormal) patterns and correlated with histology in a prospective and blinded fashion. The sensitivity, specificity, and positive predictive value of the ridge/villous pattern for the diagnosis of intestinal metaplasia without high-grade dysplasia (HGD) were 93.5%, 87%, and 95% respectively. The sensitivity, specificity, and positive predictive value of the irregular/distorted patterns for HGD were 100%, 99%, and 95% respectively. These promising results have been validated by other investigators. Similarly, Dr. Sharma has published several studies involving the use of chromoendoscopy in patients with BE. In a study of 80 patients with columnar lined esophagus using indigo carmine dye and 115x magnification endoscopy, three mucosal patterns as stated above were identified. The presence of ridge/villous pattern for detecting intestinal metaplasia had high sensitivity, specificity, and positive predictive value (97%, 76%, and 92% respectively). Six patients had an irregular/distorted pattern and biopsies revealed HGD in all these patients (11). Our group has been active in evaluating the role of ablative therapies in patients with BE.

We recently assessed the long-term efficacy of achieving complete reversal (endoscopic and histologic) between multipolar electrocoagulation and argon plasma coagulation in patients with BE and assessed factors influencing successful ablation. A total of 35 BE patients were followed for at least 2 years following ablative therapy and complete reversal of BE was achieved in 70% of the patients, regardless of the ablative technique (19). The International Working Group on the Classification of Oesophagitis has been at the forefront for the classification and grading of acid-peptic related esophageal diseases such as erosive esophagitis and BE. Sharma et al developed and validated an endoscopic grading system for BE (Prague C \& M criteria). The criteria includes assessment of the circumferential (C) and maximal (M) extent of the endoscopically visualized BE segment along with endoscopic landmarks. This grading system demonstrated a high validity for the endoscopic assessment of visualized BE lengths (20).

1. RESEARCH DESIGN AND METHODS Overview Patients with known / suspected BE Pre-registration, informed consent

Standard endoscopy followed by HRE, NBI and chromoendoscopy

Digital video-recording of endoscopy findings (Subsequent de-identification of all patient information)

Target biopsies of suspected lesions

Removal of abnormal lesions with EMR / mucosal ablation

Pathologist for histologic interpretation

This is a part of a multicenter study initiated at Kansas City. Patients will be recruited from Amsterdam, Mainz, Wiesbaden and Kansas City.

Methods Physical examination and questionnaire All patients will complete a validated GERD questionnaire (GERQ) (21) that records duration, severity, and frequency of heartburn and regurgitation with supplemental information on medications such as acid suppressive therapy, aspirin, NSAIDs, etc. Demographic information and clinical findings such as height, weight and waist circumference will be recorded .

Endoscopic procedure The endoscopic procedures will be performed with a high-resolution endoscope (Olympus GIF-Q240Z, GIF-Q160Z, or more recent versions) that has the capability of performing NBI with the push of a switch. Standard methods of conscious sedation and cardiopulmonary monitoring will be used during each procedure. The distal esophagus will be carefully inspected for the presence of erosions, nodules and plaques. The presence and size of hiatal hernia will be recorded. A plastic disposable distal attachment cap (e.g. Olympus D-201-11802) with a free distal distance of 2-3 mm will be attached to the endoscope in order to fix mucosal areas of interest until images and biopsies have been taken. The endoscopist will be blinded to the previous BE histology of the patient. The classification of mucosal and vascular patterns will be performed in a standardized manner into groups as described above. Details of endoscopy findings and digital video recordings will be entered into an endoscopy form .