Endoscopic Surveillance Overview

Given the dismal outcome among patients with symptomatic esophageal cancer, current practice guidelines recommend endoscopic surveillance of patients with Barretts esophagus in an attempt to detect cancer at an early and potentially curable stage. Several retrospective studies suggest that patients with Barretts esophagus in whom adenocarcinoma was detected in a surveillance program have their cancers detected at an earlier stage with resulting improved 5-year survival compared to similar patients not undergoing routine endoscopic surveillance. Furthermore, nodal involvement is far less likely in surveyed patients compared to nonsurveyed patients. Since esophageal cancer survival is stage-dependent, these studies suggest that survival may be enhanced by endoscopic surveillance. A recent decision-analysis study of the optimal surveillance strategy for Barrett’s esophagus with an endpoint of esophagectomy for high-grade dysplasia found that surveillance every five years was the most effective strategy to increase both length and quality of life. More frequent surveillance was more expensive and resulted in a lower life expectancy. The cost of surveillance at 5-year intervals compares favorably to screening programs for colon cancer, breast cancer and cervical cancer.

However, in this era of evidence based medicine, others argue that since most patients with Barrett’s esophagus will not die from esophageal cancer, endoscopic surveillance is not warranted until substantiated by prospective studies. While it is scientifically appealing to wait for evidence to support endoscopic surveillance, we still have the dilemma of dealing with individual patients at increased risk for the development of esophageal cancer.

The Risk of Esophageal Adenocarcinoma in an Individual Patient
While the incidence of esophageal adenocarcinoma continues to increase rapidly, most patients with Barrett’s esophagus never develop esophageal cancer. The reported incidence of esophageal cancer in Barrett’s esophagus varies widely, but recent studies suggest a risk of approximately 0.5% annually.

Development of Esophageal Adenocarcinoma from Barrett’s Esophagus
Cancer risk appears to be limited to patients with specialized columnar epithelium, a finding that may be explained by the increased rate of cellular proliferation encountered in these cells compared to cardiac or fundic type epithelium. Compelling evidence exists for a dysplasia-carcinoma sequence in Barretts esophagus whereby specialized columnar epithelium progresses to low-grade dysplasia, high-grade dysplasia, intramucosal carcinoma, and submucosal carcinoma. Foci of carcinoma typically appear adjacent to dysplasia. However, the time course for this progression is highly variable, and most patients do not even progress. Furthermore, the progression from dysplasia to carcinoma is not preordained, because many patients with low-grade dysplasia revert to no dysplasia, and some patients with high-grade dysplasia have been followed for years without developing overt carcinoma or even reverting to no dysplasia. These observations, which may simply reflect sampling error, serve to highlight the highly variable natural history of dysplasia.

It is hypothesized that cancer develops in a subset of patients who have acquired genomic instability in Barretts epithelium. This predisposes to the subsequent development of abnormal clones of cells that then accumulate more and more genetic errors. As genomic instability increases, abnormal clones of cells spread to occupy larger areas of the Barretts mucosa and one or more clones of cells may evolve the capacity for invasion. This genomic instability is associated with a variety of cell cycle abnormalities and a loss of control of cellular proliferation. However, there is no clearly predictable sequence of genetic abnormalities that leads to the development of cancer.

Candidates for Endoscopic Surveillance of Barrett’s Esophagus
Only patients at increased risk for the development of carcinoma, that is those with intestinal metaplasia, should undergo endoscopic surveillance. It is generally agreed that all otherwise healthy patients with Barrett’s esophagus should undergo surveillance, with an endpoint of either high-grade dysplasia or adenocarcinoma. Elderly patients or patients with comorbid illnesses who are not candidates for esophagectomy generally would not undergo surveillance or would be dropped from surveillance at a certain undetermined age. However, new ablation techniques currently under investigation may make more of these patients eligible for surveillance.

Surveillance Techniques
Current guidelines suggest obtaining systematic four quadrant biopsies at 2 cm intervals once inflammation related to GERD is controlled with antisecretory therapy. Such a rigorous, systematic biopsy protocol is safe and does not increase the risk of gastrointestinal bleeding or perforation. At the time of endoscopy, landmarks including the diaphragmatic hiatus, esophagogastric junction, and squamocolumnar junction should be carefully defined prior to commencing the biopsy protocol. Subtle mucosal abnormalities, no matter how trivial, such as ulceration, erosion, plaque, nodule, stricture, friability or other luminal irregularity in the Barretts segment, should be biopsied first, as they may be obscured by bleeding if bland areas of Barrett’s esophagus are first biopsied. Biopsy specimens should then be obtained in four quadrants at 2 cm intervals along the entire length of Barrett’s epithelium beginning at the end of the tubular esophagus as defined by the proximal margin of gastric folds and continuing to the squamocolumnar junction. The “turn and suction” technique allows acquisition of biopsies that are significantly larger than those obtained by the traditional technique of advancing the open forceps into the lumen and then closing it to obtain the biopsy sample.

The rationale for such a comprehensive biopsy program comes from observations that high-grade dysplasia and early carcinoma in Barretts esophagus often occur in the absence of endoscopic abnormalities and the focal nature of dysplasia. Systematic esophagectomy mapping studies demonstrate just how focal dysplasia and superficial cancer may be. In a study of 30 esophagectomy specimens from patients undergoing surgery for either high-grade dysplasia or early invasive adenocarcinoma with no endoscopic evidence of cancer, the median surface area of total Barretts esophagus was found to be 32 cm2; low-grade dysplasia 13 cm2; high-grade dysplasia 1.3 cm2; and adenocarcinoma 1.1 cm2. The three smallest cancers had surface areas of 0.02, 0.3 and 0.4 cm2. Invasive adenocarcinoma was associated with either focal or extensive areas of high-grade dysplasia. In work from the University of Washington, cancer was detected in only a single 1 cm segment of the esophagus in over 80% of patients with high-grade dysplasia who developed cancer. Thus, detection of adenocarcinoma in Barrett’s patients with high-grade dysplasia may be akin to looking for a “needle in a haystack.”

The aim of surveillance is the detection of dysplasia. The description of this condition should use a standard five tier system advocated by a number of authors:

1. negative for dysplasia
2. indefinite for dysplasia
3. low-grade dysplasia;
4. high-grade dysplasia
5. carcinoma

Dysplasia describes a change that is unequivocally neoplastic. It is characterized by nuclear pleomorphism, nuclear hyperchromatism, and an alteration in nuclear polarity. Active inflammation makes it more difficult to distinguish dysplasia from reparative changes. As such, surveillance endoscopy should not be done until any active inflammation related to GERD is controlled with antisecretory therapy, a fact that remains underappreciated by many gastroenterologists.

Surveillance Intervals
Surveillance intervals (Table 1), determined by the presence and grade of dysplasia, are based on our limited understanding of the biology of esophageal adenocarcinoma as outlined earlier. However, these intervals are arbitrary and have never been subjected to a clinical trial. Surveillance every two to three years is recommended as adequate in patients without dysplasia. If low-grade or indefinite dysplasia is diagnosed the surveillance interval is shortened to every 6 months for one year followed by annual surveillance. These patients should receive aggressive antisecretory therapy for reflux disease with a proton pump inhibitor to decrease the changes of regeneration that make pathologic interpretation of this category so difficult.

Table 1. 1998 American College of Gastroenterology Practice Guidelines for Endoscopic Surveillance of Barrett’s Esophagus

High-grade* Expert confirmation followed by either resection or continued surveillance every 3 months

*Note : For continued surveillance, current “Seattle” protocol involves
– 4 quadrant biopsies at 1 cm intervals
– endoscopy at 1 month, 3 months and then every 6 months

Adapted from Sampliner RE. Practice guidelines on the diagnosis, surveillance, and therapy of Barrett’s esophagus. Am J Gastroenterol 1998;93:1028-1032.

If high-grade dysplasia is found, the diagnosis should first be confirmed by an experienced gastrointestinal pathologist. The endoscopic biopsy protocol should then be repeated within one month to exclude an unsuspected carcinoma. Biopsies should now be obtained at 1 cm intervals with large particle (“jumbo”) forceps to maximize the ability to detect unsuspected cancer. However, unsuspected cancer may still be found in esophagectomy specimens of these patients with high-grade dysplasia if the biopsy protocol is at 2 cm intervals despite the use of a jumbo biopsy forceps. This technique requires passage of a therapeutic endoscope and its generalizability in clinical practice is problematic.

If high-grade dysplasia is confirmed, there is no agreement on the most appropriate management of these patients (Table 2). Esophagectomy is recommended by many authors to eliminate the risk of carcinoma or to detect and treat cancer at an early curable stage, because of the marked variability in the finding of unsuspected cancer in patients with high-grade dysplasia which ranges from 0% to 73%. However, problems with this approach include the potential risks of surgery, especially in older patients and the highly variable natural history of high-grade dysplasia. Surgical mortality in high volume centers is now less than 5% and in our hands at the Cleveland Clinic is 2%. Other experts recommend a continued rigorous endoscopic surveillance program utilizing the systematic biopsy protocol described above at 1 and 3 months after initial diagnosis and then every 6 months, reserving esophagectomy for patients with a preoperative diagnosis of intramucosal or submucosal carcinoma. The ultimate approach to the patient with high-grade dysplasia should consider factors such as available surgical expertise, age, length of Barretts epithelium that would require biopsy to eliminate sampling error, compliance with endoscopic surveillance, future need for multiple surveillance endoscopies, and suspicious lesions such as plaques, nodules and strictures. In the older, poor-risk patient, or in patients with a short segment of Barretts esophagus, continued surveillance may be a reasonable option. In the younger, good-risk patient, with a long segment of Barretts epithelium, and the anticipated need for numerous endoscopic procedures over the years, surgery may be a more reasonable option. The role of ablation therapy and endoscopic mucosal resection in this setting remains the subject of continued investigation. The niche of these minimally invasive approaches at present is in the elderly high-risk patients with cormorbid conditions.

Table 2. Approaches to High-Grade Dysplasia

Alternative Techniques To Sample Greater Areas Of Involved Mucosa
More efficient endoscopic surveillance is clearly a desirable goal. Currently, all Barrett’s esophagus patients are handled in a similar fashion unless dysplasia is present. However, in the future, it may be possible to make our surveillance techniques more effective by either sampling larger areas of Barrett’s mucosa or targeting our biopsies to areas with a higher probability of harboring dysplasia.

Brush cytology may be complementary to endoscopic biopsies and is recommended by some to be part of the routine endoscopic surveillance of Barrett’s patients. Cytology has a number of theoretic advantages compared to routine endoscopic biopsies: ability to sample a greater area of involved epithelium, preferential exfoliation of the less cohesive dysplastic cells, simplicity, and lower cost. There are clear cytologic criteria for dysplasia and sophisticated biomarker studies including staining for p53 and telomerase expression can be performed on cytologic specimens.

A variety of endoscopic optical techniques have the potential to obtain “light” biopsies of Barrett’s esophagus. Candidate techniques include fluorescence spectroscopy, light spectroscopy, optical coherence tomography, light scattering spectroscopy, and light induced fluorescence endoscopy. All of these techniques are based on the principle that benign and malignant tissue have different optical qualities. In theory, this would permit optical sampling of larger areas of the columnar-lined esophagus and improve the efficiency of biopsies by targeting areas thought to harbor dysplasia or cancer. While promising, these techniques need to decrease the overlap encountered between normal mucosa, Barrett’s epithelium and all grades of dysplasia. Magnification endoscopy combined with chromoendoscopy is also under investigation.

Risk Stratification in Endoscopic Surveillance Programs
One of the great dilemmas in endoscopic surveillance of Barrett’s esophagus is our inability to stratify patients without dysplasia by their risk of cancer. It is unclear what predisposes only a small subset of Barretts patients to develop cancer. A number of clinical and biologic markers may define patients at increased risk for the development of adenocarcinoma, but none have been subject to rigorous examination. Increased age, length of Barrett’s epithelium, hiatal hernia size and smoking have all been implicated in some, but not all studies as clinical risk factors for the development of adenocarcinoma.

The current goal of surveillance is the detection of dysplasia, which is still the best available marker of cancer risk. However, dysplasia is not distinguishable endoscopically, and the focal nature of dysplasia makes targeting of biopsies problematic. Furthermore there is considerable interobserver variability in the grading of dysplasia and the natural history of high-grade dysplasia is more variable than previously thought. Therefore, a less subjective marker for cancer risk that could supplement or replace the current dysplasia grading system is needed. A variety of biological markers have been studied but none provides the ultimate answer for risk stratification. For example, flow cytometric and p53 abnormalities may predate the subsequent development of dysplasia or cancer in some patients. A recent study by the University of Washington group found that among patients with negative, indefinite or low-grade dysplasia at entry with neither aneuploidy or increased 4N cell populations, the 5-year incidence of cancer was 0% compared to 28% in those patients with either aneuploidy or increased 4N fractions. A total of 35 patients developed cancer within 5 years and each of these patients had either high-grade dysplasia, aneuploidy or increased 4N at entry. These findings demonstrate the great, but to date unfulfilled potential of biomarkers in Barrett’s esophagus surveillance programs. However, follow-up is short and these flow cytometry results have not been duplicated outside of Seattle.

The ultimate answer for risk stratification may come from genomic profiling of Barrett’s patients using the new chip technology. Potentially, a genomic “fingerprint” could provide the consummate risk stratification tool for an individual patient. One could then envision lengthening endoscopic surveillance intervals for patients at low risk for developing adenocarcinoma and intensifying endoscopic surveillance intervals for patients at increased risk of developing adenocarcinoma. This would then conserve resources and allow for more aggressive screening of GERD patients or the general population for Barrett’s esophagus, the precursor lesion of most esophageal adenocarcinoma.