Screening

Although conceptually attractive, screening for BE is complicated because of a variety of factors. At present, there is no clear evidence that screening reduces mortality rates due to esophageal adenocarcinoma. In addition, although BE is most common in elderly white men with chronic gastroesophageal reflux disease (GERD), only 40% of patients in some population-based studies of BE prevalence reported recent GERD symptoms. Thus, screening using a history of recent GERD may miss over half of patients with BE. In addition, although obesity and dietary factors have been proposed as risk factors for the development of BE, there are currently no risk factors that reliably identify BE in asymptomatic patients. Furthermore, upper endoscopy, the current method of screening, is both costly and invasive. Given these difficulties, recent guidelines do not endorse population-based screening and only weakly recommend considering screening in elderly Caucasian men with frequent and long-standing GERD symptoms.

In the future, screening may become feasible via the development of less invasive, inexpensive, and more accurate screening modalities. Recently, dedicated esophageal capsules (Pillcam ESO and PillcamESO2, Given Imaging, Yokneam, Israel) have been used as a minimally invasive technique to identify BE. The second generation of these devices nearly doubles the field of view, increases the depth of field, contains multiple lenses, and obtains up to 18 frames per second. Unfortunately, despite these improvements, this technique was found to have limited sensitivity for BE detection in several studies, while being less cost effective than esophagogastroduodenoscopy (EGD)-based screening. Future advancements in capsule imaging, including single-fiber endoscopy using a tethered capsule with air insufflation capabilities, may make capsule-based BE screening a reality. It is likely, however, that the most attractive future screening tool will not depend on direct visualization of the esophageal mucosa. Instead, it would detect BE via the examination of serum, saliva, or stool specimen for biomarkers that reliably predict the presence of BE. The development of a sensitive and inexpensive test using these easily obtained specimen could provide cost-effective, population-based screening. This test could consist of a single, or more likely multiple, biomarkers that would predict the presence of BE. An early version of such a device has been developed and consists of a capsule sponge that is ingested and removed via a string attached to the sponge. Using a biomarker, the trefoil factor 3 gene, which is expressed at high levels on the luminal surface of BE, but not in the normal esophagus or stomach, a sensitivity of 78% and a specificity of 94% was achieved for the detection of BE. Using such a technique, those having a negative- or low-probability study would require no further evaluation, whereas those with a positive or high-probability test would undergo subsequent EGD to confirm the presence, length, and degree of dysplasia within the detected BE. A less optimal, but still useful, advance would be a similar test that could predict the likelihood of the development of BE in patients with clinical or endoscopic evidence of GERD.

Surveillance

At present, the frequency of endoscopic surveillance once BE has been detected is determined primarily by the highest grade of dysplasia present. In general, patients with newly diagnosed nondysplastic BE are recommended to have a repeat endoscopy in 1 year (primarily to detect dysplasia that may have been missed at the index endoscopy due to sampling error), followed by further surveillance endoscopies every 3 years. Patients with low-grade dysplasia (LGD) typically have a repeat endoscopy every 6 months, whereas patients with high-grade dysplasia (HGD) not electing immediate endoscopic or surgical therapy undergo surveillance biopsies at 3-month intervals. Although surveillance does seem to lead to increased detection of early-stage esophageal adenocarcinoma, clear evidence that this practice leads to a disease-specific mortality reduction is lacking. Furthermore, endoscopists frequently do not follow the recommended protocol, with one study showing an overall compliance rate of 51.2% in the United States. Similarly, poor adherence rates were also seen in the Netherlands, with worsening compliance with increasing length of Barrett’s epithelium seen in both countries. In addition, even if performed appropriately, only 4% to 6% of the metaplastic region is sampled, and routine surveillance of all patients with BE incurs considerable expense and inconvenience. Given these limitations, several enhancements to our current practice of endoscopic surveillance are desired.

First among these is the development of systematic reporting and the performance of a careful, standardized Barrett withdrawal technique, similar to what is currently emphasized for colonoscopy. Routine documentation of key esophageal landmarks, washing of the esophageal epithelium, usage of validated BE measuring systems (Prague Criteria), performance of a careful visual examination of the esophagus using advanced imaging technologies, and adherence to biopsy technique and protocols should be expected and used as an endoscopic quality metric. The use of his quality metric will hopefully improve dysplasia detection and reduce the need for repeat surveillance biopsies in patients being considered for endoscopic eradication therapies due to poor documentation or biopsy protocol. In addition, multicenter trials are needed to prospectively assess the impact of surveillance on disease-specific survival. Last, risk stratification to determine who would benefit most from this approach is arguably the most important future advance regarding surveillance and would markedly improve the cost effectiveness of this approach.

In addition to enhancements in the process of surveillance, several endoscopic imaging advancements have been made to improve the efficiency and diagnostic yield of surveillance biopsies in detecting dysplasia or early neoplasia. These advancements have included high-definition imaging, image magnification, and narrow band imaging (NBI) that are built in to many current endoscopes. In addition, technologies such as autofluorescence imaging and confocal laser endomicroscopy (CLE) have also been used to enhance dysplasia detection and to provide “real-time” histology, leading to the ability to provide immediate therapy via EMR. In the near future, surveillance protocols could be altered to incorporate many of these technologies. For example, it seems that NBI combined with high-definition white-light endoscopy may increase the detection of dysplastic lesions with a reduction in the number of needed biopsies. For the vast majority of patients without dysplasia, future surveillance could involve having the entire BE region being assessed with NBI/high-definition white light endoscopy, and if no high-risk lesions are seen, random 4-quadrant probe-based CLE could be performed in a fashion similar to obtaining biopsies. For patients without visualized abnormalities on CLE, biopsies could be avoided altogether, whereas those with abnormalities on either white-light, NBI, or probe-based CLE would undergo targeted biopsies of the identified potentially dysplastic regions in addition to random 4-quadrant biopsies every 2 cm. Such a strategy has the potential to reduce the number of biopsies taken in low-risk surveillance of nondysplastic BE by 10 fold, with about 2 of 3 of patients requiring no biopsies.

On the more distant horizon, techniques using in-vivo molecular imaging such as fluorescence endoscopy using proflavine or indocyanine green may supplant video endoscopy as the preferred method used to initially scan the entire region of BE for dysplasia. These methods may be used to detect highly sensitive and specific whole or partial antibodies, nanoparticles, or peptide markers of dysplasia/malignancy within the BE segment. These identified regions could then be marked and immediately interrogated via probe-based CLE and, if confirmed to contain abnormal tissue, undergo immediate EMR. This EMR would allow for a rapid, highly accurate method for immediate eradication of dysplasia/early cancer while avoiding the need for random biopsies altogether. Further advances in CLE are also anticipated, including multiphoton CLE that would avoid the need for an injected contrast agent (fluoroscein) by using autofluorescence from endogenous flavins and collagen. This device could provide subcellular imaging while adding information regarding biochemical/metabolic activity to the cellular imaging and structural data obtained by current single photon CLE devices. It is hoped that by using a combination of these approaches, a more limited, economical, and effective surveillance system can be developed and used in those most likely to benefit from close observation.

Surveillance

At present, the frequency of endoscopic surveillance once BE has been detected is determined primarily by the highest grade of dysplasia present. In general, patients with newly diagnosed nondysplastic BE are recommended to have a repeat endoscopy in 1 year (primarily to detect dysplasia that may have been missed at the index endoscopy due to sampling error), followed by further surveillance endoscopies every 3 years. Patients with low-grade dysplasia (LGD) typically have a repeat endoscopy every 6 months, whereas patients with high-grade dysplasia (HGD) not electing immediate endoscopic or surgical therapy undergo surveillance biopsies at 3-month intervals. Although surveillance does seem to lead to increased detection of early-stage esophageal adenocarcinoma, clear evidence that this practice leads to a disease-specific mortality reduction is lacking. Furthermore, endoscopists frequently do not follow the recommended protocol, with one study showing an overall compliance rate of 51.2% in the United States. Similarly, poor adherence rates were also seen in the Netherlands, with worsening compliance with increasing length of Barrett’s epithelium seen in both countries. In addition, even if performed appropriately, only 4% to 6% of the metaplastic region is sampled, and routine surveillance of all patients with BE incurs considerable expense and inconvenience. Given these limitations, several enhancements to our current practice of endoscopic surveillance are desired.

First among these is the development of systematic reporting and the performance of a careful, standardized Barrett withdrawal technique, similar to what is currently emphasized for colonoscopy. Routine documentation of key esophageal landmarks, washing of the esophageal epithelium, usage of validated BE measuring systems (Prague Criteria), performance of a careful visual examination of the esophagus using advanced imaging technologies, and adherence to biopsy technique and protocols should be expected and used as an endoscopic quality metric. The use of his quality metric will hopefully improve dysplasia detection and reduce the need for repeat surveillance biopsies in patients being considered for endoscopic eradication therapies due to poor documentation or biopsy protocol. In addition, multicenter trials are needed to prospectively assess the impact of surveillance on disease-specific survival. Last, risk stratification to determine who would benefit most from this approach is arguably the most important future advance regarding surveillance and would markedly improve the cost effectiveness of this approach.

In addition to enhancements in the process of surveillance, several endoscopic imaging advancements have been made to improve the efficiency and diagnostic yield of surveillance biopsies in detecting dysplasia or early neoplasia. These advancements have included high-definition imaging, image magnification, and narrow band imaging (NBI) that are built in to many current endoscopes. In addition, technologies such as autofluorescence imaging and confocal laser endomicroscopy (CLE) have also been used to enhance dysplasia detection and to provide “real-time” histology, leading to the ability to provide immediate therapy via EMR. In the near future, surveillance protocols could be altered to incorporate many of these technologies. For example, it seems that NBI combined with high-definition white-light endoscopy may increase the detection of dysplastic lesions with a reduction in the number of needed biopsies. For the vast majority of patients without dysplasia, future surveillance could involve having the entire BE region being assessed with NBI/high-definition white light endoscopy, and if no high-risk lesions are seen, random 4-quadrant probe-based CLE could be performed in a fashion similar to obtaining biopsies. For patients without visualized abnormalities on CLE, biopsies could be avoided altogether, whereas those with abnormalities on either white-light, NBI, or probe-based CLE would undergo targeted biopsies of the identified potentially dysplastic regions in addition to random 4-quadrant biopsies every 2 cm. Such a strategy has the potential to reduce the number of biopsies taken in low-risk surveillance of nondysplastic BE by 10 fold, with about 2 of 3 of patients requiring no biopsies.

On the more distant horizon, techniques using in-vivo molecular imaging such as fluorescence endoscopy using proflavine or indocyanine green may supplant video endoscopy as the preferred method used to initially scan the entire region of BE for dysplasia. These methods may be used to detect highly sensitive and specific whole or partial antibodies, nanoparticles, or peptide markers of dysplasia/malignancy within the BE segment. These identified regions could then be marked and immediately interrogated via probe-based CLE and, if confirmed to contain abnormal tissue, undergo immediate EMR. This EMR would allow for a rapid, highly accurate method for immediate eradication of dysplasia/early cancer while avoiding the need for random biopsies altogether. Further advances in CLE are also anticipated, including multiphoton CLE that would avoid the need for an injected contrast agent (fluoroscein) by using autofluorescence from endogenous flavins and collagen. This device could provide subcellular imaging while adding information regarding biochemical/metabolic activity to the cellular imaging and structural data obtained by current single photon CLE devices. It is hoped that by using a combination of these approaches, a more limited, economical, and effective surveillance system can be developed and used in those most likely to benefit from close observation.