Segment Length

While esophageal adenocarcinoma can develop in both short and long segments of Barrett’s esophagus (traditionally defined as >3 cm), the understanding of the relationship between segment length and the risk of progression has evolved in recent years. A 2012 meta-analysis found a lower annual incidence of esophageal adenocarcinoma in patients with short-segment Barrett’s esophagus (<3 cm) than in the overall Barrett’s esophagus population (0.19% vs 0.33% per year). Work from the Northern Ireland Barrett’s Esophagus Registry demonstrated an increased risk for progression to adenocarcinoma or high-grade dysplasia in long-segment Barrett’s esophagus (hazard ratio [HR], 7.1; 95% confidence interval [CI], 1.74–29.04). A recent case-control study from Berlin also found an association of segment length with progression to adenocarcinoma or high-grade dysplasia. Patients with long-segment Barrett’s esophagus had an increased risk of progression when compared with those with short-segment Barrett’s esophagus (odds ratio [OR], 2.69; 95% CI, 1.48–4.88).

Newer studies have examined the relationship of segment length and risk of progression not only as a binary variable of long versus short but also as a continuous variable. In a large multicenter study, increasing segment length was an independent risk factor for neoplastic progression in patients with nondysplastic Barrett’s esophagus. Patients who progressed to adenocarcinoma or high-grade dysplasia had a longer Barrett’s segment (6.1 cm vs 3.5 cm). Perhaps more importantly, the risk for neoplastic progression increased by 28% for every 1-cm increase in length of the Barrett’s segment. Similarly, a Netherlands cohort study of more than 700 patients with nondysplastic Barrett’s esophagus or low-grade dysplasia confirmed the concept of increasing risk with increasing segment length. The relative risk of neoplastic progression to adenocarcinoma or high-grade dysplasia was 1.11 (95% CI, 1.01–1.2) per 1-cm increase in segment length. The recently completed SURF (Surveillance vs Radiofrequency Ablation) trial of radiofrequency ablation in patients with Barrett’s esophagus with low-grade dysplasia also found segment length to be an independent predictor of neoplastic progression in the surveillance arm of the study (OR, 1.35 per cm; 95% CI, 1.04–1.76).

However, progression does occur in shorter segments of Barrett’s esophagus, and a population-based study of over 8000 patients in Ireland found no relationship between segment length and risk of progression. Taken together, it would seem that longer segments of Barrett’s esophagus are associated with an increased risk of progression to adenocarcinoma or high-grade dysplasia.

Hiatal Hernia

Hiatal hernia is a well-documented risk factor for the development of Barrett’s esophagus. In addition, some data suggest that a larger hiatal hernia size may increase the risk of neoplastic progression in Barrett’s esophagus (OR, 1.20 per cm hiatal hernia; 95% CI, 1.04–1.39). In a cohort study of 550 patients from the Veterans Affairs Medical Center, Kansas City, a large hiatal hernia (>6 cm) was associated with an increased risk of neoplastic progression to adenocarcinoma or high-grade dysplasia when compared with patients with no hiatal hernia. However, other studies find contrary results. A 2013 case-control study of approximately 600 patients demonstrated that although the presence of hiatal hernia increased the risk for Barrett’s esophagus, it did not increase the risk of neoplastic progression to adenocarcinoma or high-grade dysplasia. Overall, it is unclear if hiatal hernia size is an independent risk factor for neoplastic progression in Barrett’s esophagus.

Mucosal Abnormalities

Several mucosal changes within the Barrett’s segment are associated with an increased risk of progression to adenocarcinoma or high-grade dysplasia. Erosive esophagitis has emerged as a potential risk factor for esophageal adenocarcinoma. A Dutch multicenter cohort study found an increased risk of progression to adenocarcinoma or high-grade dysplasia in patients with Barrett’s esophagus with esophagitis at baseline endoscopy (risk ratio [RR], 3.5; 95% CI, 1.3–9.5). A Danish cohort study also found an increased standardized incidence ratio for esophageal adenocarcinoma among patients with Barrett’s esophagus with erosive esophagitis when compared with the general population (standardized incidence ratio, 5.2; 95% CI, 4.6–5.8).

Ulceration within a Barrett’s segment is also associated with an increased risk of neoplastic progression to adenocarcinoma or high-grade dysplasia. The above-mentioned population-based case-control study from Northern Ireland also found that patients with ulceration in the Barrett’s segment at diagnosis, but not elsewhere in the esophagus, were more likely to progress to cancer or high-grade dysplasia than those without (HR, 1.72; 95% CI, 1.08–2.76).

It is unclear if mucosal abnormalities such as erosive esophagitis or ulceration are in fact risk factors for progression or rather are markers of prevalent adenocarcinoma or high-grade dysplasia, as shown by current endoscopic eradication studies. Overall, it seems that mucosal abnormalities are associated with an increased risk of neoplastic progression in patients with Barrett’s esophagus.

Circumferential Position

Early adenocarcinoma and high-grade dysplasia seem to have a predilection to develop in the right hemisphere of the esophagus. The author’s group found that 85% of patients with adenocarcinoma or high-grade dysplasia referred for endoscopic management had these abnormalities located in the right hemisphere of the esophagus, predominantly in the area between 12-o’clock and 3-o’clock positions. Similar findings were described in an Australian study, where more than 50% of advanced lesions were found between 2-o’clock and 5-o’clock positions. Both of these studies are in line with previous data demonstrating greater chances for esophagitis to be found on the right hemisphere of the esophagus, suggesting a potential inflammatory mechanism to explain this observation.

Intestinal Metaplasia Versus Columnar Metaplasia

At present, there is some disagreement among gastrointestinal (GI) societies as to the definition of Barrett’s esophagus. The major difference is whether or not the presence of intestinal metaplasia within the columnar lined esophagus is required for the diagnosis. Early data from Scandinavia suggested that the risk of progression to adenocarcinoma or high-grade dysplasia in the columnar lined esophagus was equivalent in patients with and without intestinal metaplasia. However, multiple subsequent studies have demonstrated an increased risk of neoplastic progression in patients with intestinal metaplasia.

Work from the population-based Northern Ireland Barrett’s Esophagus Registry examined the risk of progression to high-grade dysplasia and esophageal adenocarcinoma in 8522 patients diagnosed with Barrett’s esophagus defined as a columnar lined esophagus both with and without intestinal metaplasia. The risk of cancer for patients with intestinal metaplasia at index endoscopy was increased compared with that of those without intestinal metaplasia at index endoscopy (0.38% per year vs 0.07% per year; HR, 3.54; 95% CI, 2.09–6.0).

An observational study from the University of Chicago demonstrated that among 379 patients with a columnar lined esophagus without goblet cells on pathologic findings, none progressed to adenocarcinoma or high-grade dysplasia during an average follow-up of 5 years. In contrast, 8.9% of patients with a columnar lined esophagus with goblet cells progressed to adenocarcinoma or high-grade dysplasia.

Although there are not yet enough data to definitively state that intestinal metaplasia is a necessary component to the neoplastic risk of Barrett’s esophagus, it does seem that there is an increased risk of neoplastic progression in patients with intestinal metaplasia compared with that in those without intestinal metaplasia.

Dysplasia

Dysplasia remains the single best marker for risk of progression in Barrett’s esophagus. For patients with nondysplastic Barrett’s esophagus, the risk of neoplastic progression remains low. A 2012 meta-analysis found an incidence of esophageal adenocarcinoma of 1 case per 300 patient-years in patients with nondysplastic Barrett’s epithelium. Two large, population-based studies not included in this meta-analysis also support these findings. A cohort study from Denmark evaluated the incidence of adenocarcinoma in more than 11,000 patients with Barrett’s esophagus. The incidence rate was 1 case per 1000 person-years in patients with nondysplastic Barrett’s esophagus over a median follow-up of 5.2 years. A cohort study from Ireland also had similar low rates of progression in patients with nondysplastic Barrett’s esophagus, 0.17% per year.

Recent work of a large US multicenter consortium evaluated the importance of repeated biopsy-proven nondysplastic Barrett’s esophagus during surveillance endoscopy. Patients were found to have a lower annual risk of progressing to either adenocarcinoma or high-grade dysplasia if they had multiple endoscopies documenting persistent nondysplastic Barrett’s esophagus (0.34% in patients with 5 endoscopies vs 0.75% in patients with 1 endoscopy). Overall, the neoplastic progression risk in nondysplastic Barrett’s esophagus seems to be very low.

Low-Grade Dysplasia

Low-grade dysplasia has been extensively studied as a risk factor for progression with highly variable results. Probably due to the interobserver variability in the diagnosis of this lesion, the diagnosis becomes problematic. A large, multicenter outcomes study from the United States investigated 210 patients with low-grade dysplasia to determine the rate of neoplastic progression. There was a 1.83% per year incidence of progression to adenocarcinoma or high-grade dysplasia in this cohort. While the progression to esophageal adenocarcinoma was 0.18% per year if only 1 of 3 pathologists confirmed the diagnosis of low-grade dysplasia, the incidence rate increased to 0.39% per year if all 3 pathologists agreed. Critics of the study point out that there was a low interobserver agreement among the 2 expert pathologists (κ, 0.14). In addition, approximately 1 in 4 of the original low-grade dysplasia samples were subsequently upgraded to high-grade dysplasia in this study, further calling into question the results. In a landmark study from the Netherlands, 293 patients with low-grade dysplasia were assessed. Biopsy samples were confirmed by 2 independent pathologists with extensive experience with dysplasia within Barrett’s esophagus from a panel of 6 pathologists. Upon expert review, almost three-quarters of patients were downstaged to either nondysplastic Barrett’s esophagus or indefinite for dysplasia. In the patients who were confirmed to have low-grade dysplasia on expert review, the risk of neoplastic progression was 9.1% per patient-year with a median follow-up of over 3 years. The patients with nondysplastic Barrett’s esophagus or those who were indefinite for dysplasia had a significantly lower risk of neoplastic progression (0.6%–0.9% per patient-year).

These results are similar to previous reports from the Academic Medical Center group in the Netherlands, which put the risk of progression to adenocarcinoma or high-grade dysplasia at 13.4% per year in patients with confirmed low-grade dysplasia. In the surveillance arm of the SURF study examining radiofrequency ablation for patients with confirmed low-grade dysplasia, 26.5% of patients progressed to adenocarcinoma or high-grade dysplasia during a median follow-up of 30 months. Similarly, a high progression rate was seen in the original clinical trial of radiofrequency ablation for low-grade dysplasia where expert gastrointestinal pathologist confirmation was required: 14% of patients in the sham treatment arm developed high-grade dysplasia at 1 year of follow-up. Finally, in a population-based cohort study from Denmark, the standardized incidence ratio for esophageal adenocarcinoma in patients with low-grade dysplasia was 5.1 per 1000 patient-years (95% CI, 3.0–8.6).

These rates are considerably higher than those described in the most recent meta-analysis, which found a rate of progression from low-grade dysplasia to adenocarcinoma of 0.54% per year (95% CI, 0.32%–0.76%) and to adenocarcinoma or high-grade dysplasia of 1.73% per year (95% CI, 0.99%–2.47%). However, the investigators reported significant heterogeneity between the studies and acknowledged that rates of neoplastic progression were higher in the studies when expert pathologists confirmed low-grade dysplasia.

In summary, low-grade dysplasia is a challenging lesion to diagnose but seems to be a risk factor for neoplastic progression when confirmed by multiple pathologists with expertise in GI pathology.

Biomarkers of Increased Risk

Given the low rate of neoplastic progression in Barrett’s esophagus and the inherent limitations of current endoscopic surveillance programs, there has long been interest in identifying biomarkers of increased risk in patients with Barrett’s esophagus. Although multiple biomarkers have been studied, a select few are discussed here.

One of the best studied biomarkers is the tumor suppressor gene p53. In patients with Barrett’s esophagus, loss of heterozygosity (LOH) for p53 (RR, 16; CI, 6.2–39.0) as well as overexpression of p53 (OR, 8.42; 95% CI, 2.37–30.0) are associated with progression to adenocarcinoma or high-grade dysplasia. A nested case-control study demonstrated similar results; loss of p53 (RR, 14.0; 95% CI, 5.3–37.2) and overexpression of p53 (RR, 5.6; 95% CI, 3.1–10.3) were associated with a higher risk of progression to adenocarcinoma or high-grade dysplasia. In addition, biopsies with both low-grade dysplasia and aberrant p53 expression seem to have higher rates of neoplastic progression.

Other tumor suppressor genes (ie, p16), cell-cycle-related proteins (ie, cyclin A, cyclin D1), growth factor receptors (ie, epidermal growth factor receptor), and flow cytometry for DNA abnormalities (ie, aneuploidy and tetraploidy) have been studied at length. Despite some promising results, none of these potential biomarkers are appropriate for clinical practice at this time.

Considerable efforts have also gone into developing panels of these biomarkers that may assist in the identification of patients at increased risk for progression. A combination of 17p LOH, 9p LOH, and aneuploidy/tetraploidy was found to have a relative risk of 38.7 (95% CI, 10.8–138.5) for neoplastic progression and a 10-year cumulative incidence of adenocarcinoma of nearly 80%. In patients with none of these biomarkers, only 12% progressed to adenocarcinoma. A panel of methylation biomarkers for 8 different genes (including p16, RUNX3, HPP1) had sensitivity for detection of progression of less than 50%. The panel of just p16, RUNX3, and HPP1 has also been studied. Finally, a nested case-control study from the Northern Ireland Barrett’s Esophagus Registry evaluated a combination of biomarkers and anatomic pathology. The panel of low-grade dysplasia, DNA aneuploidy/tetraploidy detected by image cytometry, and Aspergillus oryzae lectin demonstrated an increasing OR for each component of the panel that was present (OR, 3.73 per biomarker; 95% CI, 2.43–5.79).

Biomarkers and biomarker panels may in the future assist in determining who is likely to progress to adenocarcinoma or high-grade dysplasia. At the present time, there are multiple limitations that need to be addressed before these biomarkers are incorporated into daily clinical practice.

Age

The incidence of esophageal adenocarcinoma increases with age, regardless of sex or race. Between the ages of 50 and 59 years, white males have an esophageal adenocarcinoma incidence rate of 8.44 per 100,000 person-years (95% CI, 8.05–8.85), which increases to 26.31 per 100,000 person-years (95% CI, 25.27–27.38) between ages 70 and 79 years. The most recent publication of the Surveillance, Epidemiology, and End Results (SEER) registry reaffirms the increase in esophageal adenocarcinoma with age. Starting at age 40 years and continuing until 79 years, the incidence rates of esophageal adenocarcinoma continue to rise. Beyond the age of 80 years, the incidence rate seems to level off in most groups.

In patients with Barrett’s esophagus, most studies suggest an increase in the risk of neoplastic progression with increasing age. A Dutch population-based study found that the risk of progression increased with increasing age at diagnosis, with a marked increase in risk after the age of 75 years (HR, 12; 95% CI, 8.0–18). Similarly, in the Danish Barrett’s esophagus population-based study, the incidence of adenocarcinoma or high-grade dysplasia increased progressively with age and was greatest in patients older than 70 years.

Race

White race is a known risk factor for esophageal adenocarcinoma. There is a fourfold increase in the incidence rate of esophageal adenocarcinoma in white males compared with black males and twofold increase compared with that of Hispanic males. Similar patterns are also seen in women, although with lower incidence rates in all races. While the cause of this difference is unclear, a recent study of patients with Barrett’s esophagus found a higher rate of dysplasia (7% vs 0%) and of long-segment Barrett’s esophagus (26% vs 12%) at the time of endoscopy in Caucasians compared with black patients. As both dysplasia and long-segment Barrett’s esophagus are independent risk factors for esophageal adenocarcinoma, this may be an avenue worthy of further study.

Sex

Male gender is another well-documented risk factor for esophageal adenocarcinoma with a 6- to 10-fold risk increase compared with that of women. However, both genders have seen a dramatic increase in the number of esophageal cancers. Newer data from the SEER registry have demonstrated the largest gender difference exists largely in the earlier-age cancers (ie, younger than 65 years). Furthermore, the age-adjusted incidence rate in women older than 80 years continues to increase, whereas the rate in men plateaus. This difference may relate to estrogen exposure. Previously, estrogen exposure was suggested as a protective mechanism against both gastric and colon cancers and was demonstrated to reduce apoptosis and cell growth in esophageal adenocarcinoma. As women age, there is a significant decrease in estrogen, which may explain the delayed increase in esophageal adenocarcinoma incidence, as seen in the SEER database data. Studies examining how risk factors with possible male predominance, including gastroesophageal reflux disease (GERD), obesity, and tobacco, may affect neoplastic progression risk have not yet provided a clear mechanistic answer to this epidemiologic difference. Among patients with Barrett’s esophagus, male gender is also a clearly recognized risk factor for progression to esophageal adenocarcinoma.

Family History/Genetic Risk

Familial aggregation of Barrett’s esophagus and esophageal adenocarcinoma has suggested a potential genetic component to these disease entities. Much of the work in this area has come from Chak and colleagues who initially found that a positive family history (first- or second-degree relative with Barrett’s esophagus, esophageal adenocarcinoma, or esophagogastric junction carcinoma) was higher among case subjects than among GERD controls (24% vs 5%). On multivariate analysis, a positive family history for Barrett’s esophagus, adenocarcinoma of the esophagus, or esophagogastric junction was associated with an increased risk of developing these lesions when compared with patients with GERD alone (OR, 12.23; 95% CI, 3.34–44.76).

A subsequent segregation analysis by the same group found an incomplete autosomal dominant inheritance pattern for familial aggregations of Barrett’s esophagus, esophageal adenocarcinoma, and gastroesophageal junction carcinoma. Finally, in multiplex aggregations characterized by 3 or more members of a family with Barrett’s esophagus and/or esophageal adenocarcinoma, the median age for the diagnosis of adenocarcinoma was approximately 5 years less than that in duplex families or sporadic cases. Others have found an increased prevalence of Barrett’s esophagus in first-degree relatives of patients with adenocarcinoma.

Last, examination of germline mutations in patients with Barrett’s esophagus and esophageal adenocarcinoma has yielded interesting results. In a model-free linkage analysis comparing both concordant sibling pairs with Barrett’s and esophageal adenocarcinoma and discordant sibling pairs, 3 genes were identified with significant mutations (MSR1, ASCC1, and CTHRC1). MSR1 was associated with the presence of both Barrett’s esophagus and esophageal adenocarcinoma. A genome-wide association study from the BEACON (Barrett’s and Esophageal Adenocarcinoma Consortium) group of more than 1500 cases of esophageal adenocarcinoma and 2300 cases of Barrett’s esophagus found a high genetic correlation between Barrett’s esophagus and adenocarcinoma as well as between multiple shared genes underlying the development of both.

Taken together, these data suggest that there is a genetic component to neoplastic progression in some patients with Barrett’s esophagus. More data are still needed to fully characterize this risk.

Tobacco

Tobacco use is a clear risk factor for the development of both Barrett’s esophagus and esophageal adenocarcinoma. A pooled analysis evaluating approximately 3000 cases of either esophageal adenocarcinoma or esophagogastric junctional adenocarcinoma found an OR of 1.67 (95% CI, 1.04–2.67) for developing esophageal adenocarcinoma in patients with tobacco use, with a significant increase with increased number of pack-years smoking. Current or prior tobacco use also increases the risk of neoplastic progression in patients with Barrett’s esophagus. A population-based cohort study from the Northern Ireland Barrett’s Esophagus Registry demonstrated an increased risk of neoplastic progression to adenocarcinoma or high-grade dysplasia in current smokers when compared with nonsmokers (HR, 2.03; 95% CI, 1.29–3.17). Overall, tobacco is an established risk factor for progression to esophageal adenocarcinoma.

Obesity

Obesity is a well-described risk factor for esophageal adenocarcinoma. A 2013 meta-analysis of more than 8000 cancers examined the relationship between body mass index (BMI), calculated as the weight in kilograms divided by the height in meters squared, and esophageal and gastric cardia adenocarcinoma. The relative risk for developing esophageal adenocarcinoma increased with an increasing BMI; the relative risk was 1.71 (95% CI, 1.50–1.96) for a BMI in the range 25 to 30 and 2.34 (95% CI, 1.95–2.81) for a BMI greater than 30. An observational study from England showed a significant increase in esophageal adenocarcinoma with a BMI greater than 35 (HR, 4.95), but only an upward trend at lower BMIs. This finding corresponds to earlier data suggesting that the largest risk was in patients with BMI greater than 35. It would seem that male pattern central obesity is the key component of this risk, with a 2013 meta-analysis reporting an OR of 2.51 (95% CI, 1.56–4.04) for developing esophageal adenocarcinoma in patients with central adiposity.

The impact of obesity and abdominal adiposity specifically on neoplastic progression of Barrett’s esophagus to adenocarcinoma is a bit more unclear. A small cross-sectional analysis from Seattle Barrett’s Esophagus Project demonstrated an increasing risk of histologic and genetic abnormalities associated with a high likelihood of neoplastic progression in patients with a predominantly abdominal fat distribution. However, a recent cohort study of more than 400 patients with Barrett’s esophagus found no relation between higher waist-hip ratios and neoplastic progression to adenocarcinoma. Although the role of neoplastic progression in Barrett’s esophagus is not fully known, it is clear that obesity and specifically abdominal adiposity are associated with an increased overall risk of esophageal adenocarcinoma.

The exact mechanism underlying this risk is not fully known but may involve increased levels of insulinlike growth factor 1, insulin resistance, and adipokines such as leptin. These relationships are explored at length by Chandar and Iyer, elsewhere in the issue.