Barrett’s esophagus (BE) is defined by the American College of Gastroenterology (ACG) as intestinal metaplasia (IM) of the distal esophageal squamous epithelium and is a known precursor to esophageal adenocarcinoma (EAC) [1]. Although the true prevalence is difficult to ascertain, approximately two percent of adults are affected by BE, with 0.5–1% going on to develop EAC per year [2]. Gastroesophageal reflux disease (GERD) is thought to be a pathologic driver behind BE, yet 40% of patients diagnosed with BE report no symptoms of GERD at the time of diagnosis [3].

EAC is characterized by a strong male predominance, with a rapidly increasing incidence in the West, now surpassing the incidence of esophageal squamous cell carcinoma (ESCC) in several European, North American, and Oceanic countries [4]. Incidence increases with age, and certain genetic factors may play a role in the development of EAC from BE. The prognosis of EAC is strongly related to the stage at the time of diagnosis; however, due to most cases being diagnosed with the late-stage disease, the prognosis is poor, with a 20% overall survival rate at five years.

2. Definition and Diagnosis

BE is generally defined as columnar metaplasia of the distal esophagus; however, there is controversy concerning the diagnostic criteria for this disease process [5]. This is mainly in regard to whether IM is necessary for the diagnosis. IM is generally characterized by the presence of goblet cells; however, while the presence of goblet cells indicates intestinal metaplasia, not all intestinal metaplastic epithelia contain goblet cells [6]. The ACG, American Gastroenterological Association (AGA), and European Society of Gastrointestinal Endoscopy (ESGE) define BE as a change in the normal esophageal squamous mucosa, greater than 1 cm, that is visible endoscopically and demonstrates IM on biopsy, while the guidelines of the British Society of Gastroenterology (BSG), and Japan Esophageal Society (JES) do not require IM, and BE may be diagnosed with cardiac, intestinal, or oxyntic columnar mucosa on biopsy ^{[3][7][8][9][10]}[3,7,8,9,10]. It is important to consider that differentiating goblet cells from pseudogoblet cells, which occur when columnar cells become distended with mucin and acquire a shape that mimics goblet cells, can be challenging. There is no histochemical or immunohistochemical stains that can reliably identify goblet cells or differentiate them from pseudogoblet cells when they are not apparent on standard H&E stain ^[11][12][11,12]. The difficulty in differentiating goblet and pseudogoblet cells was demonstrated with poor interobserver agreement in a study between seven GI pathologists [6]. In addition to the histologic challenges in the diagnosis, the risk of sampling bias is high. Studies have demonstrated a mosaic distribution of goblet cells within the areas of columnar metaplasia ^[13][14][13,14]. This results in a variation of goblet cell density, and therefore detection rate, depending on the location within the esophagus. The probability of detecting goblet cells has been demonstrated to increase proportionally with the number of biopsies obtained at the time of endoscopy [15]. In an observational study, a minimum of eight biopsies was deemed optimal to diagnose IM in patients with a columnar-lined esophagus on endoscopy. IM was diagnosed at a rate of 68% when eight biopsies were taken compared to 35% when only four biopsies were taken [16]. In addition to a sampling error, the presence and density of goblet cells can fluctuate with time and disease progression. In a study of 43 patients with less than 3 cm of columnar mucosa and no intestinal metaplasia on initial biopsy, Jones et al. demonstrated that 23% of the patients with suspected short-segment (<3 cm) BE had intestinal metaplasia on the repeat biopsy [17]. Kim et al. found that with the repeat biopsy six weeks after study entry and the initial biopsy, 21 (18%) of the 116 patients who met the criteria for BE did so only on one of the two exams [18]. The inclusion of IM in the diagnostic criteria of BE stems from the data supporting a significantly increased risk of dysplasia and EAC in patients with IM. It is argued that IM should be required for the diagnosis until columnar metaplasia without IM is proven to have a significant risk of progression to EAC. Chandrasoma et al. examined 214 patients with columnar metaplasia and demonstrated that IM was present in all 55 cases of dysplasia and/or EAC [19]. A large study of over 8500 patients in the North Ireland Cancer Registry demonstrated a hazard ratio (HR) of 3.54 [95% confidence interval (CI) 2.09–6.00] of developing BE-related adenocarcinoma in patients with IM compared to those without [20]. Advocates for requiring IM for the diagnosis of BE also point to the serious changes in quality of life that result from being diagnosed and labelled with a neoplastic process such as BE ^[21][22][21,22]. An analysis of the University of Chicago Medical Center database over 21 years concluded that the removal of goblet cells from the diagnostic criteria of BE would result in an ~150% increase in BE diagnoses; additionally, none of the patients without goblet cells (n = 118) developed EAC or dysplasia, with a mean 5.8 years of follow-up and 2.8 repeat endoscopies [23]. On the other hand, Kelty et al. found a similar risk of EAC in patients with columnar epithelia with and without IM (4.5% versus 3.6%, respectively) in a United Kingdom (UK)-based study, with a median follow-up period of 12 years [24]. Some have argued that sampling could explain the differing results between the Chandrasoma et al. and Kelty et al. studies, with the prior having a rigorous biopsy protocol compared to the latter ^[5][25][5,25]. Takubo et al. classified the mucosa surrounding early mucosal EAC obtained by endoscopic mucosal resection in 141 cases. Overall, 70% of the primary EAC were adjacent to cardiac/fundic-type mucosa as opposed to intestinal-type mucosa [26]. Similarly, Liu et al. found no significant differences in the DNA content abnormalities between samples with and without IM [27]. In summation, due to the concerns with the accurate identification of the presence of IM and the uncertainty regarding whether other mucosal types can progress to dysplasia and EAC, the BSG concluded that the presence of IM is not a defining prerequisite for BE, but that the presence of IM should be taken into account when considering clinical surveillance and follow-up [8]. On the other hand, the AGA and ACG concluded that the diagnosis of BE should include the presence of IM due to the lack of data supporting the risk of the malignant transformation of non-intestinal mucosa ^[3][7][3,7]. EAC is a gland-forming tumor most commonly with tubular, tubulopapillary, or papillary growth patterns. Well-differentiated tumors show >95% gland formation, moderately differentiated tumors show 50–95% gland formation, and poorly differentiated tumors show <50% gland formation [28]. The diagnosis of EAC is typically confirmed via an endoscopic biopsy. Endoscopic ultrasound (EUS) is currently the procedure of choice to determine the depth of the local invasion and clinical T-staging [29]. In a systematic review, EUS had a 91% true positive rate at differentiating T1/2 from T3/4 tumors, which is critical for clinical management [30]. EUS, computed tomography (CT), and fluorodeoxyglucose positron emission tomography (FDG-PET) are the principal modalities for evaluating regional lymph node metastases [29].