Barrett’s esophagus (BE) refers to the histological transformation of the squamous epithelium of the distal esophagus into specialized intestinal metaplasia. Given its potential to progress from low-grade dysplasia (LGD) to high-grade dysplasia (HGD) to esophageal adenocarcinoma (EAC), it is important to accurately diagnose and survey these patients who have these pathologic findings. Current guidelines recommend detailed visualization by endoscopy and obtain biopsies in a random four-quadrant fashion every 1–2 cm as well as targeted biopsies of any suspicious lesions [1, 2]. In effect, histology remains the hallmark for diagnosis.

Although commonly practiced, these techniques lack standardization on account of several confounders including the increased endoscopic time required to obtain multiple biopsies as well as the variability between endoscopists in adhering to protocols, both contributing to sampling errors. Further inaccuracies can accrue from the observation that dysplasia is twice as likely to occur in the proximal half according to post hoc analysis of two randomized trials [3]. Moreover, there exists a discrepancy among pathologists interpreting tissue samples. In a study from the Netherlands of patients diagnosed with LGD, 85% were downgraded to nondysplastic Barrett’s esophagus (NDBE) after review by two expert GI pathologists [4]. Interestingly, in another international study, European pathologists had a higher concordance rate of interpretation than their US counterparts, which underlines the varied standards, experience, and clinical practice [5].

Since the histological confirmation of dysplasia including its degree is the pivotal factor that determines the type of subsequent treatment or surveillance, accuracy of histological diagnosis is essential to providing optimal care. Many newer techniques have been developed to improve diagnostic accuracy. As examples, enhanced endoscopic imaging with the acquisition of high-definition (HD) white-light images aids in “red-flagging” suspicious areas, thereby enabling the performance of targeted “smart” biopsies. Dye-based chromoendoscopy, another targeting method, has been used for many years although it has the limitations of being time-consuming and having unequal distribution of contrast over the mucosa. Compared to this, optically based electronic chromoendoscopy such as the narrowband imaging technology (NBI) is being used more widely since it offers a wider field of imaging and easy integration into standard endoscopic methods with no staining required and smooth performance. NBI works by optical filtration based on light wavelength and absorption patterns of hemoglobin, highlighting changes between mucosa and superficial vessels. Other platforms such as the flexible spectral imaging color enhancement (FICE) and I-Scan® are based on similar principles.

Several studies have reported that NBI is beneficial or complementary to white light in detecting BE, with or without dysplasia based on its ability to better distinguish vascular and mucosal patterns [6, 7]. While aiming for target the “bull’s eye” for precise diagnosis, NBI can reduce the number of biopsies needed to detect BE with resultant cost savings [8]. The American Society of Gastrointestinal Endoscopy (ASGE) endorsed the use of NBI for this application based on its meta-analysis of Barrett’s imaging using NBI, reporting a high sensitivity (94%), negative predictive value (NPV) (97%), and specificity (94%) [9]. Nevertheless, given the lack of well-defined or standardized characteristics for NBI-assisted targeting of biopsies, the Barrett’s international NBI group (BING) sought to define and validate a simpler classification system based on surface changes—“mucosal” and “vascular” [10]. Use of this system enabled the development of criteria that identify dysplasia in BE with >90% accuracy and a high level of inter-observer agreement.

With this background, Nogales et al. [11] have published an interesting study in this issue of Digestive Diseases and Sciences applying the BING criteria while evaluating its diagnostic usefulness for identifying dysplasia in BE. Using a data set obtained with an improved version of image processing (EVIS EXERA III with HD endoscopes) with accompanying NBI filters offering twice the viewable distance and with brighter images, they selected 100 images, assigning eight observers to evaluate the mucosal and vascular pattern based on the BING criteria for the prediction of dysplasia. The interpretations, when compared with histology, demonstrated a higher diagnostic accuracy and higher level of confidence (85.8%). Nonetheless, the inter-observer concordance was weak for predicting dysplasia (K = 0.40), emphasizing a steep learning curve. Surprisingly, trainees performed better (K = 0.55) than seasoned endoscopists or experts (K = 0.35) with a higher concordance rate for prediction. The study was limited by the fewer subsets of patients having confirmed dysplasia, a single pathologist interpretation, and static images rather than video real-time images for review.

What can be inferred from these studies? First, the BING criteria are certainly an important step in defining and bridging the descriptive gaps needed for effectively utilizing NBI imaging to predict histology. Targeted image-guided biopsies in real time clearly translate to better patient outcomes and cost-effectiveness. Second, while image-enhanced technology is useful when it comes to defining the subtle changes of BE, it ultimately relies on the training and competency of the interpreter resulting in poor inter-observer agreement. The performance thresholds have a higher accuracy when the diagnoses are made with “high confidence” as noted in the above study, arguing that the level of confidence is subjective. Interpretations are further confounded by differences among training modules and the lack of set guidelines with regard to the intensiveness of training, that range from the traditional classroom style to self-directed training with continuous refresher courses while ascending one’s own learning curve and ensuring proficiency. Third, appropriate interpretation requires endoscopists at specialized centers as opposed to those in the community. While the ASGE recommends that NBI should be performed by endoscopists with “expertise” in advanced procedures such as at academic centers, one might argue that it is in the community where the target population with GERD and/or Barrett’s is primarily encountered. From a survey, only a third of gastroenterologists in the community (37%) routinely used advanced imaging [12]. Therefore, ideally, knowledge and the advancement of technology should appropriately branch out to community GI practices while aiming for quality and ensuring competency. There is also a need for higher reimbursement for the additional time and effort required for widespread implementation. Finally, it is still not clear whether optical endoscopy can enhance surveillance and post-endoscopic eradication treatment (radiofrequency ablation [RFA] or resection) highlighting the need for further studies in this regards.

In the era of advanced image technology, other novel imaging tools have been used to detect Barrett’s esophagus. Using the thresholds set by the ASGE—“Preservation and Incorporation of Valuable Endoscopic Innovations” (PIVI), besides NBI, dye-based chromoendoscopy (acetic acid) and endoscope-based confocal laser endomicroscopy (eCLE) scored high, meeting their benchmark criteria. Fewer data exist for other technologies such as autofluorescence imaging (AFI) and volumetric laser endomicroscopy (VLE) [13]. Molecular imaging for BE using probes with labeled markers targeting expressed proteins in vivo currently in development shows promise in demonstrating dysplasia or cancer [14].

Given these advancements, it is nonetheless premature to suggest that dependable routine histology can be completely superseded by advanced imaging, underscoring the importance of careful and thorough endoscopic evaluation. Therefore, further studies evaluating criteria such as “BING” will kindle refinements in strengths and weaknesses. These will then fit into their matching roles as a slice in the larger imaging pie with goal of improving patient care and satisfaction. Furthermore, by collaborating with industry and obtaining support from GI societies, solutions to the barriers for implementation in clinical practice can be achieved by creating environments that motivate endoscopists to adopt newer technology which have clear scientific backing [15]. Looking through the kaleidoscope for innovation in the field of BE, the future seems exciting making one reflect—Let the light shine on!