Do you see what I see? Towards standardized reporting of endoscopic findings in eosinophilic esophagitis

The endoscopic findings in eosinophilic esophagitis are one of the hallmarks of the disease, a visible sign of an upper gastrointestinal process that has only been recognized over the past two decades [1]. The findings also contain important information. Features such as esophageal rings, linear furrows, decreased vascularity, or white plaques ([Fig. 1]) might prompt an endoscopist to obtain esophageal biopsies to assess for eosinophils if the diagnosis was unexpected, or might confirm a clinical suspicion or suggest that a prescribed therapy is not working. A diffusely narrowed esophagus or a focal esophageal stricture could indicate that the disease has been present and undiagnosed for many years [2] [3], leading to esophageal remodeling [4] and increasing the risk for food impactions [5]. However, these findings are in some way a conundrum. They are not a formal part of current diagnostic guidelines [1] [6], they are not pathognomonic for eosinophilic esophagitis [7], they may not uniformly respond to treatment [8], and reporting can be variable [7] [9] [10] [11].

Fig. 1 Typical endoscopic features in a patient with active eosinophilic esophagitis, where rings, furrows, edema/decreased vascularity, white plaques/exudate, and narrowing can be seen. This patient would have an EREFS score of E2 R2 E1 F1 S0. Specifically, the exudates cover > 10 % of the mucosa, the rings are moderate (distinct but allow passage of a standard upper scope), edema is present, furrows are present, and there are no focal strictures.

Whereas esophageal rings were originally described in early case reports and series of eosinophilic esophagitis in adults [12] [13], it took several years for reports to suggest that linear furrows were also typical and that white plaques were likely to represent eosinophilic microabscesses [14] [15] [16]. In studies during this era, a relatively high proportion of patients (at least 10 % – 20 %) were reported to have eosinophilic esophagitis with an endoscopically normal esophagus, which is one reason why these findings were not required to be present for a diagnosis of eosinophilic esophagitis [7] [9] [17].

Recently, a large meta-analysis by Kim et al. compiled data from 100 studies representing 4678 cases of eosinophilic esophagitis and 2742 controls in order to study endoscopic findings of eosinophilic esophagitis [2]. There were several conclusions. First, although the individual endoscopic findings of eosinophilic esophagitis were not universal, the vast majority of patients with eosinophilic esophagitis (83 %) had at least one finding. Second, findings differed by age, with adults more likely to have rings and strictures, and children more likely to have white plaques and decreased vascularity/edema. Third, endoscopic findings were neither sensitive nor specific enough to be used alone for eosinophilic esophagitis diagnosis. Finally, prospective studies found a higher prevalence of eosinophilic esophagitis findings than retrospective studies, suggesting that when endoscopists conducted a dedicated examination for eosinophilic esophagitis findings, these findings could be more commonly, and perhaps more consistently, reported.

There have been several attempts to increase the detection of endoscopic findings of eosinophilic esophagitis. Lucendo et al. reported that chromoendoscopy highlighted features such as rings and furrows [18]. Peery et al. hypothesized that narrow-band imaging would lead to a higher rate of recognition of the endoscopic findings, but this was not borne out [10]. Similarly, Tanaka et al. examined narrow-band imaging magnifying endoscopy and described a series of magnified structural features that were seen more commonly in eosinophilic esophagitis than in gastroesophageal reflux disease, but the clinical utility of this method has yet to be determined [19].

A vexing issue was that even though these endoscopic finings appeared to be so prominent and characteristic, a way to describe them reliably had not been developed. This was demonstrated in a study by Peery et al. in which 77 gastroenterologists were shown a series of 35 still endoscopic images and inter- and intraobserver agreements were calculated [10]. For rings, furrows, and white plaques, the kappa values for interobserver agreement were 0.56, 0.48, and 0.29, respectively, indicating at best only fair to good agreement between endoscopists. The intraobserver values were even more variable, ranging from 0 to 1.0 depending on the finding, and showing that individual endoscopists were not always reliable for reporting the same endoscopic findings.

This set the stage for a key study by Hirano et al. [20]. They showed 25 endoscopic videos to 21 gastroenterologists and trainees, calculated interobserver agreement, constructed an atlas of endoscopic findings in eosinophilic esophagitis with standardized categories, and developed and validated the eosinophilic esophagitis Endoscopic Reference Score (EREFS). This acronym, conveniently, also stands for the five main findings: Exudates, Rings, Edema, Furrows, and Strictures. The authors found that the classification scheme resulted in good interobserver agreement, with kappas of 0.46, 0.40, 0.43, 0.54, and 0.52 for the findings of exudates, rings, edema, furrows, and strictures, respectively. Additionally, they found that the findings of narrowing and felinization were not reliably described, and that crêpe-paper mucosa, while consistently identified, was rarely present; these adjunct findings, though potentially of clinical importance, were therefore not included in the final EREFS system.

In this context, the study by van Rhijn et al. in this issue of Endoscopy is an important step forward in bringing the EREFS classification system into common practice. In this study, which is the first published experience of EREFS by a group both outside of Chicago and outside of the United States, four expert and four trainee endoscopists reviewed a series of still images collected prospectively from 30 patients with eosinophilic esophagitis, 24 with active disease and 6 in remission after treatment. Prior to grading the images, all of the physicians were required to read the paper by Hirano et al., and view a slideshow with representative images and the written descriptions of the EREFS system. Both inter- and intraobserver agreements were calculated.

Because each of the 30 patient images was reviewed twice by each endoscopist, once for the interobserver data and a second time for the intraobserver data, there were a total of 480 observations available for analysis. Notably, with careful examination of the images, every patient had at least one finding of eosinophilic esophagitis detected, supporting the previous observation that in patients with eosinophilic esophagitis who were examined prospectively, endoscopic findings may be near-universal. For interobserver agreement, kappa values were all in the good range for rings, exudates, furrows, and stricture (0.70, 0.63, 0.49, 0.54, respectively) but were poor for edema (0.12). For the intraobserver agreement, there was again a wide range of kappa values, but apart from edema, these were mostly in the good range. In a subanalysis comparing the expert with the trainee endoscopists, the agreement levels were roughly comparable, regardless of level of experience, implying that the EREFS system is robust. This supports the findings in the Peery and Hirano studies, where agreement was in general similar between experts and trainees, and between those with more or less experience with eosinophilic esophagitis patients [10] [20].

These results, though strong, should be interpreted in the setting of some limitations of the study. First, this study included only a small number of endoscopists from a single center, fewer than in either of the prior studies examining this issue. Second, similar to the Peery study, but different from the Hirano study in which the EREFS was developed, only still endoscopic images were used rather than video sequences. Assessing findings during a “live” endoscopy (and potentially with a video during a study), allows for washing, pausing on certain findings, and taking into account the appearance of the entire esophagus. van Rhijn et al. attempted to overcome this limitation by using multiple (3 – 6) images per patient, which is a stronger methodology than using a single image. However, it is still not known how the EREFS system would perform with external video validation, but presumably it would be no worse than with still images. Finally, the authors also went beyond the EREFS classification itself to examine whether re-categorizing the findings might improve agreement. Particularly for edema, if the categories were changed from absent vs. present to absent or mild vs. severe, the kappa improved from 0.12 to 0.24. The authors suggest that this new classification might be preferable, but at this time, the data do not support a change from the classification proposed by Hirano et al. More experience with the original classification system should be accrued from other centers. However, if future studies confirm that the finding of edema has poor agreement, redefining how edema is classified should be considered.

In summary, the study by van Rhijn et al. has provided the first external validation of the EREFS system. Using a simple training protocol involving the reading of Hirano’s original paper and reviewing the image atlas, they found that interobserver agreement for endoscopic findings of eosinophilic esophagitis, with the exception of edema/decreased vascularity, was in the good range. In fact, the agreements for exudates and rings were even better than reported when the classification was developed. Intraobserver agreements were generally also good. There are still a number of important steps that are required to maximize the value of the EREFS system. A formal scoring algorithm is required (and this is under development), and the responsiveness of the score, while demonstrable in individual cases ([Fig. 2]), needs to be documented in prospective studies (also underway). Nevertheless, taken together, the Hirano and van Rhijn studies demonstrate that the EREFS system is a very useful tool than can be incorporated into general endoscopy practice at this time. The system is easy to apply ([Fig. 1] and [Fig. 2]), can quickly be added to report documentation, and provides a quantifiable way to assess endoscopic response to treatment rather than relying on general descriptions. Similar to the way that the Prague classification brought order to endoscopic reporting in Barrett’s esophagus [22], the EREFS system has the potential to do the same for eosinophilic esophagitis. This is a tool that all gastroenterologists who perform upper endoscopy should begin to use.

Fig. 2 Endoscopic features in a patient with eosinophilic esophagitis before and after successful treatment with a topical corticosteroid. a At baseline, there are rings, furrows, and edema, with an EREFS score of E0 R2 E1 F1 S0. b After treatment, the findings are resolved, and the improvement is reflected in the score: E0 R0 E0 F0 S0.

• References

• 1 Liacouras CA, Furuta GT, Hirano I et al. Eosinophilic esophagitis: Updated consensus recommendations for children and adults. J Allergy Clin Immunol 2011; 128: 3-20.e26
• 2 Schoepfer A, Safroneeva E, Bussmann C et al. Duration of untreated inflammation represents the main risk factor for stricture development in eosinophilic esophagitis. Gastroenterology 2013; 144: S485
• 3 Dellon ES, Kim HP, Sperry SL et al. A phenotypic analysis shows that eosinophilic esophagitis is a progressive fibrostenotic disease. Gastrointest Endosc 2014; 79: 577-585.e4
• 4 Aceves SS. Remodeling and fibrosis in chronic eosinophil inflammation. Dig Dis 2014; 32: 15-21
• 5 Kuchen T, Straumann A, Safroneeva E et al. Swallowed topical corticosteroids reduce the risk for long-lasting bolus impactions in eosinophilic esophagitis. Allergy 2014; 69: 1248-1254
• 6 Dellon ES, Gonsalves N, Hirano I et al. ACG clinical guideline: evidence based approach to the diagnosis and management of esophageal eosinophilia and eosinophilic esophagitis. Am J Gastroenterol 2013; 108: 679-692
• 7 Kim HP, Vance RB, Shaheen NJ et al. The prevalence and diagnostic utility of endoscopic features of eosinophilic esophagitis: a meta-analysis. Clin Gastroenterol Hepatol 2012; 10: 988-996.e5
• 8 Straumann A, Conus S, Degen L et al. Budesonide is effective in adolescent and adult patients with active eosinophilic esophagitis. Gastroenterology 2010; 139: 1526-1537.e1
• 9 Dellon ES, Aderoju A, Woosley JT et al. Variability in diagnostic criteria for eosinophilic esophagitis: a systematic review. Am J Gastroenterol 2007; 102: 2300-2313
• 10 Peery AF, Cao H, Dominik R et al. Variable reliability of endoscopic findings with white-light and narrow-band imaging for patients with suspected eosinophilic esophagitis. Clin Gastroenterol Hepatol 2011; 9: 475-480
• 11 Peery AF, Shaheen NJ, Dellon ES. Practice patterns for the evaluation and treatment of eosinophilic oesophagitis. Aliment Pharmacol Ther 2010; 32: 1373-1382
• 12 Feczko PJ, Halpert RD, Zonca M. Radiographic abnormalities in eosinophilic esophagitis. Gastrointest Radiol 1985; 10: 321-324
• 13 Straumann A, Spichtin HP, Bernoulli R et al. [Idiopathic eosinophilic esophagitis: a frequently overlooked disease with typical clinical aspects and discrete endoscopic findings]. Schweiz Med Wochenschr 1994; 124: 1419-1429
• 14 Gupta SK, Fitzgerald JF, Chong SK et al. Vertical lines in distal esophageal mucosa (VLEM): a true endoscopic manifestation of esophagitis in children?. Gastrointest Endosc 1997; 45: 485-489
• 15 Straumann A, Spichtin HP, Bucher KA et al. Eosinophilic esophagitis: red on microscopy, white on endoscopy. Digestion 2004; 70: 109-116
• 16 Lim JR, Gupta SK, Croffie JM et al. White specks in the esophageal mucosa: an endoscopic manifestation of non-reflux eosinophilic esophagitis in children. Gastrointest Endosc 2004; 59: 835-838
• 17 Furuta GT, Liacouras CA, Collins MH et al. Eosinophilic esophagitis in children and adults: a systematic review and consensus recommendations for diagnosis and treatment. Gastroenterology 2007; 133: 1342-1363
• 18 Lucendo AJ, De Rezende LC, Gonzalez-Castillo S et al. Chromoendoscopy with indigo-carmine improves the recognition of endoscopic mucosal findings in adult eosinophilic esophagitis. Gastroenterology 2009; 136 (Suppl. 01) S1874
• 19 Tanaka K, Rubio CA, Dlugosz A et al. Narrow-band imaging magnifying endoscopy in adult patients with eosinophilic esophagitis/esophageal eosinophilia and lymphocytic esophagitis. Gastrointest Endosc 2013; 78: 659-664
• 20 Hirano I, Moy N, Heckman MG et al. Endoscopic assessment of the oesophageal features of eosinophilic oesophagitis: validation of a novel classification and grading system. Gut 2013; 62: 489-495
• 21 van Rhijn BD, Warners MJ, Curvers WL et al. Evaluating the Endoscopic Reference Score for eosinophilic esophagitis: moderate to substantial intra- and interobserver reliability. Endoscopy 2014; 46: 1049-1055
• 22 Sharma P, Dent J, Armstrong D et al. The development and validation of an endoscopic grading system for Barrett’s esophagus: the Prague C & M criteria. Gastroenterology 2006; 131: 1392-1399