EUVIMEDOpdag EUVIMED
EditorialKommentar

Predicting the risk of developing low‐grade dysplasia in patients with Barrett's esophagus without dysplasia

Fumiaki Ishibashi iD, Sho Suzuki iD

Digestive Endoscopy · 2023

DOI: 10.1111/den.14528 ↗

At a glance

Editorial in Digestive Endoscopy, 2023. Written by Fumiaki Ishibashi and 1 co-authors. Institution: Department of Gastroenterology, International University of Health and Welfare Ichikawa Hospital, Chiba, Japan. Major MEDLINE topics: Barrett Esophagus, Esophageal Neoplasms, Precancerous Conditions. 10 references.

AbstractSubjectsBibliografiske oplysningerCiteRelated studies

Abstract (original language)

Peleg et al.1 constructed a novel model to predict the risk of histological progression, including into low-grade dysplasia (LGD), in patients with Barrett's esophagus (BE) without dysplasia. They built a model to predict the risk of histological progression in patients with BE by incorporating neutrophil-to-lymphocyte ratio (NLR), BE length, smoking history, age at BE diagnosis, and chronic kidney disease as parameters. In their study, the efficacy of the model was validated using an internal validation cohort. The results showed an area under curve (AUC) of 0.88 for the predictive ability of BE patients to develop dysplasia at 3 years. The most significant difference from a previous report2 was that this model is a predictive model of histological progression specifically for nondysplastic BE patients and does not include BE patients with dysplasia in the model construction. The novelty of this study is that the data were not collected retrospectively but rather in a prospective longitudinal cohort of nondysplastic BE patients who were not found to have dysplasia at the time of index endoscopy. Another interesting aspect of this model is that the parameter included the NLR, which reflects histological inflammation. NLR has previously been shown to be prognostically effective in Barrett's adenocarcinoma, and the authors further showed in a previous study that NLR is also effective in predicting histologic progression in BE patients.3 The prediction model constructed in the current study was an extension of the results of this study. In addition, it is unique that this model does not include male sex as a parameter, which was one of the parameters building the other model.2 The rationale is that, although the study cohort initially included 69.37% male patients, a comparison of the backgrounds of the groups with and without histological progression showed no difference by sex. Barrett's esophagus is a condition that results from chemical and mechanical irritation of the gastroesophageal junction (GEJ) because of acid reflux into the esophagus, causing irreversible columnar metaplasia. The severity of acid reflux has been shown to be associated with the degree of obesity. Conversely, a decrease in gastric acid secretion owing to Helicobacter pylori infection inhibits acid reflux. The prevalence of reflux esophagitis is increasing worldwide owing to increasing obesity and decreasing H. pylori infection rates. As a result, the prevalence of BE is also increasing.4 With increased opportunities to diagnose BE patients in clinical practice, it is important to know how to manage BE patients in terms of their risk of developing cancer. Barrett's adenocarcinoma arising in BE is thought to occur as a temporal progression to LGD and high-grade dysplasia (HGD).5 The American College of Gastroenterology (ACG) guidelines state that treatment is indicated when LGD is detected.6 Many attempts have been made to diagnose dysplasia in BE at an early stage based on gross endoscopic appearance. If the presence or absence of LGD can be accurately determined using endoscopic findings alone, the frequency of biopsies can be reduced. A recent review article summarized that magnified narrow-band imaging (M-NBI) is useful in the detection of dysplasia within BE.7 In addition, computer-aided diagnosis systems based on M-NBI have demonstrated excellent results in terms of accuracy, sensitivity, and specificity, but evidence has not yet been established at a level that can be applied in actual clinical practice.7 There are also reports showing the effectiveness of endocytoscopy, confocal laser endoscopy, optical coherence tomography, and volumetric laser endomicroscopy; however, they have not been widely used because of the narrow range of imaging capabilities, the price of equipment, and the length of time required to master the techniques.7 The biggest problem is that there is no certain answer as to which modality, including image-enhanced endoscopy, is the most effective method for detecting Barrett's epithelial dysplasia. Therefore, at this time, the ACG guideline recommends that LGD should be diagnosed histologically using random biopsy of Barrett's epithelium according to the Seattle protocol8 under surveillance endoscopy instead of optically.6 This guideline also states that surveillance endoscopy should be performed for all patients with BE. However, there are several problems with the implementation of this approach. First, the Seattle protocol is to biopsy four points every 90° at 2 cm intervals from the GEJ. Since dysplasia and cancer usually occur scattered within Barrett's epithelium, it has been noted that biopsy methods following this protocol may result in sampling errors.6 Furthermore, performing biopsies according to this protocol can be technically challenging because bleeding can interfere with achieving adequate biopsies at precise intervals. Systematically performing biopsies according to the Seattle protocol and submitting all biopsy material to pathology is time-consuming and costly compared to routine screening. Therefore, the guideline clearly notes that “current surveillance programs are expensive and time-consuming.”6 Second, the landmark of GEJ and definition of BE vary widely among guidelines.9 Specifically, the ACG and other Western guidelines define the GEJ as the upper end of the gastric mucosal folds, while the Japanese guidelines define it as the lower end of the esophageal palisade vessels.9 Therefore, the indications and methods for performing random biopsies of the Barrett's epithelium may differ among countries in real practice. Third, the diagnostic concordance rate among pathologists is not very high in the pathological diagnosis of LGD in BE, at 42%,10 which is a limitation to the pathological diagnosis. To reduce the frequency of surveillance endoscopy and minimize these shortcomings, a stratified model of the risk of developing dysplasia in patients with nondysplastic BE is needed. Ideally, the model should also be unaffected by differences in GEJ landmarks according to different guidelines. To address the problem of differences in national guidelines, the model constructed by Peleg et al. can be successfully adapted. Since the parameters used in the model do not include BE length, it is possible that this model can be adapted to Asian countries with a small proportion of long-BE patients.9 It is expected that the generalizability and transportability of this model will be demonstrated by validating the model in other cohorts, including Japan, in the future. When developing a surveillance strategy for a disease, it must be directly linked to the strongest end-point, such as a reduction in disease-specific mortality. Early detection of HGD has been shown to improve the prognosis of patients with BE, and it is important to identify precancerous lesions.6 Furthermore, since population-based studies have shown that the frequency of HGD or Barrett's adenocarcinoma differs depending on the presence or absence of LGD in BE, the detection of LGD in BE may determine the prognosis of patients.6 In fact, the ACG guidelines also recommends systematic surveillance every year for those with LGD while every 3–5 years for cases without LGD.6 Therefore, accurately determining the presence or absence of LGD in BE patients and predicting the risk of developing LGD are critical for the management of the disease. Adaptation of this model to stratify risk may allow adjustment of surveillance intervals for low-risk cases. Surveillance of the growing number of affected BE patients can be one of the causes of the tightening of the health-care economy. The use of such a predictive model will help make the best use of limited health-care resources. One limitation of this model was that the number of cases of HGD and Barrett's adenocarcinoma in the data used to construct the model was small, making it difficult to adapt the model to HGD and Barrett's adenocarcinoma. Since this model predicts histologic progression to dysplastic lesions integrating LGD, HGD, and Barrett's adenocarcinoma, it is not a model that can predict step-wise progression to any stage of dysplasia or cancer. The authors reported that the accuracy for predicting progression to HGD or Barrett's adenocarcinoma, even when adapting the other model developed in a previous report2 to the authors' dataset, was not as high (AUC 0.68). Thus, the difficulty in separately predicting advanced histology is a limitation of such models. The authors suggest that if the progression to LGD can be accurately predicted, the progression to HGD and Barrett's adenocarcinoma can also be predicted as the annual histologic progression rates in this cohort are similar to those previously reported. Ideally, however, the model should also include the prediction of progression to HGD and Barrett's adenocarcinoma. Another concern regarding the model construction is that a certain number of patients initially enrolled in this study may have missed dysplasia. The prevalence of development of dysplasia in nondysplastic patients in the study is high (133 out of 542). If the initial diagnosis for the model construction is inaccurate, the constructed model is unreliable. However, since LGD can be missed even in actual clinical practice, it may be acceptable for the model to include some degree of oversight. In fact, this study proves adequate predictive performance in validation. Future updates of this model using a larger cohort of datasets with more cases of HGD and Barrett's adenocarcinoma will bring this model closer to completion. Author S.S. is an Associate Editor of Digestive Endoscopy. The other author declares no conflict of interest for this article. None.

Subjects

MeSH terms (MEDLINE)

Barrett EsophagusEsophageal NeoplasmsPrecancerous ConditionsHumansHyperplasiaRisk Factors

Authors and institutions

  • Department of Gastroenterology, International University of Health and Welfare Ichikawa Hospital, Chiba, Japan.

ORCID profiles: Fumiaki Ishibashi, Sho Suzuki

Cite

APA

Ishibashi, F., & Suzuki, S. (2023). Predicting the risk of developing low‐grade dysplasia in patients with Barrett's esophagus without dysplasia. Digestive Endoscopy. https://doi.org/10.1111/den.14528

Vancouver

Ishibashi F, Suzuki S. Predicting the risk of developing low‐grade dysplasia in patients with Barrett's esophagus without dysplasia. Digestive Endoscopy. 2023. doi:10.1111/den.14528

BibTeX
@article{Ishibashi2023,
  author = {Ishibashi, Fumiaki and Suzuki, Sho},
  title = {Predicting the risk of developing low‐grade dysplasia in patients with Barrett's esophagus without dysplasia},
  journal = {Digestive Endoscopy},
  year = {2023},
  doi = {10.1111/den.14528}
}
Start for free →