Article Text

Download PDFPDF

Barrett’s oesophagus with indefinite for dysplasia shows high rates of prevalent and incident neoplasia in a UK multicentre cohort
  1. Maja Kopczynska1,2,
  2. Elizabeth Ratcliffe2,3,
  3. Harika Yalamanchili3,
  4. Anna Thompson1,
  5. Adib Nimri1,
  6. James Britton1,
  7. Yeng Ang1,2
  1. 1 Gastroenterology Department, Northern Care Alliance NHS Foundation Trust, Salford, Manchester, UK
  2. 2 Faculty of Medical and Human Sciences, The University of Manchester, Manchester, UK
  3. 3 Gastroenterology Department, Wrightington Wigan and Leigh NHS Foundation Trust, Wigan, UK
  1. Correspondence to Dr Elizabeth Ratcliffe, Gastroenterology Department, Wrightington Wigan and Leigh NHS Foundation Trust, Wigan, UK; elizabeth.ratcliffe{at}


Aims Barrett’s oesophagus with indefinite for dysplasia (IDD) carries a risk of prevalent and incident dysplasia and oesophageal adenocarcinoma. This study seeks to determine the risk of neoplasia in a multicentre prospective IDD cohort, along with determining adherence to British Society of Gastroenterology (BSG) guidelines for management and histology reporting.

Methods This was a cohort study using prospectively collected data from pathology databases from two centres in the North West of England (UK). Cases with IDD were identified over a 10-year period. Data were obtained on patient demographics, Barrett’s endoscopy findings and histology, outcomes and histological reporting.

Results 102 biopsies with IDD diagnosis in 88 patients were identified. Endoscopy was repeated in 78/88 (88%) patients. 12/78 progressed to low-grade dysplasia (15% or 2.6 per 100 person years), 6/78 (7.7%, 1.3 per 100 person years) progressed to high-grade dysplasia and 6/78 (7.7%, 1.3 per 100 person years) progressed to oesophageal adenocarcinoma. The overall incidence rate for progression to any type of dysplasia was 5.1 per 100 person years. Cox regression analysis identified longer Barrett’s segment, multifocal and persistent IDD as predictors of progression to dysplasia. Histology reporting did not meet 100% adherence to the BSG histology reporting minimum dataset prior to or after the introduction of the guidelines.

Conclusions IDD carries significant risk of progression to dysplasia or neoplasia. Therefore, careful diagnosis and management aided by clear histological reporting of these cases is required to diagnose prevalent and incident neoplasia.

  • Histopathology
  • Barrett Esophagus
  • Gastroenterology
  • Gastrointestinal Neoplasms
  • Esophagus

Data availability statement

Data are available upon reasonable request.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

What is already known on this topic

  • Indefinite for dysplasia (IDD) on Barrett’s oesophagus biopsies can be challenging to diagnose with the Vienna Classification, with national guidelines advising dual histological reporting.

  • Systematic review and metanalysis has shown rates of incident and prevalent neoplasia not too dissimilar to that of low grade dysplasia.

What this study adds

  • At the time of publication this is the first UK cohort study to include data from a non-tertiary referral centre and hence reflect real world experience.

  • Further to this, it is novel in the exploration of the role of the British Society of Gastroenterology (BSG) guidelines comparing the management of these cases prior to and after their publication.

  • We also explore histological reports and compare with the BSG minimum dataset for reporting pre and post the BSG guideline publication.

How this study might affect research practice or policy

  • This study shows a high rate of neoplasia comparable to other published cohorts but reflects a more real world experience and shows the importance of managing these cases appropriately.

  • Our data have highlighted some predictors of future progression to neoplasia which could be used to guide future follow up pathways.

  • Our data highlight the heterogeneity of histological reporting for this condition and future policy could include better coding and documentation of IDD cases.


Barrett’s oesophagus (BO) is the only known precursor lesion for oesophageal adenocarcinoma (OAC).1 In the Western world, rates of OAC are rising and oesophageal cancer is the fastest growing solid organ cancer.2 BO surveillance offers an opportunity for early detection of lesions amenable to minimally invasive endoscopic therapies, which have excellent outcomes.3 4 The natural history of BO associated cancer suggests a progression from the native stratified squamous epithelium to BO metaplasia, from which dysplasia progresses from low grade to high grade, to intramucosal cancer (IMC) before invasion.5 At each of these preinvasion stages, an opportunity arises for endoscopic therapy such as radiofrequency ablation (RFA) or endoscopic resection. Hence, national and international guidelines support the use of endoscopic surveillance 1 to 5 yearly to facilitate detection of preinvasive changes.6–8 Cohort studies suggest improved outcomes for cancers detected on surveillance, though large, randomised trials are yet to be published.9–12

At present, there is no recommended biomarker or panel advised for sole use to risk stratify patients with BO, and therefore the main marker for future progression is the presence of dysplasia on histology. Histology samples are assessed by pathologists using the Vienna Classification, a five stage assessment of morphological and structural changes which denote a level of atypia13: low-grade dysplasia (LGD), high-grade dysplasia (HGD), or IMC. However, when a sample cannot be attributed to one of these categories or true dysplasia cannot be distinguished from inflammatory changes the sample may be labelled as ‘indefinite for dysplasia’ (IDD). Often a confident decision of dysplasia may be affected by technical factors or the presence of concurrent inflammation—a feature favouring inflammation and regeneration is surface maturation—where atypical changes seen in deeper glands improves nearer the epithelium. Features more suggestive of dysplasia would include polarisation of the nuclei, nuclear pleomorphism and an abrupt transition between atypical and normal epithelium.6 14 Defining IDD has been challenging and there has been a wide interobserver variability reported in histology reporting for IDD.15–17 Given this variability, both British Society of Gastroenterology (BSG) and American College of Gastroenterology (ACG) guidelines advise reporting should be performed by two histopathologists, one of whom should be experienced in gastrointestinal reporting.6 7

The identification of IDD on histology is clinically significant, with recent metanalysis suggesting an incidence rate of progression to LGD of 11.4 cases per 100 person years, 1.5 pooled for HGD/OAC and 0.6 for OAC.18 The potential significance of IDD for development of dysplasia is acknowledged in the national guidelines, with BSG guidelines introduced in year 2013 advising high-dose acid suppression with proton pump inhibitors (PPI) and repeat endoscopy surveillance in 6 months, similar to the LGD pathway, after which there may be regression to non-dysplastic BO (NDBO) or alternatively, progression to any underlying dysplasia may be evident.6 The Benign Barrett’s oesophagus consensus task force, an international expert Delphi process, advises IDD is an interim diagnosis and acid suppression and early repeat endoscopy<1 year to allow for prevalent neoplasia to be detected.8 The latest update from the ACG guideline highlights the need for two histopathologists reporting and 6-monthly repeat endoscopy.19

The aim of this study was to identify the risk of prevalent and incident neoplasia in a patient cohort from a tertiary referral centre and a district general hospital to gain insight into the ‘real world’ IDD outcomes. The secondary aim was to explore the adherence to BSG guidelines for follow-up and management of IDD cases and compare the histological reports with the BSG guideline minimum dataset for histological reporting.


Study design

This was a cohort study using prospectively collected data from pathology databases from two centres in the North West of England (UK): a tertiary referral centre and a district general hospital. The pathology databases included patient demographics and coded morphology for every oesophageal biopsy undertaken between 01 January 2009 and 31 December 2018. The full histology report of all biopsies with inflammation, atypia, metaplasia, dysplasia and normal tissue morphology were reviewed (figure 1). Patients with a diagnosis of IDD were included in the study. The study was registered with the hospital audit department (local reference number 2019 188) and data were managed in accordance with Caldecott Principles to protect patient confidentiality.

Figure 1

Flow chart diagram outlining the range of coding found on screening of cases identified as indefinite for dysplasia on review of histopathology reports. NOS, not otherwise specified.

Data collection

For patients with the diagnosis of IDD further data on clinical characteristics (tobacco smoking, alcohol excess, diabetes mellitus, body mass index), index endoscopy (indication, findings), management (organisation of full multidisciplinary team (MDT), prescription of high-dose PPI, date of repeat endoscopy), and outcomes (persistent IDD, progression to dysplasia, survival) were collected using clinical notes. If more than one IDD biopsy was identified for a patient, the first diagnosis of IDD date was used as the start of the follow-up. Patients were followed up until death or the end of follow-up on 31st July 2021. Progression to dysplasia rates were categorised as prevalent neoplasia when the progression to any grade of neoplasia (LGD/HGD/OAC) occurred within 1 year of follow-up, or incident neoplasia when the progression occurred after 1 year of follow-up. Additionally, the changes in diagnostics (including histology reporting standards) and management were compared for patients who were diagnosed with IDD before and after introduction of BSG guidelines in year 2013.

Statistical analysis

Categorical variables are described as proportions and continuous variables are described as median with IQR or mean with SD. Analyses were performed using χ2 for categorical variables to compare diagnostics and management before and after introduction of BSG guidelines. Incidence of progression to LGD, HGD or OAC was calculated per 100 patient years of follow-up. To determine the risk factors of progression to any type of dysplasia (combined LGD, HGD and OAC) univariate analysis was performed using the logistic regression model. Estimates of relative risk were presented as OR with 95% CI. Two-tailed p value<0.05 was considered statistically significant. The statistical software used was R V.4.0.3.


Patient demographics

In total, 102 biopsies with IDD diagnosis in 88 patients were identified. Median follow-up was 5.3 years (IQR 3.5–8.2) with a total of 523.3 patient years. Median patient age at index biopsy was 69 years (IQR 56–76) and 58/88 (66%) patients were men. The demographics and clinical characteristics of included patients are presented in table 1.

Table 1

Patient demographics and clinical characteristics

Endoscopy findings

The most common indication for undergoing endoscopy was BO surveillance in 56/88 (64%) patients, followed by dysphagia in 11/88 (12%) and iron deficiency anaemia in 8/88 (9.1%) patients. Characteristics of index endoscopy are summarised in table 2.

Table 2

Characteristics of index endoscopy


The analysis of outcomes was performed for 78 patients who underwent a repeat endoscopy at any point since IDD diagnosis. In 51/78 (65%) patients, IDD changes regressed. In 14/78 (18%), IDD persisted on a repeat endoscopy. In total, 24/78 (31%) patients progressed to develop any type of dysplasia. Of these, 11/24 (46%) were prevalent cases and 13/24 (54%) incident cases. The relationship between persistent IDD and dysplasia progression is presented in figure 2.

Figure 2

The relationship between persistent IDD and dysplasia progression. IDD, indefinite for dysplasia.

In total, 12/78 (15.4%) patients progressed to develop LGD (in 6 cases progression was prevalent and in 6 incident), 6/78 (7.7%) progressed to develop HGD (in 2 cases progression was prevalent and in 4 incident) and 6/78 (7.7%) patients progressed to develop OAC (in 3 cases progression was prevalent and in 3 incident).

This corresponded to an incidence rate of 2.6 per 100 person years for progression to LGD, rate of 1.3 per 100 person years for progression to HGD, and 1.3 per 100 person years for progression to OAC. The overall incidence rate for progression to any type of dysplasia was 5.1 per 100 person years.

Cox regression was performed to identify predictors of progression to any type of dysplasia (table 3). Longer Barrett’s segment, multifocal and persistent IDD were associated with progression to dysplasia.

Table 3

Predictors of progression to dysplasia


In total, 13/88 (14.8%) of patients died during the follow-up, of which one died due to progression to oesophageal cancer. Other causes of death included cardiac, respiratory, unrelated cancers and infections.

Histopathology reporting

The histological diagnosis of IDD on index biopsy was confirmed by two histopathologists in 37/88 (42%) cases and p53 staining was performed in 12/88 (13.6%) cases. The standard of histology reporting is summarised in figure 3.

Figure 3

Histopathology standards in description of index indefinite for dysplasia diagnosis biopsy.

Management after IDD diagnosis

After a diagnosis of IDD, 41/88 (47%) cases were discussed at a MDT meeting. 61/88 (69%) patients were initiated on a high dose of PPI for 6 months. Endoscopy was repeated in 78/88 (88%) patients, for 57/88 (65%) cases a repeat endoscopy was undertaken within 6 months. One patient underwent endoscopic mucosal resection for an IDD lesion and two underwent RFA procedure.

Changes in practice post BSG guidelines

The changes in diagnostics and management of IDD cases were compared for patients diagnosed with IDD before and after BSG guideline’s introduction in 2013 (table 4). The use of p53 staining significantly improved after the introduction of BSG guidelines and there was a trend in more frequent confirmation of IDD by two histopathologist. There were no obvious trends in histopathology reporting with some sections improving and some deteriorating in quality (table 4). Interestingly, the incidence of IDD detection did not change before and after BSG guidelines with 38 out of 1598 (2.4%) of biopsies examined before the guideline introduction having an IDD diagnosis in comparison to 64 out of 2984 (2.1%) of biopsies examined after the guideline introduction.

Table 4

Changes in diagnostics and management of indefinite for dysplasia cases


In this study, we found that a significant proportion of patients with IDD progressed to develop neoplasia. Progression to LGD was the most common; however, progression to HGD and OAC was also noted, both as prevalent and incident cases. Moreover, this is the first study to explore the adherence to BSG guidelines for the follow-up and management of IDD cases, which identified potential challenges and areas of improvement. Additionally, the novelty of the study is the evaluation of the histological reports and comparison with the BSG guideline minimum dataset for histological reporting, highlighting the need for development of standardised wording for the diagnosis of IDD in the summary section and flagging cases for discussion in MDT in the report.

Overall, 31% of IDD cases in our cohort progressed to any type of neoplasia, with 14% cases developing prevalent and 17% incident neoplasia. In total, 15.4% patients progressed to develop LGD, 7.7% to HGD and 7.7% to OAC. The rate of progression was 2.6 per/100 person years for LGD, 1.3/100 person years for HGD, 1.3/100 person years for OAC and 5.1/100 person years for any type of neoplasia. A recent metanalysis and systematic review of 8 studies (4 in the US and 4 in Europe) showed greater incidence rate for LGD of 11.4 per 100 person years but lower for OAC (1.5 pooled for HGD/OAC, 0.6 for OAC); however, the authors did note higher OAC rate in European studies.18 Other studies published since this metanalysis have shown differing rates. The RIBBON network—Ireland BO network—found a risk of HGD of 3% per annum in IDD cases, and for OAC of 1.9% per annum,20though lower when excluding the first follow-up year (1.28% HGD, 0.77% OAC). In the study by Han et al, which followed 107 IDD cases over a 10-year period, 9.3% cases developed prevalent dysplasia, with no incident HGD/OAC. Phillips et al cohort from Cambridge, UK, and Mayo clinic, USA, showed prevalent neoplasia in 9.5% of cases, with 14.5% incident neoplasia.21 This gave a risk of any neoplasia of 3.2 cases/100 person years, and for HGD/OAC of 0.6 cases/100 person years. Our data showed a higher risk of any neoplasia (5.1/100 person years), comparing with the above from two tertiary referral centre. This may be reflective of the Northwest of the UK having higher rates of OAC; however, the histology slides for our cases were not re-reviewed for this study and some not dual reported so it may be that some of these cases were dysplastic at their index assessment.

Our study also explored ‘real world’ management of IDD cases and impact of BSG guidelines on care delivery. First, only two-thirds of cases achieved a repeat endoscopy at 6 months, though follow-up endoscopy was performed at some point in 88% of cases. Of note, there was no significant improvement in the 6-month endoscopy follow-up post the introduction of the BSG guidelines. The importance of the 6-month follow-up is to detect prevalent neoplasia in the IDD patients. Multiple cohort studies have shown high rates of neoplasia at follow-up at less than 1 year as outlined above; hence, early follow-up for these cases is currently included in all national and international guidelines, with most experts treating IDD in a similar way to LGD.20–22 Reasons for non-adherence have not been explored in our cohort but factors could include scheduling issues, clinicians not recognising the IDD cases, or patient factors. Second, use of high dose PPI was only seen in 69% of cases, which was also not improved after the introduction of BSG guidelines. Use of high-dose PPI is crucial as ongoing inflammation may be part of the natural history of progression of IDD to neoplasia.23

The novelty of our study was the investigation of the standards of histology reporting. We found that only 42% of histology reports were reviewed by two histopathologists, though this did increase after introduction of BSG guidelines. In the past studies, LGD and IDD have been shown to have fair and poor interobserver agreement, respectively, with ACG guidance quoting the low concordance in IDD diagnosis alongside the potential over and under diagnosis of these conditions in as many as 9% of cases.19 Interestingly, studies have suggested that LGD progression is more likely when more histopathologists agree with the diagnosis. In the study by Duits et al, the review of LGD cases from community referrals by an expert panel led to a downstaging of 73% of cases from LGD to IDD or NBDO.24 Though this was a LGD cohort, it emphasises the importance of dual reporting in these ambiguous conditions, as it could reduce the unnecessary follow-up for those with reactive atypia alone. Notably, even among their expert panel, there was 28% disagreement for IDD/LGD/NDBO, and they only achieved interobserver agreement of K=0.49. Overall, adherence to the BSG minimum histology reporting in our cohort was well below 100% and comparing before and after the BSG guideline’s introduction, there was no significant change in the reporting practices. Moving forward improved adherence to the BSG reporting guidelines could be beneficial particularly with standardised wording for the diagnosis of IDD in the summary section and flagging cases for discussion in MDT in the report.

Noteworthy, in our study, one of the key challenges was case finding, which was due to heterogeneity of histological reporting as there was no fixed code describing IDD (figure 1). The identification of the included cases was only possible thanks to reviewing the full histology report of the biopsy samples. This reflects a ‘real world’ issue that cases in practice may not be recognised or flagged as significant if not identified clearly in the report. Of note, in our study only 17% of the cases had a summary included the histology report. It is possible that the standard of the reporting was one of the reasons for follow-up issues, in addition to the fact that only 47% cases were discussed in MDT.

In our study, we were also interested in identifying factors associated with progression from IDD to dysplasia. We identified longer Barrett’s segment, multifocal and persistent IDD as significant factors, which is similar to previous studies.18 21 In the future, use of tissue biomarkers such as aberrant p53 staining could be useful in providing more precise risk stratification.25 This is supported by data from Januszewicz et al who showed p53 staining reduced the diagnosis of IDD by 40% and aberrance correlated with future progression.26 Notably, in our study, we found reporting of p53 staining in 13.6% of cases though this did increase after the introduction of the BSG guidelines (p=0.015). There is potential that wider use of this biomarker could streamline cases to identify those at greatest risk of progression to dysplasia and reduce unnecessary intense follow-up for low-risk cases. Additionally, in the future, DNA flow cytometry may provide further risk stratification of IDD cases. Choi et al showed inflammation was a key indicator of future risk alongside DNA methylation abnormalities.23 Using DNA flow cytometry, they graded the amount of DNA abnormalities in specimens. The combination of the amount of DNA abnormalities with the grade of inflammation was predictive of outcome with high risk of neoplasia if both results were highly abnormal and low risk of neoplasia with no inflammation and minimal DNA abnormalities. This work is supported by Kilcoyne et al who found that genomic abnormalities could predict future oesophageal cancer years before transformation.27

The limitations of this study were that although cases were identified from a prospectively maintained database, additional data were obtained retrospectively from endoscopy and histology reports. The histology slides themselves were not reviewed, and many of the cases not dual reported, so it may be the true incidence of IDD diagnosis is different from that described and may be that progressors were dysplastic at their index biopsies. The overall number of cases across two sites over 10 years was small compared with the surveillance population but comparable to other cohort studies and identified through a thorough screening of all histology reports during the study period—this emphasises the need for improved documentation and coding of cases for future studies. Strengths of this study include the inclusion of non-referral centre data which could make the data more generalisable to wider experience, and the inclusion of consecutive cases makes data more reliable though prospective case selection.


In this cohort, we have shown that IDD is a condition to watch and has comparable risk of progression to previously described cohorts. Worryingly, follow-up for these cases has been mixed and action should be taken to enhance the clarity of histological reporting and acknowledgement of the risk in IDD in order to make follow-up care for these cases more robust. There is pressure on waiting list times and services, so use of p53 or other risk stratification models may be beneficial to streamline the high-risk cases to more aggressive follow-up after the initial repeat endoscopy at 6 months which should detect prevalent neoplasia.

Data availability statement

Data are available upon reasonable request.

Ethics statements

Patient consent for publication


With thanks to the histopathology departments at each site, Dr Stephen Hayes for guidance on defining IDD. We wish to thank Dr Ayman Elmasri and the audit department team for assistance with case acquisition.



  • Handling editor Runjan Chetty.

  • Twitter @lil_ratcliffe

  • Contributors YA, MK, JB and ER were involved in the conception and design of the study. MK, ER, HY, AT and AN performed the extensive screening of the histology reports and data collection and cleaning. MK performed the statistical analysis of the data with input from ER, JB and YA. MK and ER wrote the initial draft manuscript, all authors reviewed and revised the manuscript. Final approval of the submitted manuscript was obtained from all authors. YA is the guarantor for the study.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests YA and ER receive research funding from Medtronic for other studies.

  • Provenance and peer review Not commissioned; externally peer reviewed.