Detection of high-grade dysplasia and adenocarcinoma in Barrett's esophagus using high-resolution virtual chromoendoscopy versus the Seattle protocol: the CONVERSE study

Abstract

Background

Barrett’s esophagus (BE) is a pre-cancerous condition with an increased risk of esophageal adenocarcinoma (EAC). Current surveillance involves white light endoscopy with random biopsies (Seattle Protocol, SP) but has limitations. This study explores high-resolution virtual chromoendoscopy (HRMC) as a potential alternative.

Methods

This controlled trial will enroll 110 patients with Barrett’s esophagus (BE) across 15 gastroenterology departments in France. Each patient will undergo both HRMC and SP examinations during the same endoscopic procedure. Although the trial is non-randomized, it is important to highlight that the two endoscopists performing either the HRMC or the SP will be blinded to each other’s results. Therefore, each patient will serve as their own control. Biopsy decisions will be based on both methods, with any visible lesions resected. The primary objective is to compare the detection rate of HGD and EAC lesions using HRMC with targeted biopsies versus SP with random biopsies. Secondary objectives include comparing detection rates of LGD lesions, procedure time, missed lesions, and cost-effectiveness.

Discussion

If HRMC proves superior to SP for detecting HGD and EAC, it could lead to more accurate, efficient, and cost-effective BE surveillance strategies, improving patient outcomes and resource utilization.

Trial registration

ClinicalTrials.gov NCT05229783 – First Submitted 2021-12-07.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12876-026-05038-3.

Background

Esophageal cancer ranks eleventh globally in incidence and seventh in mortality among all cancers [1]. In 2019, the United States reported 328,030 new cases of digestive diseases, resulting in 165,460 deaths. This included 17,650 new cases of esophageal cancer with 16,080 deaths [2].

Esophageal cancer comprises two main histological types: adenocarcinoma (EAC) and squamous cell carcinoma (ESCC). EAC primarily affects the lower third of the esophagus and is often preceded by Barrett’s esophagus (BE), a recognized precursor condition. ESCC typically affects the upper two-thirds of the esophagus [3]. Risk factors for ESCC include smoking and alcohol consumption, while EAC is associated with obesity, Caucasian ethnicity, gastro-esophageal reflux disease (GERD), smoking, and family history [4].

Historically, ESCC has been predominant, but in developed countries such as the UK, USA, and Australia, there has been a shift towards EAC in recent years [5]. This shift is attributed to rising GERD incidence due to obesity and poor dietary habits, contributing significantly to BE development. The relative risk of EAC in BE patients is reported as 11.3 [6].

BE involves metaplastic transformation of esophageal squamous epithelium into columnar epithelium (Fig. 1). Disease severity is assessed based on dysplasia presence and degree in biopsy pathology. BE represents a premalignant stage, with annual cancer development risks estimated at 0.33% for non-dysplastic BE, 0.54% for low-grade dysplasia (LGD), and 7% for high-grade dysplasia (HGD) [7].

Fig. 1.

Comparison of 2 images of early lesions in Barrett’s esophagus performed using the same endoscope. (A) WLE showing no lesion; (B) HRMC (Blue Light Imaging modality) showing a suspicious area for HGD

The gold standard for BE monitoring is esophagogastroduodenoscopy with random 4-quadrant biopsies (Seattle protocol SP), tailored by BE length which correlates with EAC risk [7–11]. Surveillance aims to detect early EAC, particularly pre-submucosal invasion, which elevates lymph node metastasis risk [12]. Despite debate on efficacy, systematic biopsies prolong procedures, increase costs, and risk sampling errors, impacting lesion detection and disease management adherence.

Financial burdens and resource demands on gastroenterology departments question long-term surveillance efficacy given low disease progression risks. However, large cohort studies (n = 30,000) report earlier EAC detection, improved survival, and reduced cancer-related mortality with surveillance-detected EAC cases [13].

To address biopsy limitations, advanced imaging like high-resolution endoscopy with magnification and electronic chromoendoscopy (HRMC) emerged, enhancing BE diagnosis. Narrow-band imaging (NBI), using optical filters to enhance microvascular visualization, demonstrates high accuracy in EAC and HGD detection, surpassing white light endoscopy (WLE) in sensitivity, specificity, and biopsy requirements [14–16]. Notably, no trial has compared WLE with SP against HRMC.

The CONVERSE study innovates by evaluating HRMC against standard care. Using a crossover design, blinded endoscopists compare HRMC-targeted biopsies with SP 4-quadrant biopsies in the same patient. Lesions detected are immediately resected, potentially altering French and European guidelines to simplify care, reduce biopsies, endoscopy time, sedation needs, and histological interpretation, lowering BE management costs. The French Society of Digestive Endoscopy (SFED) strongly supports this project.

The study aims to validate HRMC for BE surveillance, potentially replacing current standards involving numerous biopsies and lengthy procedures.

Methods/design

This study protocol was developed and reported in accordance with the SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) guidelines to ensure transparency and completeness in the reporting of clinical trial protocols.

Aims and objective

The primary objective of the CONVERSE study is to demonstrate the superiority of high-resolution endoscopy combined with high-resolution magnification chromoendoscopy (HRMC) and targeted biopsies over the current standard of care, which involves white light endoscopy (WLE) with targeted biopsies and systematic biopsies following the SP, for detecting esophageal adenocarcinoma (EAC) and high-grade dysplasia (HGD).

The secondary objectives of the study are:

• To demonstrate the superiority of HRMC with targeted biopsies over WLE with targeted biopsies in detecting HGD and EAC.

• To demonstrate the superiority of HRMC with targeted biopsies compared to the SP in detecting low-grade dysplasia (LGD).

• To assess whether HRMC with targeted biopsies reduces the overall procedure time compared to WLE with targeted biopsies combined with SP.

• To evaluate the potential for missed lesions through an adjudication committee review of procedure videos, divided into two phases: (A) HRMC examination and (B) WLE examination.

• To demonstrate the superiority of HRMC in guiding the resection of HGD or EAC compared to WLE.

• To assess the rate of adverse events (AEs) associated with HRMC compared to the standard procedure.

• To conduct a cost-effectiveness analysis to evaluate the economic impact of HRMC as a surveillance tool in Barrett’s esophagus (BE) management.

By achieving these objectives, the study aims to validate HRMC as a more effective, efficient, and potentially cost-saving alternative for the detection and management of dysplasia and early-stage esophageal cancer in BE patients.

Study design and setting

This study is a multicenter, controlled, and prospective investigation. Although not randomized, endoscopists will remain blinded to each other’s assessments during the examinations, ensuring that each patient serves as their own control.

This national, multicenter study will involve patients from gastroenterology departments of fifteen French hospitals and clinics, including CHU Nantes, CHU Brest, various APHP sites, CHU Bordeaux, CHU Nice, the Paris – Bercy Clinic, Lyon Civil Hospices, CHU Poitiers, CHU Limoges, CHU Rennes, Sainte Barbe Clinic, CHU Amiens, CHU Nancy, and CHU Besançon.

Study population

This study will enroll adult patients, aged 18 years or older, both male and female, who are affiliated with the social security system and have been diagnosed with non-dysplastic BE measuring greater than 3 cm in length or dysplastic Barrett's esophagus (BE) of any length, with no visible lesion identified on prior endoscopy. All eligible participants must have received and signed the informed consent form. Women of childbearing potential must use effective contraception, such as an intrauterine device, oral contraceptives, or implants, for the duration of the study.

Several exclusion criteria will be applied. Pregnant or breastfeeding women, as well as patients under legal guardianship, will be excluded. Additionally, patients with a prior history of BE treatment, those with contraindications to sedation or general anesthesia, or those currently taking clopidogrel, anticoagulants, or with coagulopathies will be ineligible. Patients with a general health status that precludes adequate follow-up for BE or those participating in another interventional study will also be excluded.

Recruitment is planned over a 36-month period. Inclusion in this protocol will be simple.BE is an asymptomatic epithelial transformation, common in the general population (approximately 1–2%), and currently requires surveillance once histologically confirmed. In France, an estimated 1,500 new cases of esophageal adenocarcinoma (EAC) are diagnosed each year.

The dramatic increase in the incidence of EAC over the past decades appears to have been preceded by a similar steep increase in the incidence of BE [17]. Indeed, a 93% increase in the incidence of BE was observed between 1993 and 2005, with a potential confounding bias avoided by counting the number of BE cases per 100 endoscopies [18]. A commonly proposed explanation for the increasing incidence of BE in Western countries is the decreasing prevalence of gastric colonization with Helicobacter pylori [19]. An alternative hypothesis, based on the prevalence of obesity, has been proposed to explain the increasing prevalence of BE and EAC and is gaining popularity [20].

Interventions description

Prior to any study-specific procedures, written informed consent will be obtained from all participants. Individuals have the right to withdraw from the study at any time without compromising their ongoing care.

Additional research biospecimens (biopsies of lesions detected only by HRMC) will be handled in the same manner as standard biospecimens (other biopsies) by pathology departments. At the conclusion of the trial, all specimens will be stored in pathology departments in accordance with local regulations.

As illustrated in the study diagram (Fig. 2), the screening visit occurs between Day − 60 and Day − 1 (D-60 to D-1) prior to endoscopy. On D0 (baseline), an upper gastrointestinal endoscopy is performed under general anesthesia. Each patient serves as their own control. Two endoscopists proceed as follows using the same endoscope:

Fig. 2.

CONVERSE study design

• Endoscopist A: Performs HRMC using Blue Light Imaging (BLI) or Narrow Band Imaging (NBI) modalities of the endoscope, meticulously inspecting and documenting all visible lesions with precise location indications (in centimeters from the incisors and clockwise orientation). Annotates on the SFED scheme the biopsies or resections that would have been performed during this procedure.

• Endoscopist B: Performs the examination using the endoscope’s WLE modality and describes all visible lesions, including precise location.

At the end of the examination, endoscopists will be unblinded to each other’s findings, and either Endoscopist A or B will perform biopsies/resection.

Each lesion detected by WLE, HRMC, or both will be transferred to the pathology department in a separate container. For quadrantic biopsies performed according to the SP, each container will hold four quadrant biopsies from the same esophageal level. The procedural schedule is summarized in the following diagram (Fig. 3).

Fig. 3.

The study’s procedure for the 2 endoscopist. Video 1 will consist in whole examination and lesions description by endoscopist A performing HRMC. Video 2 will consist in examination and lesions description by endoscopist B under WLE

Recent advancements in endoscopic technology have led to the development of high-resolution endoscopes capable of real-time, detailed visualization of mucosal morphology. These devices, equipped with both WLE and HRMC, are now widely available to endoscopists. Fujifilm and Olympus offer training courses in chromoendoscopy. WLE provides broad-spectrum, high-brightness images of the digestive mucosa, essential for lesion detection. High-resolution endoscopes from Fujifilm^® and Olympus^® incorporate BLI and NBI, respectively. These electronic chromoendoscopy modalities enhance capillary and mucosal microstructure visualization, improving lesion assessment.

All study sites are equipped with high-resolution endoscopes like Olympus Gif 190 or Fuji 760. These devices are standard equipment in these centers and commercially available for the intended use. They will not be labeled as “clinical study” devices. Participating pharmacies will manage device ordering, storage, and distribution. Endoscopic equipment information will be recorded in the e-CRF.

On Day 30 (± 7 days), a consultation will be scheduled to review the pathology results with the patient. During this session, the biopsy findings will be communicated, including any detection of precancerous or cancerous cells. Subsequent management will be determined based on these results. Videos showing discrepancies between the endoscopist’s observations and the pathologist’s findings will be forwarded to the adjudication committee. This committee comprises experts in Barrett’s esophagus who are not investigators in the study. The trial schedule is illustrated in Table 1.

Table 1.

Participant timeline: schedule of enrollment, interventions, and assessments of CONVERSE study

Activities	Pre-Inclusion Visit D-90 to D-1	Anesthesiologist Visit	Endoscopy Visit D0 (inclusion visit)	Follow-up Visit D30+/-7
Eligibility review and confirmation (Inclusion/Non-inclusion criteria)	x		x
Informed Consent Form	x
Endoscopy with videos recording			x
Clinical examination Medical History	x	x		x
Adverse Event			x	x

All treatments deemed necessary by the investigator for the well-being of the patient. At the conclusion of the clinical research, the patient will continue to be monitored by their gastroenterologist and will receive the standard care for their condition. Additionally, the sponsor has obtained insurance coverage for its civil liability in accordance with regulations.

The treatment duration per patient corresponds to one day for the endoscopy, with a follow-up period of one month. The recruitment period extends to 36 months.

Outcomes

The primary endpoint of this study is to compare the detection rates of high-grade dysplasia (HGD) and esophageal adenocarcinoma (EAC) lesions using high-resolution magnification chromoendoscopy (HRMC) with targeted biopsies, which is expected to approach 90%, against the detection rates achieved with white light endoscopy (WLE) using both targeted and random biopsies, anticipated to be around 70%. To ensure consistency in lesion identification, a margin of ± 1 cm and ± 3 ‘hours’ (equivalent to one-quarter of the circumference in a clock-face orientation) will be applied, indicating that tumors described within these parameters by both endoscopists will be considered identical.

The secondary outcomes of this study include the following:

First, we will compare the rates of HGD and EAC lesions detected by HRMC with targeted biopsies versus those detected by WLE with targeted biopsies. Second, we will evaluate the rates of low-grade dysplasia (LGD) lesions detected by HRMC with targeted biopsies compared to those identified through the Seattle Protocol (SP). Third, the duration of each procedure—WLE, HRMC, and SP—will be assessed to evaluate efficiency.

Additionally, we will analyze the rate of missed lesions (HGD and EAC) as diagnosed by an adjudication committee that reviews procedure videos. We will also determine the rate of resected lesions (HGD and EAC) identified by one endoscopist but overlooked by another. The study will record the number of adverse events (AEs) encountered during the procedures. Finally, a cost-effectiveness analysis will be conducted to compare the two strategies for detecting early EAC and HGD, specifically assessing HRMC against the SP strategy.

Sample size and statistical analyses

A total of 110 patients were required for this study. This sample size was calculated based on a projected Positive Predictive Value (PPV) of 90% for HRMC compared to 70% for WLE, considering a type-I error of 5% and a power of 90% (requiring 67 patients). An additional 50% was added to account for patients with no detected lesions. Due to the possibility of multiple detected lesions per patient, the actual power achieved is expected to exceed the calculated power. Given the nature of the study, where each patient served as their own control, a crossover design was employed. HRMC was always performed first, followed by the Seattle Protocol, with different endoscopists conducting each procedure.

To compare the paired proportions of PPVs obtained with HRMC and SP, a McNemar’s test was utilized. The PPV for each technique was calculated as the ratio of detected HGD and EAC lesions to all detected lesions. For secondary objectives, given the paired nature of the data, both McNemar’s test (for binary variables) and paired Student’s t-test (for continuous variables) were employed to compare outcomes between HRM and SP.

A health-economic analysis was conducted from a hospital perspective, considering a time horizon starting at the initiation of the study and ending upon the completion of histological results. The comparative effectiveness of the two strategies was assessed based on the number of detected HGD and EAC lesions. Costs were estimated by quantifying biopsy and histology procedures, with resource utilization valued using official NHI fees. Mean costs and their corresponding 95% confidence intervals were reported. A cost-effectiveness plane was used to graphically represent the findings.

Interim analyses were not conducted. Similarly, no additional subgroup analyses were performed. Imputation methods were not necessary as data on detection by each technique was available for every lesion identified by HRM and/or SP.

Data collection and management

Converse’s Scientific Committee was established by Professor Coron and is led by Dr. Queneherve. Its membership encompasses a biostatistician and methodologist, the medico-economist, and the project manager of the Clinical Investigation Center (CIC 1413). The Steering Committee comprises members of the Scientific Committee, augmented by the data management team, the study coordinator responsible for facilitating patient enrollment at additional sites, the sponsor’s project manager, and the monitoring Clinical Research Assistant. The sponsor’s project manager oversees this committee and produces the “Converse newsletter,” which features updates on patient enrollment and protocol amendments.

An eCRF will be created for each participant included in CONVERSE. Participants will be identified by a unique code, which will be the sole identifying information included in the eCRF, allowing for retrospective patient linkage. Upon study completion, the CRF database and safety database will undergo reconciliation prior to database lockdown. A similar reconciliation process will occur annually during the Annual Safety Report update.

Ethics and disseminations plans

All patient medical data will be disclosed exclusively to the sponsor or their authorized representatives, and where applicable, to authorized health authorities, subject to strict confidentiality. All patient medical data will be disclosed in accordance with applicable laws, including the General Data Protection Regulation (GDPR) and the French Data Protection Act. Direct access to medical records will be strictly limited to authorized representatives of the sponsor and health authorities, and will only be granted where legally permitted. Appropriate security measures will be implemented to ensure the confidentiality of participants’ personal information throughout the study. The sponsor and regulatory authorities may request direct access to medical records for verification of study procedures and/or data, as permitted by law and regulation. All trial data will be processed electronically in compliance with CNIL regulations. The CNIL is an independent regulatory authority responsible for enforcing data privacy laws governing the collection, storage, and use of personal information.

Aligned with French Law (Jardé n°2012 − 300), this intervention study poses minimal risks and constraints, mirroring routine care practices rendering a Data and Safety Monitoring Committee unnecessary.

While the protocol aims to modify BE monitoring guidelines, emphasizing visualization, the ESGE classifies diagnostic endoscopies, including surveillance with or without biopsy, as low-risk for hemorrhage [21]. Given the minimal risks and constraints, no study-related adverse events are anticipated. Any adverse reactions encountered during patient care will be reported to the appropriate vigilance system (pharmacovigilance, biovigilance, haemovigilance, medical device vigilance).

The amended protocol should be a dated, updated version. If necessary, the informed consent form and patient information sheet should be revised. The CHU Nantes sponsor project manager should notify study sites, and a copy of the revised protocol should be provided to all principal investigators for inclusion in the Investigator Site File. The current protocol version is 4, dated February 24, 2023. Requests for substantial modifications to the protocol should be submitted by the sponsor for approval or notification to the French National Agency for the Safety of Medicines and Health Products (ANSM) and to the Ile-de-France XI –Ethics Committee.

A sponsor and/or regulatory authority inspection or audit may occur during this study. Inspectors will examine documents, logistics, records, and any other trial-related materials located at the study site.

Trial results will be disseminated through publication in international oncology, medical, and scientific journals, as well as presentations at national and international conferences. The datasets analyzed in the present study along with the statistical code, are accessible upon reasonable request from the corresponding author. The complete study protocol is also available.

Trial status

This trial is still ongoing; patient inclusion is not yet complete.

The updated protocol is at version 4 on February 24 2023.

The first patient was enrolled in May 2022.

Recruitment by the investigating centers is planned to continue until May 2025.

As of October 2024, 106 of the planned 110 participants had been recruited; 7 participants were still in follow-up and 4 remained to be included.

Discussion

Following a review of the protocol, we have modified one inclusion criterion. Initially, patients with non-dysplastic Barrett’s esophagus (BE) greater than 3 cm or dysplastic BE of any size were eligible. However, based on preliminary data from the first 22 enrolled patients, we have decided to prioritize patients with dysplastic BE, specifically those with flat lesions.

Among the 22 patients initially enrolled, only 5 (23%) exhibited high-grade dysplasia or adenocarcinoma upon examination using at least one of the three techniques (HRMC, WLE, or Seattle quadrantic biopsies). This finding significantly deviates from our initial hypothesis, which anticipated that 67% of patients would present with such lesions.

Furthermore, none of the nine patients with non-dysplastic BE greater than 3 cm exhibited high-grade dysplasia or adenocarcinoma. Therefore, we have decided to exclude this patient type from future enrollment. By focusing on patients with flat lesions, which are generally not detectable by white-light endoscopy, we aim to increase the likelihood of identifying dysplastic lesions.

To compensate for the exclusion of non-dysplastic BE patients greater than 3 cm, we have requested an increase in the total enrollment number from 100 to 110. This adjustment will ensure that we maintain adequate statistical power for the study.

The findings of this study have the potential to significantly improve patient care and public health. By focusing on patients with flat lesions and utilizing HRMC, we anticipate a reduction in the number of biopsies required, thereby decreasing the risk of hemorrhagic complications. Moreover, earlier detection and treatment of lesions could lower the risk of progression to invasive cancer, aligning with the World Health Organization’s Global Action Plan for the Prevention and Control of Noncommunicable Diseases (2013–2020).

The potential benefits of this study extend beyond individual patient outcomes. By reducing procedure time, HRMC could optimize operating room scheduling for gastroenterologists and endoscopy staff. Additionally, the elimination of random biopsies through HRMC will enable pathologists to focus on targeted biopsies, increasing diagnostic accuracy while reducing costs. Furthermore, the collection of endoscopic video sequences and histological samples of early carcinoma and high-grade dysplasia in Barrett’s esophagus will facilitate the development of artificial intelligence projects to improve diagnosis and management of Barrett’s esophagus.

Abbreviations

AE

Adverse event

ANSM

Agence Nationale de Santé du Médicament et des produits de santé

BE

Barrett’s esophagus

BLI

Blue Laser Imaging

CHU

University hospital center

CIC

Clinical Investigation center

CNIL

Commission Nationale de l’Informatique et des Libertés

CRF

Case Report Form

EAC

Esophageal Adenocarcinoma

eCRF

Electronic Case Report Form

ESCC

Esophageal Squamous Cell Carcinoma

ESGE

European Society for Gastrointestinal Endoscopy

GERD

Gastro-Esophageal Reflux Disease

HGD

High Grade Dysplasia

HRMC

High Resolution endoscopy combined with Magnification and electronic Chromoendoscopy

ICER

Incremental cost-effectiveness ratio

LGD

Low Grade Dysplasia

NBI

Narrow Band Imaging

PPV

Positive Predictive Value

SP

Seattle Protocol

SFED

French Society of Digestive Endoscopy

WLE

White Light Endoscopy

Authors’ contributions

EC has the idea of the trial, EC, LQ and AP wrote the study protocol, AP wrote the manuscript. LQ, JBH, YC, VPR, MM, CL, IJA, SB, IA; CR, OP, AJ, GR, AB; DK; CY; MP; TB; IT; MS; JJ; TW; GV; SK; MB; EC assisted with the drafting of the manuscript. MM with the help of CL; CR and OP coordinated the submission of the protocol and the follow-up of (1) the Health Ministry tender and (2) the regulatory authorities and is coordinating the trial. JBH wrote the methodological/statistical analyses in the protocol. VPR and YC wrote the health economic analyses. LG; IA; SB; IJA; CR; OP; GR; AB; DK; CY; MP; TB; IT; MS; JJ; TW; GV; SK; MB and EC participated in patient enrolment and follow-up. AJ are assisting with pharmacovigilance for the trial. All authors read and approved the final manuscript.

Funding

This study was supported by a grant from the French National Cancer Institute (INCa) (programme PHRC-K, year 2020 and registered number: PHRC-K20-209). The protocol underwent independent scientific peer review as part of the funding application process. The funding body had no role in the design of the study, data collection, analysis, interpretation of data, or in the decision to publish.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This clinical study was submitted to and approved by Ile de France VIII Ethics Committee (Comité de Protection des Personnes Ile de France XI N° 2021 A02239-32) on January 9, 2022.

All patients participating in the study are given oral and written information about this trial and sign the informed consent form.

Consent for publication

Not applicable - no identifying images or other personal or clinical details of participants are presented here or will be presented in reports of the trial results. Informed consent materials are attached as supplementary materials.

Competing interests

The authors declare no competing interests.