INTRODUCTION

Barrett's esophagus (BE) is a premalignant condition that can progress from nondysplastic (NDBE) to low-grade dysplasia (LGD), high-grade dysplasia (HGD), and esophageal adenocarcinoma (EAC) (1), which has one of the fastest-growing incidences of cancer in Western countries and a 5-year survival rate of less than 20% (2,3). Societal guidelines recommend endoscopic surveillance of BE because the improvement in EAC survival has been observed with early staged disease and a previous diagnosis of BE (4). Current management decisions are largely based on the diagnostic grade; however, mucosal areas with dysplasia are easily missed because of random endoscopic sampling (5). Furthermore, there is considerable interobserver variability among pathologists in the histologic grading of BE (6). A diagnosis of LGD confirmed by an expert gastrointestinal pathologist is a significant predictor of progression to HGD/EAC, and as such expert review of LGD is recommended by all society guidelines (1,7,8). However, interobserver agreement even among expert gastrointestinal pathologists is poor, which has likely contributed to the variability in published progression rates for confirmed LGD (9,10). In addition, expert review is poorly defined, not widely available, and there are logistical challenges in transferring slides and reports.

The variability in the diagnosis of LGD makes the clinical management of BE difficult for gastroenterologists and their patients (11). There is a need for objective risk stratification tests to reduce the impact of variable pathology review and standardize the clinical management of BE. A tissue systems pathology test (TissueCypher Barrett's Esophagus Test, TSP-9) has been extensively validated in multicenter studies to objectively predict the risk of future progression of BE to HGD/EAC and to detect the presence of prevalent HGD/EAC (12–17). A recent study compared the performance of the TSP-9 test vs benchmarks of generalist and expert pathology review from 5 countries in risk-stratifying patients with a community-based diagnosis of LGD from the SURF trial (18–20). Both the expert and generalist diagnoses demonstrated significant interobserver variability. A subset of pathologists overdiagnosed dysplasia, which can result in unnecessary therapy, whereas another subset underdiagnosed dysplasia, which may result in progression to HGD/EAC during a surveillance interval. The TSP-9 test provided objective risk stratification that outperformed the benchmarks of pathology in identifying progressors, indicating that the test may be an effective tool to guide management decisions for patients with BE (18). The aims of this study were to evaluate whether the use of the test can standardize clinical management decisions for patients with BE and increase the percentage of patients who receive appropriate management in a manner consistent with improved health outcomes.

METHODS Study cohort and design

The screening cohort for the SURveillance vs RadioFrequency ablation (SURF) trial consists of patients with BE with a community-based diagnosis of LGD from 9 BE treatment centers and their referring institutions in Europe (19). Data elements collected were age, sex, segment length, collection date, original diagnosis, follow-up histologic diagnoses, and HGD/EAC-free surveillance time. All patients were on proton pump inhibitors 40 mg once daily or twice daily, titrated on their reflux symptoms. The Institutional Biobank Review Committee of the Academic Medical Center approved the study.

Pathology review

Digital slides of hematoxylin and eosin (H&E) staining (2 slides) and p53 immunohistochemistry (IHC) (1 slide) were made available to study pathologists through the Concentriq for Research platform (Proscia). Slides were independently reviewed by pathologists from 5 countries including 14 expert pathologists (Netherlands n = 3, United States n = 3, Germany n = 3, United Kingdom n = 3, and Belgium n = 2) and 16 generalist anatomic pathologists (Netherlands n = 3, United States n = 3, Germany n = 4, United Kingdom n = 3, and Belgium n = 3). US pathologists reviewed only H&E per standard practice in the United States where guidelines do not recommend use of p53 IHC as a diagnostic aid. Pathologists from Europe reviewed H&E as well as p53 IHC as a diagnostic aid per their standard practices (21). For cases with multiple parts, the result was based on the highest diagnosis.

TissueCypher testing

The TissueCypher Barrett's Esophagus Test (the “TSP-9 test”) was run on each specimen in a blinded manner at Castle Biosciences' Clinical Laboratory Improvement Amendments–certified, College of American Pathologists–accredited laboratory (Pittsburgh, PA) (13). Test results were reported as high/intermediate (int) or low risk of progression to HGD/EAC within 5 years. For cases with multiple parts, the result was based on the highest risk score.

Management decision simulations

Patient management decisions were simulated 500 times with randomly assigned pathology reviewers to determine the most likely care plan with or without the test to guide management. Five hundred iterations were observed to yield stable model results for independent model runs, despite the random sampling of pathologists for each patient within the model. This random selection of pathologists was performed 500 times to construct a distribution of management decisions for each patient. This random walk approach with resampling allowed for estimation of interquartile ranges (IQRs) and use of nonparametric statistics to compare the various management decision methods.

The standard of care (SOC) arm assumed that clinicians followed current guidelines. The models did not account for compliance. For both the SOC and TissueCypher-guided models, each patient was randomly assigned a generalist pathologist for an initial diagnosis, and an expert from the same country was randomly assigned if required as a part of the SOC. A bootstrapping technique was used in which random assignment with replacement was performed for each iteration. The patient journey simulations were assessed to determine the degree to which management decisions varied when different pathologists reviewed the patient slides.

Each simulation was scored based on whether the resulting management decision was appropriate per the known progression/nonprogression outcomes from the SURF trial. Appropriate management for progressors was endoscopic eradication therapy (EET) or short-interval surveillance in less than 1 year, whereas appropriate management for nonprogressors was long-interval surveillance in 3–5 years. EET includes any therapy performed endoscopically to eradicate dysplasia and early EACs. The percentage of iterated incidences in which patients received appropriate management was calculated and compared between the SOC arm vs arms where guidance was provided by the test. A previous study demonstrated that the rate of progression to HGD/EAC was similar in the intermediate- and high-risk classes, and therefore, the TSP-9 test was evaluated as a binary classifier (intermediate/high risk combined vs low risk) (17). Variation in the patient journey from the initial diagnosis to the disease management decision was assessed by assigning an integer value to each management decision (i: long-interval surveillance, ii: short-interval surveillance, or iii: EET), and the SD was calculated for the management decisions for all 500 simulations.

Statistical analyses

Simulations and statistical analyses were run with custom-written software and open-source packages in R version 4.2.1. Median and IQRs were used to describe variables with a skewed distribution. Wilcoxon signed-rank paired test was performed for non-normally distributed continuous variables to compare progressors vs nonprogressors. The Kolmogorov-Smirnov test was used to test for differences in the resulting variability distributions between the models. An α of 0.05 was set for statistical testing.

RESULTS Patient characteristics

One hundred fifty-four patients with BE with community-based LGD and longitudinal outcomes met the inclusion criteria; 24 progressed, and 130 did not progress to HGD/EAC during 5-year follow-up (see Supplementary Table S1, Supplementary Digital Content 3, https://links.lww.com/AJG/C976). Patients were from the screening cohort of the SURF trial and included patients with confirmed LGD who were enrolled in the surveillance arm of the SURF trial (n = 20) as well as patients who were downstaged to either NDBE or indefinite for dysplasia (IND) and followed prospectively (n = 134). Patients had a median age of 61 ± 10 years, and 83.3% of progressors and 78.5% of nonprogressors were men. All patients had long-segment BE with a median length of 5 cm (IQR, 3.0–6.3) and 4 cm (IQR, 3.0–5.5) for progressors and nonprogressors, respectively. There were no significant differences in age, sex, and segment length between progressors and nonprogressors (P > 0.05). Progression to HGD/EAC occurred at a median of 1.7 years (IQR, 0.6–2.5) after the baseline endoscopy. Nonprogressors had a median HGD/EAC-free follow-up of 7.8 years (IQR, 5.8–10.1).

Surveillance and management decision models

To understand the effectiveness of the TSP-9 test in guiding the management of patients with BE, each patient's management was simulated with varying pathology reviewers to determine the most likely care plan with or without guidance from the test results. Figure 1a describes the SOC surveillance and management model based on current guidelines (1). Figure 1b depicts a model where the TSP-9 test results were used adjunctively with the SOC to guide management. Patients who tested low risk were assigned to 3- to 5-year endoscopic surveillance if pathology diagnosis was NDBE/IND or surveillance within 12 months with an expert review diagnosis of LGD. Patients who tested high/intermediate risk were assigned for review by an expert pathologist, then assigned to short-interval surveillance if the review diagnosis was NDBE/IND, or EET if the expert review diagnosis was LGD/higher. Given that the overall predictive accuracy of expert and generalist diagnoses have been shown to be similar, an additional model was run where the test results were used with 1 pathology diagnosis, simulating a scenario where cases with an initial diagnosis of LGD are not referred for secondary review (Figure 1c). Management decisions using TSP-9 results alone were also simulated.

Figure 1.:

Surveillance and management decision models. (a) SOC management arm where patients were managed per current guidelines. (b) TissueCypher (TSP-9)-guided management decision model combining the SOC with guidance from the test. (c) TissueCypher (TSP-9)-guided management decision model where test results were used in conjunction with 1 pathology diagnosis (no secondary review of LGD). LGD, low-grade dysplasia; SOC, standard of care; TSP-9, tissue systems pathology test.

Use of the TSP-9 test in conjunction with the SOC increases the likelihood of patients receiving appropriate management and standardizes management decisions

The simulation results showed that use of the TSP-9 test with SOC pathology review significantly increased the percentage of patients receiving appropriate management from a median of 80.8% (IQR, 64–92) with SOC alone to 100% (IQR, 81–100) (P = 0.0007) when the test results were used to guide management decisions (Figure 2a). The percentage of patients with 100% of simulations resulting in appropriate management significantly increased from 9.1% for pathology alone, to 58.4% when TSP-9 results were used with pathology, and further increased to 77.3% of patients receiving appropriate management when only TSP-9 results were used (see Supplementary Table S2, Supplementary Digital Content 4, https://links.lww.com/AJG/C977). In a subanalysis of patients who progressed to HGD/EAC within 5 years, use of the test increased the percentage of patients receiving appropriate management per their known progression outcome (P = 0.013, Figure 2a) by increasing the percentage of progressors receiving EET from a median of 24.4% (IQR, 2–79) to 46.8% (IQR, 23–88) (P = 0.024) and by decreasing the percentage of progressors receiving of 3- to 5-year surveillance from a median of 33.4% (IQR, 14–74) to 0 (IQR 0–25 P = 0.012, Figure 2b). These results indicate that the use of the test with SOC pathology can increase the detection of progressors and direct more progressors to treatment with EET to prevent progression. In the subset of nonprogressors, use of the test also increased the percentage of patients receiving appropriate management per their known nonprogression outcome (P = 0.0081, Figure 2a). Use of the test increased the percentage of nonprogressors receiving a management plan of 3- to 5-year surveillance from a median of 81.7% (IQR, 69–92) to 100% (IQR, 85–100) (P = 0.0081) and reduced the percentage of nonprogressors receiving a decision for short-interval surveillance (Figure 2c). These results indicate that use of the test can help to avoid overuse of EET and surveillance in patients with BE. Use of the test to guide management decisions also significantly increased the consistency of decisions when different pathologists reviewed the patients' slides (P < 0.0001, Figure 2d). When the SOC alone was used, only 7.1% of patients had no deviation in management decisions when different pathologists reviewed their slides. By contrast, when the test was used with the SOC, 57.1% of patients had no deviation in management decisions, despite their slides being reviewed by different pathologists (Figure 2d).

Figure 2.:

Use of the TSP-9 test in conjunction with the SOC increased the likelihood of patients receiving appropriate management. (a) Likelihood of receiving appropriate management using SOC vs SOC with guidance from the TSP-9 test. (b) Percentage of simulated management decisions for surveillance or therapy in progressors (n = 24). (c) Percentage of simulated management decisions for surveillance or therapy in nonprogressors (n = 130). (d) Variability in management decisions for patients managed per SOC vs SOC with guidance from the TSP-9 test. See Methods for definition of appropriate management. SOC, standard of care; TSP-9, tissue systems pathology test.

Use of the TSP-9 test with a single pathology reviewer also increases the likelihood of appropriate management decisions in patients with BE

When the results from the TSP-9 test were used with 1 pathology review (generalist or expert) as shown in Figure 1c, there was a similar increase in the percentage of patients receiving appropriate management from a median of 80.8% (IQR, 64–92) for the SOC alone to 93.6% (IQR, 66–100) when the test results and a single pathology diagnosis were used to guide management (P = 0.018, Figure 3a). A statistically significant increase in appropriate management was observed for progressors (median 66.6%, IQR 26–86 vs median 100%, IQR 86–100, P = 0.0046, Figure 3b), and although an increase in appropriate management was observed for nonprogressors, the difference was not statistically significant (P = 0.082). In the subset of patients who progressed, the percentage of patients receiving a management decision of 3- to 5-year surveillance was significantly reduced from a median of 33.4% (IQR, 14–74) to zero (IQR, 0–14) (P = 0.0058) for progressors when the test scores were used with just 1 pathology diagnosis, indicating that most progressors were managed by EET or short-interval surveillance when care was guided by the test. There was also a significant increase in the consistency of management decisions when the test results were used with 1 pathology review compared with the SOC (P < 0.0001, Figure 3d), indicating that the test can reduce the impact of variable pathology review on management decisions for patients with BE. Similar results were obtained when the pathology review was provided by only generalist pathologists or only expert pathologists (see Supplementary Figures S1 and S2, Supplementary Digital Content 1 and 2, https://links.lww.com/AJG/C974 and https://links.lww.com/AJG/C975).

Figure 3.:

Use of the TSP-9 test in conjunction with 1 pathology diagnosis increased the likelihood of appropriate management decisions. (a) Likelihood of receiving appropriate management using SOC vs a single pathology review with guidance from the TSP-9 test. (b) Percentage of simulated management decisions for surveillance or therapy in progressors (n = 24). (c) Percentage of simulated management decisions for surveillance or therapy in nonprogressors (n = 130). (d) Variability in management decisions for patients managed per SOC vs the TSP-9 test in conjunction with a single pathology review. SOC, standard of care; TSP-9, tissue systems pathology test.

DISCUSSION

This study evaluated the ability of a tissue systems pathology (TSP-9) test to improve and standardize clinical management in a manner that will improve health outcomes for patients with BE. The study results showed that patients who progressed to HGD/EAC were statistically more likely to receive EET or short-interval surveillance based on a high-risk result by the test and less likely to be placed on long-interval surveillance when the test was used with the SOC to guide management decisions. Patients who did not progress were more likely to be placed on long-interval surveillance when the test results were used to guide management decisions. Similar overall clinical utility was observed when TSP-9 results were used with pathology review, with or without expert review of LGD. Optimal overall clinical utility was obtained when TSP-9 results were used independently, indicating the test can guide management decisions for patients with BE across a variety of practice settings. Use of the test also increased the consistency of management decisions by reducing the impact of variable pathology review across healthcare settings. Taken together, these results indicate that the test has clinical utility to improve health outcomes in 2 ways by targeting early therapeutic interventions and close surveillance of the at-risk subset of patients who progress to HGD/EAC but can be missed by the current SOC and by reducing unnecessary endoscopies and procedures in patients with BE who will not progress to HGD/EAC.

Clinical guidelines recommend endoscopic surveillance in patients with BE, but proper implementation of the recommendations remains challenging. Real-world practice, including the performance of surveillance, surveillance interval, and biopsy protocol, is frequently not aligned with society guidelines (22–24). Overall poor adherence to the guidelines is of concern because shorter surveillance intervals and unnecessary treatment in low-risk patients add significant costs and put low-risk patients at risk of adverse events, which can negatively impact the quality of life and health outcomes. Long-interval surveillance can lead to missed dysplastic and cancerous lesions, which can lead to poor health outcomes (25). In addition, diagnostic grading of BE can be challenging because of random sampling and high interobserver variability. Variable pathology review can lead to significant variability in management decisions in different settings, resulting in care plans and health outcomes ranging from poor to excellent depending on which pathologist reviewed the slides. Objective risk stratification tools are needed to identify patients harboring prevalent HGD/EAC or who will progress to HGD/EAC in the future. These patients will benefit from EET to prevent progression or close surveillance to detect EAC at the earliest possible stage. The current study demonstrated that use of the TSP-9 test to guide management decisions can increase the percentage of progressors who are managed by EET or short-interval surveillance, which is due to the test performing with higher sensitivity in detecting progressors than benchmarks of pathology review, as shown previously (18). EET has been shown to be safe and highly effective in treating dysplasia and early stage EAC and in preventing future progression to HGD/EAC, leading to improved health outcomes (26). Short-interval surveillance can also be an effective management approach for high-risk patients because surveillance can detect EAC at earlier stages, which can also improve outcomes. These findings indicate that use of the test can direct more progressors to upstaged care, leading to improved health outcomes. Risk stratification tools are also critical to identify patients at low risk of progression because these patients can avoid unnecessary treatments and can be effectively managed by long-interval surveillance. This study showed that use of the TSP-9 test to guide management decisions increased the percentage of nonprogressors who are managed by 3- to 5-year surveillance with a corresponding decrease in nonprogressors receiving surveillance within 1 year. This finding can also be attributed to the higher predictive accuracy of the test compared with benchmarks of pathology review as demonstrated previously (18). This increase in appropriate management decisions has the potential to improve the quality of life for most patients with BE who are at low risk of progression to HGD/EAC.

The additional key finding of this study was that use of the TSP-9 test increases the consistency of management decisions when different pathologists review the slides. Use of the test to guide decision-making could reduce the impact of variable pathology review, enabling patients with BE to get consistent care, regardless of which pathologist reviewed their specimen slides. The test is automated, objective, and can be an effective solution to standardize BE management, enabling more patients to get care plans that are appropriate for their individual risk regardless of practice setting. Although improved management for both progressors and nonprogressors was achieved using the TSP-9 test with standard pathology review practices, similar results were obtained when only a single pathologist reviewed the slides and cases with LGD were not referred for expert review. This finding indicates that the test has utility in clinical settings where access to and logistics of sending cases out for review are challenging.

The main strength of this study lies in modeling management decisions using results from a study that compared the predictive performance of the TSP-9 test vs benchmarks of generalist and expert pathology from 5 countries in the screening cohort of a randomized controlled trial (19,20). This cohort included patients who progressed to HGD/EAC and patients who did not progress during long-term follow-up. The study diagnoses were provided by 16 generalist and 14 expert pathologists, enabling modeling of patient journeys across diverse practice settings. Although all patients had an initial community-based diagnosis of LGD per the inclusion criteria of the SURF trial screening cohort, 81% (mean of 30 pathology reviews) were downstaged to NDBE or IND. This study modeled various ways to use the test clinically to aid physicians in recommending appropriate management plans for patients with BE. Because of the original diagnosis of LGD from a community-based generalist pathologist, some of the simulations presented here patient cases were reviewed again by generalist pathologist, which exceeds standard practice. However, this enriched cohort provided sufficient progression events to evaluate the utility of the TSP-9 test in both progressors and nonprogressors, and the review by multiple pathology reviewers enabled simulation of real-world observer variability. Although this study used a simulation-based approach, the assumptions were derived from real-world data on the predictive performance of pathology diagnosis and the TSP-9 test in patients with known outcomes. In addition, risk factors such as obesity, duration of acid reflux, reflux esophagitis, and family history were not available to include in the decision simulations, and the simulations assumed that physicians and patients adhered to the specified management decisions in each arm, and such adherence can be challenging in practice.

In conclusion, objective risk stratification provided by the TSP-9 test significantly improved the likelihood of appropriate management for patients with BE vs pathology review alone and also increased the management decision consistency across practice settings. The test may be an effective solution to standardize management decisions and improve health outcomes for both patients at high risk and for patients at low risk of progression to HGD/EAC. Using the test with the SOC can shift the current clinical practice paradigm of surveillance and treatment based on the subjective diagnosis to objective risk stratification, with therapeutic interventions and close surveillance targeted at patients who are at high risk of progression and surveillance-only approaches for low-risk patients.

CONFLICTS OF INTEREST

Guarantor of the article: Rebecca J. Critchley-Thorne, PhD.

Specific author contributions: All authors have approved the final draft submitted. Conceptualization: J.J.G.H.M.B. and R.J.C.-T. Investigation: J.J.G.H.M.B., R.J.C.-T., L.C.D., N.F.F., and R.E.P. Methodology: J.J.G.H.M.B., R.J.C.-T., L.C.D., N.F.F., R.E.P., and A.M.K. Data curation and analysis: J.J.S., R.J.C.-T., L.C.D., N.F.F., A.M.K., and C.S. Writing—original draft: M.A. Writing—review and editing: R.J.C.-T., M.A., J.J.G.H.M.B., L.C.D., N.F.F., R.E.P., A.M.K., J.J.S., and C.S. Funding acquisition: R.J.C.-T. Resources: R.J.C.-T. Project administration: R.J.C.-T. Supervision: J.J.G.H.M.B., R.J.C.-T., and L.C.D.

Financial support: This study was funded by Castle Biosciences.

Potential competing interests: R.J.C.-T. is a full-time employee of, and holds stock and stock options in Castle Biosciences, and is an inventor on patents on the TissueCypher Barrett's Esophagus Assay. C.S., M.A., and J.J.S. are full-time employees of and hold stock and stock options in Castle Biosciences. J.J.G.H.M. Bergman has received research funding from Castle Biosciences, CDx Diagnostics, and Lucid Diagnostics. A.M.K., N.F.F., R.E.P., and L.C.D. disclose no conflicts.

Study Highlights

WHAT IS KNOWN ✓ Guidelines recommend endoscopic surveillance with biopsies to enable early detection and treatment of esophageal adenocarcinoma in patients with Barrett’s esophagus (BE). ✓ The clinical management of BE is challenging due to substantial interobserver variation in the histologic evaluation and lack of standardization of care plans.

WHAT IS NEW HERE ✓ Use of a tissue systems pathology test (TissueCypher Barrett’s Esophagus Test, TSP-9) to guide management decisions significantly increases the likelihood of receiving appropriate management in patients with BE. ✓ Use of the TSP-9 test also increased the consistency of management decisions for BE patients by reducing the impact of variable pathology review across practice settings. ✓ The TSP-9 test demonstrates significant clinical utility to improve health outcomes by providing objective risk stratification to target early therapeutic intervention or close surveillance to patients who progress to HGD/EAC, and by reducing unnecessary endoscopies in BE patients who will not progress. ACKNOWLEDGMENTS

TissueCypher SURF LGD Study Pathologist Consortium: USA: John Goldblum (Cleveland Clinic, Cleveland, Ohio), Elizabeth Montgomery (University of Miami, Miami, Florida), Jon Davison (University of Pittsburgh, Pittsburgh, Pennsylvania), Jagjit Singh (University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania), Jared Szymanski (Connect Pathology, Lehi, Utah), and Anthony Perry (Connect Pathology, Lehi, Utah). Netherlands: Kees Seldenrijk (St Antonius Hospital, Nieuwegein), Fiebo ten Kate (University Medical Center Utrecht, Utrecht), G. Johan A. Offerhaus (University Medical Center Utrecht, Utrecht), Paul Drillenberg (Onze Lieve Vrouwe Gasthuis, Amsterdam), Casper Jansen (Laboratorium Pathologie Oost Nederland, Hengelo), and Natalja Leeuwis-Fedorovic (Deventer Ziekenhuis, Deventer). United Kingdom: Runjan Chetty (Deciphex Ltd, Mornington, Ireland), Roger Feakins (Royal Free Hospital, London), Marnix Jansen (University College London, London), Catherine Chinyama (Princess Elizabeth Hospital, Guernsey), Edwin Cooper (Yeovil District Hospital, Somerset), and Reza Vaziri (Worcestershire Acute Hospitals, Worcester). Germany: Gustavo Baretton (University Hospital Carl Gustav, Dresden), Andrea Tannapfel (Ruhr-Universität Bochum, Bochum), Michael Vieth (Friedrich-Alexander-Universität, Erlangen), Balint Melcher (Institute of Pathology, Koblenz), Ildiko Mesteri (Institute of Pathology, Überlingen), Heiko Müller (Institute of Pathology, Überlingen), and Philipp Wetzel (Institute of Pathology, Überlingen). Belgium: Gert de Hertogh (University Hospitals Leuven, Leuven), Anne Hoorens (UZ Gent, Gent), Stepanie Verschuere (AZ Delta, Roeselare), An Tamsin (AZ Delta, Roeselare), Kevin Wetzels (AZ Sint-Blasius, Dendermonde), and Marie-Astrid van Caillie (AZ Sint-Lucas, Ghent).

REFERENCES 1. Shaheen NJ, Falk GW, Iyer PG, et al. Diagnosis and management of Barrett's esophagus: An updated ACG guideline. Am J Gastroenterol 2022;117:559–87. 2. Thrift AP, Whiteman DC. The incidence of esophageal adenocarcinoma continues to rise: Analysis of period and birth cohort effects on recent trends. Ann Oncol 2012;23:3155–62. 3. Hur C, Miller M, Kong CY, et al. Trends in esophageal adenocarcinoma incidence and mortality. Cancer 2013;119:1149–58. 4. Codipilly DC, Chandar AK, Singh S, et al. The effect of endoscopic surveillance in patients with Barrett's esophagus: A systematic review and meta-analysis. Gastroenterology 2018;154:2068–86.e5. 5. Abrams JA, Kapel RC, Lindberg GM, et al. Adherence to biopsy guidelines for Barrett's esophagus surveillance in the community setting in the United States. Clin Gastroenterol Hepatol 2009;7:736–42; quiz 710. 6. Vennalaganti P, Kanakadandi V, Goldblum JR, et al. Discordance among pathologists in the United States and Europe in diagnosis of low-grade dysplasia for patients with Barrett's esophagus. Gastroenterology 2017;152:564–70.e4. 7. Wani S, Rubenstein JH, Vieth M, et al. Diagnosis and management of low-grade dysplasia in Barrett's esophagus: Expert review from the Clinical Practice Updates Committee of the American Gastroenterological Association. Gastroenterology 2016;151:822–35. 8. ASGE Standards of Practice Committee, Qumseya B, Sultan S, et al. ASGE guideline on screening and surveillance of Barrett's esophagus. Gastrointest Endosc 2019;90:335–59.e2. 9. Singh S, Manickam P, Amin AV, et al. Incidence of esophageal adenocarcinoma in Barrett's esophagus with low-grade dysplasia: A systematic review and meta-analysis. Gastrointest Endosc 2014;79:897–909.e4; quiz 983.e1, 983.e3. 10. Duits LC, van der Wel MJ, Cotton CC, et al. Patients with Barrett's esophagus and confirmed persistent low-grade dysplasia are at increased risk for progression to neoplasia. Gastroenterology 2017;152:993–1001.e1. 11. Katzka DA, Falk GW. Management of low-grade dysplasia in Barrett's esophagus: Incremental progress continues. Gastroenterology 2017;152:928–32. 12. Critchley-Thorne RJ, Duits LC, Prichard JW, et al. A tissue systems pathology assay for high-risk Barrett's esophagus. Cancer Epidemiol Biomarkers Prev 2016;25:958–68. 13. Critchley-Thorne RJ, Davison JM, Prichard JW, et al. A tissue systems pathology test detects abnormalities associated with prevalent high-grade dysplasia and esophageal cancer in Barrett's esophagus. Cancer Epidemiol Biomarkers Prev 2017;26:240–8. 14. Prichard JW, Davison JM, Campbell BB, et al. TissueCypher: A systems biology approach to anatomic pathology. J Pathol Inform 2015;6:48. 15. Davison JM, Goldblum J, Grewal US, et al. Independent blinded validation of a tissue systems pathology test to predict progression in patients with Barrett's esophagus. Am J Gastroenterol 2020;115:843–52. 16. Frei NF, Konte K, Bossart EA, et al. Independent validation of a tissue systems pathology assay to predict future progression in nondysplastic Barrett's esophagus: A spatial-temporal analysis. Clin Transl Gastroenterol 2020;11:e00244. 17. Frei NF, Khoshiwal AM, Konte K, et al. Tissue systems pathology test objectively risk stratifies Barrett's esophagus patients with low-grade dysplasia. Am J Gastroenterol 2021;116:675–82. 18. Khoshiwal AM, Frei NF, Duits LC, et al. 695: An objective, fully automated Barrett's risk prediction assay outperforms most pathologists in risk stratifying Barrett's esophagus with low grade dysplasia. Gastroenterology 2022;162:S-174. 19. Phoa KN, van Vilsteren FGI, Weusten BLAM, et al. Radiofrequency ablation vs endoscopic surveillance for patients with Barrett esophagus and low-grade dysplasia: A randomized clinical trial. JAMA 2014;311:1209–17. 20. Pouw RE, Klaver E, Phoa KN, et al. Radiofrequency ablation for low-grade dysplasia in Barrett's esophagus: Long-term outcome of a randomized trial. Gastrointest Endosc 2020;92:569–74. 21. Schlemper RJ, Riddell RH, Kato Y, et al. The Vienna classification of gastrointestinal epithelial neoplasia. Gut 2000;47:251–5. 22. Wani S, Williams JL, Komanduri S, et al. Over-utilization of repeat upper endoscopy in patients with non-dysplastic Barrett's esophagus: A quality registry study. Am J Gastroenterol 2019;114:1256–64. 23. Wani S, Williams JL, Komanduri S, et al. Endoscopists systematically undersample patients with long-segment Barrett's esophagus: An analysis of biopsy sampling practices from a quality improvement registry. Gastrointest Endosc 2019;90:732–41.e3. 24. Cruz JD, Paculdo D, Ganesan D, et al. Clinical variation in surveillance and management of Barrett's esophagus: A cross-sectional study of gastroenterologists and gastrointestinal surgeons. Medicine (Baltimore) 2022;101:e32187. 25. Duits LC, Klaver E, Bureo Gonzalez A, et al. The Amsterdam ReBus progressor cohort: Identification of 165 Barrett's surveillance patients who progressed to early neoplasia and 723 nonprogressor patients. Dis Esophagus 2019;32:doy037. 26. Wani S, Puli SR, Shaheen NJ, et al. Esophageal adenocarcinoma in Barrett's esophagus after endoscopic ablative therapy: A meta-analysis and systematic review. Am J Gastroenterol 2009;104:502–13.