The Evolving Role of Artificial Intelligence in Gastrointestinal Histopathology: An Update.

Significant advances in artificial intelligence (AI) over the past decade potentially may lead to dramatic effects on clinical practice. Digitized histology represents an area ripe for AI implementation. We describe several current needs within the world of gastrointestinal histopathology, and outline, using currently studied models, how AI potentially can address them. We also highlight pitfalls as AI makes inroads into clinical practice.

Rates of Recurrent Intestinal Metaplasia and Dysplasia After Successful Endoscopic Therapy of Barrett's Neoplasia by Endoscopic Mucosal Resection vs Endoscopic Submucosal Dissection and Ablation: A Large North American Multicenter Cohort.

INTRODUCTION: Endoscopic eradication therapy (EET) combining endoscopic resection (ER) with endoscopic mucosal resection (EMR) or endoscopic submucosal dissection (ESD) followed by ablation is the standard of care for the treatment of dysplastic Barrett's esophagus (BE). We have previously shown comparable rates of complete remission of intestinal metaplasia (CRIM) with both approaches. However, data comparing recurrence after CRIM are lacking. We compared rates of recurrence after CRIM with both techniques in a multicenter cohort. METHODS: Patients undergoing EET achieving CRIM at 3 academic institutions were included. Demographic and clinical data were abstracted. Outcomes included rates and predictors of any BE and dysplastic BE recurrence in the 2 groups. Cox-proportional hazards models and inverse probability treatment weighting (IPTW) analysis were used for analysis. RESULTS: A total of 621 patients (514 EMR and 107 ESD) achieving CRIM were included in the recurrence analysis. The incidence of any BE (15.7, 5.7 per 100 patient-years) and dysplastic BE recurrence (7.3, 5.3 per 100 patient-years) were comparable in the EMR and ESD groups, respectively. On multivariable analyses, the chances of BE recurrence were not influenced by ER technique (hazard ratio 0.87; 95% confidence interval 0.51-1.49; P = 0.62), which was also confirmed by IPTW analysis (ESD vs EMR: hazard ratio 0.98; 95% confidence interval 0.56-1.73; P = 0.94). BE length, lesion size, and history of cigarette smoking were independent predictors of BE recurrence. DISCUSSION: Patients with BE dysplasia/neoplasia achieving CRIM, initially treated with EMR/ablation, had comparable recurrence rates to ESD/ablation. Randomized trials are needed to confirm these outcomes between the 2 ER techniques.

Recognition of Refractory Benign Esophageal Strictures at Index Endoscopy: Creation of a Predictive Model.

BACKGROUND AND AIMS: Refractory benign esophageal strictures (RBES) are defined by inadequate response to dilation. Early recognition of RBES allows for earlier initiation of aggressive therapy potentially leading to less morbidity and cost. We sought to establish a predictive model for RBES. METHODS: Patients who underwent esophagogastroduodenoscopy (EGD) with esophageal dilation at Mayo Clinic Rochester were identified. In addition, a cohort of patients from a clinical database of patients with RBES managed with self-dilation was identified. Malignant strictures, Schatzki rings, and previously treated strictures were excluded. RBES was defined by inability to maintain a diameter ≥14mm over 5 dilation sessions. Multivariable logistic regression models were built to predict RBES. RESULTS: 128 patients with index EGD and esophageal dilation were identified, with 25 meeting RBES criteria. An additional 63 RBES patients were identified from the self-dilation cohort for a total of 88 RBES and 103 non-RBES patients. Multivariable analysis yielded a strong predictive model, with a c-statistic of 0.85, identifying stricture length ≥ 2cm, diameter ≤ 7mm and proximal/diffuse stricture location as associated with a higher risk for RBES. Patients without any of these risk factors had a 2% risk of RBES while those with all 3 risk factors had a risk of 73% for RBES. CONCLUSIONS: Risk of RBES can be predicted at index EGD based on stricture features. A predictive model for RBES was created based on readily available risk factors, which may guide an individualized therapeutic approach to patients with benign esophageal stricture, potentially reducing morbidity and cost.

Population Based Time Trends in the Epidemiology and Mortality of Gastroesophageal Junction and Esophageal Adenocarcinoma.

BACKGROUND: Limited data are available on the epidemiology of gastroesophageal junction adenocarcinoma (GEJAC), particularly in comparison to esophageal adenocarcinoma (EAC). With the advent of molecular non-endoscopic Barrett's esophagus (BE) detection tests which sample the esophagus and gastroesophageal junction, early detection of EAC and GEJAC has become a possibility and their epidemiology has gained importance. AIMS: We sought to evaluate time trends in the epidemiology and survival of patients with EAC and GEJAC in a population-based cohort. METHODS: EAC and GEJAC patients from 1976 to 2019 were identified using ICD 9 and 10 diagnostic codes from the Rochester Epidemiology Project (REP). Clinical data and survival status were abstracted. Poisson regression was used to calculate incidence rate ratios (IRR). Survival analysis and Cox proportional models were used to assess predictors of survival. RESULTS: We included 443 patients (287 EAC,156 GEJAC). The incidence of EAC and GEJAC during 1976-2019 was 1.40 (CI 1.1-1.74) and 0.83 (CI 0.61-1.11) per 100,000 people, respectively. There was an increase in the incidence of EAC (IRR = 2.45, p = 0.011) and GEJAC (IRR = 3.17, p = 0.08) from 2000 to 2004 compared to 1995-1999, plateauing in later time periods. Most patients had associated BE and presented at advanced stages, leading to high 5-year mortality rates (66% in EAC and 59% in GEJAC). Age and stage at diagnosis were predictors of mortality. CONCLUSION: The rising incidence of EAC/GEJAC appears to have plateaued somewhat in the last decade. However, both cancers present at advanced stages with persistently poor survival, underscoring the need for early detection.

Prediction of Progression in Barrett's Esophagus Using a Tissue Systems Pathology Test: A Pooled Analysis of International Multicenter Studies.

BACKGROUND & AIMS: Prediction of progression risk in Barrett's esophagus (BE) may enable personalized management. We aimed to assess the adjunct value of a tissue systems pathology test (TissueCypher) performed on paraffin-embedded biopsy tissue, when added to expert pathology review in predicting incident progression, pooling individual patient-level data from multiple international studies METHODS: Demographics, clinical features, the TissueCypher risk class/score, and progression status were analyzed. Conditional logistical regression analysis was used to develop multivariable models predicting incident progression with and without the TissueCypher risk class (low, intermediate, high). Concordance (c-) statistics were calculated and compared with likelihood ratio tests to assess predictive ability of models. A risk prediction calculator integrating clinical variables and TissueCypher risk class was also developed. RESULTS: Data from 552 patients with baseline no (n = 472), indefinite (n = 32), or low-grade dysplasia (n = 48) (comprising 152 incident progressors and 400 non-progressors) were analyzed. A high-risk test class independently predicted increased risk of progression to high-grade dysplasia/adenocarcinoma (odds ratio, 6.0; 95% confidence interval, 2.9-12.0), along with expert confirmed low-grade dysplasia (odds ratio, 2.9; 95% confidence interval, 1.2-7.2). Model prediction of progression with the TissueCypher risk class incorporated was significantly superior than without, in the whole cohort (c-statistic 0.75 vs 0.68; P < .0001) and the nondysplastic BE subset (c-statistic 0.72 vs 0.63; P < .0001). Sensitivity and specificity of the high risk TissueCypher class were 38% and 94%, respectively. CONCLUSIONS: An objective tissue systems pathology test high-risk class is a strong independent predictor of incident progression in patients with BE, substantially improving progression risk prediction over clinical variables alone. Although test specificity was high, sensitivity was modest.

Comparative Outcomes of Cap Assisted Endoscopic Resection and Endoscopic Submucosal Dissection in Dysplastic Barrett's Esophagus.

BACKGROUND & AIMS: Endoscopic resection is an important component of the endoscopic treatment of Barrett's esophagus (BE) with dysplasia and intramucosal adenocarcinoma. Endoscopic resection can be performed by cap-assisted endoscopic mucosal resection (cEMR) or endoscopic submucosal dissection (ESD). We compared the histologic outcomes of ESD vs cEMR, followed by ablation. METHODS: We queried a prospectively maintained database of all patients undergoing cEMR and ESD followed by ablation at our institution from January 2006 to March 2020 and abstracted relevant demographic and clinical data. Our primary outcomes included the rate of complete remission of dysplasia (CRD): absence of dysplasia on surveillance histology, and complete remission of intestinal metaplasia (CRIM): absence of intestinal metaplasia. Our secondary outcome included complication rates. RESULTS: We included 537 patients in the study: 456 underwent cEMR and 81 underwent ESD. The cumulative probabilities of CRD at 2 years were 75.8% and 85.6% in the cEMR and ESD groups, respectively (P < .01). Independent predictors of CRD were as follows: ESD (hazard ratio [HR], 2.38; P < .01) and shorter BE segment length (HR, 1.11; P < .01). The cumulative probabilities of CRIM at 2 years were 59.3% and 50.6% in the cEMR and ESD groups, respectively (P > .05). The only independent predictor of CRIM was a shorter BE segment (HR, 1.16; P < .01). CONCLUSIONS: BE patients with dysplasia or intramucosal adenocarcinoma undergoing ESD reach CRD at higher rates than those treated with cEMR, although CRIM rates at 2 years and complication rates were similar between the 2 groups.

Development of a deep learning model for the histologic diagnosis of dysplasia in Barrett's esophagus.

BACKGROUND AND AIMS: The risk of progression in Barrett's esophagus (BE) increases with development of dysplasia. There is a critical need to improve the diagnosis of BE dysplasia, given substantial interobserver disagreement among expert pathologists and overdiagnosis of dysplasia by community pathologists. We developed a deep learning model to predict dysplasia grade on whole-slide imaging. METHODS: We digitized nondysplastic BE (NDBE), low-grade dysplasia (LGD), and high-grade dysplasia (HGD) histology slides. Two expert pathologists confirmed all histology and digitally annotated areas of dysplasia. Training, validation, and test sets were created (by a random 70/20/10 split). We used an ensemble approach combining a "you only look once" model to identify regions of interest and histology class (NDBE, LGD, or HGD) followed by a ResNet101 model pretrained on ImageNet applied to the regions of interest. Diagnostic performance was determined for the whole slide. RESULTS: We included slides from 542 patients (164 NDBE, 226 LGD, and 152 HGD) yielding 8596 bounding boxes in the training set, 1946 bounding boxes in the validation set, and 840 boxes in the test set. When the ensemble model was used, sensitivity and specificity for LGD was 81.3% and 100%, respectively, and >90% for NDBE and HGD. The overall positive predictive value and sensitivity metric (calculated as F1 score) was .91 for NDBE, .90 for LGD, and 1.0 for HGD. CONCLUSIONS: We successfully trained and validated a deep learning model to accurately identify dysplasia on whole-slide images. This model can potentially help improve the histologic diagnosis of BE dysplasia and the appropriate application of endoscopic therapy.

Comparative outcomes of radiofrequency ablation and cryoballoon ablation in dysplastic Barrett's esophagus: a propensity score-matched cohort study.

BACKGROUND AND AIMS: Strong evidence supports the use of radiofrequency ablation (RFA) in the management of dysplastic/neoplastic Barrett's esophagus (BE). Recently, the efficacy of the cryoballoon ablation (CBA) system was demonstrated in multicenter cohort studies. We aimed to assess the comparative effectiveness and safety of these 2 ablation modalities for endoscopic eradication therapy (EET) in a cohort study. METHODS: Data were abstracted on patients with dysplastic BE or intramucosal carcinoma undergoing EET using RFA or CBA as the primary ablation modality at 2 referral centers. The primary outcome was the rate of complete remission intestinal metaplasia (CRIM). Secondary outcomes were rates of complete remission of dysplasia (CRD) and adverse events. Cox proportional hazards models and propensity scored-matched analyses were conducted to compare outcomes. RESULTS: Three hundred eleven patients (CBA, 85 patients; RFA, 226 patients) with a median follow-up of 1.5 years (interquartile range, .8, 2.5) in the RFA group and 2.0 years (interquartile range, 1.3, 2.5) in the CBA group were studied. On multivariable analyses, the chances of reaching CRD and CRIM were not influenced by ablation modality. Propensity score-matched analysis revealed a comparable chance of achieving CRIM (CBA vs RFA: hazard ratio, 1.24; 95% confidence interval, .79-1.96; P = .35) and CRD (CBA vs RFA: hazard ratio, 1.19; 95% confidence interval, .82-1.73; P = .36). The CBA group had a higher stricture rate compared with the RFA group (10.4% vs 4.4%, P = .04). CONCLUSIONS: Histologic outcomes of EET using CBA and RFA for dysplastic BE appear to be comparable. A randomized trial is needed to definitively compare outcomes between these 2 modalities.

Wide-area transepithelial sampling for dysplasia detection in Barrett's esophagus: a systematic review and meta-analysis.

BACKGROUND AND AIMS: Seattle protocol forceps biopsy sampling (FB) is currently recommended for surveillance in Barrett's esophagus (BE) but limited by sampling error and lack of compliance. Wide-area transepithelial sampling with 3-dimensional analysis (WATS3D; CDx Diagnostics, Suffern, NY, USA) is reported to increase BE dysplasia detection. We assessed the incremental yield and clinical significance of WATS3D for dysplasia detection over FB in a systematic review and meta-analysis. METHODS: We queried major scientific databases for studies using WATS3D and FB from 2000 to 2020. The primary outcome was the incremental yield of WATS3D-detected dysplasia (defined as a composite of indefinite for dysplasia, low- and high-grade dysplasia [HGD] and esophageal adenocarcinoma [EAC]) over FB. Secondary outcomes were incremental yields of HGD/EAC and rate of reconfirmation of WATS3D dysplasia on subsequent FB. RESULTS: Meta-analysis of 7 eligible studies demonstrated that FB diagnosed dysplasia in 15.9% of cases, whereas the incremental yield with WATS3D was 7.2% (95% confidence interval, 3.9%-11.5%; I2= 92.1%). Meta-analysis of 6 studies demonstrated that FB diagnosed HGD/EAC in 2.3% of patients, whereas the incremental yield with WATS3D was 2.1% (95% confidence interval, .4%-5.3%; I2= 92.7%). Notably, WATS3D was negative in 62.5% of cases where FB identified dysplasia. Two studies reported reconfirmation of WATS3D dysplasia with FB histology in only 20 patients. CONCLUSIONS: WATS3D increases dysplasia detection; however, the clinical significance of this increased dysplasia detection remains uncertain. Data from endoscopic follow-up to ascertain FB histology in patients with dysplasia based solely on WATS3D are needed to determine the optimal clinical application and significance of WATS3D-only dysplasia.

Neoplasia Detection Rate in Barrett's Esophagus and Its Impact on Missed Dysplasia: Results from a Large Population-Based Database.

BACKGROUND & AIMS: It is a challenge to detect dysplasia in Barrett's esophagus (BE) and esophageal adenocarcinomas (EACs) are missed in 25%-33% of cases. The neoplasia detection rate (NDR), defined as the rate of high-grade dysplasia (HGD) or EAC detection during initial surveillance endoscopy, has been proposed as a quality metric for endoscopic evaluation of patients with BE. However, current estimates are from referral center cohorts, which might overestimate NDR. Effects on rates of missed dysplasia are also unknown. We analyzed data from a large cohort of patients with BE to estimate the NDR and factors associated with it, and assess the effects of the NDR on the rate of missed dysplasia. METHODS: We analyzed data from 1066 patients in the Rochester Epidemiology Project-linked medical record system, a population-based cohort of patients with BE (confirmed by review of the endoscopic and histologic reports) from 11 southeastern Minnesota counties from 1991 through 2019. Biopsies reported to contain dysplasia were confirmed by expert gastrointestinal pathologists. The NDR was calculated as the rate of HGD or EAC detected by histologic analyses of biopsies collected during the first surveillance endoscopy. Patients without HGD or EAC at their initial endoscopy (n = 391) underwent repeat endoscopy within 12 months; HGD or EAC detected at the repeat endoscopy were considered to be missed on index endoscopy. Factors associated with NDR and missed dysplasia were identified using univariate and multivariate logistic regression models. RESULTS: The NDR was 4.9% (95% CI, 3.8-6.4); 3.1% of patients had HGD, 1.8% had EAC, and 10.6% had low-grade dysplasia. Factors associated with higher rates of detection of neoplasia included older age, male sex, smoking, increasing length of BE, and surveillance endoscopies by gastroenterologists. This NDR was associated with a substantially lower rate of missed dysplasia (13%). CONCLUSIONS: In an analysis of 1066 patients with BE in a population-based cohort, we found a lower NDR and lower rate of missed dysplasia than previously reported. NDR may have value as a quality metric in BE surveillance if validated in other cohorts.

Esophageal Lichen Planus Is Associated With a Significant Increase in Risk of Squamous Cell Carcinoma.

Esophageal lichen planus (ELP) is a chronic inflammatory process characterized by apoptotic activity of cytotoxic CD8+ T cells on basal keratinocytes as well as regulatory and helper T cells.1-3 Due to chronic inflammation, malignant transformation may be a concern and has been documented in case reports.4,5 We describe the occurrence of esophageal cancer in ELP patients from a large tertiary patient population.

Development of Electronic Health Record-Based Machine Learning Models to Predict Barrett's Esophagus and Esophageal Adenocarcinoma Risk.

INTRODUCTION: Screening for Barrett's esophagus (BE) is suggested in those with risk factors, but remains underutilized. BE/esophageal adenocarcinoma (EAC) risk prediction tools integrating multiple risk factors have been described. However, accuracy remains modest (area under the receiver-operating curve [AUROC] ≤0.7), and clinical implementation has been challenging. We aimed to develop machine learning (ML) BE/EAC risk prediction models from an electronic health record (EHR) database. METHODS: The Clinical Data Analytics Platform, a deidentified EHR database of 6 million Mayo Clinic patients, was used to predict BE and EAC risk. BE and EAC cases and controls were identified using International Classification of Diseases codes and augmented curation (natural language processing) techniques applied to clinical, endoscopy, laboratory, and pathology notes. Cases were propensity score matched to 5 independent randomly selected control groups. An ensemble transformer-based ML model architecture was used to develop predictive models. RESULTS: We identified 8,476 BE cases, 1,539 EAC cases, and 252,276 controls. The BE ML transformer model had an overall sensitivity, specificity, and AUROC of 76%, 76%, and 0.84, respectively. The EAC ML transformer model had an overall sensitivity, specificity, and AUROC of 84%, 70%, and 0.84, respectively. Predictors of BE and EAC included conventional risk factors and additional novel factors, such as coronary artery disease, serum triglycerides, and electrolytes. DISCUSSION: ML models developed on an EHR database can predict incident BE and EAC risk with improved accuracy compared with conventional risk factor-based risk scores. Such a model may enable effective implementation of a minimally invasive screening technology.

Squamous Cell Carcinoma of the Esophagus.

Esophageal squamous cell carcinoma (ESCC) is common in the developing world with decreasing incidence in developed countries and carries significant morbidity and mortality. Major risk factors for ESCC development include significant use of alcohol and tobacco. Screening for ESCC can be recommended in high-risk populations living in highly endemic regions. The treatment of ESCC ranges from endoscopic resection therapy or surgery in localized disease to chemoradiotherapy in metastatic disease, and prognosis is directly related to the stage at diagnosis. New immunotherapies and molecular targeted therapies may improve the dismal survival outcomes in patients with metastatic ESCC.

Making Dysphagia Easier to Swallow: A Review for the Practicing Clinician.

Evaluating a patient with dysphagia can be a complex and daunting task. In this article, we present a practical approach to the evaluation, physical examination, and subsequent work-up of dysphagia that is applicable to practicing physicians.

Celiac disease: a clinical review.

Celiac disease (CD) is an immune-mediated inflammatory enteropathy triggered by gluten exposure in genetically susceptible individuals. It has a high prevalence approaching 1% of the US population. A high index of suspicion is warranted to diagnose CD as frequently patients present with extraintestinal or atypical manifestations. CD is diagnosed by a combination of serum serologies and duodenal biopsies. The majority of patients will respond to a lifelong gluten-free diet which is the cornerstone of therapy. Complications such as refractory CD, ulcerative jejunoileitis, enteropathy associated T-cell lymphoma and small bowel adenocarcinoma occur in a minority of patients.