Predicting Incident Adenocarcinoma of the Esophagus or Gastric Cardia Using Machine Learning of Electronic Health Records.

BACKGROUND & AIMS: Tools that can automatically predict incident esophageal adenocarcinoma (EAC) and gastric cardia adenocarcinoma (GCA) using electronic health records to guide screening decisions are needed. METHODS: The Veterans Health Administration (VHA) Corporate Data Warehouse was accessed to identify Veterans with 1 or more encounters between 2005 and 2018. Patients diagnosed with EAC (n = 8430) or GCA (n = 2965) were identified in the VHA Central Cancer Registry and compared with 10,256,887 controls. Predictors included demographic characteristics, prescriptions, laboratory results, and diagnoses between 1 and 5 years before the index date. The Kettles Esophageal and Cardia Adenocarcinoma predictioN (K-ECAN) tool was developed and internally validated using simple random sampling imputation and extreme gradient boosting, a machine learning method. Training was performed in 50% of the data, preliminary validation in 25% of the data, and final testing in 25% of the data. RESULTS: K-ECAN was well-calibrated and had better discrimination (area under the receiver operating characteristic curve [AuROC], 0.77) than previously validated models, such as the Nord-Trøndelag Health Study (AuROC, 0.68) and Kunzmann model (AuROC, 0.64), or published guidelines. Using only data from between 3 and 5 years before index diminished its accuracy slightly (AuROC, 0.75). Undersampling men to simulate a non-VHA population, AUCs of the Nord-Trøndelag Health Study and Kunzmann model improved, but K-ECAN was still the most accurate (AuROC, 0.85). Although gastroesophageal reflux disease was strongly associated with EAC, it contributed only a small proportion of gain in information for prediction. CONCLUSIONS: K-ECAN is a novel, internally validated tool predicting incident EAC and GCA using electronic health records data. Further work is needed to validate K-ECAN outside VHA and to assess how best to implement it within electronic health records.

Yield of Repeat Endoscopy for Barrett's Esophagus After Normal Index Endoscopy.

INTRODUCTION: Guidelines suggest 1-time screening with esophagogastroduodenoscopy (EGD) for Barrett's esophagus (BE) in individuals at an increased risk of esophageal adenocarcinoma (EAC). We aimed to estimate the yield of repeat EGD performed at prolonged intervals after a normal index EGD. METHODS: We conducted a national retrospective analysis within the U S Veterans Health Administration, identifying patients with a normal index EGD between 2003 and 2009 who subsequently had a repeat EGD. We tabulated the proportion with a new diagnosis of BE, EAC, or esophagogastric junction adenocarcinoma (EGJAC) and conducted manual chart review of a sample. We fitted logistic regression models for the odds of a new diagnosis of BE/EAC/EGJAC. RESULTS: We identified 71,216 individuals who had a repeat EGD between 1 and 16 years after an index EGD without billing or cancer registry codes for BE/EAC/EGJAC. Of them, 4,088 had a new billing or cancer registry code for BE/EAC/EGJAC after the repeat EGD. On manual review of a stratified sample, most did not truly have new BE/EAC/EGJAC. A longer duration between EGD was associated with greater odds of a new diagnosis (adjusted odds ratio [aOR] for each 5 years 1.31; 95% confidence interval [CI] 1.19-1.44), particularly among those who were younger during the index EGD (ages 19-29 years: aOR 3.92; 95% CI 1.24-12.4; ages 60-69 years: aOR 1.19; 95% CI 1.01-1.40). DISCUSSION: The yield of repeat EGD for BE/EAC/EGJAC seems to increase with time after a normal index EGD, particularly for younger individuals. Prospective studies are warranted to confirm these findings.

Management of eosinophilic esophagitis associated food impaction in Europe and the United States.

Eosinophilic esophagitis (EoE) is the most common cause of esophageal food impaction (EFI). Approaches to management of EFI due to EoE have not been well characterized. We conducted a web-based survey to understand approaches to management of EFI due to EoE among endoscopists. Questions focused on management of patients from presentation to post-endoscopy follow-up. The survey was administered to a list of eligible candidates provided by societies of gastroenterology. A total of 308 endoscopists completed the questionnaire. The majority (83%) practiced in Europe and treated adults (78%). Most agreed patients should be advised to seek emergency care (66%) within 1 to 2 hours (41% agreement). There was agreement that medications to induce vomiting should be avoided (84%) and that blood tests or imaging studies were usually not required before endoscopy. By contrast, there was more variability in the type of sedation recommended and the need for endotracheal intubation, especially when comparing more experienced with less experienced EoE-endoscopists. Overall, fewer than half (43%) respondents recommended obtaining esophageal biopsies during the initial endoscopy. However, there were significant differences in the proportion who recommended biopsies based on level of EoE-experience (25, 52, 77%, P < 0.001; less vs. moderate vs. very experienced) and comparing pediatric and adult endoscopists (32, vs. 79%, P < 0.001; adult vs. pediatric). There exists heterogeneity among endoscopists in recommendations to manage EFI in patients with EoE. These findings support development of clinical guidelines and new studies to clarify the rationale for best practices. Key summary: Established knowledge-The optimal management of patients with esophageal food impaction due to eosinophilic esophagitis from presentation at the emergency department to postendoscopy care is unclear. New findings-Considerable recommendation variation exists in the management of EFI in patients with EoE. Our findings provide a rationale for the creation of consensus practice guidelines and further study into best practices.

Loss to follow-up after food impaction among patients with and without eosinophilic esophagitis.

Symptoms of esophageal dysfunction such as food impaction are consistent with, but not diagnostic for eosinophilic esophagitis (EoE) without obtaining histology. We conducted a retrospective study to characterize patients with food impaction at a tertiary center. We hypothesized that many patients with food impaction may be lost to follow-up and that many have features suggestive of EoE. Adult patients presenting to the emergency department with esophageal food impaction were identified from an endoscopic database. Electronic medical records were manually abstracted. We examined associations between demographics, comorbid conditions, and follow-up with biopsy findings. Of 220 patients who presented to the emergency department for food impaction, 74.1% were men. Adequate follow-up was not documented in 120 (54.5%). Those lost to follow-up did not differ significantly by gender, age at symptom onset, or distance from hospital compared to those with follow-up. Esophageal biopsies were obtained in 158 (71.8%), and those with ≥15 eos/HPF were more likely to be lost to follow-up than those with <15 eos/HPF (52.8% vs. 34.8%, P < 0.05). Of those never biopsied, 79.0% were lost to follow-up and had intermediate proportions of males, food allergy, and asthma when compared to those with and without eosinophilic inflammation. Patients with food impaction commonly have EoE but are often lost to follow-up. Among those never biopsied, demographic and clinical features suggest that many may have undiagnosed EoE. Strategies for increasing use of biopsies in patients with food impaction and improving follow-up are needed to diagnose and manage EoE.

Management of eosinophilic esophagitis is often discordant with guidelines and not patient-centered: results of a survey of gastroenterologists.

Published guidelines for the management of eosinophilic esophagitis (EoE) recommend an initial trial of proton pump inhibitors (PPI), histologic assessment for response to therapy, and tailoring treatments to patient needs and provider resources. Effectiveness studies directly comparing therapies are lacking, leaving a situation ripe for shared decision making. We aimed to assess gastroenterologists' adherence to guidelines and how they respond to EoE patients' preferences regarding management. We administered a web-based survey to practicing US gastroenterologists, assessing knowledge, and practice patterns in the management of EoE, including comfort with alternative treatments to steroids. Ninety-two providers responded, including 55% in private practice. Nearly half (47%) reported spending ≤10 minutes on initial education and counseling and 48% recommended PPI monotherapy prior to other strategies. Of those who did not start with PPI monotherapy, 55% chose topical steroids ± PPI and 26% dietary elimination ± PPI. Despite this, 90% felt comfortable allowing a patient to start dietary elimination instead of steroids, but less comfortable with dilation alone (39%) or no treatment (30%). Upon symptomatic resolution, 72% of academic providers recommended endoscopy with biopsies to demonstrate histologic response to treatment, compared to 27% in private practice. There are substantial variations in adherence to guidelines regarding PPI use and assessing response to therapy. Gastroenterologists prefer topical steroids over other treatment modalities and most spend little time educating and counseling, which may limit informed decision making. Strategies aimed at decreasing these variations in management and promoting shared decision making in EoE are needed.

Hungry for life: How the arcuate nucleus and neuropeptide Y may play a critical role in mediating the benefits of calorie restriction.

Laboratory studies consistently demonstrate extended lifespan in animals on calorie restriction (CR), where total caloric intake is reduced by 10-40% but adequate nutrition is otherwise maintained. CR has been further shown to delay the onset and severity of chronic diseases associated with aging such as cancer, and to extend the functional health span of important faculties like cognition. Less understood are the underlying mechanisms through which CR might act to induce such alterations. One theory postulates that CR's beneficial effects are intimately tied to the neuroendocrine response to low energy availability, of which the arcuate nucleus in the hypothalamus plays a pivotal role. Neuropeptide Y (NPY), a neurotransmitter in the front line of the arcuate response to low energy availability, is the primary hunger signal affected by CR and therefore may be a critical mechanism for lifespan extension.

Nrf2 mediates cancer protection but not prolongevity induced by caloric restriction.

Caloric restriction (CR) is the most potent intervention known to both protect against carcinogenesis and extend lifespan in laboratory animals. A variety of anticarcinogens and CR mimetics induce and activate the NF-E2-related factor 2 (Nrf2) pathway. Nrf2, in turn, induces a number of antioxidative and carcinogen-detoxifying enzymes. Thus, Nrf2 offers a promising target for anticarcinogenesis and antiaging interventions. We used Nrf2-disrupted (KO) mice to examine its role on the biological effects of CR. Here, we show that Nrf2 is responsible for most of the anticarcinogenic effects of CR, but is dispensable for increased insulin sensitivity and lifespan extension. Nrf2-deficient mice developed tumors more readily in response to carcinogen exposure than did WT mice, and CR was ineffective in suppressing tumors in the KO mice. However, CR extended lifespan and increased insulin sensitivity similarly in KO and WT mice. These findings identify a molecular pathway that dissociates the prolongevity and anticarcinogenic effects of CR.