Background Screening for Barrett's oesophagus relies on endoscopy, which is invasive and few who undergo the procedure are found to have the condition. We aimed to use machine learning techniques to develop and externally validate a simple risk prediction panel to screen individuals for Barrett's oesophagus.
MethodsIn this prospective study, machine learning risk prediction in Barrett's oesophagus (MARK-BE), we used data from two case-control studies, BEST2 and BOOST, to compile training and validation datasets. From the BEST2 study, we analysed questionnaires from 1299 patients, of whom 880 (67·7%) had Barrett's oesophagus, including 40 with invasive oesophageal adenocarcinoma, and 419 (32·3%) were controls. We randomly split (6:4) the cohort using a computer algorithm into a training dataset of 776 patients and a testing dataset of 523 patients. We compiled an external validation cohort from the BOOST study, which included 398 patients, comprising 198 patients with Barrett's oesophagus (23 with oesophageal adenocarcinoma) and 200 controls. We identified independently important diagnostic features of Barrett's oesophagus using the machine learning techniques information gain and correlationbased feature selection. We assessed multiple classification tools to create a multivariable risk prediction model. Internal validation of the model using the BEST2 testing dataset was followed by external validation using the BOOST external validation dataset. From these data we created a prediction panel to identify at-risk individuals.Findings The BEST2 study included 40 diagnostic features. Of these, 19 added information gain but after correlationbased feature selection only eight showed independent diagnostic value including age, sex, cigarette smoking, waist circumference, frequency of stomach pain, duration of heartburn and acidic taste, and taking antireflux medication, of which all were associated with increased risk of Barrett's oesophagus, except frequency of stomach pain, with was inversely associated in a case-control population. Logistic regression offered the highest prediction quality with an area under the receiver-operator curve (AUC) of 0·87 (95% CI 0·84-0·90; sensitivity set at 90%; specificity of 68%). In the testing dataset, AUC was 0·86 (0·83-0·89; sensitivity set at 90%; specificity of 65%). In the external validation dataset, the AUC was 0·81 (0·74-0·84; sensitivity set at 90%; specificity of 58%).
Interpretation Our diagnostic model offers valid predictions of diagnosis of Barrett's oesophagus in patients withsymptomatic gastro-oesophageal reflux disease, assisting in identifying who should go forward to invasive confirmatory testing. Our predictive panel suggests that overweight men who have been taking antireflux medication for a long time might merit particular consideration for further testing. Our risk prediction panel is quick and simple to administer but will need further calibration and validation in a prospective study in primary care.
Background and study aims The standard radiofrequency ablation (RFA) protocol for Barrett’s esophagus (BE) encompasses an intermediary cleaning phase between two ablation sessions. A simplified protocol omitting the cleaning phase is less labor-intensive but equally effective in studies based on single ablation procedures. The aim of this study was to compare efficacy and safety of the standard and simplified RFA protocols for the whole treatment pathway for BE, including both circumferential and focal devices.
Patients and methods We performed a retrospective analysis of prospectively collected data on patients receiving RFA between January 2007 and August 2017 at two institutions. Outcomes assessed were: 1) complete remission of dysplasia (CR-D) and intestinal metaplasia (CR-IM) at 18 months; and 2) rate of esophageal strictures.
Results One hundred forty-five patients were included of whom 73 patients received the standard and 72 patients received the simplified protocol. CR-D was achieved in 94.5 % and 95.8 % of patients receiving the standard and simplified protocol, respectively (P = 0.71). CR-IM was achieved in 84.9 % and 77.8 % of patients treated with the standard and simplified protocol, respectively (P = 0.27). Strictures were significantly more common among patients who received the simplified protocol (12.5 %) compared to the standard protocol (1.4 %; P = 0.008). The median number of esophageal dilations was one.
Conclusion The simplified RFA protocol is as effective as the standard protocol in eradicating BE but carries a higher risk of strictures. This needs to be taken into account, particularly in patients with higher pretreatment risk of strictures, such as those with esophageal narrowing from previous endoscopic mucosal resection (EMR).
Background
Proton-pump inhibitors (PPIs) are the mainstay of gastroesophageal reflux disease (GERD) treatment, however, up to 30% of patients have a poor symptomatic response. PH-impedance is the gold standard to assess whether this is due to persistent acid reflux. We aimed to characterize clinical predictors of persistent esophageal acid reflux on PPIs including gastric pH measured during endoscopy.
Methods
We prospectively recruited patients with GERD and/or Barrett’s esophagus (BE) on PPIs. All patients completed a symptom questionnaire (RDQ) and underwent gastroscopy with gastric pH analysis, immediately followed by ambulatory 24-hour pH-impedance. We used a modified cut-off of 1.3% for pathological esophageal acid exposure time (AET). Multiple linear regression model was used to analyze the correlation between AET and predictive variables.
Results
We recruited 122 patients, of which 92 (75.4%) were included in the final analysis [44 male (47.8%), median age 53 years (IQR: 43–66)]. Forty-four patients (47.8%) had persistent acid reflux with a median total AET of 2.2 (IQR1.2-5.0), as compared to 0.1 (IQR 0.0-0.2) in patients without persistent reflux (n=48; P<.001). There was no difference in age, gender, BMI, PPI-regimen, diagnosis of hiatus hernia or BE, and severity of symptoms between patients with normal and abnormal AET. Median gastric pH was significantly lower in patients with abnormal AET (5.8 vs 6.6, P=0.032) and it correlated with the total AET (P=.045; R2=12.0%). With a pH cut-off of 5.05, single point endoscopic gastric pH analysis had an area under the ROC curve (AUC) of 63.0% (95%CI 51.3-74.7) for prediction of pathological esophageal AET.
Conclusions
Symptoms and clinical characteristics are not useful to predict persistent acid reflux in patients on PPIs. One-point gastric pH correlates with 24-hour esophageal AET and could guide clinicians to assess response to PPIs, however, its utility needs validation in larger studies.
96 (28-327) days since HCST. The most common diagnoses were AML 55/159 (34.6%), B-ALL 30/159 (18.9%), and CML 14/159 (8.8%). Diarrhea 105/159 (66.0%), nausea/vomiting 63/159 (39.6%), and abdominal pain 31/159 (19.5%) were the most common indications for evaluation of GVHD. Abnormal pathology was most frequently found in the stomach 74/159 (46.5%), duodenum 66/159 (41.5%), and sigmoid colon 54/159 (34.0%). The majority of patients were on tacrolimus 101/159 (63.5%), sirolimus 70/159 (44.0%), and hydrocortisone 59/159 (37.1%) immunosuppression. Bleeding occurred following 5/159 (3.1%) endoscopies, with successful hemostasis in all cases. Death within 1 month of endoscopy occurred in 7/159 (4.4%). A pre-procedure platelet count above 50x10 3 /mL (OR, 0.24; 95% CI, 0.06-0.94; PZ0.04) and increased pre-procedure hemoglobin (OR, 0.54; 95% CI, 0.33-0.88; PZ0.01) were associated with decreased odds of adverse outcome following endoscopy. Discussion: Although a platelet cutoff above 50x10 3 /mL was associated with lower odds of adverse events post-endoscopy, the overall risk of endoscopic procedures was low. In our series, under 4% of endoscopic interventions resulted in bleeding, infection, or fever. As expected, patients with higher pre-procedure hemoglobin conferred a decreased risk of adverse events. Pre-procedure ANC was not significantly associated with adverse outcome. Further characterization of safety at different platelet thresholds is necessary to inform the decision to proceed with endoscopy in this at-risk population.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.