Conventional versus Rosemont endoscopic ultrasound criteria for chronic pancreatitis: comparing interobserver reliability and intertest agreement.

BACKGROUND: The Rosemont criteria (RC) were recently proposed by expert consensus to standardize endoscopic ultrasound (EUS) features and thresholds for diagnosing chronic pancreatitis (CP); however, they are cumbersome and are not validated. OBJECTIVE: To determine interobserver agreement between RC and conventional criteria (CC), and to assess intertest agreement in the diagnosis of CP. METHODS: Thirty-six consecutive patients who underwent EUS for abdominal pain or pancreatitis were retrospectively reviewed. Anonymized images were independently chosen as best representations of the pancreatic body and reviewed by three experts who recorded the presence of CC and RC features. Agreement (proportion and kappa statistic) between CC and RC was calculated. Interobserver agreement within the CC and RC was assessed. Secondary comparisons with endoscopic retrograde cholangiopancreatography were made where available. RESULTS: Using CC, 60 readings (83.3%) were negative for CP, while 12 readings (16.7%) were positive. Using RC, 59 readings (81.9%) were negative for CP, while 13 (18.1%) were positive. The weighted kappa for interobserver agreement for CC (four categories: normal/low probability, indeterminate, high probability or calcific) was 0.50, with 80.0% overall agreement, versus 0.27 and 68.1% for the four RC categories (normal, indeterminate, suggestive of and consistent with). Agreement on a positive diagnosis with CC was 86.1% (P=0.38 [McNemar's exact test]), with a kappa of 0.47; for RC, agreement was lower at 80.6% (P=0.016 [McNemar's exact test]), with a kappa of 0.38. For patients who underwent endoscopic retrograde cholangiopancreatography (n=12), false-negative and false-positive rates between CC and RC did not appear to be different. CONCLUSIONS: The RC do not appear to achieve the goals of improving accuracy and interobserver agreement for diagnosing CP.

A decision support system to facilitate management of patients with acute gastrointestinal bleeding.

OBJECTIVE: To develop a model to predict the bleeding source and identify the cohort amongst patients with acute gastrointestinal bleeding (GIB) who require urgent intervention, including endoscopy. Patients with acute GIB, an unpredictable event, are most commonly evaluated and managed by non-gastroenterologists. Rapid and consistently reliable risk stratification of patients with acute GIB for urgent endoscopy may potentially improve outcomes amongst such patients by targeting scarce healthcare resources to those who need it the most. DESIGN AND METHODS: Using ICD-9 codes for acute GIB, 189 patients with acute GIB and all available data variables required to develop and test models were identified from a hospital medical records database. Data on 122 patients was utilized for development of the model and on 67 patients utilized to perform comparative analysis of the models. Clinical data such as presenting signs and symptoms, demographic data, presence of co-morbidities, laboratory data and corresponding endoscopic diagnosis and outcomes were collected. Clinical data and endoscopic diagnosis collected for each patient was utilized to retrospectively ascertain optimal management for each patient. Clinical presentations and corresponding treatment was utilized as training examples. Eight mathematical models including artificial neural network (ANN), support vector machine (SVM), k-nearest neighbor, linear discriminant analysis (LDA), shrunken centroid (SC), random forest (RF), logistic regression, and boosting were trained and tested. The performance of these models was compared using standard statistical analysis and ROC curves. RESULTS: Overall the random forest model best predicted the source, need for resuscitation, and disposition with accuracies of approximately 80% or higher (accuracy for endoscopy was greater than 75%). The area under ROC curve for RF was greater than 0.85, indicating excellent performance by the random forest model. CONCLUSION: While most mathematical models are effective as a decision support system for evaluation and management of patients with acute GIB, in our testing, the RF model consistently demonstrated the best performance. Amongst patients presenting with acute GIB, mathematical models may facilitate the identification of the source of GIB, need for intervention and allow optimization of care and healthcare resource allocation; these however require further validation.