Citations

INTRODUCTION: Colorectal cancer remains a leading cause of cancer associated death in the United States and colonoscopy the primary screening strategy for prevention. Rates of adenomatous and serrated neoplasia detection are inversely associated with postcolonoscopy colorectal cancer. This crucial quality metric depends on accurate ascertainment of colorectal neoplasia findings from both endoscopy and histopathology records. We aimed to assess the feasibility of a random forest machine learning model to rapidly and accurately categorize colorectal neoplasia from electronic health record data. METHODS: A retrospective cohort study compared neoplasia detection rates among individuals undergoing colonoscopy at a large academic institution to develop a rule-based algorithm to categorize colorectal neoplasia from endoscopy reports and pathology systematized nomenclature of medicine – clinical terms. This cohort provided a large training set to develop a natural language processing system using a random forest approach to automatically classify unstructured pathology findings into adenoma, serrated, or advanced neoplasms. This system was manually validated through an independent holdout set. RESULTS: The training set comprised 35,953 unstructured pathology reports with matched systematized nomenclature of medicine – clinical terms from 95,188 unstructured colonoscopy reports. The final model was assessed on an independent holdout set of 337 manually annotated procedures obtaining an area under the receiver operating characteristic curve of 0.997 (confidence interval [CI] 0.994–1), 0.99 (CI 0.98–1), and 0.99 (CI 0.98–0.99) for prediction of adenoma, serrated, and advanced lesions, respectively. DISCUSSION: The random forest-based hybrid natural language processing system for classification of colonoscopy results was both accurate and explainable. NLP combined with effective machine learning algorithms can provide a scalable strategy for colonoscopy quality monitoring.

Abstract Objectives Patient education is increasingly acknowledged as an important aspect of the management of systemic lupus erythematosus (SLE). The aim of the study was to develop the SLE Knowledge Assessment score (SLAKE), a digital multilingual self-assessment tool designed to quantify essential SLE knowledge. Methods International healthcare professionals (HCPs) and patient representatives engaged in a multi-step process to: identify essential SLE knowledge domains, select key domains via rating, and generate an item bank of 394 questions across 11 domains, which was then adapted into 19 languages. For validation, participants completed 44 questions (including 33 randomly selected), with scores calculated for total knowledge and the 11 specific domains. Statistical analyses examined associations between scores and demographic, clinical, and educational variables. Results SLAKE was used by 1182 SLE participants (1120 [94.8%] women, median age: 45 years [IQR: 35–54 years]), with a median SLE duration of 10 years (IQR: 4–20 years). The median SLAKE score was 37 (IQR: 34–40) of a maximum of 44 points while the median score across the 11 SLAKE domains ranged between 3 and 4 over a maximum of 4 points. There was a significant positive association between SLAKE score and SLE duration (p= 0.006), previous participation to a patient education course or a patient training for lupus (p< 0.0001) and the education level (p< 0.0001) but not with age (p= 0.48) or gender (p= 0.39). Conclusion SLAKE is a valid, multilingual, digital self-assessment tool that effectively measures essential SLE knowledge. Its randomized question bank and domain-specific scoring enable targeted education, ultimately supporting better disease management.

Introduction: Sepsis-induced cardiac dysfunction (SICD) is a reversible type of cardiac disorder, with variable reported incidence and uncertain prognostic implications that may occur in patients with sepsis. This study aimed to identify patients with SICD, characterise their clinical and paraclinical features, analyse the different patterns of cardiac dysfunction, and determine prognostic implications. Methods: Patients admitted to the intensive care unit (ICU) for sepsis or septic shock who underwent echocardiography on admission were identified retrospectively. Exclusion criteria included acute myocardial infarction, preexisting severe left heart valve disease or systolic dysfunction. Clinical, paraclinical, and echocardiography data were documented. The primary outcome was a composite of in-hospital mortality, prolonged hospital stay, and prolonged ICU stay. Results: A total of 128 septic patients were included, with a median age of 72.5 years and a 55% male proportion. Alcohol abuse, vasopressor therapy and orotracheal intubation on admission were risk factors for developing SICD. More than a third developed SICD (37%). We identified four different types of cardiac dysfunction based on echocardiography. An NT-proBNP level of over 9000 pg/mL was a predictor of SICD, composite outcome and mortality. A total of 56 patients (44%) experienced in-hospital mortality. Even though the occurrence of SICD did not predict mortality, it was a significant predictor of the composite outcome. Conclusions: This study describes the incidence and spectrum of SICD in a group of septic patients admitted to a tertiary care hospital. The occurrence of any type of cardiac dysfunction associated with sepsis and high NT-proBNP levels had strong prognostic implications.