Surgical versus endoscopic gastroenterostomy for gastric outlet obstruction: a retrospective multicentric comparative study of technical and clinical success.

PURPOSE: Gastric outlet obstruction (GOO) is mainly due to advanced malignant disease. GOO can be treated by surgical gastroenterostomy (SGE), endoscopic enteral stenting (EES), or endoscopic ultrasound-guided gastroenterostomy (EUS-GE) to improve the quality of life. METHODS: Between 2009 and 2022, patients undergoing SGE or EUS-GE for GOO were included at three centers. Technical and clinical success rates, post-procedure adverse events (AEs), length of hospital stay (LOS), 30-day all-cause mortality, and recurrence of GOO were retrospectively analyzed and compared between SGE and EUS-GE. Predictive factors for technical and clinical failure after SGE and EUS-GE were identified. RESULTS: Of the 97 patients included, 56 (57.7%) had an EUS-GE and 41 (42.3%) had an SGE for GOO, with 62 (63.9%) GOO due to malignancy and 35 (36.1%) to benign disease. The median follow-up time was 13,4 months (range 1 days-106 months), with no difference between the two groups (p = 0.962). Technical (p = 0.133) and clinical (p = 0.229) success rates, severe morbidity (p = 0.708), 30-day all-cause mortality (p = 0.277) and GOO recurrence (p = 1) were similar. EUS-GE had shorter median procedure duration (p < 0.001), lower post-procedure ileus rate (p < 0.001), and shorter median LOS (p < 0.001) than SGE. In univariate analysis, no risk factors for technical or clinical failure in SGE were identified and abdominal pain reported before the procedure was a risk factor for technical failure in the EUS-GE group. No risk factor for clinical failure was identified for EUS-GE. In the subgroup of GOO due to benign disease, SGE was associated with better technical success (p = 0.035) with no difference in clinical success rate compared to EUS-GE (p = 1). CONCLUSION: EUS-GE provides similar long-lasting symptom relief as SGE for GOO whether for benign or malignant disease. SGE may still be indicated in centers with limited experience with EUS-GE or may be reserved for patients in whom endoscopic technique fails.

Automatic Classification of Tumor Response From Radiology Reports With Rule-Based Natural Language Processing Integrated Into the Clinical Oncology Workflow.

PURPOSE: Imaging reports in oncology provide critical information about the disease evolution that should be timely shared to tailor the clinical decision making and care coordination of patients with advanced cancer. However, tumor response stays unstructured in free-text and underexploited. Natural language processing (NLP) methods can help provide this critical information into the electronic health records (EHR) in real time to assist health care workers. METHODS: A rule-based algorithm was developed using SAS tools to automatically extract and categorize tumor response within progression or no progression categories. 2,970 magnetic resonance imaging, computed tomography scan, and positron emission tomography French reports were extracted from the EHR of a large comprehensive cancer center to build a 2,637-document training set and a 603-document validation set. The model was also tested on 189 imaging reports from 46 different radiology centers. A tumor dashboard was created in the EHR using the Timeline tool of the vis.js javascript library. RESULTS: An NLP methodology was applied to create an ontology of radiographic terms defining tumor response, mapping text to five main concepts, and application decision rules on the basis of clinical practice RECIST guidelines. The model achieved an overall accuracy of 0.88 (ranging from 0.87 to 0.94), with similar performance on both progression and no progression classification. The overall accuracy was 0.82 on reports from different radiology centers. Data were visualized and organized in a dynamic tumor response timeline. This tool was deployed successfully at our institution both retrospectively and prospectively as part of an automatic pipeline to screen reports and classify tumor response in real time for all metastatic patients. CONCLUSION: Our approach provides an NLP-based framework to structure and classify tumor response from the EHR and integrate tumor response classification into the clinical oncology workflow.