A deep learning-based model was developed for real-time recognition of perigastric blood vessels during laparoscopic radical gastrectomy. In testing, the model demonstrated high precision (mean 0.9442) and recall (mean 0.9099) for arteries and maintained robust performance across various challenging conditions. It showed potential to improve intraoperative safety and reduce the risk of accidental bleeding, particularly […]
Category: Digital Surgery and Telemedicine
Machine learning predicts early recurrence risk in cholangiocarcinoma
A novel machine learning model effectively predicts very early recurrence (VER) of perihilar cholangiocarcinoma (PCCA) post-surgery. In a study of 434 patients, 15% experienced VER, significantly reducing median overall survival to 8.4 months compared to 38.5 months for those without recurrence (p
Ethical integration of AI is essential in academic surgery.
The incorporation of generative artificial intelligence in academic surgery presents transformative opportunities but also significant ethical challenges. Key benefits include improved efficiency in tasks like manuscript writing and clinical documentation. However, concerns regarding bias, transparency, intellectual property, and privacy necessitate strict guidelines for responsible use. The paper emphasizes the importance of ethical training and governance […]
Deep learning model predicts early recurrence in gastric cancer
A new deep learning model utilizing multiphase CT images effectively predicts early recurrence in patients with locally advanced gastric cancer (LAGC). The model, integrated with clinical factors, demonstrated superior performance (AUC: 0.891) compared to previous models. Significant associations were found between higher prediction scores and tumor proliferation pathways, such as WNT and MYC signaling, as […]
AI-based scoring system enhances prognosis prediction for HCC
A novel AI-driven system effectively quantifies CD8+ tumor-infiltrating lymphocytes (TILs) in hepatocellular carcinoma (HCC) patients post-liver resection. In a study involving 514 patients, the automated CD8+ TIL scoring system (ATLS-8) revealed five-year overall survival rates of 34.05% for low-score and 65.03% for high-score cohorts. This research highlights the independent prognostic value of CD8+ TIL density, […]
Telemedicine significantly reduces outpatient waiting times.
A systematic review revealed telemedicine’s effectiveness in decreasing outpatient wait times, showcasing a substantial weighted mean reduction of 25.4 days across 270,388 patients. Notably, for clinical specialties, the reduction reached 34.7 days, and for surgical patients, it was 17.3 days. Most studies demonstrated low bias risk, underscoring telemedicine’s potential for enhancing equitable access to healthcare […]
Explainable machine learning model predicts mortality in infected pancreatic necrosis
An explainable machine learning model for predicting 90-day mortality in patients with infected pancreatic necrosis (IPN) has shown promising results. The final model, developed from a cohort of 344 patients, achieved a c-index of 0.863 in internal validation and 0.857 in external validation, outperforming nine other models. Key mortality predictors included multiple organ failure and […]
AI framework quantifies clinical influences on posthepatectomy length of stay
An innovative artificial intelligence framework was developed to quantify the impact of clinical factors on posthepatectomy length of stay (LOS), explaining 75% of its variability. The study analyzed 21,039 patients, revealing that clinical influences are significant while nonclinical factors account for the remaining 25%. Notably, major resections had a longer mean LOS of 6.9 days […]
Clinical factors explain less than 55% of postoperative stay variability
A machine-learning framework quantified the impact of clinical versus nonclinical influences on postoperative length of stay after colectomy. Analysis of 96,081 patients revealed that clinical factors accounted for only 29-54% of length of stay variability. Despite optimizing these variables, significant unexplained variance remained, indicating the influence of nonclinical factors. This study is pioneering in highlighting […]
Machine learning effectively predicts gastroparesis risk in surgery patients
Advanced machine learning techniques, particularly the xgboost algorithm, have shown exceptional predictive accuracy for identifying the risk of postoperative gastroparesis in colon cancer patients after complete mesocolic excision. From a cohort of 1,097 patients, featuring 87 gastroparesis cases, xgboost achieved an area under the curve of 0.939 for training and 0.876 for validation. This predictive […]