A study developed predictive models to aid surgical decision-making for hepatic cystic echinococcosis (CE). Analyzing data from 406 patients, the Cox regression model achieved a concordance index of 0.94 and AUC of 0.96. The decision tree model highlighted imaging findings, cyst stage, and symptoms as key predictors, with a mean AUC of 0.950. These validated […]
Category: Digital Surgery and Telemedicine
Machine learning effectively predicts bile leaks post-hepatectomy.
A machine learning model using extreme gradient boosting (XGBoost) was developed to predict bile leaks after hepatectomy for liver cancer. The model analyzed data from over 22,000 patients, achieving areas under receiver operating characteristic curves (AUROC) of 0.748, 0.719, and 0.711 in various patient cohorts. Factors such as serum alkaline phosphatase, surgical approach, and cancer […]
AI enhances anatomical recognition in hernia surgery
An artificial intelligence system, Eureka, was validated for its ability to accurately identify dissection layers, nerves, vas deferens, and microvessels during transabdominal preperitoneal inguinal hernia repair. Key metrics included mean intersection over union scores of 0.33 for connective tissue, 0.24 for nerves, 0.50 for vas deferens, and 0.30 for microvessels. The results indicated improved visualization […]
Identification of six hub genes for NAFLD treatment
Comprehensive machine learning facilitated the identification of 12 consensus predicted genes (CPGs) linked to non-alcoholic fatty liver disease (NAFLD) following metabolic and bariatric surgery (MBS). A mouse model confirmed the mRNA expression of six hub genes—PPARA, PLIN2, MED13, INSIG1, CPT1A, and ALOX5AP—showing strong correlations with data from three validation datasets. These findings highlight the potential […]
Machine learning effectively predicts early recurrence in gastric cancer.
A multicenter study identified early recurrence (ER) within two years post-surgery in 15% of 11,615 gastric cancer patients. Utilizing machine learning, researchers developed a stacking ensemble model that achieved an area under the receiver operating characteristic curve of 1.0 for training and 0.8 for testing, indicating robust predictive capability. Key predictors included tumor size, staging, […]
Surgical navigation enhances successful lymph node removal outcomes
Surgical navigation significantly improves successful retroperitoneal lymph-node dissection outcomes, achieving an 85% success rate compared to 50% in conventional methods. The trial, involving 69 participants, demonstrated that navigation aided surgeons in localizing targeted lymph nodes more effectively. Both complication rates were similar across both groups, and surgeons rated the navigation system favorably. These findings indicate […]
Machine learning predicts duodenal stump leakage risk in gastric cancer
A new predictive model utilizing machine learning has shown promise in forecasting duodenal stump leakage after laparoscopic gastrectomy for gastric cancer. Analyzing data from 4,070 patients, the support vector machine emerged as the most effective algorithm, achieving high sensitivity, accuracy, and specificity. Key factors influencing leakage risk include tumor location, tumor stage, operation time, preoperative […]
Telemedicine shows promise in improving surgical care in Africa
Telemedicine has emerged as a transformative force in African surgical care, particularly in areas with scarce access to quality treatment. Despite challenges such as infrastructure deficits and personnel training, the implementation of telemedicine has demonstrated favorable outcomes. Increased adoption is evident in postsurgical care and doctor-patient consultations since the pandemic. While the potential for telesurgery […]
Machine learning models effectively predict appendicitis in emergencies
This proof-of-concept study reveals that machine learning models can predict appendicitis in patients with acute abdominal pain more accurately than traditional methods. With AUROCs of 0.919 and 0.923 when including laboratory test results, the models outperformed the Alvarado scoring system (AUROC of 0.824) and matched or exceeded the performance of emergency department physicians. These findings […]
Deep learning model predicts early recurrence in gastric cancer
A deep learning model (DLRMLP) integrating clinical factors outperformed traditional methods in predicting early recurrence of locally advanced gastric cancer (LAGC) post-gastrectomy. In a study involving 620 patients, DLRMLP achieved an AUC of 0.891 compared to 0.797 with conventional models. This model effectively stratified early recurrence-free survival, disease-free survival, and overall survival (all p < […]