An AI pipeline effectively categorizes patient-submitted ostomy photos, improving triage efficiency. 43.9% of 538 images were deemed to have obstructed views, while 5.6% were healthy. The best model, Clip-ViT, achieved an impressive area under the curve of 0.99 and precision of 0.87. This technology has potential to streamline post-operative care by prioritizing cases requiring in-person […]
Category: Digital Surgery and Telemedicine
AI Transforms Preoperative Risk Assessment in Surgery
Artificial intelligence can enhance risk assessment in surgery by analyzing complex datasets, improving outcomes for patients. Machine learning algorithms effectively predict postoperative mortality and major adverse cardiovascular events. Predicting technical complications like anastomotic leaks is more challenging with current data. Surgeons should view AI as a tool for informed decision-making, not a replacement for clinical […]
AI Research Lags Behind Growth of Ambulatory Surgery Centers
Surgeons need to address a significant gap in AI research affecting patient care in ambulatory surgery centers (ASCs). Over 60% of elective surgeries occur in ASCs, yet fewer than 10 studies focus on AI in these settings compared to over 500 in hospitals. Existing ASC research primarily targets workflow optimization, with no focus on intraoperative […]
Automated Quality Assessment in Distal Gastrectomy
A new computer vision model shows that the Critical View of Quality (CVQ) can enhance lymphadenectomy quality in gastric cancer surgery. CVQ correlates positively with lymph node yield (r=0.485, p<0.001), with retrieval increasing from 28.9 to 47.6 nodes across quartiles (p<0.001). Lower CVQ scores link to older age, higher BMI, ASA classification, male sex, and […]
Governance Framework for AI in Surgery Established
A new framework guides the ethical adoption of AI in surgical practice. AI-enabled decision support systems can now match or exceed human performance in certain tasks. A 19-item consensus framework developed through expert consensus highlights key governance areas: technical design, implementation, ethics, and bias. Addressing these themes is crucial for ensuring AI adoption enhances care […]
Eye Tracking Reveals Surgical Resident Expertise Levels
Eye-tracking can differentiate surgical residents’ cognitive strategies, critical for tailored education. Three distinct gaze phenotypes identified among 57 residents: expert-like, structured but superficial, and exploratory. Focused-stable-systematic patterns seen in years 4-5; brief-active-exploratory in years 1-2. Utilizing these insights can enhance tailored training and competency evaluations in surgical education. Gaze patterns explain 87.7% of variance in […]
AI Triage Risks Found in Trauma Cases
AI algorithms may excel in trauma triage but show concerning undertriage risks, notably in older adults. Sensitivity for major trauma detection was 89.4% with a specificity of 86.5%. Geriatric patients experienced a 17.6% undertriage rate, significantly higher than 5.2% in younger adults (p = .012). Theoretical delays of 18.5 minutes linked to undertriage affected 14.2% […]
AI Enhances Aneurysm Detection in Surgery
AI assistance significantly boosts detection of cerebral aneurysms during surgery, a game-changer for surgical practice. Health care professionals achieved 78% accuracy with AI vs. 70% without it (odds ratio: 1.77, p < 0.001). Neurosurgical attendings improved from 77% to 92% accuracy with AI support (odds ratio: 4.24, p = 0.003). AI is not just for […]
Using Language Models for Surgical Consensus: Key Findings
Large language models may enhance consensus-building on ventral rectopexy but vary significantly in reliability. Openevidence had the highest content appropriateness score (3.5/5), outperforming Gemini (3.0/5) and ChatGPT (2.8/5; p < 0.001). ChatGPT fabricated 53% of citations compared to Gemini’s 12% and Openevidence’s 0% (p < 0.001). This suggests a need for careful selection of tools […]
New Model Accurately Predicts Postoperative Risk
A new model incorporates complex health data to better predict complications after major surgeries, enhancing surgical decision-making and patient safety. Complication prevalence: 8% ICU admission, 4% mechanical ventilation, 7% acute kidney injury, and 1% in-hospital mortality. Model performance: AUROC scores of 0.93 for ICU admission, 0.94 for mechanical ventilation, 0.92 for acute kidney injury, and […]