2026-10-05

Background: Neurological injury is common after cardiac arrest (CA), but early and accurate assessment remains challenging. Existing evaluation methods often lack sensitivity and specificity. This study aims to develop and validate an interpretable machine learning model for early and reliable assessment of brain injury and its prognostic relevance in ICU patients. Methods: In this study, we analyzed 1,419 CA patients from the MIMIC-IV database, randomly split into training and validation cohorts. Six machine learning (ML) algorithms-logistic regression (LR), XGBoost, artificial neural network (ANN), random forest (RF), support vector machine (SVM), and k-nearest neighbors (KNN) were developed and evaluated. The top-performing model was interpreted using SHapley Additive exPlanations (SHAP), identifying eight key predictors of brain injury. Finally, clinical data from cardiac arrest patients at our hospital were collected for external validation to assess the model's generalizability and practical utility in an ICU setting. Results: Among the six machine learning models, the Random Forest model achieved the highest numerical performance in both the internal test and external validation cohorts. In the internal test cohort, Random Forest achieved an ACC of 88.7%, sensitivity of 80.0%, specificity of 90.6%, F1 score of 0.714, NPV of 95.5%, PPV of 64.5%, and AUC of 0.901. In the external validation cohort, Random Forest achieved an ACC of 88.8%, sensitivity of 77.8%, specificity of 90.8%, F1 score of 0.677, NPV of 95.8%, PPV of 60.0%, and AUC of 0.889. SHAP analysis identified eight influential predictors: age, hemoglobin, OASIS, temperature, blood urea nitrogen, white blood cell count, non-invasive diastolic pressure, and non-invasive systolic pressure. Conclusion: Machine learning offers strong potential for early prediction of acute brain injury in in-hospital cardiac arrest patients. Our interpretable model addresses the "black box" concern by providing transparent insights into key predictive features.

Background: Global aging has made disability among older adults a major public health challenge. Existing prediction models often lack external validation and actionable risk stratification. This study aims to develop and externally validate an explainable, three-tier machine learning framework for disability risk prediction and stratification among older adults. Methods: The training set and internal validation set were derived from the longitudinal development cohort of the China Health and Retirement Longitudinal Study (CHARLS, n = 3,599). An independent cross-sectional survey (n = 2,066) was used for external validation of the model. Individuals 65 years of age and older were included. Disability was defined by basic activities of daily living (BADL). After feature selection, five candidate machine learning techniques were evaluated to identify the optimal model. The SHapley Additive exPlanations (SHAP) was used to interpret the optimal model. Ultimately, a three-tier risk stratification method was developed and verified using Decision Curve Analysis (DCA) and the Cochran-Armitage trend test. Results: LASSO regression identified seven critical predictors. The Extreme Gradient Boosting (XGBoost) achieved the optimal balance, achieving the optimal balance of predictive performance in the internal testing set (AUC = 0.737) and external validation cohort (AUC = 0.782). SHAP analysis revealed age, age-adjusted Charlson Comorbidity Index (ACCI), and lower limb mobility as the paramount global predictors. The established risk stratification thresholds (33.17 and 65.23%) successfully categorized older adults into three levels. This three-tier system exhibited a significant gradient increase in actual disability incidence within the external set (P trend < 0.001) and a broad positive net clinical benefit in DCA. Discussion: The XGBoost model showed satisfactory and consistent performance in external validation. SHAP provided clinically meaningful insights, emphasizing age, lower-limb mobility, psychological status, and comorbidity burden. The validated three-tier system bridges the translational gap between complex predicted probabilities and practical clinical decisions.

Writing discharge notes (DNs) is a time-intensive task in inpatient care that contributes substantially to physician workload and burnout. Large language models (LLMs) offer potential opportunities to automate clinical documentation. However, concerns regarding accuracy, reliability, data privacy protection laws and data governance, and compliance with healthcare privacy remain. This study evaluated the performance of locally deployed, domain-adapted LLMs for generating German-language DN in internal medicine. In this retrospective study, electronic health record data from a Swiss tertiary care hospital were used. Clinical information from 3 inpatient scenarios served as input for the 3 open source LLMs (Mixtral 8×7B, Mixtral 8×22B, and LLaMA 3.1 70B). Zero-shot prompting, 4-shot in-context learning, and supervised fine-tuning using QLoRA were compared. Generated DNs were evaluated using automated metrics, blinded physician assessment with direct comparison of all relevant clinical information using the modified Physician Documentation Quality Instrument, and additionally an LLM-as-a-judge framework and a Turing Test. Across all evaluation methods, the fine-tuned Mixtral 8×7B model achieved the best overall performance compared with prompting-based approaches and larger nonspecialized models. Simpler cases consistently received higher-quality ratings. Physician inter-rater agreement was low, highlighting the challenges of objective quality assessment for clinical documentation. In a Turing Test, physicians were unable to reliably distinguish artificial intelligence-generated from clinician-authored DNs, performing at chance level, demonstrating that high-quality DN generation is feasible using locally deployed LLMs within secure hospital infrastructures. Domain-specific fine-tuning substantially improves performance and may enable practical deployment. Future work should address further hallucination mitigation, robustness across medical specialties, and real-world integration.

Introduction: Precise anatomical navigation is fundamental to safe endoscopic pituitary surgery, a high-stakes procedure characterised by a challenging learning curve. While traditional navigation systems often rely on workflow-disrupting probes or static preoperative imaging, advancements in computer vision AI (CVAI) now enable dynamic, real-time pixel-level anatomical segmentation directly from live surgical video. Our group has previously conducted a series of preclinical human-computer interaction studies to refine the system's design, alongside digital and high-fidelity physical simulations demonstrating the potential benefit of AI assistance in improving surgical performance, training, and safety. Building on this foundation, the current study represents a first-in-human evaluation of real-time pixel-level CVAI anatomical segmentation in the neurosurgical operating room - assessing feasibility, human factors and clinical outcomes, while iteratively improving the system. Method: Guided by the DECIDE-AI and IDEAL frameworks, this single-centre evaluation comprises an initial proof-of-concept of CVAI anatomical segmentation in endoscopic transsphenoidal pituitary surgery. The AI model utilised a DINOv3-derived vision transformer architecture, deployed via a high-performance edge computing unit to achieve low-latency real-time inference without reliance on cloud infrastructure. Feasibility and functionality were assessed via structured questionnaire, prospective observation, and blinded retrospective review of the recordings of the endoscopic surgical video feed and wider operating room environment. Continuous multi-stakeholder feedback through validated human factors surveys drove iterative technical refinements between cases. Routine clinical outcomes, aligning with the standard pituitary surgery core outcome set, were collected. Results: Eight patients with pituitary adenomas were enrolled. The CVAI system was successfully deployed in six cases, demonstrating acceptable real-time pixel-level sella segmentation accuracy. Deployment failed pre-operatively in two cases owing to a single platform-level boot-configuration issue. Iterative refinement between cases was driven by our experience and surgical team feedback. This resulted in the integration of additional anatomical structure segmentations (e.g., carotid arteries), enhanced model accuracy via training dataset expansion, and hardware firmware upgrades. Multi-stakeholder surveys demonstrated satisfactory system feasibility, usability, and acceptability among the surgical team. Both prospective observation and retrospective video review confirmed the absence of adverse events, including no significant distraction to the primary surgeon, and there were no AI-related clinical complications. Conclusion: This first-in-human early clinical evaluation (IDEAL Stage 1) of real-time pixel-level AI anatomical segmentation in live neurosurgery demonstrates feasibility, showcases iterative system evolution, and reports clinical and human factors outcomes. Future work will include a larger single-centre case series (IDEAL Stage 2a) with more surgical teams to further iterate the system and explore its impact on safety, training and workflow. As the underpinning AI models improve and integrate with other intra-operative navigational technologies, such tools will likely be the cornerstone of intra-operative surgical decision support systems.