Plastic Surgery

Explainable XGBoost predicts textbook outcome after free flap oral cancer reconstruction

In 752 single-center free flap reconstructions, textbook outcome occurred in 38.2% of patients. An XGBoost model achieved AUC 0.843, accuracy 0.764, sensitivity 0.733, and Brier score 0.152 on internal validation. Top predictors were albumin, age, operative time, tumor T stage, and neutrophil-to-lymphocyte ratio. SHAP analysis suggested actionable thresholds: age >50, albumin >45 g/L, and operative time >400 minutes for risk stratification. 1

Augmented reality improves technical performance for surgical novices

Systematic review of 11 studies (347 trainees) found AR improved objective technical metrics in most studies. Consistent benefits included error reduction and faster early learning, especially for novices in visuospatial tasks. An expertise-reversal effect occurred: diminishing returns for experienced surgeons. Heterogeneous methods, outcomes, and short follow-up limit conclusions about long-term retention and cost-effectiveness. 2

Protocol: AI-assisted education trial for older head and neck cancer patients

This single-center RCT will randomize 100 postoperative head and neck cancer patients aged ≥60 to AI-assisted versus standard SMS education for 12 months. Planned outcomes include mental health, social support, and multiple quality-of-life measures assessed at five timepoints. Analysis will use intention-to-treat linear mixed models to handle repeated measures and missing data. Generalisability is limited by single-center design and reliance on self-reported outcomes. 3