Neurosurgery

Predictive model guides extent of resection to reduce short‑term lower cranial nerve dysfunction after medullary glioma surgery

A clinical‑imaging logistic model predicted short‑term postoperative lower cranial nerve (LCN) dysfunction with strong discrimination (training AUC 0.85, testing AUC 0.89). Independent predictors included extent of resection (OR 1.84), infiltrative growth (OR 10.46), preoperative LCN impairment (OR 6.79), and cervical cord involvement (OR 4.64). Prospective validation (n = 45) supported model calibration and decision‑curve utility for individualized risk estimates. Stratified planning suggested low‑risk patients tolerate gross‑total resection, whereas high‑risk patients may benefit from limited resection to preserve function. 1

Preoperative dementia increases 30‑day mortality and adverse postoperative functional outcomes in adults ≥75

In a US NSQIP cohort (n = 547,310), preoperative dementia (7.9%) was associated with higher 30‑day mortality (aRR 1.64) after multivariable adjustment. Dementia strongly increased postoperative delirium (aRR 3.09) and modestly increased loss of independence (aRR 1.24). Causal mediation found delirium mediated about 46% of dementia‑related mortality, while loss of independence mediated 26%. Dementia‑associated mortality was greatest in elective cases and among patients with lower measured frailty. 2

Protocol: AI‑assisted personalized education vs standard SMS after head and neck cancer surgery in older adults

This single‑center RCT will randomize 100 patients aged ≥60 post head‑and‑neck cancer surgery to AI personalized education versus standardized SMS for 12 months. Primary outcomes include multiple validated mental health, social support, and quality‑of‑life measures assessed up to 12 months. Analysis will use intention‑to‑treat linear mixed models to compare groups across five timepoints. Authors note single‑center design and self‑reported outcomes may limit external validity pending trial results. 3