Cardiothoracic Surgery

Randomized trials show ML/DL decision-support shortens workflow and lowers mortality

Systematic review of 32 randomized trials found ML/DL interventions improved workflow efficiency (SMD -0.71). Diagnostic time shortened by 30–120 s, and some trials reported 1.0–4.2 fewer hospital days. Behavioral nudges increased medication adherence (RR 1.59; NNT 12). Decision-support AI reduced all-cause mortality (RR 0.84; NNT 32), though blinding and sham-AI controls were limited. 1

Protocol: individualized intraoperative systolic targets for older hip fracture patients

Single-center randomized trial will enroll 180 patients aged 65–85 undergoing hip fracture repair under general anesthesia. Intervention targets systolic BP within ±10% of baseline versus reactive standard care with universal MAP ≥65 mmHg. Primary outcome is postoperative organ dysfunction within 7 days; secondary outcomes include ICU stay and 30-day mortality. Results will determine whether proactive individualized BP targets reduce organ dysfunction compared with standard management. 2

EXCOR Active Driver shows high short-term survival and no major device malfunctions

Prospective multicenter evaluation enrolled 40 IDE patients plus 118 continued-access patients using the EXCOR Active Driver. No major device malfunctions were reported across cohorts, supporting device reliability. At 90 days among IDE patients, 65% remained on support and 17.5% were transplanted; stroke incidence was 12.5% and mortality 0%. Overall 90-day survival in continued-access patients was 98.1%, though stroke and conversions to other support occurred. 3