Transplant Surgery

Topological features improve perioperative mortality prediction after lung transplantation

This retrospective cohort used static and time-dependent variables with topological feature engineering to update mortality risk at transplant, ICU, and early post-op. The model achieved 87.4% accuracy, 84.1% sensitivity, and 89.6% specificity, with an absolute AUC gain of 0.08 over the best non-topological baseline (p < 0.001). SHAP analysis identified early oxygenation trends, immunosuppressive load, and inflammatory markers as primary drivers of predicted risk. Results suggest dynamic topological transforms can yield more accurate, interpretable perioperative risk estimates, but external prospective validation is needed before clinical deployment. 1

Age ≥70 alone not an independent predictor of graft loss after living-donor liver transplant

Analysis of 4,802 Korean LDLT recipients (142 aged ≥70) found lower unadjusted 5‑year graft survival in older recipients (74.4% vs 87.2%). After multivariable adjustment and propensity matching, age ≥70 was not independently associated with graft loss (adjusted HR 1.24; 95% CI 0.83–1.87; P = 0.29). In recipients ≥70, pretransplant hospitalization, graft steatosis ≥5%, and vascular complications conferred higher graft-failure risk (aHRs 1.94, 1.74, and 3.11 respectively). These data support individualized risk profiling rather than categorical exclusion for LDLT candidates aged ≥70. 2

EXCOR Active Driver shows high reliability and acceptable outcomes in pediatric VAD support

A prospective multicenter trial evaluated the EXCOR Active Driver in 40 IDE and 118 continued-access pediatric patients receiving Berlin EXCOR devices. There were no major device malfunctions reported in either cohort and 90‑day mortality was 0% in IDE patients and 98.1% survival in continued-access patients. At 90 days among IDE patients, 65% remained on support, 17.5% transplanted, 15% converted, and stroke occurred in 12.5%. The driver appears safe and reliable in high-risk children, but stroke risk and longer-term performance warrant continued surveillance. 3