04Nov 2023
Deep-learning based pathological assessment of frozen procurement kidney biopsies to predict graft loss

Deep-learning based pathological assessment of frozen procurement kidney biopsies to predict graft loss

This study leverages deep-learning models to analyze hematoxylin & eosin-stained sections from 2431 deceased-donor kidneys. It establishes a Kidney Donor Quality Score (KDQS) based on features like Sclerotic Glomeruli and Arterial Intimal Fibrosis. The KDQS, combined with recipient and peri-transplant factors, predicts graft loss, aiding in organ utilization decisions. The models could mitigate unnecessary kidney discard, enhancing risk stratification.

  • #nephrology

Like

Save

Share