Particle.news
Download on the App Store

BIOPREVENT, a Biomarker‑AI Tool, Predicts Post‑Transplant cGVHD and Mortality Months Early

Validated in an independent cohort, the tool is now available as a free research web app pending clinical trials.

Overview

  • Developed by MUSC Hollings researchers with CIBMTR collaborators, the study was published Feb. 16 in the Journal of Clinical Investigation.
  • The model combines seven immune protein biomarkers measured around day 90–100 post‑transplant with nine clinical variables using Bayesian additive regression trees.
  • In testing, biomarker‑informed predictions outperformed clinical data alone, with the largest gains for transplant‑related mortality risk.
  • BIOPREVENT stratified patients into distinct low‑ and high‑risk groups with outcome differences observed up to 18 months after transplant.
  • The web application supports risk assessment and research rather than treatment decisions, with prospective trials planned to test whether earlier intervention improves outcomes.