Improved assessment of donor liver steatosis using Banff consensus recommendations and deep learning algorithms

•We developed and validated automated deep-learning algorithms for steatosis assessment based on Banff consensus recommendations.•Our algorithm allows for an unbiased automated evaluation of steatosis.•This will enable analysis of steatosis’ effect on organ viability and the identification of clinically relevant cut-offs.•Implementing our algorithm in daily clinical practice will enable more efficient and safe allocation of donor organs.