Artificial Intelligence
Artificial Intelligence (AI) technology has the potential to facilitate surgical decision making
Our initial project is to train an AI algorithm to quantitatively and reproducibly evaluate steatosis and other features of donor biopsies to assess their impact on transplantation.
Specifically, we aim to 1) test the ability of AI models to determine the degree of steatosis; 2) recalibrate the association of steatosis with post-transplantation outcomes, and 3) to train AI to identify additional features of biopsies associated with post-transplantation outcomes.
Models will be tested on additional slides and compared to the pathologists’ labeling. Scoring by the AI model will be correlated with early post-liver transplantation outcomes to build a predictive model of the impact of steatosis on outcomes. We hope that these results will be to enable identification of livers suitable for transplant and reduce wrongly discard grafts. Subsequently, we will use this AI infrastructure to address other surgical challenges.
Related Links
Narayan RR, Abadilla N, Yang L, Chen SB, Klinkachorn M, Eddington HS, Trickey AW, Higgins JP, Melcher ML. Artificial intelligence for prediction of donor liver allograft steatosis and early post-transplantation graft failure. HPB (Oxford). 2022 May;24(5):764-771. doi: 10.1016/j.hpb.2021.10.004. Epub 2021 Nov 1. PMID: 34815187.
Yang L, Ghosh RP, Franklin JM, Chen S, You C, Narayan RR, Melcher ML, Liphardt JT. NuSeT: A deep learning tool for reliably separating and analyzing crowded cells. PLoS Comput Biol. 2020 Sep 14;16(9):e1008193. doi: 10.1371/journal.pcbi.1008193. Epub ahead of print. PMID: 32925919.
Dr. Marc Melcher presents "Algorithms to Facilitate Surgical Decision Making: Kidney Matching, Trainee Matching, Liver Pathology" at Stanford Surgery grand rounds on May 19, 2020.