AI finally provides augmented intelligence to liver surgeons
This study answers to a real need as every liver surgeon knows that MS is a major predictor of early graft dysfunction and that macroscopic evaluation poorly correlates with MS content. Despite its monocentric design and small sample size, this work represents a potential breakthrough in the field of liver transplantation, as we were, so far, facing a paradox: pathological evaluation is still considered as the gold standard way to assess the MS content of grafts while it has been reported many times that frozen section analysis was not reliable [ 2] and that there was a huge inter/intra-operator variability . For these reasons, we really needed a reproducible and accurate tool for MS quantification and that may be what this US team is proposing here. Moreover, many software performing automatic calculation of MS content are already available, but none (or almost none) allows such estimation from frozen section slides, thus limiting their clinical utility. The present work clearly represents a significant advance for clinical acceptance and daily use of this equipment, regardless of the time of the day.
Deep learning quantification of percent steatosis in donor liver biopsy frozen sections
Pathologist evaluation of donor liver biopsies provides information for accepting or discarding potential donor livers. Due to the urgent nature of the decision process, this is regularly performed using frozen sectioning at the time of biopsy. The percent steatosis in a donor liver biopsy correlates with transplant outcome, however there is significant inter- and intra-observer variability in quantifying steatosis, compounded by frozen section artifact. We hypothesized that a deep learning model could identify and quantify steatosis in donor liver biopsies.