(306e) A Fitness Index for Transplantation of Perfused DCD Rat Livers

An estimated 4,000 patients perish every year due to lack of transplantable livers. The estimated pool of donor after cardiac death (DCD) livers that are marginally damaged is approximately 6,000 grafts per year, which can be used to resolve the crisis in organ availability. In DCD organs, the presence of cardiac death means that the organs recovered go through a phase of absence of blood supply and anoxia, also called ischemia. Ischemia is a very severe mode of injury and for the liver beyond 30 min graft survival in the recipient falls dramatically precluding the use of most DCD organs in transplantation. Preclinical studies with machine perfusion systems indicate that DCD livers can be successfully recovered from ischemia.

The aim of this work is the development of an index of fitness for transplantation for perfusion-recovery of marginally damaged DCD livers based on metabolic activity during machine perfusion.

A particular mode of machine perfusion, Normothermic extracorporeal Machine Perfusion (NMP) aims to mimic the in vivo conditions of the organ in order to maximize treatment efficacy during perfusion-preservation. It is exceptionally well-suited to monitor the dynamics of recovery from ischemic injury to evaluate the viability of the graft for transplantation; this is because the perfused organ is in a fully metabolically active state, and this activity can be evaluated and compared to an ideal organ to determine the likelihood of success. Such a quality-control step is a critical need in utilization of DCD organs to ensure that the perfused organ can be transplanted with minimal concerns.

Currently the graft quality is assessed based on gross morphology and histology, which are subjective and ineffective to distinguish between viable and nonviable DCD livers. A quantitative and objective index of fitness for transplantation relative to ischemic injury that is based on organ function is highly desirable. A suitable framework for developing such an index relative to ischemic condition is multivariate statistical process monitoring (SPM). The liver metabolism is a complex network which features many correlated metabolites. SPM look at the whole system, identify commonalities between different perfusions, correlations among variables and trends in time and hence are ideal.

A multiway principal component analysis liver perfusion model captures the correlation structure between the metabolites during perfusion of fresh livers that were later successfully transplanted and defines the confidence limits based on the metabolite concentrations in the perfusion medium. The differences in functioning of DCD livers are then evaluated by projecting the metabolite concentrations for the damaged livers onto the fresh liver model; this can be done on an online basis so that the recovery from ischemic injury is evaluated quantitatively during perfusion allowing interventions to the organ if necessary. The squared prediction error statistic quantifies the similarities of ischemic livers to fresh livers and hence constructs an index of ischemia.

Further, an online multiway partial least squares discriminant analysis model predicts the end-of-perfusion quality of a liver during perfusion. This complements the ischemia index by classifying any given liver as ischemic of fresh and forms the basis for index of transplantation of DCD livers. Once the quality of a given liver is determined as ischemic; its fitness for transplantation is assessed based on the confidence limits for successfully transplanted ischemic livers. The effectiveness of multivariate techniques on modeling dynamic metabolic data and the decision algorithm for the evaluation of the ischemic condition and the fitness for transplantation are demonstrated and tested on independent livers.