Noninvasive assessment of steatosis and viability of cold-stored human liver grafts by MRI

PURPOSE: A shortage of suitable donor livers is driving increased use of higher risk livers for transplantation. However, current biomarkers are not sensitive and specific enough to predict posttransplant liver function. This is limiting the expansion of the donor pool. Therefore, better noninvasive tests are required to determine which livers will function following implantation and hence can be safely transplanted. This study assesses the temperature sensitivity of proton density fat fraction and relaxometry parameters and examines their potential for assessment of liver function ex vivo. METHODS: Six ex vivo human livers were scanned during static cold storage following normothermic machine perfusion. Proton density fat fraction, T(1), T(2), and [Formula: see text] were measured repeatedly during cooling on ice. Temperature corrections were derived from these measurements for the parameters that showed significant variation with temperature. RESULTS: Strong linear temperature sensitivities were observed for proton density fat fraction (R(2) = 0.61, P < .001) and T(1) (R(2) = 0.78, P < .001). Temperature correction according to a linear model reduced the coefficient of repeatability in these measurements by 41% and 36%, respectively. No temperature dependence was observed in T(2) or [Formula: see text] measurements. Comparing livers deemed functional and nonfunctional during normothermic machine perfusion by hemodynamic and biochemical criteria, T(1) differed significantly: 516 ± 50 ms for functional versus 679 ± 60 ms for non-functional, P = .02. CONCLUSION: Temperature correction is essential for robust measurement of proton density fat fraction and T(1) in cold-stored human livers. These parameters may provide a noninvasive measure of viability for transplantation.