Low graft attenuation at unenhanced CT: Association with 1-month mortality or graft failure after liver transplantation

Abstract

Purpose: To investigate whether low graft attenuation at unenhanced computed tomography (CT) is associated with 1-month mortality or graft failure after liver transplant and determine its diagnostic performance. Materials and Methods: Included were 663 recipients who underwent CT imaging within 7 days after liver transplant between December 2014 and August 2016. Initial poor function (IPF) was diagnosed by using a combination of laboratory values within 7 days after liver transplant and subdivided patients into primary and secondary IPF. At 1 month after the operation, mortality and graft failure or survival in recipients was categorized. Two radiologists who were blinded to clinical data retrospectively and independently evaluated graft attenuation on unenhanced CT images (high or isoattenuation, graft attenuation greater than or equal to that of spleen; low, graft attenuation less than that of spleen). The interobserver agreement was evaluated by using intraclass correlation coefficient and k statics. Incidence of low graft attenuation between recipients with IPF and those with normal function was compared by using x2 test. The relationship between graft attenuation and outcome in primary and secondary IPF was evaluated by using log-rank test. Results: Of 663 recipients, 114 had IPF (80 primary; 34 secondary). After 1 month, 11 had graft failure or died, whereas 652 survived. Low graft attenuation was more common in patients with IPF than in normal-function patients (P < .001). In the primary group (those without identifiable cause), 15 patients had low graft attenuation, which led to mortality or graft failure within 1 month in seven of those patients. No recipient with high or isoattenuation had 1-month mortality or graft failure (P < .001). The secondary group (those with identifiable cause) showed no significant association between graft attenuation and 1-month mortality and graft failure (P = .181). Values of low graft attenuation for 1-month mortality and graft failure in primary IPF were positive predictive value, 46.7%; negative predictive value, 100%; sensitivity, 100%; specificity, 89.0%; and accuracy, 90.0%. There was excellent interobserver agreement in the assessment of graft attenuation (intraclass correlation coefficient, 0.957; κ = 1.00). Conclusion: Low graft attenuation can be associated with 1-month mortality or graft failure in liver graft recipients with primary IPF. © 2017 RSNA.