Accuracy of magnetic resonance imaging for preoperative detection of portal vein thrombosis in liver transplant candidates

The detection of main portal vein thrombosis (PVT) on preoperative imaging of liver transplant candidates has important technical implications for the transplantation procedure. Data are scarce regarding the accuracy of magnetic resonance imaging (MRI) at detecting PVT. The aim of our study was to compare preoperative findings of the portal vein on MRI to operative findings at liver transplantation. Abdominal MRI and clinical records of 172 consecutive patients who received liver transplants between January 1999 and September 2004 were reviewed. Two radiologists independently evaluated the last abdominal magnetic resonance examinations obtained before liver transplantation, blinded to the original reading, operative findings, and clinical data. Findings on MRI were compared with intraoperative findings at transplantation. Main PVT was detected in 12 patients, in whom 8 were found to have thrombus at surgery, with 6 requiring a jump graft or thrombectomy. Sensitivity and specificity of MRI for detecting main PVT were 100% and 98%, respectively. The cause of discordance between findings on MRI and at transplantation in 2 cases was a diminutive caliber of the main portal vein that was interpreted as recanalized chronic thrombosis on MRI. In conclusion, in our study group MRI detected PVT in all liver transplant recipients requiring jump grafts at transplantation. The major reason for a false-positive MRI was a diminutive but patent portal vein.