Optimal Doppler Ultrasonographic and CT Criteria for Right Hepatic Vein Stenosis at Anastomosis in Patients after Living Donor Liver Transplantation with Right-Lobe Graft: Receiver Operating Curve Analysis

  • Dr Hye Jeon Hwang, Asan Medical Center, University of Ulsan College of Medicine, Korea
  • Dr Kyoung Won Kim, Asan Medical Center, University of Ulsan College of Medicine, Korea, Democratic People's Rep
  • Dr Woo Kyoung Jeong, Hanyang University Guri Hospital, Korea
  • Dr Ki-Young Go, Asan Medical Center, University of Ulsan College of, Korea
  • Dr Gi-Won Song, Asan Medical Center, University of Ulsan College of Medicine, Korea
  • Dr Shin Hwang, Asan Medical Center, University of Ulsan College of Medicine, Korea
  • Dr Sung-Gyu Lee, Asan Medical Center, University of Ulsan College of Medicine, Korea

    • Purpose: To establish optimal Doppler ultrasound (DUS) and CT criteria for right hepatic vein (RHV) stenosis after living donor liver transplantation (LDLT) and to compare accuracies of these methods using receiver operating curve (ROC) analysis.
    Materials and Methods: This study was approved by our institutional review board. The study group consisted of 53 patients (M:F=34:19; 50.5±9.9 years) who underwent DUS and CT within 1 week of hepatic venography, following LDLT with right-lobe grafts, between October 2006 and December 2007. On venography, 56 RHVs were classified into stenosis and nonstenosis groups. On DUS, venous pulsatility index (VPI) was defined as difference between maximum and minimum frequency-shifts divided by maximum. On CT, diameters of anastomosis (ØA) and RHV (ØRHV) were measured, and percentage stenosis of RHV anastomosis was calculated as (ØRHV-ØA)/ØRHVX 100. Mean VPI and CT parameters in the two groups were compared, and ROC analysis was performed to compare accuracy and to determine optimal threshold of each parameter.
    Results: 25 RHVs were classified into stenosis group and 31 into nonstenosis group. Mean VPI and mean ØA were significantly lower and mean percentage stenosis was significantly higher in stenosis than in non-stenosis group (P<0.0001 each). On ROC analysis, ØA and percentage stenosis were significantly more accurate than VPI (P=.048, .017). Optimal cut-offs of VPI, ØA, and percentage stenosis were 0.16, 3.7 mm, and 41.7%.
    Conclusion: CT is more accurate than DUS for RHV stenosis in patients after LDLT. Patients suspected of RHV stenosis on DUS should undergo CT to reduce unnecessary venography.