Semi-automatic and automatic segmentation in the preoperative assessment of graft volume before living donor liver...

Accurate preoperative liver volumetry is essential for successful living donor liver transplantation (LDLT). Manual CT volumetry is time-consuming and operator-dependent, prompting the development of semi -automated and automated methods. The aim of this study is to compare the accuracy of one semi-automated (Vitrea, Canon) and two automated (Synapse, Fujifilm and United Imaging, UI) CT volumetry workstations in estimating liver graft volume of the right lobe, left lobe, and left lateral segment relative to intraoperative graft weight.

This retrospective study included 95 living liver donors (67 right lobe, 12 left lobe, and 16 left lateral segments) who underwent LDLT from 2022 to 2024. Estimated graft volumes were compared with actual graft weight (AGW). Correlation, percentage error, interobserver agreement, and potential influences of donors' age and gender on estimation accuracy were assessed.

All three methods demonstrated strong correlation with AGW (p < 0.001). Application Liver Analysis of the workstation Synapse showed significantly higher accuracy in the right lobe volume estimation compared to the others (p < 0.001), as well as in the left lateral segment estimation compared to Vitrea (p < 0.001) and UI (p = 0.006). Interobserver agreement was high, and no significant influence of donors' age or gender was found.

All three volumetry tools are reliable and user-friendly for preoperative planning for LDLT. Among them, Liver Analysis of Synapse demonstrated the highest accuracy in assessing right lobe and left lateral sector volumes, offering potential advantages in surgical decision-making.