Ability of 3-D liver reconstruction to estimate liver resection volume after anatomical minor resection: a...

Estimation of liver resection volume (LRV) is a key step to plane safe liver surgery. Modern 3D liver reconstruction software (3-D) allows to calculate LRV based on the portal blood supply, overcoming some limits of the conventional hand-trace method. The aim of this prospective study was to evaluate the ability of 3-D to estimate the LRV after minor anatomical resections (mAR). The consistency of virtual LRV (vLRV) and real weighted specimen (rLRV) was evaluated. Factors affecting the median discrepancy between vLRV and rLRV were analyzed. Exclusion criteria included inadequate contrast-enhanced computed tomography, left lateral sectionectomy, and changes in the surgical plan based on intraoperative ultrasound findings. Thirty-five consecutive mARs were analyzed: 4 subsegmentectomies, 9 segmentectomies, and 22 bisegmentectomies. A strong positive correlation was found between vLRV and rLRV (r = 0.945, p < 0.001). The median vLRV and rLRV were 236 mL and 180 mL, respectively. The median discrepancy between vLRV and rLRV was - 38 mL, indicating a slight tendency of the 3D software to overestimate LRV. The median discrepancy was greater in cases of large subglissonian lesions (> 3 cm) (65 mL vs. 22.5 mL for other lesion types, p = 0.028) and bisegmentectomies (60.5 mL vs. 16 mL for segmentectomies/subsegmentectomies, p = 0.001). Multivariate analysis confirmed that bisegmentectomy was the only factor independently associated with increased discrepancy [RR 2.724 (12.8-88.9), p = 0.010]. 3D software provided accurate predictions of liver specimen volume in patients who underwent minor anatomical resections.