Table 3 Summary of methods

Hakime et al. 2010 [19]

Radio-frequency ablation

For image registration, a set of three landmark markers were selected on CT and US. The fusion imaging display mode was then used to overlay US and CT images. The lateral, anterior-posterior, and vertical axes were measured to determine the difference in the tumor’s spatial placement between the US and CT images

Overall maximum difference (Dmax): 11.53 ± 8.38 mm. Dmax was 6.55 ± 7.31 mm with CT performed immediately before VNav versus 17.4 ± 5.18 with CT performed 1–30 days before (p \(\le\)0.0001), Dmax under general anesthesia =7.05 ± 6.95, and without anesthesia = 16.81 ± 6.77 (p \(\le\)0.0015)

Luo et al. 2020 [13]

Laparoscopic liver resection with AR-assisted navigation

An unsupervised convolutional network (CNN) framework uses stereo picture pairs from the laparoscope to assess depth and produce an intra-operative 3D liver surface. Preoperative CT images are segmented into 3D models of the patient’s surgical field utilizing end-to-end predictive V-Net architecture. Live laparoscopic pictures are merged with preoperative 3D models to give the surgeon detailed knowledge of the anatomy

Five operating room in vivo pig trials and four laboratory ex vivo porcine liver tests were used to confirm the accuracy of the suggested navigation system. Re-projection errors (RPE) in ex vivo = 6.04?±?1.85?mm, and in vivo= 8.73 ± 2.43 mm

Aribas et al. 2012 [34]

Percutaneous ultrasonography-guided liver biopsy

In a biopsy cohort study of 1300 patients, 610 biopsies were performed with small size (20G) suction needles and 690 using big size (19G) cutting needles. Needles were evaluated for safety and diagnostic accuracy for a variety of focal liver diseases

Diagnostic accuracy in metastases: small needle group: 85%, big needle group: 96.9%, hepatocellular carcinoma: small needle group: 85.5%, big needle group: 97.9%, with regenerative nodules: small needle group: 75%, and big needle group: 98.9%

Chi et al. 2017 [36]

US-assisted liver biopsies with cutting biopsy needles

All sequential liver biopsies between 2005 and 2014, were included. Severe complications were those that required intervention or hospitalization for two days or longer

A total of 1806 liver biopsies were analyzed. Total complications: 102 (5.6%), severe complications: 31 (1.7%). The number of biopsy passes had no effect on the likelihood of severe complications (Pge0.24). Risk factors: hepatic malignancy (OR: 3.21; 95% CI: 1.18–8.73; p = 0.022) and INR 4.1 or more (OR: 7.03; 95% CI: 2.74–18.08; ple0.001)