The patient is usually placed supine, and the surgeon stands either between legs or on the patient’s left side. A total-abdominal approach is mostly used, but some authors have proposed the introduction of intercostal trocars to improve the angle of surgical instruments relative to the hepatic veins, especially for dissection high in the dome area. The Osaka group, led by Inoue, published their series of 29 patients treated with (n11) and without (n18) intercostal trocars. They report no difference in patient outcomes (including overall complications, R0 resection, hospital stay, blood loss) with the exception of conversion to open surgery: 0% of patients in the intercostal group vs. 38.9% of patients in the “conventional” group (p = 0.026). Chiow et al. instead compared patients treated with and without intercostal trocars and found no difference but for hospital stay: median 2 vs. 6 days, respectively (p 0.032) (33). Both studies compare their current technique using intercostal trocars with their previous experience, and this represents a serious limitation to interpretation of results as the latter may have been affected by improvements in technical expertise or technology over time (10, 33). Analysing specific intercostal trocar-related complications, Hayashi et al. report their experience with 32 patients with only 1 clinically silent minor pneumothorax (34). A chest drain was left in place after the procedure in only 2 cases due to air infiltration in the thoracic cavity. IOUS application to the Glissonean-first approach is fundamental to reaching G8 in the transparenchymal and transfissural approaches but can also be useful for the hilar approach to guide dissection along Glissonean pedicles.

The first route is hilar: dissection starts at the hepatic hilum with identification of the portal bifurcation (7) ( Figure 2 ). The right Glissonean pedicle is followed until its own bifurcation in the right anterior and posterior pedicles. The right anterior pedicle is clamped to confirm its area of perfusion. It is then followed further until its separation into the SV and SVIII pedicles. The SVIII pedicle is identified, confirmed and divided. The approach was described in detail by Jang et al. in 2017 (12). This technique holds the advantage of unequivocal identification of G8 through a structured step-by-step, secure path. When the right and right anterior pedicles are relatively short, this technique may be ideal. Yet navigation of main Glissonean pedicles may not be risk-free and major vascular accidents are always possible. Small branches may also suffer injury, and this in turn may lead to biliary complications such as leakage or delayed biliary stricture. Furthermore, if the intrahepatic course of these two branches is long, the hilar approach may prove difficult and time-consuming.

Kim et al. have proposed a different approach, named “transparenchymal” ( Figure 3 ). In this case, the right anterior Glissonean pedicle is identified through IOUS on the ventral liver surface and a small bridge of parenchyma is divided to access the vascular space (13). G8 is then identified, and the operation proceeds with the steps already described. This approach has some theoretical advantages over the “hilar approach”: it is faster, it involves less parenchymal disruption and most importantly it avoids hilar dissection. The latter can be of particular importance for patients who are likely to undergo further liver surgery (e.g. HCC patients who might suffer recurrence or eventually undergo transplantation) as preserving an intact hilum reduces chances of hilar adhesions and reduces difficulty of dissection and risks of vascular/biliary accident. This approach may theoretically be simplified by exploiting Sugioka’s Laennec’s capsule-based technique to gain easy access to the right anterior Glissonean pedicle with no or minimal parenchymal disruption, yet it has never been described except by Sugioka himself in an untranslated Japanese article (35, 36). A recent video article describes in detail this approach to the Glissonean pedicle (37).

A third possibility is represented by the transfissural approach (38) ( Figure 4 ). The main hepatic fissure is marked on the liver surface, using the MHV as the main IOUS landmark or alternatively by clamping the right Glissonean pedicle (14). Parenchyma is divided at this level, ultimately opening the liver like a book, eventually landing on the MHV. This procedure helps in the identification of G8 (which may still be troublesome due to anatomical variations) and most of all helps with the ensuing dissection along the MHV and anterior border of the IVC, guaranteeing a comfortable access. Ogiso et al. have gone a little bit further, with their valuable anatomical and surgical studies. Their work, with 3D reconstruction and simulation modelling on CT images of 40 patients, suggests that G8 can be found quite constantly behind an SV branch of the MHV (v5) (39). Therefore, they have standardised a “reliable and reproducible” Glissonean transfissural v5-guided technique for anatomical SVIII resections (18). Their main point of reference is v5, identified with IOUS before start of dissection and constantly reassessed during dissection of the main fissure. Once v5 is reached and sacrificed, easy access to G8 will be acquired. The authors state that this renewed anatomical knowledge may indeed be key to dissemination of laparoscopic SVIII resection. The same group has found this technique to be suitable even when dealing with large tumours at the hepatocaval confluence (19). They report that complete opening of the main fissure facilitated dissection of the dorsal side of the mass.

The Glissonean approach ensures radical, anatomical resection of SVIII along correct planes by meticulous identification of vascular inflow. A further advantage is that sub-segmental resection can be achieved: dissection of vascular structures can be extended to permit identification of sub-segments/cone units. A preoperative accurate study of CT images and possibly 3D reconstruction is clue to planning these operations. When a lesion is located within anatomical borders of a sub-segment, its resection may be ideal in its dual scope of removing the whole tumour-burdened portal territory and widening margins while conserving parenchyma. Both dorsal and ventral SVIII resections have been reported by Kim (13, 14) ( Figures 3 , 4 ).

Table 2 summarises results for isolated LASVIIIR. Data are available for 71 patients for isolated LASVIIIR with a Glissonean-first approach (7–10, 12–14, 18, 19). Results are encouraging with major morbidity reported only in 3% (2 patients with bile leak), no mortality. Margin status was available in 47 patients and R0 resection achieved in 91%. No long-term oncologic results are available, for any indication.

Whatever the technique to access G8, the Glissonean approach carries some drawbacks: possible injury to small branches and consequent biliary complications, division of SV venous drainage and consequent SV congestion and the difficult dissection of RHV, MHV and anterior IVC surface from below—the caudal laparoscopic view.

The patient can be placed either supine or in the left lateral decubitus position to facilitate access to the right hemi-liver. Right subcostal trocars are placed, with the surgeon standing alternatively between legs and on the right side of the patient. Although proponents report an easier dissection, the cranio-caudal approach still results technically difficult when performed with instruments that are coming from below, approaching veins in a parallel direction. This may increase the chances of injury to branching points of the vein. To circumvent this obstacle, a further technical modification has been introduced. Ishizawa et al. were the first to describe the use of intercostal trocars to obtain a lateral view which may be more convenient to approach the hepatic veins in a cranio-caudal fashion (11) ( Figure 6 ). These trocars also allow for a non-parallel approach to major venous dissection. The patient is placed in the left lateral decubitus position, and insertion of intercostal trocars is performed under direct vision. Three main techniques are described: in any case, much attention should be given to place the trocar centrally in the intercostal space to avoid intercostal vessels, which run along the inferior costal margin. The first is to use trocars through the ribs but below the diaphragm, with direct access to the abdominal cavity. These trocars though will be quite low and be less helpful in the dissection. Another technique used for example by Ishizawa et al. is to compress the diaphragm against the chest wall with a forceps to exclude the lung from the pleural space around the trocar insertion site, without deflating the lung (11). The third option is to use an optical trocar to gain access to the thoracic cavity and then pierce through the diaphragm, while collapsing the right lung (10). Insertion is preferentially preceded by IOUS to visualize the target lesion and by liver mobilisation on demand (34). Two or three intercostal trocars may be placed, either in the same or in different intercostal spaces (usually between the 6th and 10th). Ports are usually 5 mm, to minimize complications, as the intercostal space may be as small as 15 mm and placing a 10–12-mm trocar may increase bleeding risk (10). Most authors advocate the use of balloon ports so that the inflated balloon may pull the diaphragm against the chest wall, enlarging the field of view. After hepatectomy, aspiration of gas in the pleural space is important, and so is suturing of the phrenic incisions from below. Ome et al. conclude that the use of intercostal trocars makes the cranial approach safer and more feasible (20). It seems reasonable to consider the practice safe, yet a relevant limitation of this procedure is that the hepatobiliary surgeon may not be familiar with intercostal trocar placement, increasing the likelihood of thoracic complications. Furthermore, the “needless” violation of a different anatomical district gives concerns for oncological catastrophe. IOUS is vital for the cranio-caudal approach and represents the surgeon’s companion along the whole operation. In particular, as most SVIII dissection is carried out before G8 division, it cannot be based on ischemic demarcation but is based mostly on the direction outlined by IOUS detection of RHV and MHV.

Surgery starts with right liver mobilisation, and the RHV is fully exposed from all sides by division of the right hepatocaval ligament. Dissection and encirclement of RHV and MHV roots usually follow. Extrahepatic control is not strictly necessary, and some authors do not perform this manoeuvre, although the potential vascular risk has been an object of debate (41). Once vascular outflow control is achieved, and after IOUS-guided marking of the venous plane, transection follows the hepatic veins down into the parenchyma, exposing the bare venous surface. Dissection may either start from the MHV, effectively opening the main hepatic fissure until the root(s) of G8/G8 branches (11, 15, 21), or begin from the RHV and then proceed medially along the anterior IVC surface (16). The latter approach is sometimes referred to as the “lateral” or “dorsal” approach ( Figures 5 , 6 ). G8 is identified late in the operation when at least one side of the resection has been conducted. Nonetheless, Ome et al. describe their MHV-first approach as “Glissonean” as they divide G8 before initiating other transection planes (20). Supporters of this technique argue that beginning dissection from above provides an easier plane to follow as it is “clear and shallow” and safer as hepatic veins have thick walls and fewer branches at this latitude and exposure of the IVC are also facilitated (21). Furthermore, when dissecting HVs from the root, finding a plane to preserve Laennec’s capsule seems possible, thereby conserving HV parietal resilience (40). Furthermore, due to the anatomical relationship between main hepatic veins and their branches (conflating at acute angles inferiorly), a caudal approach is prone to procuring “split-injuries” while a cranial approach is more likely to produce “pulled-up” injuries, which are considered easier to control, with simple pressure (40). Honda et al. and Ome et al., who use a caudal approach in their routine practice, suggest that using the instrument tip in a back-scoring (root to periphery) fashion may be a good trick to reduce/avoid split injuries (20, 42). Some authors argue that the transection line is easily kept blood-free as blood flows inferiorly by gravity. If haemorrhage from a hepatic vein does occur, decreasing airway pressure (by temporarily pausing ventilation) helps in reducing central venous pressure, and this aids bleeding control (43). Finally, outflow control coupled with the standard Pringle manoeuvre allows for greater confidence in haemostatic control. On the other hand, late G8 control and dissection along landmark yet unconfirmed planes may lead to less precise anatomical resections and to increased remnant liver ischemia. This may be particularly true when dealing with the 25% of cases where the classical right anterior separation into a cranial SVIII and a caudal SV is replaced by a ventro-dorsal tertiary branching (4, 5). Lopez-Ben et al. have also reported the possibility of SVIII dorsal resection with a lateral approach as G8 dorsal can be found in proximity of the RHV, although there is no clear landmark and G8dorsal identification may be problematic (16).

Table 2 summarises results for isolated LASVIIIR. Data are available for 27 patients for isolated LASVIIIR with a cranio-caudal approach (11, 16, 20, 21). Results appear superlative with no morbidity nor mortality. Margin status was available only in 19 patients and R0 resection achieved in 100%. No long-term oncologic results are available, for any indication. Ome et al. published a very interesting study comparing 7 and 19 patients treated with the caudo-cranial and cranio-caudal approaches, respectively (20). There were no significant differences between groups, although numerosity is low. Interestingly, two patients in the caudal approach had R1 resection and two had postoperative bile leak while all patients in the cranial group had R0 resections and no biliary complications. The authors report their impression of a cranial approach facilitating obtainment of free surgical margins and reducing biliary injury (either direct or ischemic).

Theoretically, the cranio-caudal hepatic vein dissection is not necessarily an alternative to a Glissonean-first approach. The two approaches could be combined with G8 controlled followed by root-to periphery dissection, although this has not yet been described for laparoscopic SVIII resection to the best of our knowledge.