Split graft size is an important factor in SLT. Splitting at the falciform ligament yields LLS and RTS grafts (Fig. 1, line A). The LLS is generally suitable for pediatric recipients. When a small infant is the recipient, graft-to-recipient weight ratio (GRWR: [liver graft weight ÷ recipient body weight] × 100) should not exceed 4–5 % to avoid large-for-size-related complications, such as open abdomen and vascular thrombosis. If this is the case, the LLS split graft has to be further reduced to avoid such problems (Kasahara et al. 2003). The RTS graft size, on the other hand, is in most instances large enough to avoid small-for-size-related graft dysfunction in adult recipients.

In hemiliver splitting for two adult-sized recipients, the liver is split on the right side of the middle hepatic vein (Fig. 1, line B). Determination of graft size is crucial to decide whether splitting is feasible and to minimize the possibility of small-for-size-related graft failure. Although the minimal graft size to meet recipient’s metabolic demand in living donor liver transplantation is considered to be as small as a GRWR of 0.6–0.8 %, the minimal ratio remains unknown. SLT appears to require a higher GRWR to compensate for suboptimal graft quality related to longer cold ischemia time and donor hemodynamic instability associated with brain death. Accordingly, a GRWR of 1.0 % seems to be the minimal requirement in SLT to avoid early graft dysfunction (Lee et al. 2013).

Imaging studies are rarely available in SLT to estimate graft weight and evaluate donor liver anatomy. Therefore, this important surgical information is most often unknown until the time of organ recovery or after a liver is taken out of a deceased donor. Using the donor body surface area, whole liver volume (mL) can be estimated using equations 1072.8 × body surface area (m²) − 345.7 for Caucasians (Heinemann et al. 1999) and 706.2 × body surface area (m²) + 2.4 for Asians (Urata et al. 1995). More simply, whole liver weight can be estimated as 2 % of donor body weight (Lee 2010). These estimated values can be divided into standard estimates for lobar distribution (35 % for the left lobe and 65 % for the right lobe) to estimate hemiliver graft size.

For successful SLT, recipient selection is as important as donor selection. In SLT using the LLS graft for pediatric recipients, graft-recipient size mismatch resulting in large-for-size complications should be avoided. When the recipient is an older child, the LLS graft might not be enough to provide adequate liver mass. In such an instance, the left lobe graft is necessary to achieve a GRWR > 1.0 %. For the RTS graft, recipients can be chosen more liberally, similar to when a whole liver graft is used.

SLT for adults has the potential risk of small-for-size syndrome, particularly with hemiliver grafts. Generally, the best SLT recipients for a hemiliver graft are an adolescent or a small adult with minimal portal hypertension and/or a relatively low MELD score, particularly for the left lobe graft. Although a recipient with a high MELD score can be transplanted with a hemiliver split graft, the data are not available to support the routine use of hemiliver grafts for high-risk recipients (Nadalin et al. 2009; Hashimoto et al. 2014). When a recipient has significant portal hypertension, a larger right lobe graft is preferred in order to lower the risk of small-for-size syndrome. In addition to examining medical history, the severity of portal hypertension can be assessed using a triphasic CT scan or MRI (Aucejo et al. 2008). These imaging studies show the recipient’s surgical anatomy and also can show portosystemic shunt, portal vein thrombosis, and stenosis of the celiac trunk, which are important pieces of surgical information. The management of portosystemic shunt is controversial (Ikegami et al. 2013). When recipients have a large spontaneous or surgical portosystemic shunt, the shunt can cause hypoperfusion of a transplanted split graft due to a steal phenomenon. In contrast, it also helps lower portal vein pressure to favorably accept a small partial graft that is damaged by portal hyperperfusion. Accordingly, a case-by-case assessment is important to determine whether to close shunts in recipients.

In addition to thorough donor and recipient selection, appropriate donor-recipient paring is crucial to achieve good outcomes in SLT. In adult SLT, split grafts are generally taken from larger donors and transplanted into smaller recipients. This graft-recipient paring enables the majority of recipients to achieve a GRWR > 1.0 % (Hashimoto et al. 2014), representing a size advantage that helps avoid small-for-size syndrome.

The use of split grafts for high-risk recipients is controversial (Nadalin et al. 2009; Hashimoto et al. 2014). Under the philosophy of the “sickest first” MELD allocation, standard criteria donors who are suitable for bipartition are allocated to those recipients with a high MELD score who are generally unsuitable for SLT.

When a splittable donor becomes available, the most important factors determining whether to proceed with SLT are when a whole donor liver is deemed to be too large to fit a primary adult candidate or a small pediatric recipient is on the waiting list. SLT has proven to be a great benefit for pediatric candidates who usually need an LLS graft without compromising survival in adult recipients receiving the RTS graft (Maggi et al. 2015). It is equally important, however, that small adults who are often bypassed on the waiting list due to size mismatch can have more opportunities by SLT. For these recipients, split grafts can provide enough liver volume to tolerate portal hyperperfusion. The remaining split graft can be used for a candidate with minimal portal hypertension and a lower MELD score. This graft-recipient matching helps achieve excellent survival after SLT under the allocation system where the MELD score regulates transplant priority. However, such ideal matching is difficult to achieve on a routine basis, and this is why many centers underutilize or do not use split grafts, particularly when hemiliver splitting is indicated. According to the Cleveland Clinic experience from April 2004 to June 2012, 137 out of 1089 deceased donors (12.6 %) met the SLT criteria and were identified as suitable for splitting. However, among these splittable donors, only 38 (3.5 %) were used for SLT because suitable recipients were not available.

An important technical challenge in SLT is a lack of consensus between transplant centers regarding surgical techniques, particularly sharing patterns of major vessels and bile ducts between two split grafts. The ideal sharing pattern was originally described by Bismuth in 1989 (Bismuth et al. 1989). The principle concept of this technique is to avoid multiple branches to be reconstructed in the recipient operation. Impeccable knowledge of surgical liver anatomy is essential to understand this sharing pattern. The left lobe frequently has a single branch of the portal vein, hepatic duct, and venous outflow that is a common channel of the left and middle hepatic veins, but multiple branches of the hepatic arteries often exist. The right lobe, on the other hand, often has a single right hepatic artery and multiple branches commonly seen in the venous drainage, hepatic duct, and portal vein. According to the sharing pattern by Bismuth, the left lobe retains the celiac trunk, and the right lobe retains the remaining major structures, including the common hepatic duct, main portal vein, and vena cava. Although typically the priority is for the primary recipient to keep necessary structures in the graft allocation, sharing should depend on actual donor anatomy and recipient needs. The final decision should be made with flexibility and agreement by both teams who each take one of the split grafts.

As long as both sides of the split grafts have a complete set of inflow and outflow vessels and biliary drainage, anatomical variations are not considered to be a contraindication to splitting. Recipient surgeons must decide on the division of these vital structures to make the liver graft safely usable in the recipients. The following are relatively common anatomical variants seen in organ recovery: