|Home | About | Journals | Submit | Contact Us | Français|
Precise clinical knowledge of liver anatomy is required to safely perform a hepatectomy, for both open and laparoscopic surgery. At the 19th meeting of the Japanese Research Society of Clinical Anatomy (JRSCA), we conducted special symposia on essential issues of liver surgery, such as the history of hepatic segmentation, the glissonean pedicle approach, application of 3-D imaging simulation and fluorescent imaging using indocyanine green solution, a variety of segmentectomies including caudate lobectomy, the associating liver partition and portal vein embolization for stage hepatectomy and harvesting liver grafts for living donor liver transplantation. The present review article provides useful information for liver surgeons and anatomic researchers.
The 19th meeting of the Japanese Research Society of Clinical Anatomy (JRSCA) was held at Sanjo-Kaikan (Hilltop Hall) at the University of Tokyo on November 14, 2015. The program consisted of two symposia on clinical anatomy of the liver, which included general remarks and specific information, one symposium on pancreatic anatomy for safe and precise pancreatectomy, a plenary session and other oral sessions on clinical anatomy for hepatobiliary and pancreatic surgery, plastic surgery and oral-maxillofacial surgery.
In this review article, the presentations which were provided in the two symposia on clinical anatomy of the liver are summarized to further the understanding of hepatic anatomy for the clinical application of safe hepatic surgery.
The history of anatomical liver segmentation began at the end of the 15th century, with morphological observations of its external shape, and evolved over 400 years toward the modern concept of liver segmentation. Since then, Rex (1888), Cantlie (1898) and Serege (1901) classified the liver into left and right sides. Rex, in 1888, advocated the concept of the hemiliver based on the anatomy of the portal venous system, and found that the boundary between the right and left hemilivers is located on the line connecting the gallbladder bed and the inferior vena cava (IVC), and the middle hepatic vein (MHV) indicates the position of the intersegmental plane (fig. (fig.1A)1A) . Cantlie independently found the same anatomical characteristics of the liver and predicted clinical application of these segmental characteristics for bleeding control during hepatectomy or portal vein ligation. Therefore, the intersegmental line between either side of the hemiliver has been called ‘Rex-Cantlie's line’ (fig. (fig.1B)1B) .
The intersegmental borders within the liver were then classified by Hjortsjö (1948), Healey (1953) and Couinaud (1954) et al. In 1951, Hjortsjö classified the liver into 6 parts: 3 on the right side and 3 on the left side . Healey and Schroy established the current basic concept of gross liver segmentation, which consists of the left lateral, medial, anterior and posterior areas. In 1953, Healey and Schroy further divided each area into 2 parts, using 100 cast models of biliary trees (fig. (fig.2)2) . His findings were used by Couinaud in 1954 to further the concept of liver segmentation into 8 parts. Couinaud proposed clinically applicable liver segmentation for systemic liver surgery (fig. (fig.3)3) .
The first right hepatectomy was performed by Honjo in Japan in 1949, and then by Lortat-Jacob in France in 1952. These hepatectomies were performed after division of the right hepatic artery and portal vein individually in the glissonean sheath, which is called the controlled method. On the other hand, in the glissonean approach, the glissonean pedicle can be secured and divided without dissection of each vascular anatomy in the hepatoduodenal ligament. The intrahepatic Glissonean approach was first reported by Lin in Taiwan (1960) and Tung in Vietnam (1962). The extrahepatic Glissonean approach was introduced by Couinaud (1985) and Takasaki (1986)  (fig. (fig.4)4) and is applied to laparoscopic anatomical hepatectomy. The root of the third order branches of the glissonean pedicles can be secured by approaching via the hepatic hilum after securing the second order glissonean pedicles, which is defined as corn unit resection (fig. (fig.55).
The intraoperative identification method and functional evaluation for veno-occlusive regions have yet to be established. Kawaguchi used the indocyanine green (ICG)-fluorescence imaging technique to study 45 patients and 18 donors who developed venous occlusion in the remnant livers. Veno-occlusive regions were identified as being territories with lower fluorescence intensity than non-veno-occlusive regions [fig. [fig.6;6; median (range), 99 (39-173) vs. 182 (71-255), respectively, p < 0.001]. The portal uptake of ICG in veno-occlusive regions decreased to approximately 40% of that in non-veno-occlusive regions. The ICG concentration ratio of veno-occlusive regions to non-veno-occlusive regions (median ± standard deviation) was 0.36 ± 0.17.
It is well known that the major hepatic veins run along the liver segment borders, but there are few studies on the drainage area of the hepatic vein and its boundary.
Cho et al. [7,8] advocated the concept of the ‘anterior fissure’, where the anterior segment is divided into ventral and dorsal subsegments. They also proposed that the anterior fissure coincides with the boundary between the drainage area of the MHV and the right hepatic vein (RHV). However, their proposition seems inconsistent because the anterior fissure vein (AFV) runs along anterior fissure and mainly flows into the MHV and sometimes into the RHV.
To investigate the boundary between the MHV and RHV, we assessed computed tomography (CT) data from 32 living liver donors using three-dimensional simulation software (OVA; Hitachi Medico). The AFV flows into the MHV (n = 24, group M) and the RHV (n = 6, group R). In the other 2 donors (group MR), there were two AFVs that separately flow into the MHV and RHV.
In all donors, the anterior portal branch ran along the boundary plane between the drainage area of the MHV and RHV. In group M, the dorsal branch of the portal vein ran along the boundary (fig. (fig.7).7). In contrast, the ventral branch of the portal vein ran along the boundary in group R. In group MR, both the ventral and dorsal branches partially ran along the boundary. The findings are summarized in figure figure8.8. The boundary line on the liver surface between the MHV and RHV drainage area did not coincide with the anterior fissure. In summary, the anterior portal vein and its tributaries run along the boundary plane between the MHV and RHV drainage areas. The boundary does not coincide with the anterior fissure.
Laparoscopic hepatectomy has limitations for visual inspection and manual palpation. Ishizawa showed that fluorescence imaging can be applied more efficiently for laparoscopic hepatectomy. That is because surgeons do not need to turn their eyes from the operative field to a monitor as they do during open hepatectomy. ICG fluorescence imaging enables tumor identification and anatomic segment visualization (fig. (fig.9)9) and, thus, is helpful to intraoperatively understand liver anatomy.
The anatomy or boundary of the caudate lobe has been discussed for several decades, and anatomical resection of the caudate lobe has been considered to be technically demanding compared with resection of other hepatic segments.
In the 1980s, Dr. Kumon began studies on liver anatomy, with a focus on the caudate lobe, when he was a resident at the National Cancer Center Hospital in Tokyo. He made 5 casts at the Cancer Center Hospital and found that the caudate lobe had several branches from the first-order branch of the portal veins and hepatic arteries at the hepatic hilum. In the liver casts, blue, red and yellow silicon rubber was injected in the portal, arterial and biliary systems. The caudate lobe was classified into three parts: the Spiegel lobe, the paracaval portion and the caudate process. Dr. Kumon continued his research at the Kochi Medical School in Kochi, Japan, and studied variations of the portal branches to the caudate lobe using a total of 23 liver casts .
Although the caudate lobe accounts for 2-10% of the total liver volume in adults, Dr. Kumon found that the volume of the caudate lobe was larger in the fetus, probably because the caudate branches have a direct relationship with the umbilical veins. The right-side boundary of the caudate lobe is controversial, but he showed that the caudate limbs should branch off the first branches of the portal veins (i.e. the right and left portal veins). In some cases, thick branches from the anterior portal pedicle run beside the IVC, but they cannot be defined as paracaval branches.
The caudate vein was first described by Rex in 1888. In Japan, Dr. Arai in 1949 and Dr. Toyohara in 1954 used liver casts to show that the number of the caudate veins is 1-5. In 1981, Dr. Nakamura classified the branching pattern of the short hepatic veins in the Spiegel lobe into 3 types . Kogure revealed the relationship among the portal segmentation, proper caudate vein and the external notch of the caudate lobe using 88 liver casts. The notch of the caudate lobe was found in 54% of livers and had a close relationship with the portal segmentation of the caudate lobe and the proper caudate vein (i.e. the main venous branch in the caudate lobe)  (fig. (fig.10).10). He further studied the hepatic veins on the border between the caudate lobe and the right liver using 54 liver casts, and found 3 types of hepatic veins running into the IVC or the inferior RHV  (fig. (fig.1111).
The boundary of the caudate lobe and the right liver is controversial. Maki studied the right-side boundary of the caudate lobe using 3-D simulation customized with CT images of 67 liver transplant donors. He found that on the right-side boundary of the caudate lobe, a small paracaval vein runs along the boundary of the caudate lobe in 89% of livers in addition to the RHV vein, inferior RHV and caudate processus hepatic vein, as defined by Kogure (fig. (fig.1212).
Knowledge of clinical anatomy of the caudate lobe, and especially the boundary of the caudate lobe between the surrounding structures, is required for precise and safe caudate resection. Three points can be raised:
The caudate hepatic artery supplies the dorsal part of the left medial segment in two thirds of the livers. During extended hemihepatectomy for hilar cholangiocarcinoma, some of the arterial branches from the caudate lobe are divided behind the MHV.
The ramification of the portal branches in the caudate lobe does not always coincide with that of biliary branches. Ninety percent of the Spiegel portal branches are from the left portal vein, while 50% of the portal branches in the caudate process are from the right portal vein.
The thickest hepatic vein in the caudate lobe is located at the 1-2 o'clock position around the IVC. The Spiegel lobe surrounds the IVC beyond the dorsal midline in 20% of livers (fig. (fig.1313 and and14)14) .
Left trisectionectomy is indicated in 11% of patients with perihilar cholangiocarcinoma and is associated with greater blood loss, longer operative time and more frequent surgical morbidity than conventional right or left hemihepatectomies (table (table1)1) . Based on the Japanese national registry, surgical mortality associated with this procedure was reported to be 12.1% in 2013 by the Japanese Society of Hepato-Biliary-Pancreatic Surgery, which was the highest mortality rate among all of the hepatectomy procedures.
Left trisectionectomy is associated with: (1) wider transactional area and greater blood loss; (2) less volume of the future liver remnant; (3) fewer bile duct orifices to be reconstructed compared with other hemihepatectomies; and (4) difficulty exposing and preserving the RHV, which is the only major hepatic vein that is preserved. However, in patients with an independent right posterior portal vein, division of the anterior and left portal vein is easy to accomplish, and hepaticojejunostomy can be performed more safely than in patients with conventional anatomy.
Anatomical resection of Couinaud's segment can be recommended in patients with hepatocellular carcinoma. Anatomical resection of segment 8 (S8) consists of 28% of all the other segmentectomies, which shows that S8 is the most popular segment that is resected. S8 accounts for 24.2% of the entire liver volume, which is the largest segment .
During anatomical resection of S8, the ventral and dorsal glissonean branches of segment 8 are divided, and the MHV and RHV are exposed on the transactional plane as the landmark veins. To reveal the boundary of S8, the dye injection method  and the counterstaining technique  are useful. However, it is sometimes difficult to use the glissonean approach to reach the root of G8. Recent introduction of dye-staining using ICG under fluorescence imaging facilitated the visualization of each segment even in livers with insufficient visualization of segments using conventional dye injection techniques .
After anatomical S8 resection, the landmark hepatic veins were exposed on the transactional area in 38% of patients at the University of Tokyo Hospital  (fig. (fig.15).15). The boundary between S8 and S5 is controversial, and the branching variants of the portal and biliary system do not always coincide .
A new two-stage hepatectomy, called associating liver partition and portal vein embolization for stage hepatectomy (ALPPS), to induce hypertrophy of the future liver remnant in patients with advanced hepatic malignancies has recently been developed in Germany . One complication associated with ALPPS is necrotic change of the de-vascularized liver segments after the first stage . During conventional ALPPS, in which liver transection is performed just to the right of the falciform ligament, A4 is completely divided, and collateral arterial supply to segment 4 is key to avoid ischemic complications, but it is not always reliable. Therefore, Tanaka and colleagues proposed a modified ALPPS procedure that preserves all the portal pedicles at the transection plane (fig. (fig.16).16). They found that the modified ALPPS provided comparable hypertrophy of the future liver remnant to the conventional ALPPS.
Right lateral sector (RLS) grafting has contributed to the expansion of the donor pool for living donor liver transplantation (LDLT). At the University of Tokyo Hospital, 28 RLS grafts were used in 251 LDLTs between 2000 and 2013. Procurement of an RLS graft requires caution to dissect and cut the thinner hepatic artery, portal vein and bile duct compared with procurement of a conventional right or left liver graft. In donors with a supraportal hepatic artery, dissection of the thin hepatic artery between the right posterior portal vein and right posterior bile duct is technically demanding. This is more difficult in donors with a supraportal hepatic artery in segment 7 and infraportal hepatic artery in segment 6 (6.5% in all donors; fig. 17A) . Another concern for the bifurcation of the hepatic artery is donors in whom A8 branches from the posterior hepatic artery. Four donors who underwent procurement of an RLS graft at the University of Tokyo Hospital had this anomaly. Posttransplant hepatic artery thrombosis (HAT) occurred in two of the RLS grafts, and A6 and A7 were reconstructed individually to preserve A8 for the donor (fig. 17B1). However, no HAT occurred in two of the grafts in the posterior hepatic artery that was reconstructed when A8 was sacrificed. No ischemic complications occurred in these donors (fig. 17B2) .
Anatomical knowledge of the liver is essential for safe and precise liver surgery. Hepatic segmentation has been proposed by many researchers, and Couinaud's segmentation was adopted after a long discussion that lasted for several decades. This segmentation can be simulated using 3-D CT images before surgery, or it can be visualized using fluorescent images with ICG during surgery. The caudate lobe is one of the most complicated segments of the liver, surrounded by other segments and landmark vessels, and anatomical resection of the caudate lobe requires proper understanding of its three components: the Spiegel lobe, the paracaval portion and the caudate process. Left trisectionectomy, anatomical resection of segment 8, ALPPS and procurement of right lateral liver grafts in LDLT are challenging for surgeons. Detailed knowledge of the liver's anatomical structure is important for safe completion of these technically demanding hepatectomies.
The authors declare no conflicts of interest.
This work was supported in part by Grant-in-Aid for scientific research from the Ministry of Health and Welfare of Japan.
The authors acknowledge the valuable contribution of the following colleagues to the 19th annual meeting of Japanese Research Society of Clinical Anatomy: Junichi Shindoh, Hepatobiliary-Pancreatic Surgery Division, Department of Digestive Surgery, Toranomon Hospital; Masakazu Yamamoto, Department of Surgery, Institute of Gastroenterology, Tokyo Women's Medical University; Shouichi Satou, Department of Hepatobiliary-Pancreatic Surgery, Saitama Medical Center, Saitama Medical University; Takeaki Ishizawa, Department of Gastrointestinal Surgery, Cancer Institute Hospital, Japanese Foundation for Cancer Research; Kimitaka Kogure, Laboratory of Cell Physiology, Institute for Molecular and Cellular Regulation, Gunma University; Junji Yamamoto, Department of Surgery, National Defense Medical College; Minoru Esaki & Yoji Kishi, Hepatobiliary and Pancreatic Surgery Division, National Cancer Center Hospital; Kuniya Tanaka, Department of Surgery, Teikyo University Chiba Medical Center, and Takashi Kokudo, Division of Gastroenterological Surgery, Saitama Cancer Center.
The authors give special thanks to Prof. Tatsuo Sato, President for Tokyo Ariake University of Medical and Health Science, for his instructive and beneficial comments on the sessions on liver anatomy.
All of the abstract of the 19th congress of Japanese Research Society of Clinical Anatomy, November 14th, 2015, are going to be published in another journal .