Fumitake HATA, Gen MURAKAMI, Koichi HIRATA, Shingo KITAGAWA and Mitsuhiro MUKAIYA

Transcription

1 Okajimas Folia Anal Jpn., 76(1): 1-16, May, 1999 Configuration of Hepatic Veins in the Right Surgical Lobe of the Human Liver with Special Reference to Their Complementary Territorial Relationships: Morphometric Analysis of Controlled Specimens with Clearly Defined Portal Segmentation By Fumitake HATA, Gen MURAKAMI, Koichi HIRATA, Shingo KITAGAWA and Mitsuhiro MUKAIYA Departments of Surgery and Anatomy, Sapporo Medical University School of Medicine, South 1, West 16, Sapporo, , Japan -Received for Publication, December 7, Key Words: Hepatic veins, Portal veins, Right surgical lobe, Liver, Human anatomy Summary: The configurations of hepatic veins, particularly the complementary territorial relationships between the veins, in the right surgical lobes of 156 human livers, the segments of which were identified clearly according to the portal ramification, were studied. In order to assess the functional roles of the vessels, we compared the diameter4 values of the vessels in the upper region of the lobe under the assumption that this parameter corresponds very closely to the actual blood flow volume through the vessel. The right hepatic vein (RHV) sometimes (17%) developed poorly and drained neither segment V (S5) nor VI (S6); instead, the inferomedial part of the S6 was often (26%) drained by the middle hepatic vein (MHV). However, most thick short hepatic veins (SHVs) did not drain S6 specifically instead of a poorly-developed RHV, but usually drained both S6 and segment VII (S7), irrespective of the configurations of other veins. Sometimes (8%), the RHV, rather than the MHV, drained a large part of S5. Overall, in the lower region of the lobe, the RHV, SHV and MHV showed complementary venous drainage relationships. Furthermore, the RHV usually (75%) ran not between S5 and S6 but through S6, corresponding to its usual territory, and could not be regarded as the intersectorial landmark in the lower region. Although various names have been given to the thick venous tributaries in the upper region of the right lobe, we systematically classified these tributaries into 5 types of "right superior radicles" (RSR), according to their topographical relationships with the right superior portal branches (P7 and P8), RHV, MHV, SHVs and their tributaries. When just thick RSRs were considered, the 5 types (anterior, posterior, lateral, medial and intersegmental) were observed in 90%, 88%, 89%, 38% and 34%, respectively. Notably, all but the intersegmntal RSRs almost always corresponded to a specific segmental territory (S7 or segment VIII (S8)) and had specific terminals (at the RHV or MHV), suggesting that these 4 RSRs were proper segmental or intrasegmental veins, whereas, the 5th type showed a different configuration, i.e., an intersegmental vein that drained both S7 and S8, ran along and above the RHV and merged with it. Our morphometric examination of the upper region revealed that the sizes of P7 and P8 increased or decreased simultaneously, and suggested that the RSRs played limited roles in venous return. Rather than the RSRs, the SHVs appeared to drain the overflow in excess of the venous return capacity of the RSRs. Despite their limitations and possible complementary relationships with some veins, however, all the superior veins tended to developed evenly well or poorly. Overall, the venous return did not seem to converge into a single particular vein, but was carried away by multiple superior veins: the RHV, MHV, SHVs, and 5 RSRs. The right surgical lobe of the liver is drained by 3 hepatic venous systems: the right, middle and short hepatic veins (Fig. 1). Excluding caudate lob, the right surgical lobe is composed of 4 portal segments, V, VI, VII and VIII (S5, S6, S7 and S8; Couinaud, 1957), and is divided into 2, the anterior (S5 and S8) and posterior (S6 and S7) liver sectors. Although the term "sector" is replaced by "segment" and the term "segment" established by Couinaud is described as "Segment" in the Classification of Primary Liver Cancer (Liver Cancer Study Group of Japan, 1997), we have used "sector" instead of "segment" and "segment" instead of "Segment" in this paper to avoid possible con- Address correspondence to: Fumitake Hata, M.D.

2 2 F. Hata et al. fusion when we have to discriminate "intersectorial" from "intersegmental". The right hepatic vein (RHV) is situated along the intersectorial border between the 2 sectors, drains 4 segments, which are included in the 2 sectors, subsequently, passes between the anterior and posterior segmental portal branches. The middle hepatic vein (MHV) runs along the Cantlie's line/ plane between the right and left surgical lobes (Cantlie, 1898) and accordingly, it drains multiple segments (S4, S5, S6, S8 and S9 in the lobe) adjacent to and/or near it. The short hepatic vein (SHV), named first by Hardy (1972), is a comprehensive entity: the number per specimen, length, thickness (diameter), courses and territories of SHVs vary. Sometimes, the term SHV includes the proper drainage veins of the caudate lobe (Si and S9). In this study, however, we defined the SHVs as those hepatic veins running a posteriorly-located course behind the posterior segmental portal branches (P6s supplying S6 and P7s supplying 57) and that entered directly the inferior vena cava (IVC). Consequently, the hepatic vessels in the right surgical lobe were considered to be located from the posterolateral to anteromedial sides as follows: SHVs, P6s and P7s, RHV and its tributaries, P5s supplying 55 and P8s supplying S8, MHV and its tributaries (Fig. 1). However we need to confirm whether the vein running behind P6 and P7 was consistently a direct tributary of the IVC. Therefore, the firstly aim of this study was to elucidate the territories, courses and terminals of the SHVs, excluding the veins of the caudate lobe, in selected specimens in which the segmental portal branches had been identified. The SHVs and RHV in the right surgical lobe show complementary (compensatory) territorial relationships (Couinaud, 1981). When the territory of the RHV is reduced to the extent it dose not drain S6, a thick SHV and/or multiple SHVs drain S6 instead and, in turn, thick RHVs tend to accompany thin and/or small numbers of SHVs (Couinaud, 1981; Nakamura and Tsuzuki, 1981; Champetier et al., 1993; Masselot and Leborgne, 1978; Makuuchi et al, 1983; Hata et al., 1989). This complementary relationships has already led to the development of modified right hepatectomy procedures (Makuuchi et al, 1987). In the same context, "thick" SHVs have been observed and called various names depending on the researchers and the locations: retrohepatic (Masselot and Leborgne, 1978, territory: 56 and/or S7), dorsal hepatic (Nakamura and Tsuzuki, 1981, territory: 56 and/or S7), posteroinferior (Nakamura and Tsuzuki, 1981, territory: maybe S6), inferior right hepatic (Champetier et al, 1993; Makuuchi et al., 1983, territory: maybe S6), right inferior hepatic (Couinaud, 1981, terriroty: maybe S6), right middle hepatic (Couinaud, 1981, territory: maybe 57) and accessory hepatic (Williams and Warwick, 1980, territory: maybe 56) veins. In contrast to the RHV-SHV complementary relationships, we have less information about the role of the MHV and its complementary relationships in the right surgical lobe. However, Masselot and Leborgne (1978) found a highly-developed MHV draining 56 in 23% of the specimens they examined. Their report suggested the MHV makes a much bigger contribution to the drainage of the right hepatic lobe than indicated by our general hypothesis that the MHV partially drains 55 and 58 during its course along the lobar border. Therefore, the second aim of this study was to elucidate fully the complementary relationships of the 3 venous systems (RHV, MHV and SHVs) draining the lower region of the right surgical lobe (55 and S6). In the upper region of the right surgical lobe (S7 and S8), the hepatic venous configuration is more complex than that in the lower region, as other thick veins, irrespective of whether they are direct tributaries of the IVC, usually coexist with the RHV, MHV and SHVs. These thick superior veins have been called variously: right anterosuperior (Nakamura, 1981), anterior superior (Hardy, 1972) and postero-internal (Masselot and Leborgne, 1978) veins, which were thought to be anterior tributaries of the RHV, and right superior (Hardy, 1972; Nakamura, 1981) and postero-external (Masselot and Leborgne, 1978) veins, which were thought to be posterior tributaries of the RHV. Makuuchi et al. (1986) also reported a consistent intrasegmental vein draining S8 properly and running between the ventral and dorsal subsegmental branches (P8v and P8d, respectively) of the P8. However, the origins, courses and/or terminals of most of these reported veins, particularly their topographical relationships with the portal branches, are still unclear. The first problem we encountered when we began to study the veins in the upper region of the lobe was the terminology. No matter at which of the RHV, MHV, SHV and IVC these superior veins terminated, we considered that these veins could be called collectively "right superior radicles" (RSRs). Therefore, how do we classify the RSRs? In this study, according to the topographical relationships between P7, P8 and their subsegmental branches and the hepatic veins, we divised a nomination system for the RSRs and a hypothetical model or their locations and interrelationships (Fig. 2 and Table 1). In summary, the vessels are located in the upper region of the lobe from the postero-

3 Configuration of Hepatic Veins in the Right Lobe 3 Table 1. Hepatic venous configuration in the upper half of the right surgical lobe Courses a g are shown in Figure 2. RHV, right hepatic vein; MHV, middle hepatic vein; LHV, left hepatic vein; IVC, inferior vena cava; SHV, short hepatic vein; RSR, right superior radicle lateral to anteromedial aspect as follows: SHV- P7-posterior RSR-RHV-anterior RSR-P8d-lateral RSR-P8v-medial RSR. These names depended on neither the territories nor the terminals but only on the courses (topographical relationships with other vessels). We also hypothesized that these RSRs and 3 hepatic venous systems showed complementary territorial relationships, in other words, their volumes of venous return were interrelated. Therefore, the third aim of this study was to elucidate the venous configuration in the upper region according to our RSR nomination system in order to establish the complementary venous drainage relationships. Can we define the "functional" complementary relationships between veins in terms of the actual venous return on the basis of anatomical observations? The portal segmental branches, i.e. P7 and P8, and corresponding segmental hepatic arteries supply blood to the upper region of the lobe, and the same volume would be expected to return through the 3 hepatic venous systems, including the RSRs. Therefore, we hypothesized that there is a positive correlation between the "sizes" of the portal segmental branches and the hepatic veins, including the RSRs, and positive or negative correlations between the "sizes" of the veins. How do we estimate the size of vessels? In other words, on which factors does/do the blood flow depend? According to Poiseuille's law, the blood flow volume per unit of time is equal to a x diameter4/length of a vessels. Therefore, in order to compare blood flow volumes, we tried to compare the diameter4 values of various vessels with each other and analyzed them statistically without taking the length of each vessel into considerration, because the lengths seemed to vary over a range of x0.1 to x10. Then, the morphometric data of the vessels in the upper region were used to estimate the "functional" relationships between vessels, particularly between the P7, P8, RHV, SHVs and RSRs. Thus, the fourth aim of study was to evaluate the venous function in the upper region. In this study, we used carefully selected specimens (controlled specimens) in which the portal segmentation was clear and could be identified unequivocally in order to identify the segments definitively. In previous studies of hepatic veins, the territories of each hepatic vein and its tributaries were described on the basis of the portal segmentation system, but intrahepatic portal ramification variations did not seem to be taken into consideration when selecting specimens. Sometimes, it is difficult to identify the portal segmentation system due to particular portal ramification variations (Hasegawa et al., 1986; Takayasu et al., 1985; Kinoshita et al., 1988; Hata et al., 1999). Therefore, at the beginning of this study, we examined the intrahepatic portal ramification in liver specimens in order to exclude several variations, according to certain criteria (see Materials and Methods). Materials and Methods 1. Preparation of materials Two-hundred-and-fifty-two livers that were free of macroscopic pathology were removed from donated Japanese adult cadavers (110 males and 142 females, aged years), which had been fixed by perfusion with % v/v formol solution through the femoral artery. The sizes of the liver specimens varied considerably, e.g., the range of their transverse lengths was mm (205 mm on average). The specimens, particularly their right halves,

4 4 F. Hata et al. were dissected minutely from their visceral (inferior) surfaces in order to preserve the topographical relationships between the vessels. The SHVs were cut at their terminals at the IVC in order to create a space to approach the deep regions of the lobe. The entire SHVs on the liver side (their peripheral side) were preserved during dissection. The diaphragmatic surface of each specimen was also preserved as far as possible in order to maintain the original shape of the specimen. 2. Selection of specimens After dissection, we selected 156 specimens (75 males and 81 females, aged years) in which P5 and P6 were clearly identified. Specimes showing variations involving sectorial trunk formation, such as a common trunk of P5 and P6 or trifurcation of P6, P7 and the anterior sectorial trunk, were excluded. In each of the 156 specimens, P8 formed a single trunk apart for one or two thin proximal twigs. 3 hepatic venous systems in the lower region of the right lobe and the thicknesses, entrances and territories of the SHVs in these 156 specimens were examined. Next, in order to study the upper region of the right lobe, we selected 82 of the 156 specimens described above. The serection criteria were that the segmental borders between all 4 right segments (S5, S6, 57 and 58) were clearly visible and could be identified unequivocally, i.e., we selected specimens showing simple bifurcation of the posterior sectorial trunk dividing into P6 and P7. In each of these 82 specimens, P8 was divided clearly into a dorsal and a ventral branch as reported by Makuuchi (1986). 3. Measurement and evaluation of venous drainage The inner diameters of P6 and P7 were measured immediately after these branches arose from the posterior sectorial trunk. In the upper region of the lobe, the inner diameters of the SHVs at their entrances into the INC and of the RSRs into the RHV, MHV or IVC were measured. The inner diameters of the RHV at its entrance into the IVC (we called this the "RHV root") and immediately distal to the point at which the RHV merged with the RSRs (we called this the "proper RHV") were measured. In the lower region of the lobe, the inner diameters of the MHV tributaries at their crossings with P5 and P6, if they crossed and where they drained S6, were measured and those of the SHVs were measured as described above. These diameter data were compared and analyzed statistically. According to Poiseuille's law, the blood flow volume per unit of time is equal to a x diameter4/length of a vessel. In order to cornpare the blood flow values particularly the flow through the RSRs, we compared the diameter4 values (or the sum of diameter4 values if several vessels needed to be considered together, e.g., 2 veins with diameters of 7 and 8 mm: ) with each other and analyzed them statistically. However, we did not take the length of teach vessel into account, because since the lengths varied by factor of about x 0.1 x10, e.g., from the minimum SHV length of about 2 cm to the maximum RHV length of about 30 cm. Statistical processing was based on Pearson's correlation assessment. In Results, we define the "diameter" or "diameter4" values analyzed case by case. It should be noted that a vein 10 mm in diameter can drain almost same volume of blood as 100 veins each 3 mm in diameter if their lengths are same. From the functional viewpoint of blood flow, we usually describe veins over 3 or 5 mm in diameter as thicker depending on the particular aim (the criteria are stated in each section of the Results). Results A. SHVs The SHVs, veins passing behind the posterior segmental portal branches (P6s and P7s), were frequently observed (84.6%, 132/156) when the diameter considered was limited to over 3 mm. The total number of SHVs over 3 mm in diameter was 189 (1-5 veins per specimen, 1.43 on average). Of these 189 SHVs examined, 21.2% (40) were over 10 mm in diameter, the territories of 19 (10.0%) were restricted to S6, those of 131 (69.4%) were restricted to S7 and 39 (20.6%) drained both segments. Therefore, most the SHVs drained S7. However, notably, most (31/40) of the thick SHVs over 10 mm in diameter drained both S6 and S7. Moreover, such thick SHVs often (22/40) drained the deep parenchyma of 56 and/or 57, in contrast to the thinner SHVs that consistently drained the superficial (inferior) parenchyma of the lobe. Accordingly, only thick veins were able to interdigitate 3- dimensionally with the portal branches. The SHVs consistently entered the IVC without merging with the RHV or RSRs. Although the SHVs most frequently drained 57, which was located at a higher position than S6, the entrances of the 189 veins into the IVC were distributed evenly along 94 mm of the IVC ranging from 26 mm above to 68 mm below (on average: 26.9 mm below) the entrance of the RHV into the IVC. An even distribution of the SHV terminals was ensured by the random courses of the SHVs: upward, transverse and slightly downward. In other words, SHVs draining 57

5 Configuration of Hepatic Veins in the Right Lobe 5 sometimes ran slightly downward to enter the IVC, whereas, corresponding to theis territories that often included S6, the 40 thickest SHVs (over 10 mm in dimeter) terminated at the lower level: 31.9 mm (on average) below the entrance of the RHV into the IVC. No SHV drained either S5 or S8, not even very thick ones located at the most inferior or superior regions of the lobe. B. Lower region of the right surgical lobe 1. In general We classified the RHVs into 4 types according to their drainage territories (Fig. 3) as follows: 1) S7 and S8 (poorly-developed type), 2) S7, S8 and part of S6 (intermediate type), 3) S7, S8, almost all of S6 and a part of S5 (usual type), and 4) S6, S7, S8 and a large part of S5 (well-developed type). The RHV ranged from 5 to 15 mm in diameter at the point immediately distal to the entrance of the RSRs (proper RHV). The most proximal region of the RHV between the joining of the RSRs and the entrance into IVC was called the RHV root (the diameter ranged from 13 to 28 mm), the configurations of which in relation to the RSRs are described in section C of Results. The MHV (a tributary over 1 mm in diameter) sometimes (25.6%, 40/156) drained S6 (Fig. 3). In these 40 specimens, a right and inferior tributary of the MHV originated from S6 (usually its deep, but sometimes its superficial, parenchyma), ran to the left side, crossed P5 and then turned upward. Although the thin veins that were less than 1 mm in diameter at their crossing P6 were not counted, their diameters ranged from pin-hole sized to 7 mm at their crossing points. Figure 4B shows the thickest MHV (7 mm in diameter at the crossing P6) observed. The configurations of the SHVs, depending on the 4 types of RHV, are also described below. 2. The four types of RHV Poorly-developed type (Figs. 4 and 5A): The proper RHVs of this type did not drain S5 and S6 (17%, 26/156). Twenty-four thick SHVs over 10 mm in diameter were observed in 22 of the 26 specimens. Three of 24 thick SHVs drained S7, one drained S6 and 20 drained both segments. Thus, 21 SHVs in 21 specimens drained part or all of S6. Futhermore, a tributary of the MHV was often well developed and drained S6 (46%, 12/26). In these 12 specimens, average diameters of the MHVs at the points where they crossed P5 and P6 were 5.8 and 3.5 mm, respectively. An MHV tributary and a thick SHV shared the drainage of S6 in 9 of the 12 specimens, in which the SHV diameters ranged from 5 to 13 mm (8.8 mm on average). In the other 3 specimens, the venous drainage of S6 depended entirely on the MHV, although the role of the thin SHVs (less than 3 mm in diameter) was not taken into consideration. Intermediate type: We found RHVs with a morphology intermediate between those of the usual and poorly-developed types. This type of RHV, found in 2% (4/156) specimens, drained S6, S7 and S8 and, at its most peripheral course, the proper RHV was situated intrasegmentally in S6. In 2 of these 4 specimens, the MHV drained S6 and it was 3 or 6 mm in diameter at the point where it crossed P6, whereas one thick SHV that drained both S6 and S7, was seen in one specimen. Usual type: This type of RHV, seen in 73% (113/ 156) of the specimens originated at the intersegmental border between S5 and S6 and then passed between S7 and S8. In these specimens, SHVs over 3 mm in diameter were observed frequently (83%, 94/113). The total number of SHVs was 132, 11 of which were thick SHVs over 10 mm in diameter and a single SHV drained S6 in one specimen, S7 in one and both in 9. Futhermor, a tributary of the M1-IV sometimes (23%, 26/113) drained S6 and a thick tributary over 3 mm in diameter was found in 12 of these specimens. However, we found only 3 specimens with both an MHV and a thick SHV that drained S6. Notably, the usual type of RHV ran not between S5 and S6 but through S6, corresponding to its venous territory. However, most of the inferior tributaries of the RHVs did not interdigitate with the subsegmental branches of P6 but ran parallel to them. Well-developed type (Fig. 5B): In contrast to specimens with the poorly-developed type of RHV, the MHV never drained the S6 when the proper RHV was well developed and drained a large part of the S5, seen in 8% (13/156) of the specimens. In these 13 specimens, the right half of the gall bladder bed of the liver was also drained by the RHV, although the segmental proportion of the bed between S4 and S5 varied significantly among the diameter of the specimens. In 4 of these 13 specimens, a thick SHV over 10 mm in diameter coexisted, one of which drained both S6 and S7 and the other 3 drained S7. C. Upper region of the right lobe 1. Anterior and posterior RSRs and the RHV root The anterior RSR drained S8 and passed between the RHV and P8, whereas the posterior RSR drained $7 and ran between the RHV and P7. The anterior RSR never drained S7 and the posterior RSR never drained 58. Notably, the anterior and posterior RSRs, when only those over 5 mm in dimeter were considered, consistently entered

6 6 F. Hata et al. the RHV root (the most proximal part of the RHV), i.e., their terminals were located at points 13.4 ± 8.6 and 11.3 ± 7.4 mm, respectively, distal to the entrance of the RHV into the IVC. Accordingly, the length of the RHV root ranged from about 5 to 20 mm. The anterior RSR, when only those over 5 mm in diameter at the entrance into the RHV were considered, was found frequently (90%, 74/82) and was sometimes double (14/74). Likewise, the posterior RSR was found frequently (88%, 72/82) and was also sometimes double (14/ 72). The maximum diameter of both the posterior and anterior RSRs was 10 mm (6.6 mm and 5.9 mm on average, respectively, when only thick veins over 5 mm in diameter were considered). The observations and statistical assessment showed that the RHV root was thicker when the anterior and posterior RSRs, particularly the latter, were thicker. Moreover, both these RSRs simultaneously developed well or poorly in the same specimen. For instance, when absence of the anterior RSR (8/82), the posterior RSR was also frequently absent (7/8). In turn, when absence of the posterior RSR (10/82), the anterior RSR was either absent (6/10) or also relatively thin. In specimens with the thin anterior and posterior RSRs, a thick SHV drained S7 and notably, the proper RHVs fequently did not drain S5 (i.e., they were of the poorly-developed or intermediate type). 2. Lateral RSR and P8 Dorsal and ventral subsegmental branches of the P8 were observed consistently and their average diameters were 5.5 and 4.5 mm, respectively, at the point immediately peripheral to the division of the P8 trunk (Fig. 6A). One or two lateral RSR(s) passed between the dorsal and ventral branches of P8 (Fig. 6B). When the diameters of the lateral RSRs considered were limited over 5 mm, 77 lateral RSRs were observed in 73 of 82 specimens (89%), in 4 of which, the vein was double. The lateral RSR dimeter ranged from pin-hole sized to 10 mm (6.3 mm on average, when the 77 RSRs over 5 mm in diameter were considered). The lateral RSRs frequently (89.6%, 69/77) entered the proximal re-, gion of the MHV, but sometimes (9.1%, 7/77) terminated at the RHV root or rarely (1.3%, 1/77) drained directly into the IVC. When the lateral RSR did not enter the MHV (8 specimens), the diameter of the dorsal branch of P8 increased to over 6.0 mm. 3. Medial RSR The medial RSR was often (38%, 31/82) a thick vein over 5 mm in diameter that drained the medial and ventral parts of S8. The medial RSR consistently entered the proximal region of the MHV and at their terminals, the medial RSRs ranged in diameter from pin-hole sized to 7 mm (5.6 mm on average when the 31 RSRs over 5 mm were considered). Co-existent of the lateral and medial RSRs were seen in 25 specimens, but the absence both radicles was rare (5/82). Strangely, when a thick SHV (over 5 mm in diameter) was absent (31 specimens), the medial RSR was seen more frequently (21/31) than usual, despite the fact that SHVs and medial RSRs did not share territories. 4. Other comments All 4 RSRs, when only those over 5 mm in diameter were considered, coexisted in only (4/82) specimens. Anterior, posterior and lateral RSRs over 5 mm in diameter coexistent in 31 specimens (31/82). Notably, in all these specimens, P8 was thin with an average diameter of 7.6 mm. Except for the 4 RSRs, another course was observed in 34% of the specimens examined. This 5th type of RSR(s) drained both S7 and 58, ran intersegmentally superior to the RHV, entered the RHV root without forming a common trunk with other 4 RSRs and was sometimes double or triple (Figs. 6 and 7). 5. Statistical evaluation of the correlations between the various hepatic veins and between the hepatic veins and P7 and P8: comparisons of their diameter4 values (Table 2) 1) P7 versus P8 The diameter of the P8s ranged from 5 to 12 mm (8.0 mm on average) at the points immediately peripheral to the issue of the P5s, whereas those of the P7s ranged from 4 to 11 mm (7.4 mm on average) at the points immediately peripheral to the issue of the P6s. The diameter of P8 was usually (58.5%, 48/82) larger than that of P7 and the average P8/P7 diameter ratio was Notably, the diameter4 values of P7 and P8 showed a significant positive (0.313) correlation. This positive correlation between the P7 and P8 was more highly significant evident if the population considered was limited to 26 specimens in which P7 was dominant (P8/P7 < 1.0, 0.829) and 48 specimens in which P8 was dominant (P8/P7 > 1.0, 0.759). 2) P7 and P8 versus the hepatic veins, including the RSRs The diameter4 value of the proper RHV, i.e., the peripheral RHV distal to the merging of the RSRs, showed a significant positive correlation with that of P7 (0.480), but no significant correlation with that of P8 (0.147). In particular, in 6 specimens in which P7 was highly dominant (P8/P7 < 0.6), the

7 Configuration of Hepatic Veins in the Right Lobe 7 Table 2. Statistical assessment of the correlations between the diameter4 values of the posterior segmental portal branches and/or hepatic veins in the upper half of the right surgical lobe * Rr, RHV root; pr, proper RHV; prsr, posterior RSR; arsr, anterior RSR; lizsr, lateral RSR; mrsr, medial RSR. Minus values indicate negative correlations. More than of the absolute values indicate the statistically significant correlations in this system. correlation between the diameter4 value of the proper RHV and P7 was more highly significant (0.926), in spite of the coexistence of a thick SHV over 10 mm in diameter in all 6. However, neither the anterior nor posterior RSR diameter4 values correlated with that of either P7 or P8, respectively. In contrast, when the P7 or P8 diameter4 value increased, that of either the SHV or lateral RSR also increased (0.276 and 0.260, respectively). 3) RHV and SHV versus the RSRs The diameter4 values of the posterior and anterior RSRs showed a significantly positive correlation with each other (0.400). The diameter4 values of the RHV root proximal to the merging of the RSRs depended significantly and positively on the diameter4 value of the posterior RSR (0.449) and the anterior RSR (0.439). In contrast, the diameter4 values of the SHV and RHV root showed a significant negative correlation (-0.362). Significant negative correlations between 1) the medial and lateral RSRs (-0.230) and 2) the medial RSR and SHV (-0.315, see also section A of the Results). Discussion In the lower region of the right surgical lobe of the liver, RHV, SHVs and MHV showed complementary venous drainage relationships. When the RHV was poorly developed, the complementary roles of the MHV and SHVs in draining 55 and 56 were particularly evident. However, the configuration of the SHVs can not be explained simply from the viewpoint of a role compensating for the RHV. Although thick SHVs draining S6 have been noted and influenced clinical procedure (see Introduction), our results demonstrated that thick SHVs usually drain both 56 and 57 and that if thin SHVs are also considered, they drain S7 more frequently than S6. Moreover, even if a well developed RHV was present, it did not always drain the posterior superficial parenchyma, which was drained by the SHVs instead. The SHVs rather than the MHV were usually dominant in S6, although a part of S6 as well as a large part of 55 were often drained by the MHV. We concluded that the distance from the root of the RHV and/or MHV to the superficial parenchyma near the visceral (inferior) surface of the right lobe is too long for successful drainage. Therefore, in general, the SHVs seemed to be required, although not in certain specific cases, for drainage of the peripheral parenchyma. As a consequence of its additional role in venous drainage, the SHV did not seem to interdigitate with the portal branches. Notably, a well-developed RHV draining a large part of S5, including the gall bladder bed of the liver, was only observed sometimes. Moreover, corresponding to this rightward-shifted territory, the RHV did not usually run between S5 and 56 but through 56. We had expected the RHV to run between S5 and S6, as stated in many text books as well as by most surgeons. However, we found the RHV could not be regarded as the intersectorial landmark in the lower region and we have prepared another paper about the limitations of the hepatic veins for liver segment identification (Hata et, press). We successfully identified 5 RSRs according to their courses and territories. Four RSRs were intrasegmental or proper segmental veins, whereas the other one was found to be an intersegmental veins, which ran a course between S7 and S8 and was nearly parallel and superior to the RHV. We consider that this intersegmental RSR corresponds

8 8 F. Hata et al. to the double or collateral RHV reported by Couinaud (1989). The terminology based on the concept of "RSRs" we have adopted seems simple, and specifically desceibes particular morphologial feature better than the various names that have been reported. Notably, the diameter4 value of P8 and P7 increased simultaneously in accordance with territories and declined in the same manner. This relationships suggested that an increase in volume of the upper region of the right lobe resulted from increases in both S7 and 58, not on either one alone. The postive correlation between S7 and S8 seemed to correspond to the positive correlation (simultaneous increase or reduction) between the diameter4 value of the anterior and posterior RSRs, which properly drained S8 and S7, respectively. Inequivalent increases or reductions in the blood flow volumes of S7 and S8 appeared unusual. The strange relationships of a thick medial RSR (draining S8) in the absence of thick SHV (draining the S7), might be attributable to inequivalent changes in the flow volume of S7 and S8. However, this strange relationships might be a limitation of diameter4 assessment, i.e., it might be an example of the non-specific false positive relationships. A clear complementary relationships between the RHV and SHV diameter4 values in the upper region of the right lobe, as well as in the lower region, was also evident, but strangely neither the anterior nor posterior RSR diameter4 values seem to correlate positively with either the P7 or P8 in diameter4 values. Does the extent of the territory of these RSRs depend on the magnitude of the blood flow supplied by P7 and P8? We consider that when both the liver mass and portal blood flow in the upper region increase, the RSRs develop simultaneously up to a certain limit, beyond which only the SHVs is/are responsible for draining the overflow, possibly resulting from an increase in the mass of the most posterior region of S7. Further morphometric analysis of the external shape of the liver might reveal the limited roles of the RSRs. Indeed, the proper RHV and the anterior and posterior RSRs joined at the thick RHV root, and the lateral and medial RSRs usually merged with the MHV at its root. However, taken together with the venous configurations in the upper region of the right lobe, we conclude that the venous return did not seem to converge into a single particular veins but was carried away by multiple superior veins or radicles: the SHVs, the proper RHV and its intersegmental (collateral) tributary and the anterior, posterior, lateral and medial RSRs. Acknowledgments We are grateful to the following professors. and their laboratory staff for the use of their specimens: (alphabetically) Professor Y. Dodo of the Tohoku University School of Medicine, Professor Y. Fukui of the Tokushima University School of Medicine, Professor N. Goto of the Showa University School of Medicine, Professor K. Hirata of the St. Marianna University School of Medicine, Professors H. Hoshi and T. Arikuni of the Nihon University School of Medicine, Professor H. Ito of the Nippon Medical School, Professor T. Kachi of the Hirosaki University School of Medicine, Professor M. Kikuchi of the Tohoku University School of Dentistry, Professor K. Kishi of the Toho University School of Medicine, Professor S. Kitamura of the Tokushima University School of Dentistry, Professor K. Kohno of the University of Tsukuba Institute of Basic Medical Sciences, Professor K. Kumaki of the Niigata University School of Medicine, Professor J. Matsumura of the Kyorin University School of Medicine, Professor T. Nakao of the Akita University School of Medicine, Professor T. Sato of the Tokyo Medical and Dental University Faculty of Medicine, Professors H. Yamashita and S. Kato of the Jikei University School of Medicine. This study was supported in part by a Grant-inaid for Scientific Research from the Ministry of Education, Science and Culture (No ) References 1) Cantlie J. On a new arrangement of the right and left lobes of the liver. Proceeding of the Anatomical Society of Great Britain and Ireland 1898; 32:4-9. 2) Champetier J, Haouari H, Le Bas JF, Letoublon C, Alnaasan I and Farah I. Large inferioright hepatic vein. Surg Radiol Anat 1993; 15: ) Couinaud C. Le foie. Etudes anatomiques et chirurgicales. Masson, Paris, ) Couinaud C. Controlled hepatectomies and exposure of the intrahepatic bile ducts. Anatomical and technical study. pp , Couinaud, Paris, ) Couinaud C. Surgical Anatomy of the Liver Revisited. pp , Couinaud, Paris, ) Hardy KJ. The hepatic veins. Aust NZJ Surg 1972; 42: ) Hasegawa H. Liver segments-their clinical implications. In Progress in Gastroenterology, pp (in Japanese) Nihon-Igakukan, Tokyo, ) Hata F, Hirata K, Murakami G and Mukaiya M. Identification of segments VI and VII of the liver based on the ramification patterns of the intrahepatic portal and hepatic veins. 1999, in press. 9) Hata Y, Uchino J, Une Y and Morita Y. Surgical aspects of hepatic segmentation based on hepatic venographies. Surg Radiol Anat 1989; 11: ) Kinoshita H, Sakai K, Hirohashi K, Tsuji Y, Inoue T, Kubo S and Nakatsuka H. Branching patterns of the intrahepatic

9 Configuration of Hepatic Veins in the Right Lobe 9 portal vein and hepatic segments identified by percutaneous transhepatic portography. Geka Gakkai Zasshi (J. Jpn. Surg. Soc.) 1988; 89:55-62 (in Japanese with English abstract). 11) Liver Cancer Study Group of Japan. Classification of Primary Liver Cancer. lth ed, pp Kanehara & Co, Tokyo, ) Makuuchi M, Hasegawa H, Yamazaki S, Bandai Y, Watanabe G and Ito T. The inferior right hepatic vein: ultrasonic demonstration. Radiol 1983; 148: ) Makuuchi M. Intrahepatic architecture of the portal vein in the right anterior sector of the liver by an ultrasound imaging approach. Kanzo 1986; 27:526 (in Japanese). 14) Makuuchi M, Hasegawa H, Yamazaki S and Takayasu K. Four new hepatectomy procedures for resection of the right hepatic vein and preservation of the inferior right hepatic vein. Surg Gynecol Obstet 1987; 164: ) Masselot R and Leborgne J. Anatomical study of the hepatic veins. Anat Clinica 1978; 1: ) Murakami G, Trunk 1, Viscera. In: Sato T and Sakai T (eds.) Anatomy of the Human Body. Vol 13 in Basis of the Recent Medicine. Iwanami-shoten, Tokyo, pp , ) Nakamura S and Tsuzuki T. Surgical anatomy of the hepatic veins and the inferior vena cava. Surg Gynecol Obstet 1981; 152: ) Takayasu K, Moriyama N, Muramatsu Y, Shima Y, Goto H and Yamada T. Intrahepatic portal vein branches studied by percutaneous transhepatic portography. Radiology 1985; 154: ) Williams PL and Warwick R. Gray's Anatomy. 36 th ed, pp. 763, Churchill Livingstone, London, 1980.

10 10 F. Hata et al. Plate I Explanation of Figures Plate I Fig. 1. Diagram showing the topographical relationships between the hepatic veins and portal branches (modified from Murakami, 1998). P1 P9 are the segmental portal branches supplying the segments corresponding to the numbers. The hepatic vessels in the right surgical lobe, are located from the posterolateral to the anteromedial sides as follows: short hepatic vein (SHY), posterior segmental portal branches (P6 and P7), right hepatic vein (RHV), anterior segmental portal branches (P5 and P8) and middle hepatic vein (MHV). P6 is often multiple (two in this figure) and P8 usually divides into 2 subsegrnental branches. In the caudate lobe, P1 supplies the Spiegel's lobe surrounding the inferior vena cava (IVC), whereas the P9 supplies the paracaval portion in front of the IVC and the caudate process. LHV, left hepatic vein; P, portal trunk; LP, left portal trunk; RP, right portal trunk.

11 Configuration of Hepatic Veins in the Right Lobe 11 Plate II Plate II Fig. 2. Hypothetical model of 4 types right superior radicle and their topographical relationships. We hypothesized that the hepatic veins in the upper region of the right surgical lobe ran a long 7 courses (a g): a, short hepatic veins (SHVs); b, posterior right superior radicle (RSR); c, right hepatic vein (RHV); d, anterior RSR; e, lateral RSR; f, medial RSR; g, middle hepatic vein (MHV). There is a specific topographical relationship between each of these 4 RSRs and the hepatic veins (SHV, RHV, MHV) and segmental (P7) and subsegmental (P8d, P8v) portal branches: e.g., the lateral RSR running along course is located between P8d and P8v.

12 12 F. Hata et al. Plate III Plate III Fig. 3. Diagram showing the complementary relationships between the right, middle and short hepatic veins in the right surgical lobe. A, Poorly-developed type of right hepatic vein (RHV); B, Intermediate type of RHV; C, Usual type of RHV; D, Welldeveloped type of RHV. The short hepatic veins (SHVs) are classified into 3 types according to their territories: SHV1 drains S7; SHV 2 drains both of S6 and S7; SHV 3 drains S6. Each number parentheses indicates the number of veins observed. The numbers of MI-IV tributaries originating from S6 are also indicated in parentheses.

13 Configuration of Hepatic Veins in the Right Lobe 13 Plate IV Plate IV Fig. 4. Photographs showing the middle hepatic vein (MHV) draining S6 viewed from the inferoposterior aspect. A. A thick middle hepatic vein (MHV) interdigitating with P5 and P6. The short hepatic veins (SHV1, SHV2) were partly removed and they drained the superficial parenchyma of the lobe. The right hepatic (RHV) vein is poorly developed and cannot be seen behind P7. IVC, inferior vena cava; PT, portal trunk. B. A thick MHV originating from the deep parenchyma of both S6 and the border zone between S6 and S7. The RHV (poorly-developed type) is restricted to the upper region.

14 14 F. Hata et at Plate V Plate V Fig. 5. Photographs showing two types of right hepatic vein (RHV) viewed from the inferoposterior aspect. A. Poorly-developed type. The posterior sectorial trunk simply divides into P6 and P7 and the RHV can be seen interdigitating with the subsegmental branches of P7. Instead of an RHV, a well developed short hepatic vein (SHV) rains S6, part of which is also drained by the middle hepatic vein (MHV). B. Well-developed type. The posterior segmental branches are also of the simple bifurcation type shown in Figure A. The RHV can be seen interdigitating with P5, P6 and P7 and it drains all of S6 and the posteroinferior part of S5 in the lower region of the right surgical lobe. An asterisk indicates the connective tissue have been removed from the vessels; PT, portal trunk; P9, segmental branches supplying the right caudate lobe.

15 Configuration of Hepatic Veins in the Right Lobe 15 Plate VI Plate VI Fig. 6. Photographs showing the right superior radicles (RSRs) viewed from the superior aspect. A and B show a same specimen. Each vein indicated by a, d, e, f or g is a right superior radicle or hepatic vein following a specific course in the upper region (see Figure 2). The vein indicated by c with an asterisk is an intersegmental RSR following course c. The right hepatic vein lies under vein c and cannot be seen. In B, P8 has been retracted to the posterior side in order to show the medial RSR following course f. IVC; inferior vena cava.

16 16 F. Hata et at Plate VII Plate VII Fig. 7. Photographs showing the right superior radicles (RSRs) viewed from the superior aspect. Each vein indicated by b, c, d, e or g is a right superior radicle or hepatic vein following a specific course in the upper region (see Figure 2). Two veins indicated by c with an asterisk which have been partly removed are intersegmental RSRs following course c. A right hepatic vein following course c can be seen under the intersegmental radicles. IVC; inferior vena cava.