Search

Background: Right hemihepatectomy (RH) for proximal bile duct cancer occasionally results in liver failure. We report the feasibility of left hemihepatectomy (LH) with vascular reconstruction (VR) of the right-sided hilar vessels to preserve hepatic reserve., Methods: Among 110 patients with proximal bile duct cancer (PBC) treated between January 1980 and December 1998, 11 patients underwent LH with VR of eight portal veins and nine hepatic arteries, and 14 underwent RH with VR of four portal veins and one hepatic artery. Microsurgical techniques were used in 80% (8/10) of the hepatic arterial reconstructions., Results: Although operation time was significantly longer in the LH group, hospital mortality, blood loss and incidence of histologically cancer positive margin at the bilioenteric anastomotic site were similar in the two groups. Peak serum liver enzyme concentration was significantly higher in the LH group with longer inflow occlusion time, whereas peak serum total bilirubin concentration was significantly higher in the RH group, which had smaller liver remnant. No liver abscess occurred in any patients who underwent microvascular reconstructions. The cumulative survival of the LH group was worse than that of the RH group, in which the proportion of vascular invasion was lower, but not significantly., Conclusion: LH with right-sided microvascular reconstruction is technically possible and a feasible option when RH is likely to result in postoperative liver failure.

Background and Aims: Prognosis after hepatectomy for hepatocellular carcinoma (HCC) has been improved by progress in the evaluation of hepatic functional reserve, surgical techniques and perioperative management. However, even when curative resection is performed at a relatively early stage, a considerable number of patients develop early intrahepatic and/or extrahepatic recurrence postoperatively. This study analyzed the clinicopathologic features of HCC with early recurrence., Methods: We reviewed records of 513 consecutive patients who had undergone liver resection for HCC. There were 48 deaths within a year after surgery from recurrence, including 21 patients with stage II or III HCC (group I). Clinicopathologic parameters of group I patients were compared with those of 188 patients (group II) who developed recurrence following resection of stage II or III HCC and died more than 1 year after surgery., Results: On univariate analysis, age, tumor diameter (phi), alpha-fetoprotein (AFP):phi and protein induced by vitamin K absence or antagonist II (PIVKA-II):phi were significantly greater in group I than in group II. Macroscopic portal vein invasion, microscopic vascular invasion, intrahepatic metastasis, poor differentiation, pleomorphism, sarcomatous change, vascular lake, and angiographic condensed pooling were more frequently observed in group I than group II. Five independent determinants were selected by multivariate analysis: AFP:phi, histologic pleomorphism, sarcomatous change, vascular lake and angiographic condensed pooling., Conclusions: Highly malignant HCC with extremely poor prognosis exhibits peculiar clinicopathologic characteristics, particularly histologic immaturity, and can be predicted by preoperative indicators such as markedly elevated tumor marker concentrations and condensed pooling on angiography.

Background: Although there have been extensive studies to determine risk factors affecting survival after resection for hepatocellular carcinoma (HCC), we still do not know which patients can survive disease free after curative resection. This study was undertaken to determine independent risk factors affecting the length of disease-free survival., Methods: 171 patients, who survived disease free more than 1 year after curative resection, were divided into four groups according to the disease-free period: Group I (n = 96) for the patients with intrahepatic recurrence between 1 and 3 postoperative years, Group II(n = 27) for those between 3 and 5 years, Group III (n = 40) for those between 5 and 10 years and Group IV (n = 8) for those without recurrence within 10 years. The 37 variables (host factors, tumor factors, non-tumor liver factors, surgical factors) were compared among the four groups. Activity of hepatitis and hepatic fibrosis was scored by the Histological Activity Index (HAI). In a multivariate study, possible prognostic variables with a statistical difference in the disease-free survival rate among each category were preliminarily selected from the 37 variables and the independent variables were finally selected using a proportional hazard analysis., Results: The patients' age, indocyanine green retention rate, microscopic tumor capsular invasion and portal invasion, cell differentiation, extent of hepatectomy, aggressiveness of chronic hepatitis and inflammatory activity assessed by HAI score were significantly different among the four groups. In the multivariate analysis, the following variables were selected as the independent determinants favorable for achieving a longer disease-free period: younger age, lower indocyanine green retention rate, solitary HCC with expansive growth, no microscopic portal invasion and lower activity of co-existing hepatitis., Conclusions: The importance of inflammatory activity in addition to aging, hepatic reserve and tumor characteristics was confirmed as a risk factor for recurrence by multivariate analysis. To achieve better disease-free survival, not only early detection but also suppression of co-existing hepatitis is necessary.

Little data are available regarding the effects of hepatitis virus serostatus and the severity of coexisting chronic inflammation on intrahepatic recurrence after hepatectomy for hepatocellular carcinoma (HCC). We investigated the extent to which these factors modified the prognosis of hepatectomized patients. A total of 274 patients treated in the period January 1981 to December 1996 were divided into three groups: antihepatitis C-positive (HCV; n = 144), hepatitis B surface antigen-positive and HCV antibody (Ab)-negative (HBsAg; n = 106), and HBsAg-negative and HCV Ab-negative (NBNC; n = 20). Positivity for HBV-related antibody in the HCV group was 76%. Histologic grading of inflammatory activity from coexisting hepatitis was determined according to Knodel's histological activity index (HAI) scoring system. Post-hepatectomy crude survival rates and disease-free survival (DFS) rates were compared, according to tumor characteristics, between the three groups. In the patients overall and also in the patients with a single nodular HCC, the HCV group had significantly higher HAI scores and preoperative serum aspartate aminotransaminase (AST) levels than the other two groups. When the patients were limited to those with a single nodular HCC, the crude survival was similar in the three groups with comparable tumor characteristics; however, the DFS was different (NBNC > HBsAg > HCV). When the patients were further limited to those with a single nodular HCC without microscopic extracapsular spread, in whom removal of the tumor was expected to be microscopically complete, the difference in the DFS became more marked. Irrespective of the viral serostatus, better crude and disease-free survivals were observed in the patients with lower AST levels (< or =50 IU/ 1) than in those with higher AST levels (>50IU/l). In contrast, there were no differences in survivals and HAI scores according to the presence or absence of HBV-related antibody in the HCV group. From our univariate analysis, we can conclude that the severity of virally induced inflammation, which was well correlated with viral serostatus, may be a factor that affects intrahepatic recurrence, which is more likely to originate from metachronous carcinogenesis. Prior co-infection of HBV in HCV patients may not be an adverse risk factor for intrahepatic recurrence.

This study was designed to clarify what differences the last 25 years have made in surgical results for patients with hepatocellular carcinoma (HCC). We examined results for 716 hepatectomized patients in four treatment eras: first era (1973-1980; n = 58), second era (1981-1985; n = 155), third era (1986-1990; n = 243), and fourth era (1991-1997; n = 260). Patient background, tumor characteristics, type of hepatectomy, treatment for intrahepatic recurrences, and surgical results in the four eras were compared by univariate analysis to clarify the factors that have contributed to or impeded progress in the surgical treatment of HCC. Although there were no significant chronological differences in liver pathology and surgical resectability, operative mortality was reduced to 2% in the fourth era, from 29% in the first era. With an increasing proportion of early-stage HCCs (TNM, stages I and II), the cumulative survival rate at 5 years improved in the course of the eras in our overall population of patients (12%, 31%, 38%, and 51%, respectively, for the first, second, third, and fourth eras) and in a subset of the population divided according to tumor stage. Also, we found a chronological improvement in the survival rate at 3 years after intrahepatic recurrence (10%, 28%, 36%, and 44%, respectively in the first second, third, and fourth eras). This improvement was associated with the establishment of an early detection program for intrahepatic recurrences. However, the recurrence rate was similar in any subset of the population through the four eras. Although this univariate study could not determine independent factors that contributed to the chronological progress in results for HCC surgery in the four eras, it is conceivable that the establishment of indication criteria for hepatectomy, an early detection program for primary and recurrent lesions, and the introduction of multimodal treatment for recurrence were contributory factors in this improvement. A strategy for alleviating the frequent recurrences originating from posthepatectomy metachronous carcinogenesis remains to be established.

This study was undertaken to evaluate whether topical cooling can alleviate ischemia/reperfusion injury, after continuous inflow occlusion during hepatectomy. Using a canine model of 70% partial liver ischemia (60 min), alteration in the subcellular (cytoplasm, mitochondria, nucleus) elements calcium, sodium, potassium, and chloride, and liver functions following reperfusion were compared between control livers and livers subjected to topical cooling down to 23 degrees +/- 4.9 degrees C by seeding ice slush over the ischemic lobe. The elements were determined by X-ray microanalysis using liver biopsy specimens. A similar clinical study was undertaken examining ten patients with hepatocellular carcinoma and chronic liver disease who underwent right-sided segmentectomy under continuous right inflow occlusion, five of whom were given topical cooling and five of whom were not. In the experimental study, postreperfusion worsening of liver function tests was significantly suppressed in the cooling group, which was associated with the suppression of subcellular Ca, Na, and Cl increases and K decreases after reperfusion. In the clinical study, the occlusion time was significantly longer in the hypothermic patients than in the normothermic patients, but no significant differences in postoperative liver function or postischemic increases in Ca, Na, or Cl and decreases in K were observed. These experimental and clinical findings suggest that topical cooling alleviates ischemic insult and enhances safe prolonged inflow occlusion.

Background: The liver and portal circulation contribute to production and clearance of endothelin-1 (ET-1). This study was undertaken to investigate what variables relate to the dynamics of ET-1 in hepatic resection and its clinical implication., Patients and Methods: On 20 patients with (n = 8) or without (n = 12) chronic liver disease who underwent hepatic resection, peripheral arterial and portal venous ET-1 were serially measured to determine a correlation with pre-, intra-, and postoperative variables., Results: The preoperative factors with which the portal ET-1 showed a positive correlation were the indocyanine green retention rate at 15 min (ICG R15) and portal venous pressure. The ET-1 clearance, as calculated from the difference between the portal and the peripheral ET-1 concentrations, was also correlated with the ICG R15. The peripheral ET-1 elevated significantly in the patients with increasing intraoperative blood loss or hepatic inflow occlusion. An increase in the portal ET-1 was correlated with an elevation of portal venous pressure after hepatectomy. Postoperative increase in serum bilirubin was closely correlated with the peripheral ET-1 at closure., Conclusion: The peripheral and portal ET-1 are correlated with not only preoperative hepatic reserve and portal venous pressure but also invasiveness of hepatectomy and postoperative course., (Copyright 1999 Academic Press.)

Oval cells that develop in the rat 2-acetylaminofluorene/partial hepatectomy (AAF/PH) model express the c-kit receptor tyrosine kinase (KIT) and its ligand, stem cell factor (SCF). We investigated the role of the SCF/KIT system in the development of oval cells using Ws/Ws rats, whose c-kit kinase activity was severely impaired owing to a small deletion in the kinase domain. On days 7, 9, and 13 after PH in the AAF/PH model, the development of oval cells was remarkably suppressed in Ws/Ws rats when compared with that of the control normal (+/+) rats. However, oval cells that developed in Ws/Ws rats expressed marker proteins of oval cells, such as alpha-fetoprotein (AFP), cytokeratin-19 (CK-19), and flt-3 receptor tyrosine kinase, similar to those of +/+ rats. Furthermore, labeling with [3H]-thymidine and immunostaining of Ki-67 showed that the proliferative activity of oval cells that developed in Ws/Ws rats was comparable with that of +/+ rats. The present results indicate that the signal transduction of the SCF/KIT system plays a crucial role in the development of oval cells, at least, in the rat AAF/PH model, and suggest that KIT-mediated signal transduction plays only a small role in determining the phenotype and in the proliferative activity of oval cells.

Background/aims: The crucial points of hepatic segmentectomy include an accurate mapping of proposed segment(s) to be resected and minimization of intra-operative blood loss. This study reports a surgical technique of hepatic resection for segment 7 and 8, and a benefit of supplemental use of topical cooling during consecutive right hepatic inflow occlusion., Methodology: From January 1993 to December 1996, ten hepatocellular carcinoma patients with pathologic livers underwent hepatic segmentectomy for segment 7 and 8. The mapping of the target segments was guided by en-bloc test occlusion of the relevant portal pedicles and subsequent discoloration, often after the opening hilum method. Of these, five tumors were resected with topical cooling by ice slush seeding during consecutive right inflow occlusion (cooling group) and the remainder without cooling (Normothermic group). In the normothermic group, inflow occlusion was carried out by cyclic clamping and unclamping method., Results: The inflow occlusion time was 54 +/- 9.4 min in the cooling group, much longer than in the normothermic group. Although the post-operative peak transaminase values were twice as high as those in the normothermic group, the blood loss was significantly less, and there was no detrimental effect of prolonged, consecutive ischemia on the other intra- and post-operative data., Conclusions: En-bloc taping of the target portal pedicles with or without the opening hilum method is useful in the mapping of segments, and the use of topical cooling is beneficial in prolonged inflow occlusion during complicated right-sided segmentectomy.

The clinical application of hypothermia dates back to the surgical treatment of blue babies (1949) and the early days of open heart surgery (1952), when generalized cooling was employed. The induction of hepatic hypothermia began with whole-body cooling in experimental models in 1953 and clinically in 1961. It was designed to minimize the ischemia-reperfusion injury associated with hepatic inflow occlusion. Body surface cooling and cooling via an extracorporeal circuit, however, were not widely accepted for hepatic surgery because of the adverse effects on the extrahepatic organs. Consequently, with the introduction of improved venovenous bypass techniques, in situ cold hepatic perfusion has been used in selected patients since 1971. In situ hypothermic hemihepatic perfusion, introduced in 1995, prevents an ischemic insult to the contralateral hepatic lobe. Topical cooling using ice slush under total or hemihepatic inflow occlusion was reported in 1993. This technique does not require cumbersome hypothermic perfusion equipment. In attempts to minimize intraoperative bleeding by vascular occlusion, the liver surgeon must consider the benefits and technical demands of hepatic hypothermia.

Hypothermia of the liver was developed to alleviate the ischemic insult induced by prolonged hepatic inflow occlusion. The use of hypothermia in liver surgery dates back to an experiment in 1953, which showed that the safe normothermic ischemic time of 20 minutes could be prolonged to 60 minutes with generalized cooling. Generalized cooling was first used clinically as an adjunct to hepatectomy with inflow occlusion in 1961. This method has not been widely used due to its adverse effect on systemic hemodynamics. Subsequently, methods of inducing hypothermia evolved to encompass in situ isolated hypothermic perfusion under total vascular exclusion, and then to hemihepatic hypothermic perfusion without the need for total vascular exclusion but with topical surface cooling. These technically feasible modifications were newly devised by our group. Our series of 39 hepatectomized patients, who underwent right-sided hepatectomy under hemihepatic inflow occlusion combined with topical surface cooling (October 1990 to April 1997), demonstrated that even in livers associated with chronic liver disease, if surface topical cooling was interposed, a consecutive hemihepatic inflow occlusion time of as long as 60 +/- 23 minutes was acceptable without cyclic recirculation. In addition, no further ischemic insult occurred compared with a normothermic patient series with a significantly shorter occlusion time. Our data lead to the conclusion that the topical surface cooling method can be a useful and feasible adjunct to complicated hepatectomy in which inflow occlusion time is anticipated to prolong.

This article describes the surgical techniques and indications of laparoscopic partial hepatectomy, which is not as widely available as laparoscopic cholecystectomy. Three patients with hepatocellular carcinoma and associated severe liver cirrhosis were candidates for this technique from July 1993 to August 1994. The tumor size was 4 cm or less and all the tumors were located in segment 5 or 8 which had grown nodularly and protruded from the liver surface. A microwave tissue coagulator was used for parenchymal dissection under ultrasonographic guidance in a gas-less method with or without low-pressure pneumoperitoteum of 4 mmHg. The principle of dissection consists of tissue coagulation and fragmentation with dissecting forceps. Three hepatectomies were performed uneventfully without blood transfusion and the patients rapidly returned to their preoperative conditions. The laparoscopic partial hepatectomy can be an option of treatment in selected cases where the tumor can be removed by minor, superficial resection.

Background: A safe limit for normothermic consecutive portal triad occlusion in hepatectomy for the cirrhotic liver was believed to be around 30 minutes. Possibly, the occlusion time can be prolonged by cooling the ischemic liver in vivo. We describe the technique of segmentectomy assisted by topical cooling and its usefulness in prolonging the hepatic inflow occlusion time for cirrhotic livers without causing further ischemic injury., Study Design: Fifty patients with hepatocellular carcinoma and chronic hepatic disease who underwent right-sided segmentectomy under hemihepatic inflow occlusion were divided into two groups: normothermic (n = 27), and hypothermic with hemihepatic topical cooling using ice slush (n = 23). Segmentectomies were carried out in the same way in both groups, guided by initial enbloc ligation of the corresponding portal pedicles., Results: The mean right hepatic inflow occlusion time was significantly longer in the hypothermic group than in the normothermic group (53 +/- 22 minutes compared with 17 +/- 9.3 minutes). Despite the significant difference in occlusion time, no differences were found in the recovery of hepatic functions and the incidence of postoperative complications between the groups. Intraoperative blood loss was significantly less in the hypothermic group., Conclusions: The hepatic right-sided partial inflow occlusion time can safely be prolonged to 60 to 90 minutes in the presence of cirrhosis without causing another injury from ischemia and intermittent reperfusion.

Hepatectomy was performed under in situ right lobar hypothermic perfusion combined with hepatoprotective agents in six patients who had hepatocellular carcinoma and coexisting liver disease. Following occlusion of the right hepatic vein and the right portal pedicle, in situ cold perfusion was initiated using chilled Ringer's lactate infused through a cannula placed in the right main portal vein. The right superior segments were resected in a bloodless field. The liver was cooled to 22-26 degrees C for 40 to 80 minutes with no significant changes in systemic hemodynamics or body temperature. Postoperative liver functions showed no marked derangement; the mean peak GPT was 221 U and the mean peak total bilirubin 2.3 mg d/l. Local cooling minimizes the risk of ischemia/reperfusion injury in this very vulnerable population, yet gives the surgeon adequate time to perform a challenging resection in a bloodless field.

Objective: This study reports further refinement of a prediction scoring system, which was established in 1980 as a guide to determine a safe limit for hepatectomy, based on 10 years of use., Summary Background Data: In the past, whether major resection was safe was judged empirically from the net resection volume or the residual hepatic volume combined with the patient's liver function. However, such judgment was not based on objectively defined criteria., Methods: Patients with hepatocellular carcinoma (HCC; n = 376) and metastatic cancer (n = 58) who had hepatectomy at some time from 1981 through 1990 were entered into this study. A prediction score (PS) was computed using a multiple regression equation that consists of computed tomographic scan-estimated resection rate, indocyanine green retention rate, and the patient's age. A PS greater than 55 was classified as a risky zone, a PS of 45 to 55 was considered borderline and a PS less than 45 was a safe zone., Results: With HCC and chronic liver disease, all patients in the risky zone died, whereas 33% in the borderline zone died and 7.3% died who were in the safe zone. With metastatic cancer with normal liver, all patients in the risky zone died, whereas no patient in either the borderline or safe zones died. The major cause of death in the risky zone was liver failure due to excessive resection. In the borderline and safe zones, liver failure developed primarily after abdominal sepsis or pulmonary infection, particularly for those with adverse prognostic factors such as disturbed glucose tolerance, lower platelet count, and higher indocyanine green retention rate., Conclusion: Prediction scores can eliminate deaths related to excessive resection for patients with normal or injured livers. When patients have adverse prognostic factors, careful surgery and postoperative management is mandatory to avoid liver failure triggered by intra- or extra-abdominal sepsis, even if the score remains in a borderline or safe zone.

Purpose: Although autotransfusion is used in a wide variety of surgical procedures, it is largely unexplored for use in oncologic operations. This study evaluates the impact of autotransfusion in hepatectomy for hepatocellular carcinoma., Patients and Methods: Fifty-four patients with hepatocellular carcinoma underwent hepatectomy by receiving autologous blood transfusion (group 1). Fifty patients receiving homologous blood without any autologous blood were matched for comparison (group 2)., Results: No statistical differences were found between group 1 and group 2 patients in cumulative recurrence rates (62.8% vs 67.3%, respectively) or cumulative survival rates (61.9% vs 52.8%). Autologous transfusion reduced the mean volume of homologous blood used (814 mL vs 3466 mL). The mean postoperative hemoglobin level, platelet count, prothrombin time, and partial thromboplastin time were comparable between groups., Conclusions: Autotransfusion is a safe and effective procedure in patients with hepatocellular carcinoma undergoing hepatectomy.

We compared liver volume and function kinetics after partial hepatectomy according to extent of resection and severity of coexisting liver disease in 57 adults with uneventful postoperative courses. Liver volume and massiveness of resection, or resection rate, were estimated on computed tomography. Patients were categorized into three groups on the basis of reaction rate: small (< 30%), medium (30%-50%) and large (> 50%). The regenerative patterns of normal livers in the medium and large groups consisted of three phases: a rapid increase during the first month, some decrease in the second month and a final, slower increase. This contrasted with the pattern of injured livers with chronic hepatitis or cirrhosis, which generally showed a phase of less rapid, gradual increase. The regeneration rate (volume gain, cm3/day) during the first month was found to be proportional to resection rate in the presence or absence of liver disease. Normal livers regenerated at least twice as rapidly as injured livers in patients with comparable resection rates. Normal livers reached plateau levels within 1 to 2 mo regardless of the massiveness of resection, whereas regeneration took 3 to 5 mo in injured livers. Liver function (albumin, bilirubin) recovered concomitantly with liver volume in the medium group, whereas in the large group they generally returned to their initial values behind volume restoration, particularly in cirrhotic patients. In conclusion, human liver regeneration is strongly influenced by the massiveness of the resection and presence of coexisting liver disease. However, we found that some cirrhotic livers can regenerate, albeit more slowly and less completely, as long as the extent of hepatectomy remains within safe functional limits.

The techniques and results of nine patients who underwent hepatic resections with surface and/or in situ perfusion cooling with hepatoprotective agents of PGE1 and Ulinastatin are described. Eight hepatocellular carcinomas associating chronic liver disease and one hemangioma are included in this report. Complete vascular exclusion technique with the centrifugal blood pump driven veno-venous bypass was employed on the three patients with giant liver tumors. The bypass technique with liver cooling made it possible to carry out otherwise impossible and risky resections, particularly in case with neoplastic involvement of the inferior vena cava. The other patients were subjected to segmentectomy of the right lobe under hemihepatic surface cooling or in situ hypothermic perfusion by chilled Lactate Ringer for vascular arrested right lobe. Liver tissue temperature dropped down to 19 degrees C to 26 degrees C during cooling. Postoperative physiologic changes were minimal irrespective of prolonged continuous vascular interruption of the cirrhotic liver for up to 60 min. There was no mortality and major morbidity in the present series. Liver cooling can be an useful hepatoprotective method for the cirrhotic liver in the hepatic resection with vascular occlusion.

Background: The current study was undertaken to investigate whether or not tumor cells are dislodged into the portal venous stream during hepatic resection for hepatocellular carcinomas., Methods: A catheter was placed using echo guidance into the portal branch through the mesenteric vein in 31 patients. Cytologic examinations were done on multiple blood samples at various operative stages., Results: Tumor cells were recovered in 7 of 31 patients in whom the tumor sizes were more than 5 cm and portal invasions were found microscopically and/or macroscopically. By contrast, the remaining 24 tumors were less than 5 cm in size and showed negative portal invasions. Recovery of the tumor cells was found, not during the earlier operative stage of mobilization or rotation of the hepatic lobe, but during the later stages of hilar dissection or hepatic parenchymal dissection., Conclusions: The portal pedicles should be divided before hepatic dissection in segmentectomy and lobectomy to lessen the chance of dissemination of intravasated tumor cells.

Experimental data of autotransfusion in hepatectomy for hepatocellular carcinoma and clinical data from forty-eight patients were analyzed. By a density gradient medium, cancer cells banded at a lower density (1.056 g/ml: mean) than red blood cells (1.105 g/ml: mean). The centrifuge autotransfusor separated the red blood cells from cancer cells and all sections of cell blocks of the blood harvested intraoperatively in 20 cases revealed no malignant cells. The average requirement of bank blood was significantly reduced in these patients than in patients without autotransfusion during one year before, 1573 ml to 478 ml (whole blood of concentrated red blood cells) (p less than 0.05). These findings indicate that autotransfusion in hepatectomy for hepatocellular carcinoma is safe and practical.

The present study was performed to clarify the regeneration and splenic changes following hepatectomy. The processes were compared between the three classes (L greater than 50%, 30 less than M less than 50%, S less than 30%) categorized by the resection rate (%) calculated by CT scan in the 26 cirrhotics (LC) and 22 non-cirrhotics (N). Hepato-splenic volumes were serially measured by CT scan. The regenerative speed (cm3/day) of the remaining liver were significantly higher in the N group than LC and in the larger resection class. Liver functions tended to return to the initial levels behind the time of restoration of the liver volumes. The delay was partially caused by the posttransfusion hepatitis (PTH), which developed more frequently with an increase of fresh frozen plasma transfusion given. The changing pattern of splenic size was strongly regulated by the massiveness of hepatectomy. Newly developed posthepatectomy esophageal variceal ruptures were endoscopically determined in the 10% of the 129 late deaths after hepatectomies for hepatomas (1973-1987) with or without recurrences and its occurrence was found to be enhanced in the massive resection in the LC group ended with the poor regeneration of the liver and persistent postoperative splenic enlargement.

By use of computed tomography measurements, preoperative estimations of the safe limits of hepatic resection have been attempted in 38 patients with primary hepatocellular carcinoma who underwent hepatectomies of varying degrees. Twenty eight (73.7%) of these 38 patients had cirrhosis. Preoperative computed tomography estimations of volume of liver and hepatocellular carcinoma have been compared with measured volume of the resected specimens. The average differences between the estimated and actual volumes of resected liver and hepatocellular carcinoma were within 10% and 7%, respectively. To quantify the extent of hepatectomy, the "parenchymal hepatic resection rate" (formula: see text) was provided. As an indicator of functional impairment of the liver, the retention rate 15 minutes after an intravenous injection of indocyanine green dye (ICG) at 0.5 mg/kg (ICG R15) was used. A close correlation was observed between the parenchymal hepatic resection rate and ICG R15 relative to the patients' outcome, indicating their potential use in the prediction of the safe limits of hepatectomy and the probable development of posthepatectomy liver failure.

The present study reports how we indicate safe and beneficial hepatectomy for hepatocellular carcinomas from the experiences of the consecutive 335 patients treated during 1973 and Dec. 1988. The first question was to appreciate the preoperative prediction of a safe limit of hepatectomy. We established a multiple regression formula consisting of several factors such as CT estimated resection rate, ICG retention rate and patient's age at the end of 1980. Since the beginning of 1981, we have been employing the formula as a major indicator guiding treatment option, leading to a significant drop of postoperative, liver failure. The influencing prognostic factors have been determined to be vascular invasion, intrahepatic metastasis, extent of resection regulated by hepatic reserve and tumor size in the decreasing order. From the viewpoint of the biological aspect, the tumors with DNA diploid pattern showed significantly better prognosis than those with non-diploid ones. The current surgical techniques of hepatectomies with en-bloc ligation of Glisson's vessels or with vascular exclusion supported by veno-venous bypass have greatly encouraged the recent surgical results achieving 56% of 4-year survival rate (1985-1988).

This article reports a multiple regression equation for prediction of posthepatectomy liver failure. In phase I, using the correlations between 17 preoperative parameters (Xi) and the postoperative course scored (Y) of the past 36 hepatectomized patients, we proposed the following multiple regression equation: Y = -110 + 0.942 X resection rate (%) + 1.36 X ICG retention rate (%) + 1.17 X patient's age + 5.94 X ICG maximal removal rate (mg/kg/min). With the equation, the calculated Y value (prediction score) of these patients revealed that prediction scores of the eight nonsurvivors with liver failure were more than 50 points while those of the 28 survivors were 50 points or less. In phase II, the relationships between early prognosis and a precalculated prediction score were prospectively found the same as that seen in phase I. These findings indicate that our formula is a useful prognostic index for prediction of posthepatectomy liver failure.

Since most patients with primary hepatocellular carcinoma (HCC) in Japan have cirrhosis, indication of hepatectomy is now tending to be limited to a single segment or subsegment for small HCC. However, pathologic examinations on the resected specimens have revealed that intrahepatic metastases were associated in more than 50% of HCC smaller than 5 cm, indicating that hepatic resection is required in these patients to be as extensive as possible within the limit of hepatic reserve. Our multiple regression equation provided for the preoperative prediction of an early prognosis of hepatectomized patient has successfully reduced the incidence of postoperative fatal liver failure from 17% before 1980 to 6% after 1981. The sequential volumetric CT measurements of the remaining liver after hepatectomy in humans has revealed an ability of the cirrhotic liver to regenerate nearly as normal liver in patient who has well tolerated hepatic resection. Ligation or embolization of the portal vein of an affected side in patients with unresectable HCC has rapidly induced a compensatory hypertrophy in the contralateral lobe to be remained, indicating the potential usefulness of this procedure to relatively reduce the loss of functioning liver mass by the proposed lobectomy.

In order to avoid postoperative liver failure after resection of the liver associated with chronic injury, an accurate evaluation of resectability should be established based on available preoperative factors. This paper reports our multiple regression analysis of the correlation between the prognosis (Y) after hepatectomy and the seventeen sorts of preoperative factors (X) to obtain a multiple regression equation predicting the prognosis. The analysis was employed on 36 patients who underwent hepatectomy of varying extent. The prognosis (Y) was represented in our scoring system (0-100 point) determined by the degree of postoperative liver failure. As a result, the following multiple regression equation, consisting of four prognostic factors, was derived. Y = -110 + 0.942X1 + 1.36 X2 + 1.17 X X3 + 5.94 X X4 (X1 = resection rate of the liver tissue calculated by C.T. Scan, X2 = retention rate of ICG (15)%, X3 = age, X4 = maximal removal rate of ICG (mg/kg/min). The Y-score of 36 patients computed by this equation revealed that: Y-scores of the 8 patients with fatal liver failure had wholly more than fifty points, and Y-scores of the remaining 28 patients with a favourable course had less than fifty points. The validity of this equation was ascertained through the predetermined Y-score, in which critical point is set about fifty points or more, for other 49 patients. By this equation, accurate evaluation of resectability for liver tumors has been made possible.

The present study was aimed at answering the question of what patients with hepatocellular carcinomas could obtain long-term benefits from resectional therapy. The 239 patients hepatectomized from 1973 through 1986, with 33 tumor-free, 3-year survivors, were the subjects of the study. The following criteria for long-term survival were determined. (a) It is essential that there are no macroscopic tumor thrombi in any of the vessels. (b) It is desirable that there is no daughter lesion and a tumor size of less than 5 cm. (c) When there are daughter lesions, they should be confined to the region adjacent to the main tumor and only few in number. (d) Resection should be performed on a large scale within the limit of safety, especially for the tumors with daughter lesions. (e) The surgical margin should be free of tumor though the exposure of the tumor capsule is not always incompatible with long survival.