Kunihiko Ohnishi

Natural history of hepatocellular carcinoma and prognosis in relation to treatment. Study of 850 patients

A total of 850 patients with hepatocellular carcinoma seen during the last 8 years were analyzed retrospectively for survival in relation to treatment and disease stage. A new staging scheme based on tumor size, ascites, jaundice and serum albumin was used. Clearly, the prognosis depended on disease stage. The median survival of 229 patients who received no specific treatment was 1.6 months, 0.7 month for Stage III patients, 2.0 months for Stage II, and 8.3 months for Stage I. The median survival of Stage I patients who had hepatic resection (n = 115) was 25.6 months and Stage II patients with resection (n = 42) was 12.2 months. In patients who had a small cancer (≤25% of liver area in size) the median survival was 29.0 months. Survival of the surgically treated patients, which represented a highly selected group, was better than that of medically treated patients of a comparable stage. Median survival of Stage I medically treated patients (n = 124) was 9.4 months, for Stage II (n = 290) 3.5 months, and for Stage III (n = 50) 1.6 months. Medical treatment prolonged survival in Stage II and III patients, but not in Stage I. Transcatheter arterial embolization gave a better survival compared with chemotherapy, whether intra‐arterial bolus administration of mitomycin C, systemic mitomycin C, or oral/rectal tegafur, in Stage II. Among various chemotherapeutic modalities, intra‐arterial bolus injection was superior to systemic chemotherapy in survival in Stage II. In Stage III, chemotherapy improved survival as compared with no specific treatment. The major causes of death were hepatic failure and gastrointestinal bleeding, probably due to the coexistent advanced cirrhosis. These results in survival are much improved over the past reports, and the differences are probably a result of earlier diagnosis and frequent hepatic resections.

Clinical features and prognosis of hepatocellular carcinoma with reference to serum alpha-fetoprotein levels. Analysis of 606 patients.

Clinical features of hepatocellular carcinoma (HCC) with normal serum alpha‐fetoprotein (AFP) levels were compared with those of AFP‐positive cases. A total of 606 patients were divided into four groups based on their serum AFP levels at the time of diagnosis: group 1 (≤20 ng/ml, N = 125), group 2 (20–1000 ng/ml, N = 256), group 3 (1000–10000 ng/ml, N = 149), and group 4 (>10000 ng/ml, N = 76). Increasing prevalence of group 1 patients and decreasing prevalence of group 4 were noted during the last 9 years. The proportion of hepatitis B virus‐positive cases was significantly lower in group 1 than in group 4. The serum glutamic oxaloacetic transaminase/serum glutamic pyruvic transaminase ratio was found to be significantly higher in group 4 than in group 1 regardless of the size of the tumors. The survival rates were compared among the four groups in size matched cases since the size of the tumor significantly influenced their prognosis. The median survival in relatively small HCC patients (≤5 cm in diameter, N = 199) were 24.6 months for group 1, 20.6 months for group 2, and 13.7 months for group 3 patients; and those in large HCC (≤50% in the proportion of the tumor area, N = 200) were 15.1 months for group 1, 6.3 months for group 2, 5.8 months for group 3, and 5.2 months for group 4. Thus, serum AFP values at the time of presentation are not only of diagnostic value, but also of prognostic significance in patients with HCC.

Incidence of portal vein thrombosis in liver cirrhosis

Portal vein thrombosis was thought to be a common complication of liver cirrhosis in the past. The incidence of angiographically demonstrable portal vein thrombosis was studied in 708 consecutive patients with unequivocal cirrhosis seen in the past 10 yr in whom either transhepatic portography or superior mesenteric arterial portography clearly delineated the major portal vein system. Excluding 2 cases that were thought to be associated with past splenectomy, there were 4 cases of portal vein thrombosis related to cirrhosis, all in a decompensated stage. The calculated incidence of portal vein thrombosis was 0.573% of all cirrhotic patients without splenectomy in the past. They constituted 23.5% of the 17 cases of extrahepatic portal vein obstruction encountered during the same period. There were 78 cases of idiopathic portal hypertension similarly studied angiographically, and the incidence of portal vein thrombosis unrelated to splenectomy was 2.86%. A statistical survey based on 247,728 necropsies recorded in the Japan Autopsy Registries of 1975-1982 showed a 0.05489% incidence of portal vein thrombosis and a 6.58857% incidence of cirrhosis of all types among them, suggesting that portal vein thrombosis is not a common complication of cirrhosis in Japan in recent years.

Clinical Study of Eighty-six Cases of Idiopathic Portal Hypertension and Comparison With Cirrhosis With Splenomegaly

The clinical features of 86 cases of idiopathic portal hypertension, the equivalent of hepatoportal sclerosis in the United States and of noncirrhotic portal fibrosis in India, are presented. This disease is characterized by overt splenomegaly with pancytopenia, portal hypertension, and relatively mild abnormalities in liver function tests. Although differential diagnosis from liver cirrhosis is not always easy, liver histology, laparoscopy, portography, hepatic venography, and measurement of wedged hepatic vein pressure are useful in diagnosis. Prognosis is relatively benign if variceal bleeding is controlled or prevented, and the disease does not progress to cirrhosis. The etiology is still undetermined, but the liver pathology characterized by occlusive changes of the intrahepatic portal radicles, portal and periportal fibrosis, and irregularly distributed parenchymal atrophies suggests some sort of portal venopathy that causes decreased portal perfusion of peripheral liver parenchyma. These patients with idiopathic portal hypertension were compared with 63 cases of cirrhosis with splenomegaly and 80 cases of cirrhosis without splenomegaly. There was some similarity in hematologic findings between idiopathic portal hypertension and cirrhosis with splenomegaly, but the basic disease process seemed distinctly different. The cause of marked splenomegaly does not seem to be simply congestion, and remains an enigma.

Role of portal and splenic vein shunts and impaired hepatic extraction in the elevated serum bile acids in liver cirrhosis.

To study the mechanism for the elevation of serum bile acids in liver cirrhosis, bile acid concentrations were measured in the portal, superior mesenteric, and splenic veins, using percutaneous transhepatic catheterization, and compared with those of peripheral veins in 41 patients with mild to moderately advanced cirrhosis. The demonstrated gradient of bile acid concentration was superior mesenteric vein greater than portal vein greater than peripheral vein nearly equal to splenic vein, suggesting that the superior mesenteric vein is the main route of transport for the intestinally absorbed bile acids. Bile acid concentrations in peripheral vein were correlated with the measured portal and splenic vein shunt indexes. The ursodeoxycholic acid oral tolerance test carried out in 10 patients during portal vein catheterization demonstrated that hepatic extraction of this bile acid was correlated with indocyanine green clearance and that it was inversely correlated with portal vein shunt index. These findings are consistent with the view that distorted hepatic blood flow has a significant role in elevating serum bile acid, at least in patients with moderately advanced liver cirrhosis.

Formation of Hilar Collaterals or Cavernous Transformation After Portal Vein Obstruction by Hepatocellular Carcinoma: Observations in Ten Patients

A total of 155 patients with hepatocellular carcinoma were studied by celiac and superior mesenteric angiography. Complete (9 patients) or near complete (1 patient) obstruction of the portal vein and formation of hepatopetal collateral veins in the porta hepatis, or the so-called cavernous transformation of the portal vein, were seen in 10 patients. In 4 patients, the first angiogram did not show cavernous transformation, but on the follow-up angiograms cavernous transformation was present. The suggested interval between obstruction of the portal vein and formation of cavernous transformation was no more than 5 wk. The mechanism of cavernous transformation and its clinical implications are briefly discussed.

Experimental Portal Fibrosis Produced by Intraportal Injection of Killed Nonpathogenic Escherichia coli in Rabbits

An attempt was made to develop an animal model for the study of the etiology of noncirrhotic portal fibrosis or idiopathic portal hypertension based on the assumption that it is related to chronic abdominal infection. Rabbits were given killed nonpathogenic Escherichia coli intraportally or intravenously. The animals to which a mixture of killed E. coli and rabbit antiserum (aggregated E. coli) was given intraportally developed remarkable histologic changes in the liver. The early inflammatory reactions in the portal area and parenchyma were followed by rapid disappearance of inflammation and development of portal fibrosis with bile duct proliferation. Three intraportal challenges with aggregated E. coli were sufficient to produce pronounced portal fibrosis, although there was considerable variation in response among individual animals. This procedure also produced splenomegaly, and in some animals marked portal hypertension. Injection of nonaggregated killed E. coli into the portal vein or aggregated E. coli into the ear vein also caused similar hepatic changes, but they were milder in degree. These histologic changes resemble portal fibrosis seen in idiopathic portal hypertension and, less closely, pericholangitis associated with inflammatory bowel disease in humans.

Portal hemodynamics in idiopathic portal hypertension (Banti's syndrome): Comparison with chronic persistent hepatitis and normal subjects

A comparative study of portal hemodynamics was made in 17 patients with idiopathic portal hypertension, 5 patients with chronic persistent hepatitis having no portal hypertension, and 21 healthy adults who served as the control for certain measurements. Venous pressures were measured by portal and hepatic vein catheterizations, blood flow by the pulsed Doppler flowmeter, organ volume by computed tomography, and intrahepatic shunt index by 99mTc-macroaggregated albumin instilled in the portal vein. The patients with idiopathic portal hypertension were divided into two groups: group A (n = 8) and group B (n = 9), consisting of those who respectively had portal venous flow per liver volume above and below the mean + 2 SD of healthy adults. In group A, portal vein pressure was moderately elevated, portal venous flow was significantly increased compared with the control, and portal vascular resistance was not much altered. In group B, portal vein pressure was markedly elevated above that of control, portal venous flow was comparable, and portal vascular resistance was significantly elevated. Splenic venous flow measured in the splenic vein between the left and short gastric veins was markedly increased in groups A and B, the increase being greater in the former. It was concluded that in some patients with idiopathic portal hypertension, increased portal venous flow, partly a result of increased splenic venous flow secondary to splenomegaly of an undetermined process, is the main contributor initially to the elevation of portal vein pressure; in others, possibly later, increased portal vascular resistance plays an important role.

Transformation of inferior vena caval thrombosis to membranous obstruction in a patient with the lupus anticoagulant

A 24-yr-old woman with hemolytic anemia developed multiple thrombosis of the hepatic vein and inferior vena cava. She was found to have circulating lupus anticoagulant that could have been causally related to the thrombosis and hence the Budd-Chiari syndrome. On her first admission to the hospital vena cava and hepatic vein catheterizations revealed partial thrombotic occlusion of the cava at the level of the diaphragm, which was subsequently transformed into complete membranous obstruction. The right hepatic vein, which was patent on the first admission, was also completely occluded. These observations support the theory that membranous obstruction of the inferior vena cava is a sequela to inferior vena caval thrombosis rather than a congenital anomaly.

Direction of splenic venous flow assessed by pulsed Doppler flowmetry in patients with a large splenorenal shunt

We studied the direction of blood flow in the splenic vein, using a combined ultrasonic system consisting of an electronic sector scanner and a pulsed Doppler flowmeter, in 21 patients with a large spontaneous splenorenal shunt demonstrated by angiography. Pulsed Doppler flowmetry revealed hepatofugal flow in the splenic vein in all 11 patients with chronic spontaneous hepatic encephalopathy, and hepatopetal flow in 10 patients without encephalopathy. In the former, hepatofugal flow of part of the superior mesenteric venous blood into the splenorenal shunt was corroborated by the venogram obtained after superior mesenteric arteriography. In 5 patients without a history of hepatic encephalopathy, superior mesenteric arteriography demonstrated hepatofugal flow of part of the superior mesenteric venous blood into the splenorenal shunt. Pulsed Doppler flowmetry, however, revealed hepatopetal flow in all of these patients, suggesting that the angiographic finding of hepatofugal flow may have represented an artifact.