Statistical analysis of prognosis in liver cirrhosis with reference to prediction of the high risk group for hepatocellular carcinoma

A method for determining the high risk group of hepatocellular carcinoma (HCC) and the factor related to long-term prognosis in patient with liver cirrhosis (LC) was developed, using a kind of multivariated analysis. Eighty-six histologically proven cirrhotic patients were retrospectively classified into three groups: Group 1, 18 cases with a complication of HCC during the more than one year follow-up after the diagnosis of LC was confirmed, Group 2, 27 cases died of other causes than HCC during the less than 3-year follow-up and Group 3, 41 cases alive more than 3 years in the absence of HCC. Twenty-two parameters were selected from the routine liver function tests, physical examination and the history, obtained at the time of diagnosis. A discriminant function for differentiating the above mentioned groups was derived using the theory of quantification. An investigation was also undertaken so as to determine which one of the 22 parameters was effective for differentiation. Employing this function, correct differentiation was made in 84 out of 86 cases (97.7%), and only 2 cases were erroneously differentiated. Among the parameters, age, sex, albumin, GOT, GPT, total protein, ascites and edema contributed greatly to differentiating Group 1 from Group 2 and 3. Likewise, LDH, albumin, ZTT, TTT, RBC, ascites and the family history of liver disease influenced largely on differentiation of Group 2 from Group 3.