Multivariate analysis of factors affecting postoperative survival in malignant astrocytoma

Summary Fifty-eight patients with supratentorial malignant astrocytoma were analyzed statistically to evaluate the actors most important for predicting postoperative survival. Clinical information such as age, sex, duration of preoperative symptom, Karnofsky score at admission and at discharge, location of tumor, amount of tumor removal, number of operations, and postoperative survival in months, together with data on radiation and chemotherapy were analyzed by chi square test, t-test, and multivariate analysis. Cytofluorometric DNA quantification using paraffin embedded specimens was also performed in 20 cases and these data were also evaluated.Multiple correlation coefficient, and therefore the total statistical accuracy, increased to 0.824 when data of DNA quantification, percentages of S phase cells and of polyploid cells, were included. Multivariate analysis revealed that 6 items were the major factors for predicting postoperative survival, i.e. the location of tumor, the Karnofsky score at discharge, the percentage of S phase cells, the number of operations, the percentage of polyploid cells, and the amount of tumor removed. Based on this analysis, the estimated survival time could be expressed as a formula.This work was partly supported by a Grant-in Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan.