Prediction of Prognosis in Patients with Gastric Cancer by Quantitative Morphology and Multivariate Analysis

Abstract

In 76 patients, curatively resected gastric carcinoma was studied by quantitative morphologic methods. The relationship between the morphologic data and the survival time was calculated by univariate and multivariate statistical methods. The univariate analysis by means of the Kaplan-Meier test shows that besides the variable of the TNM schedule, the Laurén classification, stromal fibrosis, and neutrophilic infiltration present significant correlation to survival time. The multivariate analysis by multivariate survival analysis with covariates in accordance with Cox results in an estimated survival curve, nearly identical to the actual survival curve, by the two morphological variables ‘N-staging’ and ‘Laurtn classification’.

Key Words:

• Prediction
• prognosis
• quantitative morphology
• survival time
• statistics
• TNM