Prognostic factors of thyroid cancers in Ramathibodi Hospital

The need to create different prognostic models for different populations is obvious after the findings that significant prognostic factors can change from one population to another. The purpose of this study is to fit a linear logistic model to a data set of Thai patients with thyroid cancers to find important prognostic factors and their effects on short-term mortality. Among the 89 cases selected, 20 died within 4 years after initial diagnosis and 69 survived more than 4 years. Age at diagnosis, sex, histological type and differentiation, microscopic vascular invasion, and extent of tumor spreading were included as independent variables. These variables were selected into the model by step-down selection procedure. Age at diagnosis was found to be the most significant prognostic factor followed by differentiation, vascular invasion and extent of spreading, respectively. However, vascular invasion appeared to exert the strongest impact on the prognosis with the risk of death increasing 62 fold for those showing microscopic evidence of this parameter. The odds ratios for the extent of spreading, age at diagnosis and differentiation were 7.5, 1.1 and 0.02, respectively.