| Summary: | 碩士 === 國立中山大學 === 應用數學系研究所 === 106 === Endometrial cancer is the third largest gynecologic cancer, among which the common risk factors are age, stage, grade, proportion of myometrium invasion and tumor size, and so on. Therefore this study used endometrial cancer patient data of kaohsiung veterans general hospital to find related risk factors which infulenced by death and lymph node metastasis. Many clinical gulidelines for the doctor diagnosis in the clinical medicine, such as proportion of myometrium invasion by 0.5, tumor size by 2 centimeters, but do these clinical gulidelines conform to our endometrial cancer data? We need further anal- ysis. Therefore we aimed at finding the optimal number of cut points and their locations in survival cox proportional hazard model and lymph node metastasis logistic regression model. We used likelihood ratio test to find number of cut-off points and location of myometrium invasion proportion and tumor size,then compared cut points from clinical gulideline with our best cut-off points by the performace of model fitting. We provided the function which found optimal cut-off points of risk factors in logistical regression model for furture researchers.
In addition to research of endometrial cancer, endometrial tissue growth in other organs, the so-called endometriosis. The patients with endometriosis has the possibility of canceration; therefore used Taiwan health insurance database in endometriosis patients data. We analyzed the related risk factors for breast cancer by Cox proportional hazards model.
|
|---|
