Παραμετρικά μοντέλα επιβίωσης ασθενών για διάφορα είδη καρκίνου
Parametric survival models for various types of cancer
View/ Open
Keywords
Ανάλυση επιβίωσης ; Παραμετρικά μοντέλα επιβίωσης ; ΚαρκίνοςAbstract
This thesis investigates the factors influencing the survival rates of patients with brain, breast, and pancreatic cancer through the application of descriptive statistics and parametric survival models, utilizing the SEER database. The primary objective is to identify determinants of mortality in these cancer types.
The first chapter provides a comprehensive introduction to brain, breast, and pancreatic cancers, highlighting their clinical significance and the role of statistical analysis in elucidating survival patterns. It sets the stage for understanding how statistical methodologies contribute to insights into patient outcomes.
The second chapter delineates the methodological framework employed in the study, focusing on the theoretical underpinnings and practical applications of proportional hazard models and parametric survival models. This section elucidates the analytical techniques utilized to interpret survival data effectively.
In the third chapter, a detailed descriptive analysis of the datasets is presented. This includes various graphical representations and statistical summaries that illustrate the distribution and characteristics of the data related to the survival of patients across the three cancer types.
The fourth chapter is dedicated to the calculation and testing of correlations within the datasets. It explores the relationships between quantitative and qualitative variables with respect to the "Survival Months" variable, accompanied by appropriate graphical depictions. Subsequently, the analysis extends to investigate correlations with the "Survival Status" variable, incorporating similar methodological approaches.
The fifth chapter focuses on the adaptation and application of proportional hazard models and accelerated failure time models. The response variables examined include survival months and survival status. The chapter aims to identify the most appropriate statistical models based on rigorous statistical testing and evaluation.
The final chapter summarizes the principal findings of the research, synthesizing the results from the statistical analyses and model fitting procedures. It also proposes avenues for future research and recommendations for enhancing strategies in the management and treatment of brain, breast, and pancreatic cancers.