We are interested in a specific aspect related to modelling disease onset, i.e. the study of family-specific risk. Generally, quantifying the family-specific risk of developing a disease is crucial to improve the patient’s survival, and healthcare data provide an essential tool to estimate such quantity. Specifically for breast cancer, early detection is particularly important to increase the chances of a successful treatment. In this setting, risk estimation allows for the identification of subjects at higher risk of developing breast cancer, so that they can be the target of tailored and more intensive screening and prevention strategies. We take families as our statistical units of interest, and we assume that these units belong to different risk groups of developing the disease. We investigate models for the age at breast cancer onset, with a structure of familiar dependence among survival times through the risk. In Chapter 1, we compare cure rate models with a continuous Gamma frailty to the conventional Cox model in terms of risk prediction accuracy. In Chapter 2, we focus exclusively on cure rate models using a binary frailty to model the age at onset. In Chapter 3 we move to the study of the heritability of longevity. Similarly to the previous chapters, we assume that families have different “risk” in terms of life duration. We carry out our analysis through simulation studies and apply the continuous frailty model from Chapter 1 to the available data sourced from the Multi-Generational Breast Cancer Swedish registry. We conclude with an exploration of most powerful tests for survival data with right censoring in Chapter 4.
Risk prediction and family history effect in survival models for disease onset, with a specific focus on breast cancer.
VINATTIERI, MARIA VERONICA
2024
Abstract
We are interested in a specific aspect related to modelling disease onset, i.e. the study of family-specific risk. Generally, quantifying the family-specific risk of developing a disease is crucial to improve the patient’s survival, and healthcare data provide an essential tool to estimate such quantity. Specifically for breast cancer, early detection is particularly important to increase the chances of a successful treatment. In this setting, risk estimation allows for the identification of subjects at higher risk of developing breast cancer, so that they can be the target of tailored and more intensive screening and prevention strategies. We take families as our statistical units of interest, and we assume that these units belong to different risk groups of developing the disease. We investigate models for the age at breast cancer onset, with a structure of familiar dependence among survival times through the risk. In Chapter 1, we compare cure rate models with a continuous Gamma frailty to the conventional Cox model in terms of risk prediction accuracy. In Chapter 2, we focus exclusively on cure rate models using a binary frailty to model the age at onset. In Chapter 3 we move to the study of the heritability of longevity. Similarly to the previous chapters, we assume that families have different “risk” in terms of life duration. We carry out our analysis through simulation studies and apply the continuous frailty model from Chapter 1 to the available data sourced from the Multi-Generational Breast Cancer Swedish registry. We conclude with an exploration of most powerful tests for survival data with right censoring in Chapter 4.File | Dimensione | Formato | |
---|---|---|---|
Reviewed-Phd-thesis-after-minor-changes.pdf
accesso aperto
Descrizione: Reviewed Phd Thesis
Tipologia:
Tesi di dottorato
Dimensione
6.74 MB
Formato
Adobe PDF
|
6.74 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.