A power series beta Weibull regression model for predicting breast carcinoma

Area of Science:

• Biostatistics
• Oncology
• Medical Statistics

Background:

• Current postmastectomy survival rates lack individual patient accuracy.
• Pathologic factors like tumor size and lymph node status significantly impact prognosis.
• Existing models may not fully capture the complexities of breast carcinoma survival.

Purpose of the Study:

• To develop a new cure rate survival regression model for predicting breast carcinoma survival in women post-mastectomy.
• To improve the accuracy of individual patient survival predictions beyond traditional statistical methods.
• To incorporate competing risks and specific distributional assumptions for enhanced predictive power.

Main Methods:

• Proposed a novel compounding regression model for cure rate survival analysis.
• Assumed power series distribution for competing causes and beta Weibull for metastasis time.
• Employed maximum likelihood estimation for model parameter estimation.
• Conducted simulations to evaluate model performance under various conditions.
• Derived matrices for assessing local influences on parameter estimates.

Main Results:

• The proposed model offers a flexible framework, encompassing several existing cure rate models.
• Simulations demonstrated the model's robustness across different parameter settings, sample sizes, and censoring levels.
• Local influence assessment methods were developed to understand model sensitivity.
• The model's potential for accurate breast carcinoma mortality prediction was validated using real patient data.

Conclusions:

• The new regression model provides a more accurate approach to predicting breast carcinoma survival post-mastectomy.
• The model's flexibility and incorporation of key prognostic factors enhance its clinical applicability.
• This statistical advancement holds promise for personalized patient management and treatment strategies.