Development of a prediction model for the association between thyroid dysfunction and breast cancer

Area of Science:

• Endocrinology
• Oncology
• Biostatistics

Background:

• Thyroid dysfunction is linked to breast tumor risk, yet predictive models are scarce.
• Existing research highlights associations but lacks robust predictive tools for benign and malignant breast tumors.
• This study addresses the need for a predictive model for thyroid dysfunction's role in breast cancer.

Purpose of the Study:

• To develop and validate a predictive model for breast tumor malignancy.
• To identify key factors, including thyroid function, associated with breast tumor risk.
• To assess the model's performance in distinguishing benign from malignant breast tumors.

Main Methods:

• Retrospective analysis of breast tumor patient data (2015-2019).
• Utilized Python for data processing, feature selection, and model construction.
• Employed logistic regression and LASSO analysis for model development and evaluation.

Main Results:

• Significant differences in thyroid hormones (TSH, T4, FT4, FT3, T3) between benign and malignant tumors.
• Key predictive factors identified: age, ethnicity, thyroid function, and hormone levels (FSH, LH, prolactin, testosterone).
• The model achieved 83.70% accuracy, 90.69% precision, 84.74% sensitivity, 81.50% specificity, and an AUC of 0.9012.

Conclusions:

• A predictive model incorporating age, ethnicity, thyroid function, and estrogen levels demonstrates strong performance.
• The model effectively predicts the occurrence and development of benign and malignant breast tumors.
• This tool can aid in early detection and risk stratification for breast cancer.