A Logistic Regression Approach for Breast Cancer Prediction in Sri Lanka

Abstract

Breast cancer is the most prevalent type of cancer among Sri Lankan women, with relatively low survival rates compared to developed countries. Over 3,000 new cases are diagnosed annually, according to the Sri Lanka National Cancer Registry. This study aims to model and predict breast cancer occurrence using Logistic Regression (LR). Data were obtained from the National Cancer Institute, Maharagama, and the Teaching Hospital, Ratnapura, with 250 patient records analyzed, each containing 15 features, including demographic, laboratory, and mammography attributes. The study identified key risk factors: married women had slightly higher odds of developing breast cancer (OR = 1.076), while a family history of breast cancer slightly reduced the risk (OR = 0.94848). Women with positive Estrogen Receptor (OR = 1.4437) and Progesterone Receptor (OR = 1.8456) statuses had a higher risk. Increasing age was associated with a small increase in risk (OR = 1.03418), and those with a history of breast cancer had a higher likelihood of recurrence (OR = 1.4398). Women who had given birth had a slightly lower risk compared to those who had not (OR = 0.946). Breast cancer was less likely to occur in the right breast (OR = 0.839), and positive HER-2 status was linked to decreased odds (OR = 0.645). The findings can help healthcare policymakers in identifying high-risk individuals, improving screening programs, and promoting awareness campaigns. Recommendations include enhancing screening, integrating hormone receptor testing, targeted awareness programs, and further research using larger datasets and advanced models. Additionally, lifestyle modifications such as regular physical activity, maintaining a healthy diet, and reducing exposure to known risk factors should be emphasized in public health initiatives. By implementing these measures, Sri Lanka’s healthcare system can enhance early detection and prevention strategies, ultimately improving breast cancer survival rates and reducing the overall burden of the disease.