Predicting Timely Diagnosis of Metastatic Breast Cancer: A Stack Ensemble Classification Approach

Healthcare inequity is a global issue that affects women's health significantly. The WiDS Datathon 2024 Challenge #1 focused on predicting a timely diagnosis of metastatic Triple Negative Breast Cancer (TNBC), a critical indicator of healthcare access disparities. TNBC is one of the most progressive cancer types for which timely treatment is of utmost importance for the survival of the patient. This paper presents a stack ensemble classification approach developed by the team “The Synergistic Solvers”, which secured <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$7^{th}$</tex> position in the competition and won the Top First-Timer Team award. The solution takes advantage of a combination of data pre-processing techniques, feature engineering, and ensemble learning. The key findings highlight the importance of patient age, diagnosis of malignant neoplasms, and careful handling of missing values. The developed model contributes to understanding the factors influencing timely cancer diagnosis and promoting equitable healthcare access.