To investigate how the choice of numerical optimization method affects the training of logistic regression models for binary cancer classification.
Approach:
Web Application Development: A Flask-React web application is developed to facilitate cancer detection, allowing clinicians to visualize predictions based on patient features.
Key Findings:
Trade-offs exist between runtime, number of iterations, and predictive quality among the optimization methods.
The choice of optimization technique affects convergence speed and computational efficiency.
Interpretation:
The findings highlight the importance of selecting appropriate optimization methods for enhancing logistic regression models in cancer detection applications.
Limitations:
The study is limited to the analysis of five specific optimization algorithms.
Results may vary with different datasets or cancer types not covered in the analysis.
Conclusion:
The integration of logistic regression with suitable optimization strategies can improve cancer screening tools in digital health systems.
Harold Burstein, MD, PhD; Sara Tolaney, MD, MPH; Erica L. Mayer, MD, MPH; Nancy Lin, MD; and Ana C. Garrido-Castro, MD discuss important breast cancer research presented at the 2026 ESMO Breast Cancer Congress.
