Response Prediction to Neoadjuvant Systemic Treatment in Breast Cancer-Yet Another Algorithm?

摘要

Meti et al1 developed machine learning algorithms for the upfront response prediction to neoadjuvant systemic treatment (NST) in breast cancer and compared the predictive performance of these modern models with a standard statistical model. We congratulate the authors for their innovative article especially with respect to the underlying vision to improve patient care through individualized risk predictions by modern, computational predictive modeling and to educate clinical scientists about the differences between machine learning and standard statistical models (eg, statistically significant variables versus predictive variables for complex, nonlinear models). To make this vision come true, building up trust in these new innovative models will be key. Over the past few years, our group has focused on identifying exceptional re-sponders to NST23 using traditional statistical methods4 and machine learning algorithms5 and as such, we would like to make some considerations for future research in this area.