[Correspondence] Primary ovarian insufficiency prediction in adult survivors of childhood cancer: model concerns

We read the Article by Cindy Im and colleagues1 with great interest. They developed four statistical regression and advanced machine learning models to predict primary ovarian insufficiency among adult survivors of childhood cancer. After external validation, the prediction performance of the age-specific logistic regression model and the XGBoost algorithm did not differ appreciably,1 bringing in to question the clinical use of machine learning algorithms.