A predictive model for chemotherapy-related diminished ovarian reserve in reproductive-age women.

Objective To develop and internally validate a clinical predictive tool to assess the likelihood that a young cancer patient will experience diminished ovarian reserve (DOR) after chemotherapy. Design Prospective cohort study. Setting University hospitals. Patient(s) Postpubertal adolescent and young adult women with a new diagnosis of cancer requiring chemotherapy. Intervention None. Main Outcome Measure(s) Diminished ovarian reserve after completion of and recovery from chemotherapy, defined as serum antimullerian hormone (AMH) Result(s) A multivariable logistic regression model which includes age, cancer type, exposure to an alkylating agent, and baseline AMH value accurately predicts the diagnosis of DOR after chemotherapy with an area under the receiver operating characteristic curve of 0.89. Conclusion(s) Pretreatment information on age, cancer type, use of an alkylating agent, and baseline AMH levels make up a clinically useful predictive tool to identify which women are most at risk for DOR caused by chemotherapy.