ASPO Abstracts
Simulation Modeling of Breast Cancer Clinical Trials: Long-term Harms and Benefits of Chemotherapy among Women with Low-risk Hormone-Sensitive Early-stage Breast Cancer
Category: Early Detection & Risk Prediction, Survivorship & Health Outcomes/Comparative Effectiveness Research
Conference Year: 2018
Abstract Body:
Purpose: Simulation modeling can play a key role in extending clinical trial data in low-risk adjuvant breast cancer settings. Oncotype DX® test is useful in the identification of women who have a low risk of recurrence and can safely avoid chemotherapy. However, there are persistent gaps in knowledge on the use of Oncotype testing in treatment decisions to minimize harms and maximize benefits of chemotherapy among women with favorable prognosis breast cancer. We simulated long-term chemotherapy effects on recurrence and mortality among early-stage breast cancer women with Oncotype scores 0-10, using an established Cancer Intervention and Surveillance Modeling Network (CISNET) breast cancer model. Methods: A Monte Carlo microsimulation was conducted to project long-term (up to 15+ years) outcomes for any first event, recurrence, distant recurrence, and all-cause mortality among women diagnosed with HER2 negative, node—negative breast cancer tumors of 1.1-5.0 cm with Oncotype scores of 0-10. The input parameters were derived from the Oxford Overview and published NSABP trial data. We assessed model-fit by comparing Kaplan-Meier survival estimates for each trial endpoint up to 5-years with Trial Assigning IndividuaLized Options for Treatment (TAILORx) published results. Results: Women with favorable gene-expression profiles had high recurrence-free interval rates for any recurrence (92%) and distant recurrence (97%) at 15-years of follow-up with omission of adjuvant chemotherapy. However, rates of invasive disease-free survival (74%) and overall survival (88%) at 15-years were comparatively low. The results for the preliminary model showed excellent concordance with the initial published results from TAILORx for all trial outcomes in the first five years for patients with Oncotype scores 0-10. Discussion: Although more recurrences are expected with longer follow-up, our model projections show that long-term recurrence rates remain low among women with Oncotype scores of 0-10. Modeling and simulation are useful in extending clinical trial data by synthesizing information from historical trials.
Keywords: early-stage breast cancer, recurrence, Oncotype testing