Optimal timing of delivery for women with breast cancer, according to cancer stage and hormone status: a decision-analytic model

Objective: To compare strategies for the timing of delivery in women with breast cancer and known cancer stage or hormone receptor subtype, and to determine the optimal gestational age for induction in regards to maternal-fetal outcomes. Study design: A decision-analytic model was designed comparing eight different strategies for scheduled delivery at 30, 31, 32, 33, 34, 35, 36, and 37 weeks gestation. Optimal breast cancer treatment was assumed to be delayed until after delivery. Baseline estimates of the stage- and subtype-specific mortality and the impact of delayed cancer treatment on 5-year survival rates were obtained from the literature. Outcomes factored into the model included the risk of intrauterine fetal demise, spontaneous delivery, respiratory distress syndrome, cerebral palsy, and neonatal demise at each gestational age. Univariate sensitivity analyses and Monte Carlo simulations were performed to test the robustness of our model. Results: For women with stage I–II breast cancer, delivery at 36 weeks yielded the highest number of overall quality-adjusted life years (QALYs), while maternal QALYs were maximized with delivery at 34 weeks. For stage III and IV disease, maternal QALYs were maximized at 31 and 30 weeks, respectively. For women with estrogen or progesterone receptor-positive, human epidermal receptor-2 negative breast cancer, both maternal QALYs and overall QALYs were maximized with delivery at 36 weeks. More aggressive biological phenotypes were similarly associated with optimal delivery at decreasing gestational age. Our model was heavily driven by the baseline probability of maternal death within 5 years, in addition to the expected progression of disease and decreases in survival rates with each week of non-treatment, and remained robust across reasonable ranges for all variables of interest. Conclusions: For women with breast cancer diagnosed during pregnancy, decisions regarding timing of delivery should take into consideration both cancer stage and hormone receptor subtype.