(182r) Customized Robust Optimal Dosage Determination in the Face of Uncertainties for IVF Practices

In Vitro Fertilization (IVF) treatment, one of the most advanced and highly recommended procedure for treating infertility in women; remains experimental even with recent advancements in technology. The increasing levels of infertility in women per year, uncertain outcome, and the high cost of this procedure; necessitate the need to develop a patient-specific treatment protocol. IVF comprises of four stages; namely, Superovulation, Egg-retrieval, Insemination/Fertilization and Embryo transfer. Superovulation stage involves stimulation and growth of multiple follicles or oocytes through externally injected hormones which makes it a highly controllable process.

In this work, we used the data from 50 IVF patients and developed customized models for drug dosage prediction. Based on this, the analysis is conducted to find patient dependent and independent parameters in the IVF model for customized dosage profiles in the given data. In real time scenario, data for two days is required to model which poses character uncertainties.

We derived robust optimal drug dosage profiles in the face of uncertainties using optimal control. Further, our analysis predicts starting optimal dosage profiles for diverse age distributions.

The resulting customized dosage profiles of the externally injected hormones differ significantly in comparison to the prescribed dosage and show a better final day follicle size distribution, lower overall hormonal dosages, and less testing. The dosage profile is robust in the presence of uncertainties.