Description

Gestational diabetes is a growing health concern posing the neonate in the risk of adverse outcomes, such as macrosomia, cesarean delivery, hypoglycemia, respiratory morbidity, and need for neonatal intensive care. Most of these outcomes are either closely or causally related to fetal hyperinsulinemia, which is induced by maternal hyperglycemia. However, the accurate and non-invasive means of estimating fetal hyperinsulinemia are currently lacking.

Given that maternal diabetes and fetal hyperinsulinemia are associated with profound changes in fetal liver blood supply, biometry, metabolism, and lipid content, it was hypothesized that fetal hyperinsulinemia would be detectable by ultrasound using modern computer-aided technologies, i.e., radiomics and machine learning. In this prospective pilot study, the aim is to recruit 20 pregnant women with type I diabetes and compare them to 40 healthy controls. Patients with type 1 diabetes were chosen because their pregnancies are known to be associated with fetal hyperinsulinemia. Study subjects will be recruited at Turku University Hospital and the City of Turku maternity welfare clinics. It is assumed that radiomics analysis of fetal liver ultrasound pictures will indicate fetal hyperinsulinemia in pregnancies complicated by type 1 diabetes at 34 gestational weeks. Simultaneously, the inter- and intraobserver variability for fetal liver ultrasound radiomic features will be characterized.

In the future, non-invasive evaluation of fetal hyperinsulinemia could provide a useful tool in managing pregnancies complicated by maternal diabetes - either pre-gestational or gestational.