Diagnostic Accuracy of a Novel Machine Learning Algorithm to Estimate Gestational Age

Z 32104 - Diagnostic Accuracy of a Novel Machine Learning Algorithm to Estimate Gestational Age

This is a prospective cohort study of women enrolled early in pregnancy, with randomization to determine the timing of three follow-up visits in the second and third trimester. At each of these follow-up visits, investigators will assess gestational age with the FAMLI technology and compare that estimate to the known gestational age established early in pregnancy.

The primary purpose of this research is to assess the diagnostic accuracy of the FAMLI Technology, a novel machine learning-based tool for gestational age assessment that can run on a smart phone or tablet. Study staff will enroll 400 pregnant volunteers prior to 14 completed gestational weeks and obtain accurate "ground truth" gestational age dating with standard ultrasound biometry, using the crown-rump length. These participants will then be asked to return for three follow-up visits, which will include a routine sonogram performed by a trained sonographer and the collection of a set of blind sweep cineloop videos using a low-cost, battery-operated device. The research will be conducted in Chapel Hill, North Carolina (at the University of North Carolina Hospital and/or sites associated with UNC OBGYN) and in Lusaka, Zambia (at the University Teaching Hospital or Kamwala District Health Centre). Approximately equal numbers of participants will be enrolled from each country.

Inclusion Criteria:

• 18 years of age or older
• viable intrauterine pregnancy at less than 14 0/7 weeks of gestation
• ability and willingness to provide written informed consent
• intent to remain in current geographical area of residence for the duration of study
• willingness to adhere to study procedures

Exclusion criteria:

• maternal body mass index = 40 kg/m^2
• multiple gestation (i.e., twins or higher order)
• major fetal malformation or anomaly
• any other condition (social or medical) that, in the opinion of the study staff, would make study participation unsafe or complicate data interpretation.