Machine learning from fetal flow waveforms to predict adverse perinatal outcomes: a study protocol.

Zahra Hoodbhoy,Babar Hasan,Bart Bijnens,Fyezah Jehan,Devyani Chowdhury

doi:10.12688/gatesopenres.12796.1

Zahra Hoodbhoy, Babar Hasan + Show 3 more

Open Access

https://doi.org/10.12688/gatesopenres.12796.1

Copy DOI

Abstract

Background: In Pakistan, stillbirth rates and early neonatal mortality rates are amongst the highest in the world. The aim of this study is to provide proof of concept for using a computational model of fetal haemodynamics, combined with machine learning. This model will be based on Doppler patterns of the fetal cardiovascular, cerebral and placental flows with the goal to identify those fetuses at increased risk of adverse perinatal outcomes such as stillbirth, perinatal mortality and other neonatal morbidities. Methods: This will be prospective one group cohort study which will be conducted in Ibrahim Hyderi, a peri-urban settlement in south east of Karachi. The eligibility criteria include pregnant women between 22-34 weeks who reside in the study area. Once enrolled, in addition to the performing fetal ultrasound to obtain Dopplers, data on socio-demographic, maternal anthropometry, haemoglobin and cardiotocography will be obtained on the pregnant women. Discussion: The machine learning approach for predicting adverse perinatal outcomes obtained from the current study will be validated in a larger population at the next stage. The data will allow for early interventions to improve perinatal outcomes.

Highlights

A pregnant woman and her newborn are most vulnerable around the time of delivery and the postnatal period[1]
Stillbirths and early neonatal deaths within 7 days of life are a major cause of perinatal mortality
Though UmbiFlow (jointly developed by Medical Research Council (MRC) Unit for Perinatal Mortality, the MRC and the Centre for Scientific and Industrial Research (CSIR) South Africa), a portable device used to detect IUGR by assessing continuous wave Doppler blood flow across the placental end of the umbilical artery reported reduced number of unnecessary referrals to a higher level health facility in cases suspected of IUGR8, it studies waveform across one arterial bed while our model takes into account 6 arterial beds which play a critical role in the fetal circulation

Summary

Introduction

A pregnant woman and her newborn are most vulnerable around the time of delivery and the postnatal period[1]. Though UmbiFlow (jointly developed by Medical Research Council (MRC) Unit for Perinatal Mortality, the MRC and the Centre for Scientific and Industrial Research (CSIR) South Africa), a portable device used to detect IUGR by assessing continuous wave Doppler blood flow across the placental end of the umbilical artery reported reduced number of unnecessary referrals to a higher level health facility in cases suspected of IUGR8, it studies waveform across one arterial bed while our model takes into account 6 arterial beds which play a critical role in the fetal circulation. The aim of this study is to obtain proof of concept for using a computational model of fetal hemodynamics, combined with machine learning based on Doppler patterns of the fetal cardiovascular, cerebral and placental flows, to identify those at increased risk of adverse perinatal outcomes such as stillbirth, perinatal mortality, and other neonatal morbidities. The results of this study will be shared through newspaper articles in the local media, continuous medical education seminars, conference proceedings and publications in well reputed journals

Discussion

NIPS I

Findings

Young Infants Clinical Signs Study Group