An intelligent medical decision: antenatal fetal risk assessment by neuro-fuzzy technique using Doppler blood flow velocity waveforms from umbilical and cerebral artery

Abstract

The study defines an intelligent neuro-fuzzy system for antepartum fetal evaluation. The task is to investigate the Doppler ultrasound measurements of umbilical artery (UA) and cerebral artery to relate the health conditions of fetuses. We use the UA blood flow velocity waveforms including the pulsality index, resistance index and systolic/diastolic ratio and the ratios of cerebral-umbilical resistance indices in terms of weeks. We then make a decision on the basis of fuzzy-rule based system combined with data-based learning strategies such as the radial basis function network and multilayer perceptron for assessing the hypoxia suspicion. A fuzzy grade of membership is used for the evaluation of the seriousness of the situation of the fetus and the diagnostic interpretations. The results show that intelligent data analysis methods are effective supportive medical tools for physicians during intensive surveillance of fetuses.