Computerised intrapartum diagnosis of fetal hypoxia based on fetal heart rate monitoring and fetal pulse oximetry recordings utilising wavelet analysis and neural networks

Abstract

Objective To develop a computerised system that will assist the early diagnosis of fetal hypoxia and to investigate the relationship between the fetal heart rate variability and the fetal pulse oximetry recordings. Design Retrospective off-line analysis of cardiotocogram and FSpO 2 recordings.Setting The Maternity Unit of the 2nd Department of Obstetrics and Gynaecology, Aretaieion Hospital, University of Athens. Population Sixty-one women of more than 37 weeks of gestation were monitored throughout labour.Methods Multiresolution wavelet analysis was applied in each 10-minute period of second stage of labour focussing on long term variability changes in different frequency ranges and statistical analysis was performed in the associated 10-minute FSpO 2 recordings. Self-organising map neural network was used to categorise the different 10-minute fetal heart rate patterns and the associated 10-minute FSpO 2 recordings. Main outcome measures Umbilical artery pH of 7.20 and Apgar score at 5 minutes of 7 formed the inclusion criteria of the risk group. Results After using k-means clustering algorithm, the two-dimensional output layer of the self-organising map neural network was divided into three distinct clusters. All the cases that mapped in cluster 3 belonged in the risk group except one. The sensitivity of the system was 83.3% and the specificity 97.9% for the detection of risk group cases. Conclusions A relationship between the fetal heart rate variability in different frequency ranges and the time in which FSpO 2 is less than 30% was noticed. Fetal pulse oximetry seems to be an important additional source of information. Computerised analysis of the fetal heart rate monitoring and pulse oximetry recordings is a promising technique in objective intrapartum diagnosis of fetal hypoxia. Further evaluation of this technique is mandatory to evaluate its efficacy and reliability in interpreting fetal heart rate recordings.

Classification of Fetal Heart Rate Using Scale Dependent Features and Support Vector Machines

“…They are comparable to those reported by other similar approaches (Salamalekis et al, 2002) where they achieved an overall classification rate of 95% using a cut-off value of 7.2 for the umbilical artery pH but this classification was slightly unbalanced since the classification rate for fetuses with pH<7.2 was 83.3% while the classification rate for fetuses with pH>7.2 was 97.9%. However, the fact that we did not include cases with umbilical artery pH in the range (7.1, 7.2) makes it a bit risky to directly compare our approach with the others found in the literature.…”

“…In our experiments no single feature (or even combination of features) was found capable to completely discriminate the two classes. Moreover, in the work of Salamalekis (2002) only scales 2-4 were used. However, in our experiments we achieved the best classification performance using all six features.…”

“…Wavelets have also been used for the analysis of FHR during the second stage of labour (Salamalekis et al, 2002). The most appealing characteristic of wavelets is that they can decompose a signal into a number of scales, each scale representing a particular "coarseness" of the signal under study (Mallat, 1998).…”

Scattering Transform for Intrapartum Fetal Heart Rate Variability Fractal Analysis: A Case-Control Study

“…First, because of the timeevolving nature of the delivery process (baseline variations, occurrence of decelerations), intrapartum F-HRV signals are much less stationary than adult HRV signals are. This lead to the use of joint time-frequency representations to account for the time varying nature of the frequency content of intrapartum F-HRV [11], or, along another line, to adaptive (or data-driven) characterization (cf. e.g., [12]).…”

Foetal heart rate variability frequency characteristics with respect to uterine contractions

“…Spectral analysis provides a tool for quantifying rather small changes in FHRV in response to internal or external stimuli that may remain undetected if only visual interpretation of FHR tracings is used. Among most common methods employed to estimate PSD, parametric and non-parametric, we can mention Short-time Fourier transform (STFT); Auto Regressive methods (AR); Fast Recursive least square algorithms (RLS) [2,12,14,19], wavelet transform [20] and Lomb method [21].…”