Baseline fetal heart rate analysis: eleven automatic methods versus expert consensus.

Area of Science:

• Perinatal medicine
• Biomedical signal processing
• Obstetrics

Background:

• Visual analysis of fetal heart rate (FHR) during labor is prone to significant inter- and intra-observer variability.
• This variability is especially problematic for anomalous FHR recordings, impacting clinical decision-making.
• Automatic FHR analysis offers a potential solution to reduce observer-dependent variability.

Purpose of the Study:

• To evaluate and compare the performance of eleven different automatic FHR analysis methods.
• To assess the accuracy of automatically determined FHR baselines against expert consensus.
• To identify which automatic methods demonstrate superior performance in baseline determination.

Main Methods:

• Eleven distinct automatic FHR analysis algorithms were implemented based on published literature.
• These methods were applied to a dataset of 66 FHR recordings from the first stage of labor.
• The automatically derived FHR baselines were compared to a gold standard baseline established by a panel of three expert clinicians.

Main Results:

• Several automatic methods, including those by Mongelli, Lu, Wrobel, and Pardey, exhibited better performance in baseline determination.
• Despite improvements, none of the evaluated automatic methods consistently produced baselines that matched the expert-defined reference across all recordings.
• Differences in signal processing approaches did not uniquely predict superior performance among the tested methods.

Conclusions:

• Automatic FHR analysis holds promise for reducing variability in labor monitoring.
• Accurate determination of the FHR baseline remains a critical challenge for current automatic analysis techniques.
• Further research and refinement of algorithms are necessary to achieve reliable and clinically applicable automatic FHR baseline detection.