Intrapartum cardiotocography trace pattern pre-processing, features extraction and fetal health condition diagnoses...

Area of Science:

• Medical Informatics
• Obstetrics and Gynecology
• Signal Processing

Background:

• Cardiotocograph (CTG) analysis is vital for monitoring fetal well-being during pregnancy and childbirth.
• Computer-assisted fetal heart rate (FHR) feature extraction and CTG classification aid obstetricians in accurate interpretation.
• Current methods focus on enhancing diagnostic precision for fetal health tracking.

Purpose of the Study:

• To develop a reliable FHR feature extraction algorithm and a systematic CTG classifier using MATLAB.
• To ensure the developed system adheres to Royal College of Obstetricians and Gynecologists (RCOG) recommendations.
• To provide a robust computational tool for obstetricians in interpreting fetal monitoring data.

Main Methods:

• Data preprocessing involved artifact cleaning and missing data compensation using RCOG guidelines and MATLAB.
• FHR fundamental features (baseline, acceleration, deceleration, variability) were extracted.
• CTG signals were classified into normal, abnormal (suspicious), or pathological categories based on RCOG guidelines.
• The computerized system's performance was validated by comparing its output with visual interpretations from five obstetricians on 150 CTG signals.

Main Results:

• The computerized CTG signal categorization showed minor variability compared to expert opinions, with differences up to +/-4-6 beats per minute.
• Obstetrician agreement was high for deceleration type and number, with slight deviations in acceleration values (+/-4).
• The study demonstrated the effectiveness and robustness of the developed computerized procedure.

Conclusions:

• The developed computerized approach for CTG interpretation is suitable for clinical application in infirmaries and home care settings.
• The system's classification, based on FHR attributes and RCOG guidelines, was validated by expert obstetricians.
• The study confirms the potential of computer-assisted methods to support fetal health monitoring.