Multi-layer self-calibrated algorithm for transabdominal fetal pulse oximetry: simulation and in vivo validation.

Area of Science:

• Biomedical Engineering
• Medical Physics
• Obstetrics

Background:

• Non-invasive fetal pulse oximetry aims to monitor fetal arterial oxygen saturation (SaO2).
• Accurate estimation is challenging due to maternal-fetal tissue layers and low fetal oxygen levels.
• Existing methods struggle to differentiate maternal and fetal signals effectively.

Purpose of the Study:

• To develop and validate a multi-layer self-calibrated algorithm for non-invasive fetal SpO2 estimation.
• To improve accuracy by distinguishing maternal and fetal tissue contributions.
• To provide a foundation for clinical translation of fetal pulse oximetry.

Main Methods:

• Developed a multi-layer self-calibrated algorithm combining the modified Beer-Lambert law and a photon partial pathlength model.
• Validated using Monte Carlo photon simulations with varying fetal depths and optical properties.
• Further validated with in vivo sheep data comparing transabdominal near-infrared spectroscopy with CO-oximetry.

Main Results:

• Simulations showed a mean absolute error (MAE) below 5% and R=0.98 with optimal parameters.
• In vivo sheep experiments yielded MAE=10.3% and R=0.91.
• Algorithm performance demonstrated sensitivity to accurate optical properties and tissue thickness inputs.

Conclusions:

• The multi-layer self-calibrated algorithm shows proof-of-concept feasibility for non-invasive fetal SpO2 monitoring.
• Further refinement is needed for accurate optical property and fetal depth estimation in clinical settings.
• This work establishes a framework for advancing fetal pulse oximetry.