Non-invasive Fetal Blood pH Estimation for Enhanced Intrapartum Monitoring using Near-Infrared Spectroscopy

Fetal monitoring remains a significant challenge within the medical field due to the lack of clinical tools for assessing fetal health. Blood pH is a critical indicator of the acid-base balance and may provide crucial insight into the fetal well-being. Current monitoring techniques for the fetal environment are limited to measuring heart rate or spatial information without supplying information into blood and tissue health. This study proposes a strategy for continuous fetal blood pH monitoring by using near-infrared (NIR) spectroscopy and extracting features in the spectral fluctuations. These methods are applied to data collected from a hypoxic lamb model and viability assessed with a convolutional neural network (CNN). Results showed that the machine learning (ML) model was capable of extracting features relating to changes in the fetal blood pH with a mean estimation error below 0.05 pH units.Clinical relevance-Application of the proposed methods would enable clinicians to further evaluate fetal health by measuring the blood acidity. Currently, there are no methods for measuring fetal blood pH, and the acid-base balance may be affected by present or previous health risks that are not observable with current medical tools. Not only does this study address diagnosing fetal acidosis, but also allows for continuous fetal blood pH measurements.