Computer-assisted Fetal Monitoring - Obstetrics

Overview

In the context of fetal heart monitoring, the SurFAO project offers an alternative to current clinical routines. The challenge is to extract, from few non-invasive sensors on the maternal abdomen, a fetal electrocardiogram (ECGf) of great quality allowing a clinical diagnosis (follow-up of the FHR (fetal heart rate)) and extraction of ECG waveforms). The proposed approach proposes a technological breakthrough shared by a consortium of researchers and clinicians. The originality is driven by innovative methodological choices: the use of a multimodal system (ECG coupling with PCG (phonocardiographic)) for the signal acquisition in order to increase the robustness of information extraction, by taking into account clinical uses and the need for the monitoring process support, and by setting up a reference for this multimodal database. The objective is to feed a database that will be used in the future to develop ECGf extraction methods estimating the FHR.

To monitor the well-being of a fetus or for a clinical diagnosis, a challenge is to extract a fetal electrocardiogram (fECG) signal of high quality from a limited number of non-invasive sensors on the maternal abdomen. During childbirth, fetal monitoring is assessed by the fetal heart rate (FHR), classically from cardiotocography (CTG). The aim is to monitor FHR variability anomalies that reflect a too high fetal malaise. It is very common to have difficulties to reliably record FHR due to confusion of rhythms of the mother and the fetus that can lead to unnecessary caesareans. With 800000 childbirths in France per year, improving the reliability of FHR estimation by fECG analysis is therefore of high clinicalinterest. To deal with, the proposed approach aims at coupling two complementary cardiac information acquired at the best-located position. Therefore, the proposed solution is (i) to combine the use of electrophysiological sensors ECG and acoustic sensors of microphonic type that can register phonocardiographic signals (PCG), (ii) to assist by computer the clinical expertise to choose the best sensors location. Several methodological aspects will be considered: to improve the robustness of the FHR estimation with degraded signals based on multimodality, to overcome some practical limitations to extract the fECG waveforms by enhancing the process extraction with PCG and to assist the user in the sensors choice and placement by predicting their best locations and their kinds (ECG or PCG). The SurFAO project proposes then an original approach of computer-assisted fetal monitoring and thus differs from previous published works of fECG extraction solutions, only based on algorithmic performances mostly in well-controlled situations. The original aspects of the investigator's proposal are the innovative methodological choices based on multimodality and informed recording process, the consideration of the medical uses and of the necessity to help the clinical monitoring to design the algorithms, the new reference multimodal database, and the academic-clinical consortium, that will guarantee the relevance of all proposed parts.