Use of Deep Learning to Detect the Maternal Heart Rate and False Signals on Fetal Heart Rate Recordings.

We have developed deep learning models for automatic identification of the maternal heart rate (MHR) and, more generally, false signals (FSs) on fetal heart rate (FHR) recordings. The models can be used to preprocess FHR data prior to automated analysis or as a clinical alert system to assist the practitioner. Three models were developed and used to detect (i) FSs on the MHR channel (the FSMHR model), (ii) the MHR and FSs on the Doppler FHR sensor (the FSDop model), and (iii) FSs on the scalp ECG channel (the FSScalp model). The FSDop model was the most useful because FSs are far more frequent on the Doppler FHR channel. All three models were based on a multilayer, symmetric, GRU, and were trained on data recorded during the first and second stages of delivery. The FSMHR and FSDop models were also trained on antepartum recordings. The training dataset contained 1030 expert-annotated periods (mean duration: 36 min) from 635 recordings. In an initial evaluation of routine clinical practice, 30 fully annotated recordings for each sensor type (mean duration: 5 h for MHR and Doppler sensors, and 3 h for the scalp ECG sensor) were analyzed. The sensitivity, positive predictive value (PPV) and accuracy were respectively 62.20%, 87.1% and 99.90% for the FSMHR model, 93.1%, 95.6% and 99.68% for the FSDop model, and 44.6%, 87.2% and 99.93% for the FSScalp model. We built a second test dataset with a more solid ground truth by selecting 45 periods (lasting 20 min, on average) on which the Doppler FHR and scalp ECG signals were recorded simultaneously. Using scalp ECG data, the experts estimated the true FHR value more reliably and thus annotated the Doppler FHR channel more precisely. The models achieved a sensitivity of 53.3%, a PPV of 62.4%, and an accuracy of 97.29%. In comparison, two experts (blinded to the scalp ECG data) respectively achieved a sensitivity of 15.7%, a PPV of 74.3%, and an accuracy of 96.91% and a sensitivity of 60.7%, a PPV of 83.5% and an accuracy of 98.24%. Hence, the models performed at expert level (better than one expert and worse than the other), although a well-trained expert with good knowledge of FSs could probably do better in some cases. The models and datasets have been included in the Fetal Heart Rate Morphological Analysis open-source MATLAB toolbox and can be used freely for research purposes.

Fetal heart rate baseline computation with a weighted median filter.

BACKGROUND: Automated fetal heart rate (FHR) analysis removes inter- and intra-expert variability, and is a promising solution for reducing the occurrence of fetal acidosis and the implementation of unnecessary medical procedures. The first steps in automated FHR analysis are determination of the baseline, and detection of accelerations and decelerations (A/D). We describe a new method in which a weighted median filter baseline (WMFB) is computed and A/Ds are then detected. METHOD: The filter weightings are based on the prior probability that the sampled FHR is in the baseline state or in an A/D state. This probability is computed by estimating the signal's stability at low frequencies and by progressively trimming the signal. Using a competition dataset of 90 previously annotated FHR recordings, we evaluated the WMFB method and 11 recently published literature methods against the ground truth of an expert consensus. The level of agreement between the WMFB method and the expert consensus was estimated by calculating several indices (primarily the morphological analysis discordance index, MADI). The agreement indices were then compared with the values for eleven other methods. We also compared the level of method-expert agreement with the level of interrater agreement. RESULTS: For the WMFB method, the MADI indicated a disagreement of 4.02% vs. the consensus; this value is significantly lower (p<10-13) than that calculated for the best of the 11 literature methods (7.27%, for Lu and Wei's empirical mode decomposition method). The level of inter-expert agreement (according to the MADI) and the level of WMFB-expert agreement did not differ significantly (p=0.22). CONCLUSION: The WMFB method reproduced the expert consensus analysis better than 11 other methods. No differences in performance between the WMFB method and individual experts were observed. The method Matlab source code is available under General Public Licence at http://utsb.univ-catholille.fr/fhr-wmfb.

Development of a Smart Mobile Data Module for Fetal Monitoring in E-Healthcare.

The fetal heart rate (FHR) is a marker of fetal well-being in utero (when monitoring maternal and/or fetal pathologies) and during labor. Here, we developed a smart mobile data module for the remote acquisition and transmission (via a Wi-Fi or 4G connection) of FHR recordings, together with a web-based viewer for displaying the FHR datasets on a computer, smartphone or tablet. In order to define the features required by users, we modelled the fetal monitoring procedure (in home and hospital settings) via semi-structured interviews with midwives and obstetricians. Using this information, we developed a mobile data transfer module based on a Raspberry Pi. When connected to a standalone fetal monitor, the module acquires the FHR signal and sends it (via a Wi-Fi or a 3G/4G mobile internet connection) to a secure server within our hospital information system. The archived, digitized signal data are linked to the patient's electronic medical records. An HTML5/JavaScript web viewer converts the digitized FHR data into easily readable and interpretable graphs for viewing on a computer (running Windows, Linux or MacOS) or a mobile device (running Android, iOS or Windows Phone OS). The data can be viewed in real time or offline. The application includes tools required for correct interpretation of the data (signal loss calculation, scale adjustment, and precise measurements of the signal's characteristics). We performed a proof-of-concept case study of the transmission, reception and visualization of FHR data for a pregnant woman at 30 weeks of amenorrhea. She was hospitalized in the pregnancy assessment unit and FHR data were acquired three times a day with a Philips Avalon® FM30 fetal monitor. The prototype (Raspberry Pi) was connected to the fetal monitor's RS232 port. The emission and reception of prerecorded signals were tested and the web server correctly received the signals, and the FHR recording was visualized in real time on a computer, a tablet and smartphones (running Android and iOS) via the web viewer. This process did not perturb the hospital's computer network. There was no data delay or loss during a 60-min test. The web viewer was tested successfully in the various usage situations. The system was as user-friendly as expected, and enabled rapid, secure archiving. We have developed a system for the acquisition, transmission, recording and visualization of RCF data. Healthcare professionals can view the FHR data remotely on their computer, tablet or smartphone. Integration of FHR data into a hospital information system enables optimal, secure, long-term data archiving.

Thirty-Year Trends of Survival and Time-Varying Effects of Prognostic Factors in Patients With Metastatic Breast Cancer-A Single Institution Experience.

BACKGROUND: Metastatic breast cancer is generally considered an incurable disease. In our study we aimed to detect a time trend of survival over the past 30 years and account for time-varying effects of the prognostic factors. PATIENTS AND METHODS: A total of 446 patients diagnosed with breast cancer at Saint Vincent de Paul Hospital, Lille, France between 1977 and 2013 who developed metastatic disease after a disease-free interval longer than 3 months and were followed-up for outcome. Data were analyzed using the Cox proportional hazards model and presented as hazard ratios (HRs). RESULTS: A monotonic time trend of survival was detected: a 2.6% lower risk of death for each increasing year over the past 30 years. Three prognostic factors had time-varying effects; the liver first metastasis (HR during the first 16 months of follow-up: 2.26; 95% confidence interval [CI], 1.65-3.11), the bone first metastasis (HR during the first 24 months of follow-up: 0.56; 95% CI, 0.43-0.74), and the disease-free interval (HR during the first 16 months of follow-up: 0.90; 95% CI, 0.85-0.95). The brain first metastasis, multiple first metastases, the lymph node ratio, and estrogen receptor status had a constant effect over time. CONCLUSION: In our study we detected a constant time trend of improvement in prognosis of metastatic breast cancer patients over the past 30 years and identified prognostic factors with time-varying effects.

[Comparison of immunochromatographic tests Actim(®) Prom and Amnioquick(®) Duo for the rapid detection of premature rupture of membranes]. / Comparaison des tests immunochromatographiques Actim(®) Prom et Amnioquick(®) Duo pour le dépistage rapide de la rupture prématurée des membranes.

Premature rupture of membranes (PRM) affects 5 to 15% of pregnancies, leading to prematurity and neonatal infection. PRM can be identified by through various amniotic fluid proteins in vaginal secretions. The aim of this study is to compare two immunochromatographic tests based on the detection of insulin-like growth factor binding protein (IGFBP-1) and alpha-foeto protein (AFP) for one of the two tests in cervico-vaginal secretions. Two tests, Actim(®) Prom and Amnioquick(®) Duo were performed on 80 pregnant women with suspected PRM. Amnioquick(®) Duo allows the simultaneous detection of IGFBP-1 and AFP with an automated incubation and reading. The number of positive results is similar (Khi-deux = 0.173, p = 0.6773) for IGFBP-1 between the two tests and there is a good agreement (K = 0.621), with a proportion of negative results of 86%. The number of positive results for AFP is more important in comparison to IGFBP-1. Results positive/positive (Actim(®) Prom/Amnioquick(®)) for IGFBP-1 seems to be related to the time when tests have been performed, that is to say in the last weeks of pregnancy. In conclusion, both tests have similar performance, but there is a risk of false positive results with AFP, this can be explained by the presence of non-visible blood in samples. An automated incubation and reading allows a good reproducibility. Moreover, the computer data storage improve the post-analytical quality.

Acute hepatitis C during the third trimester of pregnancy.

A pregnant woman presented at 32 weeks of amenorrhea with jaundice secondary to acute hepatitis C. Spontaneous delivery took place 3 days later. The infant's serum tested negative for C viral RNA 6 months after delivery. Treatment with high doses of interferon-alpha for a period of 4 weeks was begun 4 days after delivery. Although a virological response was noted at the end of the treatment, the hepatitis relapsed and progressed toward chronicity. Case reports of acute hepatitis C during pregnancy are very rare, as the methods used for the follow-up of pregnant women render the diagnosis of asymptomatic forms difficult. In one case, the acute hepatitis C was severe. The occurrence of acute hepatitis C during pregnancy seems to increase the risk of premature delivery, but not that of vertical transmission. Given the frequency of side effects, it seems preferable not to begin interferon treatment until after delivery.

Prenatal diagnosis of an aorto-pulmonary window by fetal echocardiography.

Congenital aorto-pulmonary window or congenital aorto-pulmonary septal defect is a rare fetal malformation usually diagnosed after birth by echocardiography and usually associated with other congenital cardiovascular abnormalities (interrupted aortic arch, ventricular septal defect, atrial septal defect, tetralogy of Fallot). The authors report the first case of prenatal diagnosis of an aorto-pulmonary window associated with a ventricular septal defect identified by fetal ultrasonography at 28 weeks of pregnancy. The diagnosis was based on the echocardiographic images of normal semilunar aortic and pulmonic valves with evidence of a septal defect between the ascending aorta and pulmonary artery. The purpose of this report is to demonstrate the feasibility of antenatal diagnosis of this fetal malformation and help professionals who would be faced with such an unexpected prenatal image.