Jiri Spilka

Assessment of features for automatic CTG analysis based on expert annotation

Cardiotocography (CTG) is the monitoring of fetal heart rate (FHR) and uterine contractions (TOCO) since 1960s used routinely by obstetricians to detect fetal hypoxia. The evaluation of the FHR in clinical settings is based on an evaluation of macroscopic morphological features and so far has managed to avoid adopting any achievements from the HRV research field. In this work, most of the ever-used features utilized for FHR characterization, including FIGO, HRV, nonlinear, wavelet, and time and frequency domain features, are investigated and the features are assessed based on their statistical significance in the task of distinguishing the FHR into three FIGO classes. Annotation derived from the panel of experts instead of the commonly utilized pH values was used for evaluation of the features on a large data set (552 records). We conclude the paper by presenting the best uncorrelated features and their individual rank of importance according to the meta-analysis of three different ranking methods. Number of acceleration and deceleration, interval index, as well as Lempel-Ziv complexity and Higuchis fractal dimension are among the top five features.

Discriminating Normal from "Abnormal" Pregnancy Cases Using an Automated FHR Evaluation Method

Electronic fetal monitoring has become the gold standard for fetal assessment both during pregnancy as well as during delivery. Even though electronic fetal monitoring has been introduced to clinical practice more than forty years ago, there is still controversy in its usefulness especially due to the high inter- and intra-observer variability. Therefore the need for a more reliable and consistent interpretation has prompted the research community to investigate and propose various automated methodologies. In this work we propose the use of an automated method for the evaluation of fetal heart rate, the main monitored signal, which is based on a data set, whose labels/annotations are determined using a mixture model of clinical annotations. The successful results of the method suggest that it could be integrated into an assistive technology during delivery.

Information retrieval from hospital information system

This paper details the process of mining information from a hospital information system that has been designed approximately 15 years ago. The information is distributed within database tables in large textual attributes with a free structure. Information retrieval from these information is necessary for complementing cardiotocography signals with additional information that is to be implemented in a decision support system.The basic statistical overview (n-gram analysis) helped with the insight into data structure, however more sophisticated methods have to be used as human (and expert) processing of the whole data were out of consideration: over 620,000 text fields contained text reports in natural language with (many) typographical errors, duplicates, ambiguities, syntax errors and many (nonstandard) abbreviations.There was a strong need to efficiently determine the overall structure of the database and discover information that is important from the clinical point of view. We have used three different methods: k-means, self-organizing map and a self-organizing approach inspired by ant-colonies that performed clustering of the records. The records were visualized and revealed the most prominent information structure(s) that were consulted with medical experts and served for further mining from the database.The outcome of this task is a set of ordered or nominal attributes with a structural information that is available for rule discovery mining and automated processing for the research of asphyxia prediction during delivery. The proposed methodology has significantly reduced the processing time of loosely structured textual records for both IT and medical experts.

Evaluation of feature subsets for classification of cardiotocographic recordings

Electronic fetal monitoring - continuous recording of the cardiotocogram (CTG) - consisting of fetal heart rate (fHR) and tocographic signals, is a method used for intrapartal evaluation of the well being of the fetus. In this paper we evaluate several subsets of features for outcome classification of the pregnancy based on the CTG recording of the last 20 minutes preceding actual delivery. The features subsets are created based on PCA, information gain and Group of Adaptive Models Evolution (GAME) neural network feature selection algorithm. Our data set consisted of 104 intrapartum recordings including 18 pregnancies with acidemia reflected in umbilical artery pH<7.20. The results show that the best subset consisting of mix of time-domain and non-linear features performs consistently over the whole data set with sensitivity and specificity over 70%, which is well comparable to interobserver variations.

A three class treatment of the FHR classification problem using latent class analysis labeling

Electronic Fetal Monitoring in the form of cardiotocography is routinely used for fetal assessment both during pregnancy and delivery. However its interpretation requires a high level of expertise and even then the assessment is somewhat subjective as it has been proven by the high inter and intra-observer variability. Therefore the scientific community seeks for more objective methods for its interpretation. Along this path, presented work proposes a classification approach, which is based on a latent class analysis method that attempts to produce more objective labeling of the training cases, a step which is vital in a classification problem. The method is combined with a simple logistic regression approach under two different schemes: a standard multi-class classification formulation and an ordinal classification one. The results are promising suggesting that more effort should be put in this proposed approach.

p-leader based classification of first stage intrapartum fetal HRV

Interpretation and analysis of intrapartum fetal heart rate, enabling early detection of fetal acidosis, remains a challenging signal processing task. Recently, a variant of the wavelet-based multifractal analysis, based on p-exponents and p-leaders, which provides a rich framework for data regularity analysis, has been proposed. The present contribution aims at studying the benefits of using the p-leader multifractal formalism for discrimination of intrapartum fetal heart rate. First, a dependence on p of the multifractal properties of data is evidenced and interpreted. Second, classification between healthy subjects and fetuses suffering from acidosis is shown to have satisfactory performance that increases when p is decreased.

An adaptive method for the recovery of missing samples from FHR time series

Missing data cause serious problem for automatic evaluation of the fetal heart rate(FHR) series. In this work we present an algorithm to surpress this problem. More specifically, an adaptive approach is proposed based on two steps. The first step concerns the reconstruction step where we obtain an estimate of the missing data using an empirical dictionary. The second step consists from the construction of the dictionary using the updated values from the first step. The above two steps are applied iteratively until convergence. The method adapts each time the dictionary and the reconstructed time series to the new information that we gain. Results on real and simulated experiments have shown the usefullness of our approach. More specifically, a comparison with cubic spline interpolation is performed and have shown that the proposed approach achieved 4 to 9dB better reconstruction ability.

Least Squares Support Vector Machines for FHR Classification and Assessing the pH Based Categorization

Cardiotocography (CTG) is the major monitoring tool for fetal well-being surveillance during labor. It consists of two distinctive signals: the Fetal Heart Rate (FHR) and the Uterine Contractions signal. The CTG interpretation is classically performed by obstetricians with visual inspection for reassuring or ominous patterns, which are associated with fetus’ condition. Deviations of the CTG and especially of the (FHR) from normality can be an indication of oxygen deprivation during the stressful labor process, which can lead to major neurological damage to the fetus or even death. This compromise is usually reflected at the pH level of newborn’s blood. Therefore pH levels are usually used for the discrimination between healthy and compromised fetuses. In this work we present our preliminary results of the application of a machine learning approach, using least squares support vector machines, to FHR classification using the largest CTG open-access database so far.

Intrapartum fetal heart rate classification from trajectory in Sparse SVM feature space.

Intrapartum fetal heart rate (FHR) constitutes a prominent source of information for the assessment of fetal reactions to stress events during delivery. Yet, early detection of fetal acidosis remains a challenging signal processing task. The originality of the present contribution are three-fold: multiscale representations and wavelet leader based multifractal analysis are used to quantify FHR variability ; Supervised classification is achieved by means of Sparse-SVM that aim jointly to achieve optimal detection performance and to select relevant features in a multivariate setting ; Trajectories in the feature space accounting for the evolution along time of features while labor progresses are involved in the construction of indices quantifying fetal health. The classification performance permitted by this combination of tools are quantified on a intrapartum FHR large database (≃ 1250 subjects) collected at a French academic public hospital.

Practical Problems and Solutions in Hospital Information System Data Mining

Information mining from textual data becomes a very challenging task when the structure of the text record is very loose without any rules. Doctors often use natural language in medical records. Therefore it contains many ambiguities due to non-standard abbreviations and synonyms. The medical environment itself is also very specific: the natural language used in textual description varies with the personality creating the record (there are many personalized approaches), however it is restricted by terminology (i.e. medical terms, medical standards, etc.). Moreover, the typical patient record is filled with typographical errors, duplicates, ambiguities, syntax errors and many nonstandard abbreviations.