Fetal Brain Abnormality Classification from MRI Images of Different Gestational Age

Abstract

Magnetic resonance imaging (MRI) is a common imaging technique used extensively to study human brain activities. Recently, it has been used for scanning the fetal brain. Amongst 1000 pregnant women, 3 of them have fetuses with brain abnormality. Hence, the primary detection and classification are important. Machine learning techniques have a large potential in aiding the early detection of these abnormalities, which correspondingly could enhance the diagnosis process and follow up plans. Most research focused on the classification of abnormal brains in a primary age has been for newborns and premature infants, with fewer studies focusing on images for fetuses. These studies associated fetal scans to scans after birth for the detection and classification of brain defects early in the neonatal age. This type of brain abnormality is named small for gestational age (SGA). This article proposes a novel framework for the classification of fetal brains at an early age (before the fetus is born). As far as we could know, this is the first study to classify brain abnormalities of fetuses of widespread gestational ages (GAs). The study incorporates several machine learning classifiers, such as diagonal quadratic discriminates analysis (DQDA), K-nearest neighbour (K-NN), random forest, naïve Bayes, and radial basis function (RBF) neural network classifiers. Moreover, several bagging and Adaboosting ensembles models have been constructed using random forest, naïve Bayes, and RBF network classifiers. The performances of these ensembles have been compared with their individual models. Our results show that our novel approach can successfully identify and classify numerous types of defects within MRI images of the fetal brain of various GAs. Using the KNN classifier, we were able to achieve the highest classification accuracy and area under receiving operating characteristics of 95.6% and 99% respectively. In addition, ensemble classifiers improved the results of their respective individual models.

Deep Learning Techniques for Automatic Detection of Embryonic Neurodevelopmental Disorders

“…The results in the table show that the proposed framework is competitive with other methods based on standard machine learning techniques. The highest accuracy achieved using the proposed framework was 88.6% which was greater than that achieved in Reference [10], but slightly lower than that achieved in Reference [11].…”

“…Our approach outperformed the method proposed in Reference [10]. However, in some cases, when the dataset was relatively small, standard machine learning techniques might outperform deep learning methods [36,37], which was the case when comparing with Reference [11]. Reference [11] applied handcrafted feature extraction methods such as a Gabor filter and Gray Level Co-occurrence Matrix (GLCM).…”

“…However, the use of machine learning and deep learning techniques was not extensively explored. There are a limited number of research articles [8][9][10][11] that investigated the use of such techniques for detecting ENDs. These articles utilized conventional machine learning methods for feature extraction and the classification of ENDs.…”

“…Machine learning (ML) approaches are used extensively in medical applications due to their interesting capability for extracting valuable information from medical datasets [7]. Previous work done in embryonic neurodevelopmental disorder classification (ENDC) is very limited [8][9][10][11]. These articles applied standard machine learning methods to MRI images for feature extraction and classification of embryonic neurodevelopmental disorders.…”

“…On the other hand, classifying neurodevelopmental disorders of the brain of the embryo was not extensively investigated, and very limited work was done to solve this problem. Such related work is limited to References [8][9][10][11] only. To our own knowledge, this is the first study that uses deep learning techniques to detect ENDs.…”