Probably yes. -- Supervised Deep Learning dominates performance scores for many computer vision tasks and defines the state-of-the-art. However, medical image analysis lags behind natural image applications. One of the many reasons is the lack of well annotated medical image data available to researchers. One of the first things researchers are told is that we require significant expertise to reliably and accurately interpret and label such data. We see significant inter- and intra-observer variability between expert annotations of medical images. Still, it is a widely held assumption that novice annotators are unable to provide useful annotations for use by clinical Deep Learning models. In this work we challenge this assumption and examine the implications of using a minimally trained novice labelling workforce to acquire annotations for a complex medical image dataset. We study the time and cost implications of using novice annotators, the raw performance of novice annotators compared to gold-standard expert annotators, and the downstream effects on a trained Deep Learning segmentation model's performance for detecting a specific congenital heart disease (hypoplastic left heart syndrome) in fetal ultrasound imaging.
Lung ultrasound (LUS) imaging is becoming popular in the intensive care units (ICU) for assessing lung abnormalities such as the appearance of B-line artefacts as a result of severe dengue. These artefacts appear in the LUS images and disappear quickly, making their manual detection very challenging. They also extend radially following the propagation of the sound waves. As a result, we hypothesize that a polar representation may be more adequate for automatic image analysis of these images. This paper presents an attention-based Convolutional+LSTM model to automatically detect B-lines in LUS videos, comparing performance when image data is taken in Cartesian and polar representations. Results indicate that the proposed framework with polar representation achieves competitive performance compared to the Cartesian representation for B-line classification and that attention mechanism can provide better localization.
Fetal ultrasound screening during pregnancy plays a vital role in the early detection of fetal malformations which have potential long-term health impacts. The level of skill required to diagnose such malformations from live ultrasound during examination is high and resources for screening are often limited. We present an interpretable, atlas-learning segmentation method for automatic diagnosis of Hypo-plastic Left Heart Syndrome (HLHS) from a single `4 Chamber Heart' view image. We propose to extend the recently introduced Image-and-Spatial Transformer Networks (Atlas-ISTN) into a framework that enables sensitising atlas generation to disease. In this framework we can jointly learn image segmentation, registration, atlas construction and disease prediction while providing a maximum level of clinical interpretability compared to direct image classification methods. As a result our segmentation allows diagnoses competitive with expert-derived manual diagnosis and yields an AUC-ROC of 0.978 (1043 cases for training, 260 for validation and 325 for testing).
The subject of "fairness" in artificial intelligence (AI) refers to assessing AI algorithms for potential bias based on demographic characteristics such as race and gender, and the development of algorithms to address this bias. Most applications to date have been in computer vision, although some work in healthcare has started to emerge. The use of deep learning (DL) in cardiac MR segmentation has led to impressive results in recent years, and such techniques are starting to be translated into clinical practice. However, no work has yet investigated the fairness of such models. In this work, we perform such an analysis for racial/gender groups, focusing on the problem of training data imbalance, using a nnU-Net model trained and evaluated on cine short axis cardiac MR data from the UK Biobank dataset, consisting of 5,903 subjects from 6 different racial groups. We find statistically significant differences in Dice performance between different racial groups. To reduce the racial bias, we investigated three strategies: (1) stratified batch sampling, in which batch sampling is stratified to ensure balance between racial groups; (2) fair meta-learning for segmentation, in which a DL classifier is trained to classify race and jointly optimized with the segmentation model; and (3) protected group models, in which a different segmentation model is trained for each racial group. We also compared the results to the scenario where we have a perfectly balanced database. To assess fairness we used the standard deviation (SD) and skewed error ratio (SER) of the average Dice values. Our results demonstrate that the racial bias results from the use of imbalanced training data, and that all proposed bias mitigation strategies improved fairness, with the best SD and SER resulting from the use of protected group models.
Lung ultrasound (LUS) imaging is used to assess lung abnormalities, including the presence of B-line artefacts due to fluid leakage into the lungs caused by a variety of diseases. However, manual detection of these artefacts is challenging. In this paper, we propose a novel methodology to automatically detect and localize B-lines in LUS videos using deep neural networks trained with weak labels. To this end, we combine a convolutional neural network (CNN) with a long short-term memory (LSTM) network and a temporal attention mechanism. Four different models are compared using data from 60 patients. Results show that our best model can determine whether one-second clips contain B-lines or not with an F1 score of 0.81, and extracts a representative frame with B-lines with an accuracy of 87.5%.
Congenital heart disease is considered as one the most common groups of congenital malformations which affects $6-11$ per $1000$ newborns. In this work, an automated framework for detection of cardiac anomalies during ultrasound screening is proposed and evaluated on the example of Hypoplastic Left Heart Syndrome (HLHS), a sub-category of congenital heart disease. We propose an unsupervised approach that learns healthy anatomy exclusively from clinically confirmed normal control patients. We evaluate a number of known anomaly detection frameworks together with a new model architecture based on the $\alpha$-GAN network and find evidence that the proposed model performs significantly better than the state-of-the-art in image-based anomaly detection, yielding average $0.81$ AUC \emph{and} a better robustness towards initialisation compared to previous works.
One of the challenges in developing deep learning algorithms for medical image segmentation is the scarcity of annotated training data. To overcome this limitation, data augmentation and semi-supervised learning (SSL) methods have been developed. However, these methods have limited effectiveness as they either exploit the existing data set only (data augmentation) or risk negative impact by adding poor training examples (SSL). Segmentations are rarely the final product of medical image analysis - they are typically used in downstream tasks to infer higher-order patterns to evaluate diseases. Clinicians take into account a wealth of prior knowledge on biophysics and physiology when evaluating image analysis results. We have used these clinical assessments in previous works to create robust quality-control (QC) classifiers for automated cardiac magnetic resonance (CMR) analysis. In this paper, we propose a novel scheme that uses QC of the downstream task to identify high quality outputs of CMR segmentation networks, that are subsequently utilised for further network training. In essence, this provides quality-aware augmentation of training data in a variant of SSL for segmentation networks (semiQCSeg). We evaluate our approach in two CMR segmentation tasks (aortic and short axis cardiac volume segmentation) using UK Biobank data and two commonly used network architectures (U-net and a Fully Convolutional Network) and compare against supervised and SSL strategies. We show that semiQCSeg improves training of the segmentation networks. It decreases the need for labelled data, while outperforming the other methods in terms of Dice and clinical metrics. SemiQCSeg can be an efficient approach for training segmentation networks for medical image data when labelled datasets are scarce.
Prenatal screening with ultrasound can lower neonatal mortality significantly for selected cardiac abnormalities. However, the need for human expertise, coupled with the high volume of screening cases, limits the practically achievable detection rates. In this paper we discuss the potential for deep learning techniques to aid in the detection of congenital heart disease (CHD) in fetal ultrasound. We propose a pipeline for automated data curation and classification. During both training and inference, we exploit an auxiliary view classification task to bias features toward relevant cardiac structures. This bias helps to improve in F1-scores from 0.72 and 0.77 to 0.87 and 0.85 for healthy and CHD classes respectively.
Prenatal screening with ultrasound can lower neonatal mor-tality significantly for selected cardiac abnormalities. However, the needfor human expertise, coupled with the high volume of screening cases,limits the practically achievable detection rates. In this paper we discussthe potential for deep learning techniques to aid in the detection of con-genital heart disease (CHD) in fetal ultrasound. We propose a pipelinefor automated data curation and classification. During both training andinference, we exploit an auxiliary view classification task to bias featurestoward relevant cardiac structures. This bias helps to improve in F1-scores from 0.72 and 0.77 to 0.87 and 0.85 for healthy and CHD classesrespectively.