Statistical atlases and computational models of the heart. STACOM (Workshop)最新文献

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.

Segmentation of the left ventricle and quantification of various cardiac contractile functions is crucial for the timely diagnosis and treatment of cardiovascular diseases. Traditionally, the two tasks have been tackled independently. Here we propose a convolutional neural network based multi-task learning approach to perform both tasks simultaneously, such that, the network learns better representation of the data with improved generalization performance. Probabilistic formulation of the problem enables learning the task uncertainties during the training, which are used to automatically compute the weights for the tasks. We performed a five fold cross-validation of the myocardium segmentation obtained from the proposed multi-task network on 97 patient 4-dimensional cardiac cine-MRI datasets available through the STA-COM LV segmentation challenge against the provided gold-standard myocardium segmentation, obtaining a Dice overlap of 0.849 ± 0.036 and mean surface distance of 0.274 ± 0.083 mm, while simultaneously estimating the myocardial area with mean absolute difference error of 205 ± 198 mm².

Cardiovascular simulations using patient-specific geometries can help researchers understand the mechanical behavior of the heart under different loading or disease conditions. However, to replicate the regional mechanics of the heart accurately, both the nonlinear passive and active material properties must be estimated reliably. In this paper, automated methods were used to determine passive material properties while simultaneously computing the unloaded reference geometry of the ventricles for stress analysis. Two different approaches were used to model systole. In the first, a physiologically-based active contraction model [1] coupled to a hemodynamic three-element Windkessel model of the circulation was used to simulate ventricular ejection. In the second, developed active tension was directly adjusted to match ventricular volumes at end-systole while prescribing the known end-systolic pressure. These methods were tested in four normal dogs using the data provided for the LV mechanics challenge [2]. The resulting end-diastolic and end-systolic geometry from the simulation were compared with measured image data.

3D echocardiographic (3DE) imaging is a useful tool for assessing the complex geometry of the aortic valve apparatus. Segmentation of this structure in 3DE images is a challenging task that benefits from shape-guided deformable modeling methods, which enable inter-subject statistical shape comparison. Prior work demonstrates the efficacy of using continuous medial representation (cm-rep) as a shape descriptor for valve leaflets. However, its application to the entire aortic valve apparatus is limited since the structure has a branching medial geometry that cannot be explicitly parameterized in the original cm-rep framework. In this work, we show that the aortic valve apparatus can be accurately segmented using a new branching medial modeling paradigm. The segmentation method achieves a mean boundary displacement of 0.6 ± 0.1 mm (approximately one voxel) relative to manual segmentation on 11 3DE images of normal open aortic valves. This study demonstrates a promising approach for quantitative 3DE analysis of aortic valve morphology.

Automatic segmentation of the heart's left atrium offers great benefits for planning and outcome evaluation of atrial ablation procedures. However, the high anatomical variability of the left atrium presents significant challenges for atlas-guided segmentation. In this paper, we demonstrate an automatic method for left atrium segmentation using weighted voting label fusion and a variant of the demons registration algorithm adapted to handle images with different intensity distributions. We achieve accurate automatic segmentation that is robust to the high anatomical variations in the shape of the left atrium in a clinical dataset of MRA images.