Sandro F. Queiros

Fast automatic myocardial segmentation in 4D cine CMR datasets.

A novel automatic 3D+time left ventricle (LV) segmentation framework is proposed for cardiac magnetic resonance (CMR) datasets. The proposed framework consists of three conceptual blocks to delineate both endo and epicardial contours throughout the cardiac cycle: (1) an automatic 2D mid-ventricular initialization and segmentation; (2) an automatic stack initialization followed by a 3D segmentation at the end-diastolic phase; and (3) a tracking procedure. Hereto, we propose to adapt the recent B-spline Explicit Active Surfaces (BEAS) framework to the properties of CMR images by integrating dedicated energy terms. Moreover, we extend the coupled BEAS formalism towards its application in 3D MR data by adapting it to a cylindrical space suited to deal with the topology of the image data. Furthermore, a fast stack initialization method is presented for efficient initialization and to enforce consistent cylindrical topology. Finally, we make use of an anatomically constrained optical flow method for temporal tracking of the LV surface. The proposed framework has been validated on 45 CMR datasets taken from the 2009 MICCAI LV segmentation challenge. Results show the robustness, efficiency and competitiveness of the proposed method both in terms of accuracy and computational load.

Detailed Evaluation of Five 3D Speckle Tracking Algorithms Using Synthetic Echocardiographic Recordings

A plethora of techniques for cardiac deformation imaging with 3D ultrasound, typically referred to as 3D speckle tracking techniques, are available from academia and industry. Although the benefits of single methods over alternative ones have been reported in separate publications, the intrinsic differences in the data and definitions used makes it hard to compare the relative performance of different solutions. To address this issue, we have recently proposed a framework to simulate realistic 3D echocardiographic recordings and used it to generate a common set of ground-truth data for 3D speckle tracking algorithms, which was made available online. The aim of this study was therefore to use the newly developed database to contrast non-commercial speckle tracking solutions from research groups with leading expertise in the field. The five techniques involved cover the most representative families of existing approaches, namely block-matching, radio-frequency tracking, optical flow and elastic image registration. The techniques were contrasted in terms of tracking and strain accuracy. The feasibility of the obtained strain measurements to diagnose pathology was also tested for ischemia and dyssynchrony.

Automatic 3D aortic annulus sizing by computed tomography in the planning of transcatheter aortic valve implantation

BACKGROUND Accurate imaging assessment of aortic annulus (AoA) dimension is paramount to decide on the correct transcatheter heart valve (THV) size for patients undergoing transcatheter aortic valve implantation (TAVI). We evaluated the feasibility and accuracy of a novel automatic framework for multidetector row computed tomography (MDCT)-based TAVI planning. METHODS Among 122 consecutive patients undergoing TAVI and retrospectively reviewed for this study, 104 patients with preoperative MDCT of sufficient quality were enrolled and analyzed with the proposed software. Fully automatic (FA) and semi-automatic (SA) AoA measurements were compared to manual measurements, with both automated and manual-based interobserver variability (IOV) being assessed. Finally, the effect of these measures on hypothetically selected THV size was evaluated against the implanted size, as well as with respect to manually-derived sizes. RESULTS FA analysis was feasible in 92.3% of the cases, increasing to 100% if using the SA approach. Automatically-extracted measurements showed excellent agreement with manually-derived ones, with small biases and narrow limits of agreement, and comparable to the interobserver agreement. The SA approach presented a statistically lower IOV than manual analysis, showing the potential to reduce interobserver sizing disagreements. Moreover, the automated approaches displayed close agreement with the implanted sizes, similar to the ones obtained by the experts. CONCLUSION The proposed automatic framework provides an accurate and robust tool for AoA measurements and THV sizing in patients undergoing TAVI.

Cardiac motion and deformation estimation from tagged MRI sequences using a temporal coherent image registration framework

Non-rigid image registration has been proposed to extract myocardial motion and deformation from tagged Magnetic Resonance Imaging (t-MRI). Initial efforts focused on finding a set of pairwise registrations, while more recent methods proposed to perform a joint image alignment to exploit temporal information. However, the latter methods usually measure image similarity with respect to the first phase, which may not be optimal due to tag fading. In the present study, we therefore propose a sequential 2D+t registration method exploiting temporal information based on a frame-by-frame image similarity. The method was first tested on synthetic data to fine-tune its parameters, and its applicability was illustrated in human patient data. Furthermore, the sequential 2D+t method was able to detect dysfunctional regions corresponding to delayed-enhanced MRI areas in a database consisting of 8 pig datasets. While differences with respect to traditional 2D methods are limited in terms of end-systolic strain accuracy, including temporal information lead to both smoother trajectories and smoother strain curves.

Fast fully automatic segmentation of the myocardium in 2D cine MR images

A novel automatic initialization procedure for left ventricle (LV) cardiac magnetic resonance (CMR) segmentation is proposed through the combination of a LV localization method based on multilevel Otsu thresholding and an elliptical annular template matching algorithm. We then propose to adapt the recent B-spline Explicit Active Surfaces (BEAS) framework to the properties of CMR images by integrating two dedicated energy terms: a weighted localized Chan-Vese region-based energy to explicitly control the equilibrium point between the two regions around each interface and a combined local and global region-based formulation for the myocardial region. The proposed method has been validated on 45 mid-ventricular images taken from the 2009 MICCAI LV segmentation challenge. Results show the efficiency of our method both in terms of shape accuracy and computational times.

A laparoscopic surgery training interface

Laparoscopy is a surgical procedure on which operations in the abdomen are performed through small incisions using several specialized instruments. The laparoscopic surgery success greatly depends on surgeon skills and training. To achieve these technical high-standards, different apprenticeship methods have been developed, many based on in vivo training, an approach that involves high costs and complex setup procedures. This paper explores Virtual Reality (VR) simulation as an alternative for novice surgeons training. Even though several simulators are available on the market claiming successful training experiences, their use is extremely limited due to the economic costs involved. In this work, we present a low-cost laparoscopy simulator able to monitor and assist the trainees surgical movements. The developed prototype consists of a set of inexpensive sensors, namely an accelerometer, a gyroscope, a magnetometer and a flex sensor, attached to specific laparoscopic instruments. Our approach allows repeated assisted training of an exercise, without time constraints or additional costs, since no human artificial model is needed. A case study of our simulator applied to instrument manipulation practice (hand-eye coordination) is also presented.

A competitive strategy for atrial and aortic tract segmentation based on deformable models

HighlightsA novel competitive strategy to perform atrial region segmentation is presented.The proposed method was embedded on the B‐spline Explicit Active Surface framework.The methods accuracy was shown in three databases of two different modalities.No merging between contours was found, allowing correct evaluation of thin mid walls.The methods added value was highlighted for missing walls and noisy structures. Graphical abstract Figure. No caption available. ABSTRACT Multiple strategies have previously been described for atrial region (i.e. atrial bodies and aortic tract) segmentation. Although these techniques have proven their accuracy, inadequate results in the mid atrial walls are common, restricting their application for specific cardiac interventions. In this work, we introduce a novel competitive strategy to perform atrial region segmentation with correct delineation of the thin mid walls, and integrated it into the B‐spline Explicit Active Surfaces framework. A double‐stage segmentation process is used, which starts with a fast contour growing followed by a refinement stage with local descriptors. Independent functions are used to define each region, being afterward combined to compete for the optimal boundary. The competition locally constrains the surface evolution, prevents overlaps and allows refinement to the walls. Three different scenarios were used to demonstrate the advantages of the proposed approach, through the evaluation of its segmentation accuracy, and its performance for heterogeneous mid walls. Both computed tomography and magnetic resonance imaging datasets were used, presenting results similar to the state‐of‐the‐art methods for both atria and aorta. The competitive strategy showed its superior performance with statistically significant differences against the traditional free‐evolution approach in cases with bad image quality or missed atrial/aortic walls. Moreover, only the competitive approach was able to accurately segment the atrial/aortic wall. Overall, the proposed strategy showed to be suitable for atrial region segmentation with a correct segmentation of the mid thin walls, demonstrating its added value with respect to the traditional techniques.

Development of a patient-specific atrial phantom model for planning and training of inter-atrial interventions

Background: Several authors have presented cardiac phantoms to mimic the particularities of the heart, making it suitable for medical training and surgical planning. Although the initial models were mainly focused on the ventricles, personalized phantoms of the atria were recently presented. However, such models are typically rigid, the atrial wall is not realistic and they are not compatible with ultrasound (US), being sub‐optimal for planning/training of several interventions. Methods: In this work, we propose a strategy to construct a patient‐specific atrial model. Specifically, the target anatomy is generated using a computed tomography (CT) dataset and then constructed using a mold‐cast approach. An accurate representation of the inter‐atrial wall (IAS) was ensured during the model generation, allowing its application for IAS interventions. Two phantoms were constructed using different flexible materials (silicone and polyvinyl alcohol cryogel, PVA‐C), which were then compared to assess their appropriateness for US acquisition and for the generation of complex anatomies. Results: Two experiments were set up to validate the proposed methodology. First, the accuracy of the manufacturing approach was assessed through the comparison between a post‐production CT and the virtual references. The results proved that the silicone‐based model was more accurate than the PVA‐C‐based one, with an error of 1.68 ± 0.79, 1.36 ± 0.94, 1.45 ± 0.77 mm for the left (LA) and right atria (RA) and IAS, respectively. Second, an US acquisition of each model was performed and the obtained images quantitatively and qualitatively assessed. Both models showed a similar performance in terms of visual evaluation, with an easy detection of the LA, RA, and the IAS. Furthermore, a moderate accuracy was obtained between the atrial surfaces extracted from the US and the ideal reference, and again a superior performance of the silicone‐based model against the PVA‐C phantom was observed. Conclusions: The proposed strategy proved to be accurate and feasible for the correct generation of complex personalized atrial models.

Dense motion field estimation from myocardial boundary displacements.

Minimally invasive cardiovascular interventions guided by multiple imaging modalities are rapidly gaining clinical acceptance for the treatment of several cardiovascular diseases. These images are typically fused with richly detailed pre-operative scans through registration techniques, enhancing the intra-operative clinical data and easing the image-guided procedures. Nonetheless, rigid models have been used to align the different modalities, not taking into account the anatomical variations of the cardiac muscle throughout the cardiac cycle. In the current study, we present a novel strategy to compensate the beat-to-beat physiological adaptation of the myocardium. Hereto, we intend to prove that a complete myocardial motion field can be quickly recovered from the displacement field at the myocardial boundaries, therefore being an efficient strategy to locally deform the cardiac muscle. We address this hypothesis by comparing three different strategies to recover a dense myocardial motion field from a sparse one, namely, a diffusion-based approach, thin-plate splines, and multiquadric radial basis functions. Two experimental setups were used to validate the proposed strategy. First, an in silico validation was carried out on synthetic motion fields obtained from two realistic simulated ultrasound sequences. Then, 45 mid-ventricular 2D sequences of cine magnetic resonance imaging were processed to further evaluate the different approaches. The results showed that accurate boundary tracking combined with dense myocardial recovery via interpolation/diffusion is a potentially viable solution to speed up dense myocardial motion field estimation and, consequently, to deform/compensate the myocardial wall throughout the cardiac cycle. Copyright

Fast left ventricle tracking using localized anatomical affine optical flow

In daily clinical cardiology practice, left ventricle (LV) global and regional function assessment is crucial for disease diagnosis, therapy selection, and patient follow-up. Currently, this is still a time-consuming task, spending valuable human resources. In this work, a novel fast methodology for automatic LV tracking is proposed based on localized anatomically constrained affine optical flow. This novel method can be combined to previously proposed segmentation frameworks or manually delineated surfaces at an initial frame to obtain fully delineated datasets and, thus, assess both global and regional myocardial function. Its feasibility and accuracy were investigated in 3 distinct public databases, namely in realistically simulated 3D ultrasound, clinical 3D echocardiography, and clinical cine cardiac magnetic resonance images. The method showed accurate tracking results in all databases, proving its applicability and accuracy for myocardial function assessment. Moreover, when combined to previous state-of-the-art segmentation frameworks, it outperformed previous tracking strategies in both 3D ultrasound and cardiac magnetic resonance data, automatically computing relevant cardiac indices with smaller biases and narrower limits of agreement compared to reference indices. Simultaneously, the proposed localized tracking method showed to be suitable for online processing, even for 3D motion assessment. Importantly, although here evaluated for LV tracking only, this novel methodology is applicable for tracking of other target structures with minimal adaptations.