Abstract Choosing optimal revascularization strategies for patients with obstructive coronary artery disease (CAD) remains a clinical challenge. While randomized controlled trials offer population-level insights, gaps remain regarding personalized decision-making for individual patients. We applied off-policy reinforcement learning (RL) to a composite data model from 41,328 unique patients with angiography confirmed obstructive CAD. In an offline setting we estimated optimal treatment policies and evaluated these policies using weighted importance sampling. Our findings indicate that RL guided therapy decisions generally outperformed physician-based decision making, with the best RL policy achieving about a 38% improvement in expected rewards based on a composite major cardiovascular events outcome. Additionally, we introduced methods to ensure that RL CAD treatment policies remain compatible with locally achievable clinical practice models, presenting an interpretable RL policy with a limited number of states. Overall, this novel RL-based clinical decision support tool, called RL4CAD, demonstrates potential to provide clinical decision support for the optimization of care in patients with obstructive CAD referred for invasive coronary angiography.
While three-dimensional (3D) late gadolinium-enhanced (LGE) magnetic resonance (MR) imaging provides good conspicuity of small myocardial lesions with short acquisition time, it poses a challenge for image analysis as a large number of axial images are required to be segmented. We developed a fully automatic convolutional neural network (CNN) called cascaded triplanar autoencoder M-Net (CTAEM-Net) to segment myocardial scar from 3D LGE MRI. Two sub-networks were cascaded to segment the left ventricle (LV) myocardium and then the scar within the pre-segmented LV myocardium. Each sub-network contains three autoencoder M-Nets (AEM-Nets) segmenting the axial, sagittal and coronal slices of the 3D LGE MR image, with the final segmentation determined by voting. The AEM-Net integrates three features: (1) multi-scale inputs, (2) deep supervision and (3) multi-tasking. The multi-scale inputs allow consideration of the global and local features in segmentation. Deep supervision provides direct supervision to deeper layers and facilitates CNN convergence. Multi-task learning reduces segmentation overfitting by acquiring additional information from autoencoder reconstruction, a task closely related to segmentation. The framework provides an accuracy of 86.43% and 90.18% for LV myocardium and scar segmentation, respectively, which are the highest among existing methods to our knowledge. The time required for CTAEM-Net to segment LV myocardium and the scar was 49.72 ± 9.69s and 120.25 ± 23.18s per MR volume, respectively. The accuracy and efficiency afforded by CTAEM-Net will make possible future large population studies. The generalizability of the framework was also demonstrated by its competitive performance in two publicly available datasets of different imaging modalities.
Abstract : The Marine Voyage Safety System (MARVSS) was conceived and proposed by Commander, Coast Guard Pacific Area. The Operations Analysis Branch, Commander, Atlantic Area was directed to conduct a technical analysis and feasibility study of the proposed float plan system. To determine the estimated costs of implementing MARVSS, a model was developed and evaluated for one, two and three regional data collection points. The study shows that the cost of such a float plan system cannot be justified by the limited benefits that such a system can provide. (Author)
Purpose: Accurate three‐dimensional (3D) reconstruction of myocardial infarct geometry is crucial to patient‐specific modeling of the heart aimed at providing therapeutic guidance in ischemic cardiomyopathy. However, myocardial infarct imaging is clinically performed using two‐dimensional (2D) late‐gadolinium enhanced cardiac magnetic resonance (LGE‐CMR) techniques, and a method to build accurate 3D infarct reconstructions from the 2D LGE‐CMR images has been lacking. The purpose of this study was to address this need. Methods: The authors developed a novel methodology to reconstruct 3D infarct geometry from segmented low‐resolution (Lo‐res) clinical LGE‐CMR images. Their methodology employed the so‐called logarithm of odds (LogOdds) function to implicitly represent the shape of the infarct in segmented image slices as LogOdds maps. These 2D maps were then interpolated into a 3D image, and the result transformed via the inverse of LogOdds to a binary image representing the 3D infarct geometry. To assess the efficacy of this method, the authors utilized 39 high‐resolution (Hi‐res) LGE‐CMR images, including 36 in vivo acquisitions of human subjects with prior myocardial infarction and 3 ex vivo scans of canine hearts following coronary ligation to induce infarction. The infarct was manually segmented by trained experts in each slice of the Hi‐res images, and the segmented data were downsampled to typical clinical resolution. The proposed method was then used to reconstruct 3D infarct geometry from the downsampled images, and the resulting reconstructions were compared with the manually segmented data. The method was extensively evaluated using metrics based on geometry as well as results of electrophysiological simulations of cardiac sinus rhythm and ventricular tachycardia in individual hearts. Several alternative reconstruction techniques were also implemented and compared with the proposed method. Results: The accuracy of the LogOdds method in reconstructing 3D infarct geometry, as measured by the Dice similarity coefficient, was 82.10% ± 6.58%, a significantly higher value than those of the alternative reconstruction methods. Among outcomes of electrophysiological simulations with infarct reconstructions generated by various methods, the simulation results corresponding to the LogOdds method showed the smallest deviation from those corresponding to the manual reconstructions, as measured by metrics based on both activation maps and pseudo‐ECGs. Conclusions: The authors have developed a novel method for reconstructing 3D infarct geometry from segmented slices of Lo‐res clinical 2D LGE‐CMR images. This method outperformed alternative approaches in reproducing expert manual 3D reconstructions and in electrophysiological simulations.
Abstract Background Ischemic heart disease (IHD) is the most common cause of heart failure (HF); however, the role of revascularization in these patients is still unclear. Consensus on proper use of cardiac imaging to help determine which candidates should be considered for revascularization has been hindered by the absence of clinical studies that objectively and prospectively compare the prognostic information of each test obtained using both standard and advanced imaging. Methods/Design This paper describes the design and methods to be used in the Alternative Imaging Modalities in Ischemic Heart Failure (AIMI-HF) multi-center trial. The primary objective is to compare the effect of HF imaging strategies on the composite clinical endpoint of cardiac death, myocardial infarction (MI), cardiac arrest and re-hospitalization for cardiac causes. In AIMI-HF, patients with HF of ischemic etiology (n = 1,261) will follow HF imaging strategy algorithms according to the question(s) asked by the physicians (for example, Is there ischemia and/or viability?), in agreement with local practices. Patients will be randomized to either standard (SPECT, Single photon emission computed tomography) imaging modalities for ischemia and/or viability or advanced imaging modalities: cardiac magnetic resonance imaging (CMR) or positron emission tomography (PET). In addition, eligible and consenting patients who could not be randomized, but were allocated to standard or advanced imaging based on clinical decisions, will be included in a registry. Discussion AIMI-HF will be the largest randomized trial evaluating the role of standard and advanced imaging modalities in the management of ischemic cardiomyopathy and heart failure. This trial will complement the results of the Surgical Treatment for Ischemic Heart Failure (STICH) viability substudy and the PET and Recovery Following Revascularization (PARR-2) trial. The results will provide policy makers with data to support (or not) further investment in and wider dissemination of alternative ‘advanced’ imaging technologies. Trial registration NCT01288560
Visceral adiposity is increased in those with Metabolic Syndrome (MetS) and atherosclerotic disease burden. In this study we evaluate for associations between intra-thoracic fat volume (ITFV) and myocardial infarction (MI) in patients with MetS. Ninety-four patients with MetS, MI or both were identified from a cardiovascular CMR clinical registry. MetS was defined in accordance to published guidelines; where-as MI was defined as the presence of subendocardial-based injury on late gadolinium enhancement imaging in a coronary vascular distribution. A healthy control group was also obtained from the same registry. Patients were selected into the following groups: MetS+/MI- (N = 32), MetS-/MI + (N = 30), MetS+/MI + (N = 32), MetS-/MI- (N = 16). ITFV quantification was performed using signal threshold analysis of sequential sagittal CMR datasets (HASTE) and indexed to body mass index. The mean age of the population was 59.8 ± 12.5 years. MetS+ patients (N=64) demonstrated a significantly higher indexed ITFV compared to MetS- patients (p = 0.05). Patients in respective MetS-/MI-, MetS+/MI-, MetS-/MI+, and MetS+/MI + study groups demonstrated a progressive elevation in the indexed ITFV (22.3 ± 10.6, 28.6 ± 12.6, 30.6 ± 12.3, and 35.2 ± 11.4 ml/kg/m2, (p = 0.002)). Among MetS+ patients those with MI showed a significantly higher indexed ITFV compared to those without MI (p = 0.02). ITFV is elevated in patients with MetS and incrementally elevated among those with evidence of prior ischemic myocardial injury. Accordingly, the quantification of ITFV may be a valuable marker of myocardial infarction risk among patients with MetS and warrants further investigation.
Evidence with regard to the incidence of injury to forwards and backs in the game of rugby league is extremely limited. A four year prospective study of all the injuries from one professional Rugby League club was conducted. All injuries that were received during match play were recorded, and those for forwards and backs compared. Forwards had a higher overall rates of injury than backs (139.4 [124.2-154.6] vs. 92.7 [80.9-104.6] per 1000 player hours, P < 0.00006). Forwards had a higher rate of injuries to all body sites with the exception of the ankle and the 'others' category of injury. They had significantly higher rates for the arm (11.6 [6.9-16.3] vs. 3.9 [1.4-6.4] per 1000 player hours, P = 0.005) and, the head and neck (53.9 [43.9-63.8] vs. 25.0 [18.7-31.4] injuries per 1000 player hours, P < 0.00006). Forwards had significantly more injuries than backs for contusions (17.1 vs. 7.3 per 1000 player hours, z = 2.85, P = 0.0044), lacerations (26.7 vs. 13.8 per 1000 player hours, z = 2.92, P = 0.0035) and haematomas (20.6 vs. 11.6 per 1000 player hours, z = 2.29, P = 0.02). Forwards were also more likely to be injured when in possession of the ball (70.5 [59.2-81.7] vs. 38.0 [30.2-45.7]), and also when tackling (33.2 [25.3-41.1] vs. 16.8 [11.6-22.1]). The higher rates of injury experienced by forwards were most likely as a result of their greater physical involvement in the game, both in attack and in defence.
