Background: Diagnosing cardiac amyloidosis (CA) from cine-CMR (cardiac magnetic resonance) alone is not reliable. In this study, we tested if a convolutional neural network (CNN) could outperform the visual diagnosis of experienced operators. Method: 119 patients with cardiac amyloidosis and 122 patients with left ventricular hypertrophy (LVH) of other origins were retrospectively selected. Diastolic and systolic cine-CMR images were preprocessed and labeled. A dual-input visual geometry group (VGG ) model was used for binary image classification. All images belonging to the same patient were distributed in the same set. Accuracy and area under the curve (AUC) were calculated per frame and per patient from a 40% held-out test set. Results were compared to a visual analysis assessed by three experienced operators. Results: frame-based comparisons between humans and a CNN provided an accuracy of 0.605 vs. 0.746 (p < 0.0008) and an AUC of 0.630 vs. 0.824 (p < 0.0001). Patient-based comparisons provided an accuracy of 0.660 vs. 0.825 (p < 0.008) and an AUC of 0.727 vs. 0.895 (p < 0.002). Conclusion: based on cine-CMR images alone, a CNN is able to discriminate cardiac amyloidosis from LVH of other origins better than experienced human operators (15 to 20 points more in absolute value for accuracy and AUC), demonstrating a unique capability to identify what the eyes cannot see through classical radiological analysis.

According to European Society of Cardiology guidelines (ESC2015) for infective endocarditis (IE) management, modified Duke criteria (mDC) are implemented with a degree of clinical suspicion degree, leading to grades such as “possible” or “rejected” IE despite a persisting high level of clinical suspicion. Herein, we evaluate the ¹⁸F-FDG PET/CT diagnostic and therapeutic impact in IE suspicion, with emphasis on possible/rejected IE with a high clinical suspicion. Excluding cases of definite IE diagnosis, 53 patients who underwent ¹⁸F-FDG PET/CT for IE suspicion were selected and afterwards classified according to both mDC (possible IE/Duke 1, rejected IE/Duke 0) and clinical suspicion degree (high and low IE suspicion). The final status regarding IE diagnosis (gold standard) was based on the multidisciplinary decision of the Endocarditis Team, including the ‘imaging specialist’. PET/CT images of the cardiac area were qualitatively interpreted and the intensity of each focus of extra-physiologic ¹⁸F-FDG uptake was evaluated by a maximum standardized uptake value (SUVmax) measurement. Extra-cardiac ¹⁸F-FDG PET/CT pathological findings were considered to be a possible embolic event, a possible source of IE, or even a concomitant infection. Based on the Endocarditis Team consensus, final diagnosis of IE was retained in 19 (36%) patients and excluded in 34 (64%). With a sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and global accuracy of 79%, 100%, 100%, 89%, and 92%, respectively, PET/CT performed significantly better than mDC (p = 0.003), clinical suspicion degree (p = 0.001), and a combination of both (p = 0.001) for IE diagnosis. In 41 patients with possible/rejected IE but high clinical suspicion, sensitivity, specificity, PPV, NPV, and global accuracies were 78%, 100%, 100%, 85%, and 90%, respectively. Moreover, PET/CT contributed to patient management in 24 out of 53 (45%) cases. ¹⁸F-FDG PET/CT represents a valuable diagnostic tool that could be proposed for challenging IE cases with significant differences between mDC and clinical suspicion degree. ¹⁸F-FDG PET/CT allows a binary diagnosis (definite or rejected IE) by removing uncertain diagnostic situations, thus improving patient therapeutic management.

The automatic classification of various types of cardiomyopathies is desirable but has never been performed using a convolutional neural network (CNN). The purpose of this study was to evaluate currently available CNN models to classify cine magnetic resonance (cine-MR) images of cardiomyopathies. Method: Diastolic and systolic frames of 1200 cine-MR sequences of three categories of subjects (395 normal, 411 hypertrophic cardiomyopathy, and 394 dilated cardiomyopathy) were selected, preprocessed, and labeled. Pretrained, fine-tuned deep learning models (VGG) were used for image classification (sixfold cross-validation and double split testing with hold-out data). The heat activation map algorithm (Grad-CAM) was applied to reveal salient pixel areas leading to the classification. Results: The diastolic–systolic dual-input concatenated VGG model cross-validation accuracy was 0.982 ± 0.009. Summed confusion matrices showed that, for the 1200 inputs, the VGG model led to 22 errors. The classification of a 227-input validation group, carried out by an experienced radiologist and cardiologist, led to a similar number of discrepancies. The image preparation process led to 5% accuracy improvement as compared to nonprepared images. Grad-CAM heat activation maps showed that most misclassifications occurred when extracardiac location caught the attention of the network. Conclusions: CNN networks are very well suited and are 98% accurate for the classification of cardiomyopathies, regardless of the imaging plane, when both diastolic and systolic frames are incorporated. Misclassification is in the same range as inter-observer discrepancies in experienced human readers.

The aim of this work was to compare the classification of cardiac MR-images of AL versus ATTR amyloidosis by neural networks and by experienced human readers. Cine-MR images and late gadolinium enhancement (LGE) images of 120 patients were studied (70 AL and 50 TTR). A VGG16 convolutional neural network (CNN) was trained with a 5-fold cross validation process, taking care to strictly distribute images of a given patient in either the training group or the test group. The analysis was performed at the patient level by averaging the predictions obtained for each image. The classification accuracy obtained between AL and ATTR amyloidosis was 0.750 for cine-CNN, 0.611 for Gado-CNN and between 0.617 and 0.675 for human readers. The corresponding AUC of the ROC curve was 0.839 for cine-CNN, 0.679 for gado-CNN (p < 0.004 vs. cine) and 0.714 for the best human reader (p < 0.007 vs. cine). Logistic regression with cine-CNN and gado-CNN, as well as analysis focused on the specific orientation plane, did not change the overall results. We conclude that cine-CNN leads to significantly better discrimination between AL and ATTR amyloidosis as compared to gado-CNN or human readers, but with lower performance than reported in studies where visual diagnosis is easy, and is currently suboptimal for clinical practice.

With the increasing availability of high-performance medical imaging for the management of patients with neuroendocrine tumors (NETs), a progressive growth of asymptomatic and incidentally detected cardiac metastases (CMs) has been observed in the recent years. In clinical practice, CMs of NENs are often incidentally detected by whole-body ⁶⁸Ga-labeled somatostatin analogs or ¹⁸F-fluorodihydroxyphenylalanine positron emission tomography/computed tomography, and afterwards accurately characterized by cardiac magnetic resonance (CMR) and/or gated cardiac computed tomography when CMR is contraindicated or not available. The interpreting physician should familiarize with the main imaging features of CM, a finding that may be encountered in NETs patients more than previously thought. Herein, we present a case series of four patients with CMs from small-intestine NETs highlighting strengths and weaknesses of a multimodality imaging approach in clinical practice.

Objectives: We aimed to study classical, publicly available convolutional neural networks (3D-CNNs) using a combination of several cine-MR orientation planes for the estimation of left ventricular ejection fraction (LVEF) without contour tracing. Methods: Cine-MR examinations carried out on 1082 patients from our institution were analysed by comparing the LVEF provided by the CVI42 software (V5.9.3) with the estimation resulting from different 3D-CNN models and various combinations of long- and short-axis orientation planes. Results: The 3D-Resnet18 architecture appeared to be the most favourable, and the results gradually and significantly improved as several long-axis and short-axis planes were combined. Simply pasting multiple orientation views into composite frames increased performance. Optimal results were obtained by pasting two long-axis views and six short-axis views. The best configuration provided an R² = 0.83, a mean absolute error (MAE) = 4.97, and a root mean square error (RMSE) = 6.29; the area under the ROC curve (AUC) for the classification of LVEF < 40% was 0.99, and for the classification of LVEF > 60%, the AUC was 0.97. Internal validation performed on 149 additional patients after model training provided very similar results (MAE 4.98). External validation carried out on 62 patients from another institution showed an MAE of 6.59. Our results in this area are among the most promising obtained to date using CNNs with cardiac magnetic resonance. Conclusion: (1) The use of traditional 3D-CNNs and a combination of multiple orientation planes is capable of estimating LVEF from cine-MRI data without segmenting ventricular contours, with a reliability similar to that of traditional methods. (2) Performance significantly improves as the number of orientation planes increases, providing a more complete view of the left ventricle.