Feasibility of free-breathing, GRAPPA-based, real-time cardiac cine assessment of left-ventricular function in cardiovascular patients at 3T

Evaluation of Early Gadolinium Enhancement (EGE) and Cardiac Functional Parameters in Cine-Magnetic Resonance Imaging (MRI) on Artificial Intelligence in Patients with Acute Myocarditis: A Case-Controlled Observational Study

<strong>BACKGROUND</strong> The diagnosis of myocarditis is challenging, and the treatment is generally delayed due to misdiagnosis or missed diagnosis. Endomyocardial biopsy (EMB) is not a specific or sensitive method. A case-controlled observational study was conducted to evaluate early gadolinium enhancement (EGE) and left ventricular functional parameters on Artificial Intelligence in cine-MRI in patients with acute myocarditis. <strong>MATERIAL AND METHODS</strong> We selected 21 patients with pathologically proven acute myocarditis. We analyzed the EGE findings (total/serial number and location of positive-segments using the 17-segment model according to the American Heart Association) and clinical characteristics (symptoms, arrhythmias in ECG, coronary angiography, and EMB). All patients were divided into positive EGE and negative EGE groups to analyze left ventricular functional parameters (LVEF, FS, LVEDD, LVEDV, LVESV, LVMM, LVSV, CO, and CI) on Artificial Intelligence. <strong>RESULTS</strong> We enrolled 21 patients (11 males) with a mean age of 32.6±9.8 years (range, 16 to 51 years). Abnormalities on EGE were found in 2/3 of patients, involving 41 segments among multiple locations on the myocardium. The differences in LVEF (40.2±10.2% <i>vs.</i> 51.3±3.6%), LVESV (69.0±16.1ml <i>vs.</i> 52.5±10.6ml) and LVSV (42.6±11.4 <i>vs.</i> 52.8±2.8 ml) on Artificial Intelligence was statistically significant between the positive EGE and negative EGE groups (p<0.05). <strong>CONCLUSIONS</strong> Our results suggest a significant role of EGE on the basis of Lake Louise criteria in evaluating patients with clinical suspicion of acute myocarditis. Parameters, including LVEF, LVESV, and LVSV, on Artificial Intelligence, may be useful independent predictors for capillary leakage and microcirculatory disturbance in myocarditis.

Real-time SPARSE-SENSE cine MR imaging in atrial fibrillation: a feasibility study

Background Cardiac magnetic resonance imaging (MRI) relies on correct ECG-gating, which is hindered in arrhythmia. Purpose To examine whether a prototype free-breathing real-time cine sequence using SPARSE-SENSE (SPARSE) improves left ventricular quantification in atrial fibrillation. Material and Methods On a 1.5T MR system left ventricular short-axis stacks were acquired of the SPARSE sequence and of a "reference" steady-state free precession (SSFP) sequence with arrhythmia rejection in 20 patients with atrial fibrillation. Two radiologists independently rated arrhythmia-caused artifact severity in both sequences using a 4-point scale. Coefficients of variation of myocardial signal intensity for both sequences were acquired. Volumetry was performed twice by one reader and once by another reader. Correlation between artifact severity and employed sequence was analyzed by modified Fisher's exact test. Coefficients of variation and volumetric data were compared by paired t-test and intraclass correlation. Results Median arrhythmia-caused artifact severity was 2 in both readers for SSFP and 0 (reader 1)/1 (reader 2) for SPARSE, being significantly lower in SPARSE ( P < 0.001). Mean coefficient of variance was significantly smaller in SPARSE (0.11 ± 0.04) compared to SSFP (0.22 ± 0.13, P = 0.003), which was interpreted as a hint for fewer artifacts in SPARSE. Only a small difference of 9 ± 15 mL was seen for end-systolic volume ( P = 0.019) between sequences, otherwise no significant difference was detected (end-diastolic volume, P = 0.200; stroke volume, P = 0.554; ejection fraction, P = 0.136; myocardial mass, P = 0.353). Intraclass correlation between sequences was good to excellent (range, 0.80-0.97). Conclusion Real-time MRI with SPARSE data sampling is promising in atrial fibrillation because it reduces arrhythmia-caused artifacts.

Convolutional Neural Network for the Detection of End-Diastole and End-Systole Frames in Free-Breathing Cardiac Magnetic Resonance Imaging

Free-breathing cardiac magnetic resonance (CMR) imaging has short examination time with high reproducibility. Detection of the end-diastole and the end-systole frames of the free-breathing cardiac magnetic resonance, supplemented by visual identification, is time consuming and laborious. We propose a novel method for automatic identification of both the end-diastole and the end-systole frames, in the free-breathing CMR imaging. The proposed technique utilizes the convolutional neural network to locate the left ventricle and to obtain the end-diastole and the end-systole frames from the respiratory motion signal. The proposed procedure works successfully on our free-breathing CMR data, and the results demonstrate a high degree of accuracy and stability. Convolutional neural network improves the postprocessing efficiency greatly and facilitates the clinical application of the free-breathing CMR imaging.