Purpose To develop a deep learning model for increasing cardiac cine frame rate while maintaining spatial resolution and scan time. Materials and Methods A transformer-based model was trained and tested on a retrospective sample of cine images from 5840 patients (mean age, 55 years ± 19 [SD]; 3527 male patients) referred for clinical cardiac MRI from 2003 to 2021 at nine centers; images were acquired using 1.5- and 3-T scanners from three vendors. Data from three centers were used for training and testing (4:1 ratio). The remaining data were used for external testing. Cines with downsampled frame rates were restored using linear, bicubic, and model-based interpolation. The root mean square error between interpolated and original cine images was modeled using ordinary least squares regression. In a prospective study of 49 participants referred for clinical cardiac MRI (mean age, 56 years ± 13; 25 male participants) and 12 healthy participants (mean age, 51 years ± 16; eight male participants), the model was applied to cines acquired at 25 frames per second (fps), thereby doubling the frame rate, and these interpolated cines were compared with actual 50-fps cines. The preference of two readers based on perceived temporal smoothness and image quality was evaluated using a noninferiority margin of 10%. Results The model generated artifact-free interpolated images. Ordinary least squares regression analysis accounting for vendor and field strength showed lower error (P < .001) with model-based interpolation compared with linear and bicubic interpolation in internal and external test sets. The highest proportion of reader choices was "no preference" (84 of 122) between actual and interpolated 50-fps cines. The 90% CI for the difference between reader proportions favoring collected (15 of 122) and interpolated (23 of 122) high-frame-rate cines was -0.01 to 0.14, indicating noninferiority. Conclusion A transformer-based deep learning model increased cardiac cine frame rates while preserving both spatial resolution and scan time, resulting in images with quality comparable to that of images obtained at actual high frame rates. Keywords: Functional MRI, Heart, Cardiac, Deep Learning, High Frame Rate Supplemental material is available for this article. © RSNA, 2024.

Background Fully automatic analysis of myocardial perfusion MRI datasets enables rapid and objective reporting of stress/rest studies in patients with suspected ischemic heart disease. Developing deep learning techniques that can analyze multi-center datasets despite limited training data and variations in software (pulse sequence) and hardware (scanner vendor) is an ongoing challenge. Methods Datasets from 3 medical centers acquired at 3T (n = 150 subjects; 21,150 first-pass images) were included: an internal dataset (inD; n = 95) and two external datasets (exDs; n = 55) used for evaluating the robustness of the trained deep neural network (DNN) models against differences in pulse sequence (exD-1) and scanner vendor (exD-2). A subset of inD (n = 85) was used for training/validation of a pool of DNNs for segmentation, all using the same spatiotemporal U-Net architecture and hyperparameters but with different parameter initializations. We employed a space-time sliding-patch analysis approach that automatically yields a pixel-wise “uncertainty map” as a byproduct of the segmentation process. In our approach, dubbed Data Adaptive Uncertainty-Guided Space-time (DAUGS) analysis, a given test case is segmented by all members of the DNN pool and the resulting uncertainty maps are leveraged to automatically select the “best” one among the pool of solutions. For comparison, we also trained a DNN using the established approach with the same settings (hyperparameters, data augmentation, etc.). Results The proposed DAUGS analysis approach performed similarly to the established approach on the internal dataset (Dice score for the testing subset of inD: 0.896 ± 0.050 vs. 0.890 ± 0.049; p = n.s.) whereas it significantly outperformed on the external datasets (Dice for exD-1: 0.885 ± 0.040 vs. 0.849 ± 0.065, p < 0.005; Dice for exD-2: 0.811 ± 0.070 vs. 0.728 ± 0.149, p < 0.005). Moreover, the number of image series with “failed” segmentation (defined as having myocardial contours that include bloodpool or are noncontiguous in ≥1 segment) was significantly lower for the proposed vs. the established approach (4.3% vs. 17.1%, p < 0.0005). Conclusions The proposed DAUGS analysis approach has the potential to improve the robustness of deep learning methods for segmentation of multi-center stress perfusion datasets with variations in the choice of pulse sequence, site location or scanner vendor.

Aims:Artificial intelligence (AI) techniques have been proposed for automating analysis of short-axis (SAX) cine cardiac magnetic resonance (CMR), but no CMR analysis tool exists to automatically analyse large (unstructured) clinical CMR datasets. We develop and validate a robust AI tool for start-to-end automatic quantification of cardiac function from SAX cine CMR in large clinical databases.Methods and results:Our pipeline for processing and analysing CMR databases includes automated steps to identify the correct data, robust image pre-processing, an AI algorithm for biventricular segmentation of SAX CMR and estimation of functional biomarkers, and automated post-analysis quality control to detect and correct errors. The segmentation algorithm was trained on 2793 CMR scans from two NHS hospitals and validated on additional cases from this dataset (n = 414) and five external datasets (n = 6888), including scans of patients with a range of diseases acquired at 12 different centres using CMR scanners from all major vendors. Median absolute errors in cardiac biomarkers were within the range of inter-observer variability: <8.4 mL (left ventricle volume), <9.2 mL (right ventricle volume), <13.3 g (left ventricular mass), and <5.9% (ejection fraction) across all datasets. Stratification of cases according to phenotypes of cardiac disease and scanner vendors showed good performance across all groups.Conclusion:We show that our proposed tool, which combines image pre-processing steps, a domain-generalizable AI algorithm trained on a large-scale multi-domain CMR dataset and quality control steps, allows robust analysis of (clinical or research) databases from multiple centres, vendors, and cardiac diseases. This enables translation of our tool for use in fully automated processing of large multi-centre databases.

Left ventricular (LV) ischemia has been variably associated with functional mitral regurgitation (FMR). Determinants of FMR in patients with ischemia are poorly understood.This study sought to test whether contractile mechanics in ischemic myocardium underlying the mitral valve have an impact on likelihood of FMR.Vasodilator stress perfusion cardiac magnetic resonance was performed in patients with coronary artery disease (CAD) at multiple centers. FMR severity was confirmed quantitatively via core lab analysis. To test relationship of contractile mechanics with ischemic FMR, regional wall motion and strain were assessed in patients with inducible ischemia and minimal (≤5% LV myocardium, nontransmural) infarction.A total of 2,647 patients with CAD were studied; 34% had FMR (7% moderate or greater). FMR severity increased with presence (P < 0.001) and extent (P = 0.01) of subpapillary ischemia: patients with moderate or greater FMR had more subpapillary ischemia (odds ratio [OR]: 1.13 per 10% LV; 95% CI: 1.05-1.21; P = 0.001) independent of ischemia in remote regions (P = NS); moderate or greater FMR prevalence increased stepwise with extent of ischemia and infarction in subpapillary myocardium (P < 0.001); stronger associations between FMR and infarction paralleled greater wall motion scores in infarct-affected territories. Among patients with inducible ischemia and minimal infarction (n = 532), wall motion and radial strain analysis showed impaired subpapillary contractile mechanics to associate with moderate or greater FMR (P < 0.05) independent of remote regions (P = NS). Conversely, subpapillary ischemia without contractile dysfunction did not augment FMR likelihood. Mitral and interpapillary dimensions increased with subpapillary radial strain impairment; each remodeling parameter associated with impaired subpapillary strain (P < 0.05) independent of remote strain (P = NS). Subpapillary radial strain (OR: 1.13 per 5% [95% CI: 1.02-1.25]; P = 0.02) and mitral tenting area (OR: 1.05 per 10 mm2 [95% CI: 1.00-1.10]; P = 0.04) were associated with moderate or greater FMR controlling for global remodeling represented by LV end-systolic volume (P = NS): when substituting sphericity for LV volume, moderate or greater FMR remained independently associated with subpapillary radial strain impairment (OR: 1.22 per 5% [95% CI: 1.02-1.47]; P = 0.03).Among patients with CAD and ischemia, FMR severity and adverse mitral apparatus remodeling increase in proportion to contractile dysfunction underlying the mitral valve.

Abstract Introduction Tricuspid regurgitation imposes a volume overload on the right ventricle (RV) that can lead to progressive RV dilation and dysfunction. Overt RV dysfunction is associated with poor prognosis and increased operative risk. Abnormalities of myocardial strain may provide the earliest evidence of ventricular dysfunction. CMR feature-tracking techniques now allow assessment of strain from routine cine-images, without specialized pulse sequences. Whether abnormalities of RV strain measured using CMR feature-tracking have prognostic value in patients with tricuspid regurgitation is unknown Purpose To evaluate the prognostic value of CMR feature-tracking derived RV free wall longitudinal strain (RVFWLS) in a large multicenter population of patients with severe tricuspid regurgitation. Methods Consecutive patients with severe tricuspid regurgitation undergoing CMR at four US medical centers were included in this study. Feature-tracking RVFWLS was calculated from 4 chamber cine-views (Figure-left panel). The primary endpoint was all-cause death. Cox proportional hazards regression modeling was used to examine the independent association between RVFWLS and death. The incremental prognostic value of RVFWLS was assessed in nested models. Results Of the 406 patients in this study,115 died during a median follow-up of 8.8 years. By Kaplan-Meier-analysis, patients with RVFWLS ≥median (−16%) had significantly reduced event free survival compared to those with RVFWLS < median (log-rank p<0.001) (Figure-right panel). By Cox multivariable regression modeling, each 1% worsening in RVFWLS was associated with a 13% increased risk-of-death after adjustement for clinical and imaging risk factors (HR=1.13 per %; p<0.001). Addition of RVFWLS in this model resulted in significant-improvement in the global-chi-square (26 to 65; p<0.0001). Conclusions CMR feature-tracking derived RVFWLS is an independent predictor of mortality in patients with severe tricuspid regurgitation, incremental to common clinical and imaging risk factors. Funding Acknowledgement Type of funding source: None

Abstract Introduction Ejection fraction is the principal measure used clinically to assess cardiac mechanics and provides significant prognostic information. However, echocardiographic strain imaging has shown significant abnormalities of myocardial deformation can be present despite preserved ejection fraction, which maybe associated with adverse prognosis. Cardiac-Magnetic-Resonance (CMR) feature-tracking techniques now allow assessment of strain from routine cine-images, without specialized pulse sequences. Whether abnormalities of strain measured using CMR feature-tracking have prognostic value in patients with preserved ejection fraction is unknown. Purpose To evaluate the prognostic value of CMR feature-tracking derived global longitudinal strain (GLS) in a large multicenter population of patients with preserved ejection fraction. Methods Consecutive patients with preserved ejection fraction (EF ≥50%) and a clinical indication for CMR at four US medical centers were included in this study. Feature-tracking GLS was calculated from 3 long-axis-cine-views. The primary endpoint was all-cause death. Cox proportional hazards regression modeling was used to examine the independent association between GLS and death. The incremental prognostic value of GLS was assessed in nested models. Results Of the 1274 patients in this study, 115 died during a median follow-up of 6.2 years. By Kaplan-Meier analysis, patients with GLS ≥ median (−20%) had significantly reduced event free survival compared to those with GLS < median (log-rank p<0.001) (Figure, top panel). The continuous relationship between GLS and the hazard of death is shown in the cubic spline (Figure, lower panel). By Cox multivariable regression modeling, each 1% worsening in GLS was associated with a 23.6% increased risk-of-death after adjustment for clinical and imaging risk factors (HR=1.236 per %; p<0.001). Addition of GLS in this model resulted in significant-improvement in the global-chi-square (67 to 168; p<0.0001) and Harrel's C-statistic (0.716 to 0.825; p<0.0001). Conclusions CMR feature-tracking derived GLS is a powerful independent predictor of mortality in patients with preserved ejection fraction, incremental to common clinical and imaging risk factors.

Advanced heart failure: the touch ahead / Deformation imaging and outcome 1047 LVAD thrombus patient characteristics exchange are necessary to improve mortality in patients presenting with LVAD thrombosis.

Background Assessment of viability in patients with ischemic cardiomyopathy (ICM) prior to possible revascularization has fallen out of favor after the STITCH trial, which failed to demonstrate a benefit of echocardiography or nuclear viability testing. CMR assessment of viability by directly visualizing the transmural extent of both viable and nonviable myocardium offers unique advantages which have not been explored in large trials. Several small studies have shown significant contractile improvement after revascularization of dysfunctional segments with residual viability by CMR. The objective of this study was to determine the prevalence of residual viability over the entire range of severity of dysfunction in patients with ICM.