This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.

Purpose Cardiac MRI represents the gold standard to determine myocardial function. However, the current clinical standard protocol, a segmented Cartesian acquisition, is time‐consuming and can lead to compromised image quality in the case of arrhythmia or dyspnea. In this article, a machine learning–based reconstruction of undersampled spiral k‐space data is presented to enable free breathing real‐time cardiac MRI with good image quality and short reconstruction times. Methods Data were acquired in free breathing with a 2D spiral trajectory corrected by the gradient system transfer function. Undersampled data were reconstructed by a variational network (VN), which was specifically adapted to the non‐Cartesian sampling pattern. The network was trained with data from 11 subjects. Subsequently, the imaging technique was validated in 14 subjects by quantifying the difference to a segmented reference acquisition, an expert reader study, and by comparing derived volumes and functional parameters with values obtained using the current clinical gold standard. Results The scan time for the entire heart was below 1 min. The VN reconstructed data in about 0.9 s per image, which is considerably shorter than conventional model‐based approaches. The VN furthermore performed better than a U‐Net and not inferior to a low‐rank plus sparse model in terms of achieved image quality. Functional parameters agreed, on average, with reference data. Conclusions The proposed VN method enables real‐time cardiac imaging with both high spatial and temporal resolution in free breathing and with short reconstruction time.

The purpose of the current study was to implement and validate joint real‐time acquisition of functional and late gadolinium‐enhancement (LGE) cardiac magnetic resonance (MR) images during free breathing. Inversion recovery cardiac real‐time images with a temporal resolution of 50 ms were acquired using a spiral trajectory (IR‐CRISPI) with a pre‐emphasis based on the gradient system transfer function during free breathing. Functional and LGE cardiac MR images were reconstructed using a low‐rank plus sparse model. Late gadolinium‐enhancement appearance, image quality, and functional parameters of IR‐CRISPI were compared with clinical standard balanced steady‐state free precession breath‐hold techniques in 10 patients. The acquisition of IR‐CRISPI in free breathing of the entire left ventricle took 97 s on average. Bland–Altman analysis and Wilcoxon tests showed a higher artifact level for the breath‐hold technique (p = 0.003), especially for arrhythmic patients or patients with dyspnea, but an increased noise level for IR‐CRISPI of the LGE images (p = 0.01). The estimated transmural extent of the enhancement differed by not more than 25% and did not show a significant bias between the techniques (p = 0.50). The ascertained functional parameters were similar for the breath‐hold technique and IR‐CRISPI, that is, with a minor, nonsignificant (p = 0.16) mean difference of the ejection fraction of 2.3% and a 95% confidence interval from −4.8% to 9.4%. IR‐CRISPI enables joint functional and LGE imaging in free breathing with good image quality but distinctly shorter scan times in comparison with breath‐hold techniques.