Coronary atherosclerosis is not necessarily stenotic due to outward remodelling of the vessel wall. Coronary plaque burden has been shown to correlate with risk of future coronary events¹ and thus direct, non-invasive visualization of both coronary lumen and vessel wall is desired. A 3D flow independent approach for vessel wall imaging was proposed recently², based on an interleaved acquisition and subtraction of data with (T2prep(+), lumen) and without (T2prep(-)) T2-preparation prepulse. This approach requires dual gating of both datasets to compensate for respiratory motion, resulting in long and unpredictable scan times. Here, we propose a novel nonrigid motion correction framework for free-breathing 3D whole-heart imaging using an interleaved scanning framework to produce co-registered coronary lumen and vessel wall images at near 100% scan efficiency. The proposed approach was compared with a 2D translational correction (TC) and no motion correction (NMC) in 9 subjects. Additionally, the proposed method was compared against a navigator gated (6mm) and tracked coronary lumen scan.Data was acquired using an interleaved scanning framework³ with three scans acquired in an alternated fashion: a set of 3D segmented whole-heart with and without T2 preparation and a 2D single shot golden-radial (GR) coronal image navigator (2D iNAV) (Fig. 1a). Motion correction was performed in two steps: beat-to-beat translational correction and bin-to-bin nonrigid correction (Fig. 1b). iNAVs were reconstructed and rigidly registered to estimate superior-inferior (SI) and right-left (RL) translational motion, allowing beat-to-beat translational correction. A respiratory signal was derived from the SI component and data was binned according to respiratory position. Bins were reconstructed with soft-gated⁴ iterative SENSE⁵ and registered to estimate 3D nonrigid motion. These motion fields were used in a General Matrix Description⁶ (GMD) reconstruction by solving: $$$\widehat{I} = arg min_I\left\{||\sum_b \bf A_b \bf F \bf S \bf U_b \bf I - \bf K||_2^2\right\}$$$, where $$$\widehat{I}$$$ is the motion corrected volume, $$$K$$$ the translational corrected k-space, $$$A_b$$$ the sampling matrix for bin b, $$$F$$$ the Fourier transform, $$$S$$$ the coil sensitivities and $$$U_b$$$ the nonrigid motion fields. Outliers due to deep breaths were removed prior to reconstruction. Both T2prep(+) and T2prep(-) were reconstructed with the framework described and nonrigidly registered to guarantee alignment. Vessel wall images were obtained via image subtraction as described by Andia et al².Nine healthy subjects were scanned free-breathing on a 1.5T Philips scanner using a 32-channel coil. T2prep(+) and T2prep(-) data were acquired with an ECG-triggered 3D Cartesian b-SSFP (1x1 mm in-plane resolution, 2 mm slice thickness, 300x300x90 mm FOV, TR/TE = 5.3/2.6 ms, flip angle = 70°) interleaved with a 2D golden radial iNAV spoiled gradient echo sequence (4x4 mm in-plane resolution, 25mm slice thickness, 300x300 FOV, TR/TE = 2.4/1.07 ms, flip angle = 5°). Non-motion correction (NMC), 2D translational correction (TC) and the proposed translation plus nonrigid (TC+GMD) reconstructions were obtained from these data. Additionally a 6mm diaphragmatic respiratory gated and tracked lumen scan (T2prep(+)) was performed for comparison. Coronary lumen was evaluated with vessel length and vessel sharpness (normalized to Gated), coronary vessel wall was evaluated by vessel wall thickness and vessel wall sharpness (normalized to TC+GMD); both images were evaluated by two experts on a scale of 0 (extreme blurring) to 4 (no blurring).Reformatted lumen images for Gated, TC+GMD, TC and NMC for 2 subjects are shown in Fig. 2. Motion artifacts in NMC are reduced with TC and further reduced with TC+GMD. The corresponding set of vessel wall images is shown in Fig. 3, along with cross sectional views. Vessel wall appears blurred in NMC; delineation is significantly improved by TC and sharpness further improved by TC+GMD. Fig. 4 shows metrics for coronary lumen evaluation. Lumen vessel length sharpness is reduced for NMC and increases with TC, TC+GMD and Gated. Similar results were obtained from visual scores. Analogous results were found for the vessel wall in Fig. 5: blurring increases the measured wall thickness for NMC and is reduced with TC and TC+GMD; correspondingly, wall sharpness is significantly improved with TC and TC+GMD from NMC. Visual score of the vessel wall was in agreement with the remaining metrics. TC+GMD and Gated show similar image quality and non-significant differences in any metric studied. The proposed TC+GMD improved scan time by 1.8x (on average) compared to Gated. A novel framework for simultaneous nonrigid correction for 3D coronary lumen and vessel wall was introduced. The proposed approach allows near 100% scan efficiency, while maintaining similar image quality to a 6mm gated and tracked lumen acquisition. Significant improvements were found with the proposed approach compared to translational correction alone.