Search

Background: Cardiac MRI feature‐tracking (FT) with breath‐holding (BH) cine balanced steady state free precession (bSSFP) imaging is well established. It is unclear whether FT‐strain measurements can be reliably derived from free‐breathing (FB) compressed sensing (CS) bSSFP imaging. Purpose: To compare left ventricular (LV) strain analysis and image quality of an FB CS bSSFP cine sequence with that of a conventional BH bSSFP sequence in children. Study Type: Prospective. Subjects: 40 children able to perform BHs (cohort 1 [12.1 ± 2.2 years]) and 17 children unable to perform BHs (cohort 2 [5.2 ± 1.8 years]). Field Strength/Sequence: 3T, bSSFP sequence with and without CS. Assessment: Acquisition times and image quality were assessed. LV myocardial deformation parameters were compared between BH cine and FB CS cine studies in cohort 1. Strain indices and image quality of FB CS cine studies were also assessed in cohort 2. Intraobserver and interobserver variability of strain parameters was determined. Statistical Tests: Paired t‐test, Wilcoxon signed‐rank test, intraclass correlation coefficient (ICC), and Bland–Altman analysis. A P‐value <0.05 was considered statistically significant. Results: In cohort 1, the mean acquisition time of the FB CS cine study was significantly lower than for conventional BH cine study (15.6 s vs. 209.4 s). No significant difference were found in global circumferential strain rate (P = 0.089), global longitudinal strain rate (P = 0.366) and EuroCMR image quality scores (P = 0.128) between BH and FB sequences in cohort 1. The overall image quality score of FB CS cine in cohort 2 was 3.5 ± 0.5 with acquisition time of 14.7 ± 2.1 s. Interobserver and intraobserver variabilities were good to excellent (ICC = 0.810 to 0.943). Data Conclusion: FB CS cine imaging may be a promising alternative technique for strain assessment in pediatric patients with poor BH ability. Level of Evidence: 1 Technical Efficacy: Stage 1 [ABSTRACT FROM AUTHOR]

Training data enrichment is a key factor in artificial intelligence (AI) technology development. At present, the bottleneck problem is that the quantity and type of labeled training data in valid samples are unable to meet the requirements of AI+MRI aided diagnosis. In this paper, an effective approach to solve the problem was presented. High resolution isotropic multi-dimensional data of regions of interests from patients or healthy volunteers were first acquired via a series of scanning on clinical MRI scanners, including quantitative T1, T2, proton density (Pd) and apparent diffusion coefficient (ADC) measurements. These data were then used as the ground truth, from which different types of images associated with different imaging sequences and parameters were obtained with a virtual MRI technology. The type of the images with the best boundary resolution were then selected manually by experienced doctors, on which three-dimensional mask matrix was obtained by manual contouring and labeling, serving as the template for other types of images. This enrichment method was developed as a software platform, which could provide sufficient quantity of image data from a small number of positive cases, thus meeting the data training enrichment requirement of AI+MRI diagnosis at low cost and with high efficiency.