Toward Replacing Late Gadolinium Enhancement With Artificial Intelligence Virtual Native Enhancement for Gadolinium-Free Cardiovascular Magnetic Resonance Tissue Characterization in Hypertrophic Cardiomyopathy.

BACKGROUND: Late gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR) imaging is the gold standard for noninvasive myocardial tissue characterization but requires intravenous contrast agent administration. It is highly desired to develop a contrast agent-free technology to replace LGE for faster and cheaper CMR scans. METHODS: A CMR virtual native enhancement (VNE) imaging technology was developed using artificial intelligence. The deep learning model for generating VNE uses multiple streams of convolutional neural networks to exploit and enhance the existing signals in native T1 maps (pixel-wise maps of tissue T1 relaxation times) and cine imaging of cardiac structure and function, presenting them as LGE-equivalent images. The VNE generator was trained using generative adversarial networks. This technology was first developed on CMR datasets from the multicenter Hypertrophic Cardiomyopathy Registry, using hypertrophic cardiomyopathy as an exemplar. The datasets were randomized into 2 independent groups for deep learning training and testing. The test data of VNE and LGE were scored and contoured by experienced human operators to assess image quality, visuospatial agreement, and myocardial lesion burden quantification. Image quality was compared using a nonparametric Wilcoxon test. Intra- and interobserver agreement was analyzed using intraclass correlation coefficients (ICC). Lesion quantification by VNE and LGE were compared using linear regression and ICC. RESULTS: A total of 1348 hypertrophic cardiomyopathy patients provided 4093 triplets of matched T1 maps, cines, and LGE datasets. After randomization and data quality control, 2695 datasets were used for VNE method development and 345 were used for independent testing. VNE had significantly better image quality than LGE, as assessed by 4 operators (n=345 datasets; P<0.001 [Wilcoxon test]). VNE revealed lesions characteristic of hypertrophic cardiomyopathy in high visuospatial agreement with LGE. In 121 patients (n=326 datasets), VNE correlated with LGE in detecting and quantifying both hyperintensity myocardial lesions (r=0.77-0.79; ICC=0.77-0.87; P<0.001) and intermediate-intensity lesions (r=0.70-0.76; ICC=0.82-0.85; P<0.001). The native CMR images (cine plus T1 map) required for VNE can be acquired within 15 minutes and producing a VNE image takes less than 1 second. CONCLUSIONS: VNE is a new CMR technology that resembles conventional LGE but without the need for contrast administration. VNE achieved high agreement with LGE in the distribution and quantification of lesions, with significantly better image quality.

Endogenous T1ρ cardiovascular magnetic resonance in hypertrophic cardiomyopathy.

BACKGROUND: Hypertrophic cardiomyopathy (HCM) is characterized by increased left ventricular wall thickness, cardiomyocyte hypertrophy, and fibrosis. Adverse cardiac risk characterization has been performed using late gadolinium enhancement (LGE), native T1, and extracellular volume (ECV). Relaxation time constants are affected by background field inhomogeneity. T1ρ utilizes a spin-lock pulse to decrease the effect of unwanted relaxation. The objective of this study was to study T1ρ as compared to T1, ECV, and LGE in HCM patients. METHODS: HCM patients were recruited as part of the Novel Markers of Prognosis in Hypertrophic Cardiomyopathy study, and healthy controls were matched for comparison. In addition to cardiac functional imaging, subjects underwent T1 and T1ρ cardiovascular magnetic resonance imaging at short-axis positions at 1.5T. Subjects received gadolinium and underwent LGE imaging 15-20 min after injection covering the entire heart. Corresponding basal and mid short axis LGE slices were selected for comparison with T1 and T1ρ. Full-width half-maximum thresholding was used to determine the percent enhancement area in each LGE-positive slice by LGE, T1, and T1ρ. Two clinicians independently reviewed LGE images for presence or absence of enhancement. If in agreement, the image was labeled positive (LGE + +) or negative (LGE --); otherwise, the image was labeled equivocal (LGE + -). RESULTS: In 40 HCM patients and 10 controls, T1 percent enhancement area (Spearman's rho = 0.61, p < 1e-5) and T1ρ percent enhancement area (Spearman's rho = 0.48, p < 0.001e-3) correlated with LGE percent enhancement area. T1 and T1ρ percent enhancement areas were also correlated (Spearman's rho = 0.28, p = 0.047). For both T1 and T1ρ, HCM patients demonstrated significantly longer relaxation times compared to controls in each LGE category (p < 0.001 for all). HCM patients also showed significantly higher ECV compared to controls in each LGE category (p < 0.01 for all), and LGE -- slices had lower ECV than LGE + + (p = 0.01). CONCLUSIONS: Hyperenhancement areas as measured by T1ρ and LGE are moderately correlated. T1, T1ρ, and ECV were elevated in HCM patients compared to controls, irrespective of the presence of LGE. These findings warrant additional studies to investigate the prognostic utility of T1ρ imaging in the evaluation of HCM patients.

Characterization of subclinical diastolic dysfunction by cardiac magnetic resonance feature-tracking in adult survivors of non-Hodgkin lymphoma treated with anthracyclines.

BACKGROUND: The use of anthracycline-based chemotherapy is associated with the development of heart failure, even years after the end of treatment. Early detection of cardiac dysfunction could identify a high-risk subset of survivors who would eventually benefit from early intervention. Cardiac magnetic resonance feature-tracking (CMR-FT) analysis offers a practical and rapid method to calculate systolic and diastolic strains from routinely acquired cine images. While early changes in systolic function have been described, less data are available about late effects of chemotherapy in diastolic parameters by CMR-FT. The main goal of this study was to determine whether left ventricular (LV) early diastolic strain rates (GDSR-E) by CMR-FT are impaired in long-term adult survivors of non-Hodgkin lymphoma (NHL). Our secondary objective was to analyze associations between GDSR-E with cumulative anthracycline dose, systolic function parameters and myocardial tissue characteristics. METHODS: This is a single center cross-sectional observational study of asymptomatic patients in remission of NHL who previously received anthracycline therapy. All participants underwent their CMR examination on a 3.0-T scanner, including cines, T2 mapping, T1 mapping and late gadolinium enhancement imaging. Derived myocardial extracellular volume fraction was obtained from pre- and post-contrast T1 maps. CMR-FT analysis was performed using Trufi Strain software. The data obtained were compared between anthracycline group and volunteers without cardiovascular disease or neoplasia. RESULTS: A total of 18 adult survivors of NHL, 14 (77.8%) males, at mean age of 57.6 (± 14.7) years-old, were studied 88.2 (± 52.1) months after exposure to anthracycline therapy (median 400 mg/m2). Compared with controls, anthracycline group showed impaired LV global early diastolic circumferential strain rate (GCSR-E) [53.5%/s ± 19.3 vs 72.2%/s ± 26.7, p = 0.022], early diastolic longitudinal strain rate (GLSR-E) [40.4%/s ± 13.0 vs 55.9%/s ± 17.8, p = 0.006] and early diastolic radial strain rate (GRSR-E) [- 114.4%/s ± 37.1 vs - 170.5%/s ± 48.0, p < 0.001]. Impaired LV GCSR-E, GLSR-E and GRSR-E correlated with increased anthracycline dose and decreased systolic function. There were no correlations between GDSR-E and myocardial tissue characteristics. CONCLUSIONS: Left ventricular early diastolic strain rates by CMR-FT are impaired late after anthracycline chemotherapy in adult survivors of non-Hodgkin lymphoma.

Normal values of native T1 and T2 relaxation times on 3T cardiac MR in a healthy pediatric population aged 9-18 years.

BACKGROUND: Native myocardial T1 and T2 relaxation times are diagnostic tools used in clinical practice for adult and pediatric populations. Use of a mapping technique requires accurate knowledge of normal ranges in healthy patients, which is lacking in pediatric populations. PURPOSE: To establish normal values for native T1 and T2 mapping in healthy pediatric subjects of different ages and sex. STUDY TYPE: Prospective. POPULATION: Thirty-eight healthy children (9-18 years; mean age 14.0 ± 2.7). FIELD STRENGTH: Cardiac MR with a 3T scanner. T1 and T2 mapping using MyoMaps software. ASSESSMENT: T1 and T2 relaxation times were calculated from a 0.7-1.0 cm2 region of interest placed at the mid-ventricular short-axis slice in the interventricular septum by two observers. Inter- and intraobserver variability was assessed. STATISTICAL TESTS: The Student's t-test or the Mann-Whitney test for unpaired samples was applied to compare one continuous variable between two category groups. One-way analysis of variance (ANOVA) or a Kruskal-Wallis test was applied to compare one continuous variable between three category groups. Correlation between two continuous variables was assessed with a Pearson or Spearman test. RESULTS: The mean native T1 relaxation time was 1223 ± 29 msec and T2 relaxation time was 43 ± 4.5 msec. There was no correlation between T1 /T2 values and age or body surface area (for T1 P = 0.94 and 0.90 and for T2 P = 0.19 and 0.64, respectively). There was weak correlation between T1 values and body mass index (BMI) (r = 0.448, P = 0.005). T2 values were significantly higher in females compared with males (44.6 ± 4.2 vs. 40.4 ± 3.8 msec, P = 0.002). We found a significant rise of T2 relaxation time in the pubertal period (age 13-15 years) comparing to prepubertal (age 9-12 years). Inter- and intraobserver agreement of T1 (r = 0.93; r = 0.99) and T2 (r = 0.96; r = 0.95) were high. DATA CONCLUSION: We report normal values of native T1 and T2 relaxation times obtained with Myomaps software for 3T cardiac MR in a healthy pediatric population. LEVEL OF EVIDENCE: 2 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2020;51:912-918.

Use of natural language processing to improve predictive models for imaging utilization in children presenting to the emergency department.

OBJECTIVE: To examine the association between the medical imaging utilization and information related to patients' socioeconomic, demographic and clinical factors during the patients' ED visits; and to develop predictive models using these associated factors including natural language elements to predict the medical imaging utilization at pediatric ED. METHODS: Pediatric patients' data from the 2012-2016 United States National Hospital Ambulatory Medical Care Survey was included to build the models to predict the use of imaging in children presenting to the ED. Multivariable logistic regression models were built with structured variables such as temperature, heart rate, age, and unstructured variables such as reason for visit, free text nursing notes and combined data available at triage. NLP techniques were used to extract information from the unstructured data. RESULTS: Of the 27,665 pediatric ED visits included in the study, 8394 (30.3%) received medical imaging in the ED, including 6922 (25.0%) who had an X-ray and 1367 (4.9%) who had a computed tomography (CT) scan. In the predictive model including only structured variables, the c-statistic was 0.71 (95% CI: 0.70-0.71) for any imaging use, 0.69 (95% CI: 0.68-0.70) for X-ray, and 0.77 (95% CI: 0.76-0.78) for CT. Models including only unstructured information had c-statistics of 0.81 (95% CI: 0.81-0.82) for any imaging use, 0.82 (95% CI: 0.82-0.83) for X-ray, and 0.85 (95% CI: 0.83-0.86) for CT scans. When both structured variables and free text variables were included, the c-statistics reached 0.82 (95% CI: 0.82-0.83) for any imaging use, 0.83 (95% CI: 0.83-0.84) for X-ray, and 0.87 (95% CI: 0.86-0.88) for CT. CONCLUSIONS: Both CT and X-rays are commonly used in the pediatric ED with one third of the visits receiving at least one. Patients' socioeconomic, demographic and clinical factors presented at ED triage period were associated with the medical imaging utilization. Predictive models combining structured and unstructured variables available at triage performed better than models using structured or unstructured variables alone, suggesting the potential for use of NLP in determining resource utilization.

Quantification of mitral regurgitation in patients with hypertrophic cardiomyopathy using aortic and pulmonary flow data: impacts of left ventricular outflow tract obstruction and different left ventricular segmentation methods.

BACKGROUND: Cardiovascular magnetic resonance (CMR) imaging in patients with hypertrophic cardiomyopathy (HCM) enables the assessment of not only left ventricular (LV) hypertrophy and scarring but also the severity of mitral regurgitation. CMR assessment of mitral regurgitation is primarily based on the difference between LV stroke volume (LVSV) and aortic forward flow (Ao) measured using the phase-contrast (PC) technique. However, LV outflow tract (LVOT) obstruction causing turbulent, non-laminar flow in the ascending aorta may impact the accuracy of aortic flow quantification, leading to false conclusions regarding mitral regurgitation severity. Thus, we decided to quantify mitral regurgitation in patients with HCM using Ao or, alternatively, main pulmonary artery forward flow (MPA) for mitral regurgitation volume (MRvol) calculations. METHODS: The analysis included 143 prospectively recruited subjects with HCM and 15 controls. MRvol was calculated as the difference between LVSV computed with either the inclusion (LVSVincl) or exclusion (LVSVexcl) of papillary muscles and trabeculations from the blood pool and either Ao (MRvolAoi or MRvolAoe) or MPA (MRvolMPAi or MRvolMPAe). The presence or absence of LVOT obstruction was determined based on Doppler echocardiography findings. RESULTS: MRvolAoi was higher than MRvolMPAi in HCM patients with LVOT obstruction [47.0 ml, interquartile range (IQR) = 31.5-60.0 vs. 35.5 ml, IQR = 26.0-51.0; p < 0.0001] but not in non-obstructive HCM patients (23.0 ml, IQR = 16.0-32.0 vs. 24.0 ml, IQR = 15.3-32.0; p = 0.26) or controls (18.0 ml, IQR = 14.3-21.8 vs. 20.0 ml, IQR = 14.3-22.0; p = 0.89). In contrast to controls and HCM patients without LVOT obstruction, in HCM patients with LVOT obstruction, aortic flow-based MRvol (MRvolAoi) was higher than pulmonary-based findings (MRvolMPAi) (bias = 9.5 ml; limits of agreement: -11.7-30.7 with a difference of 47 ml in the extreme case). The differences between aortic-based and pulmonary-based MRvol values calculated using LVSVexcl mirrored those derived using LVSVincl. However, MRvol values calculated using LVSVexcl were lower in all the groups analyzed (HCM with LVOT obstruction, HCM without LVOT obstruction, and controls) and with all methods of MRvol quantification used (p ≤ 0.0001 for all comparisons). CONCLUSIONS: In HCM patients, LVOT obstruction significantly affects the estimation of aortic flow, leading to its underestimation and, consequently, to higher MRvol values than those obtained with MPA-based MRvol calculations.

Cine dyscontractility index: A novel marker of mechanical dyssynchrony that predicts response to cardiac resynchronization therapy.

PURPOSE: To investigate whether magnetic resonance imaging (MRI) cine-derived dyssynchrony indices provide additional information compared to conventional tagged MRI (tMRI) acquisitions in heart failure patients undergoing cardiac resynchronization therapy (CRT). MATERIALS AND METHODS: Patients scheduled for CRT (n = 52) underwent preprocedure MRI including cine and tMRI acquisitions. Segmental strain curves were calculated for both cine and tMRI to produce a range of standard indices for direct comparison between modalities. We also proposed and evaluated a novel index of "dyscontractility," which detects the presence of focal areas with paradoxically positive circumferential strain. RESULTS: Across conventional strain indices, there was only moderate-to-poor (R = 0.3-0.6) correlation between modalities; eight cine-derived indices showed statistically significant (P < 0.05) relations to CRT outcome compared to just two tMRI-based counterparts. The novel dyscontractility index calculated on basal slice cine images (cine dyscontractility index, "CDI") was the single best predictor of clinical response to CRT (area under the curve AUC = 0.81, P < 0.001). While poorly correlated to its tMRI counterpart (R = 0.33), CDI performed significantly better in predicting response to CRT (P < 0.005), and was also numerically better than all other tMRI indices (AUC 0.53-0.76, all P for AUC comparisons <0.17). CONCLUSION: Cine-derived strain indices offer potentially new information compared to tMRI. Specifically, the novel CDI is most strongly linked to response to cardiac resynchronization therapy in a contemporary patient cohort. It utilizes readily available MRI data, is relatively straightforward to process, and compares favorably with any conventional tagging index. J. Magn. Reson. Imaging 2016;44:1483-1492.

Repaired tetralogy of Fallot: ratio of right ventricular volume to left ventricular volume as a marker of right ventricular dilatation.

PURPOSE: To compare indexed right ventricular (RV) end-diastolic volume (RVEDVi) and the ratio of RV volume to left ventricular (LV) volume (RV/LV ratio) in prediction of significant pulmonary regurgitation (PR) after tetralogy of Fallot (TOF) repair and to assess sex differences in the RV/LV ratio. MATERIALS AND METHODS: The ethics committee approved this retrospective single-center study, and patients or their parents or guardians signed written informed consent. RVEDVi, RV/LV ratio, and PR were measured with the use of magnetic resonance imaging in 155 consecutive patients with repaired TOF (mean age, 29.2 years±10.9 [standard deviation]; 98 [63.2%] male and 57 [36.8%] female patients). PR fraction of 20% or greater was considered significant. The capability of the RVEDVi and that of the RV/LV ratio for prediction of significant PR were compared by using logistic regression analysis and receiver operating characteristic curve analysis. RESULTS: RVEDVi was significantly higher in male (162.8 mL/m2±50.4) than in female (138.2 mL/m2±37.5) patients (P=.001). Conversely, the RV/LV ratio was similar in both sexes (1.82±0.56 [male] vs 1.69±0.46 [female], P=.13) both in the entire cohort and after excluding patients with significant (≥30 mm Hg) RV outflow tract gradient and/or other residual hemodynamic abnormalities (P=.63). Receiver operating characteristic analysis revealed better discrimination of significant (≥20%) from insignificant (<20%) PR with the use of the RV/LV ratio than with RVEDVi (area under the receiver operating characteristic curve, 0.937 [model 4] vs 0.849 [model 1], P=.01). In multivariate analysis, the only independent predictor of PR fraction was the RV/LV ratio. CONCLUSION: The RV/LV ratio is more accurate than the RVEDVi in differentiation of significant from insignificant PR. After TOF repair, female and male patients have similar RV/LV ratios despite significant differences in RVEDVi between the sexes.

Standardization of T1-mapping in cardiovascular magnetic resonance using clustered structuring for benchmarking normal ranges.

BACKGROUND: Cardiovascular magnetic resonance T1-mapping is increasingly used for tissue characterization, commonly based on Modified Look-Locker Inversion recovery (MOLLI). However, there are numerous MOLLI variants with differing normal ranges. This lack of standardization presents confusion and difficulty in inter-center comparisons, hindering widespread adoption of T1-mapping. METHODS: To address this, we performed a structured literature search for native left ventricular myocardial T1-mapping in healthy humans measured using MOLLI variants at 1.5 and 3 Tesla, across scanner vendors. We then used k-means clustering to structure normal MOLLI-T1 values according to magnetic field strength, and investigated correlations between common imaging parameters: repetition time (TR), echo time (TE), flip angle (FA). RESULTS: We analyzed data from 2207 healthy controls in 76 independent reports. Normal MOLLI-T1 standard deviations varied by 11-fold, and dependencies on TE, TR, and FA differed between 1.5 T and 3 T, thwarting meaningful T1 standardization even within a single field strength, including the use of Z-score. However, divergent MOLLI-T1 norms may be structured using data clustering. For 1.5 T, two clusters emerged: Cluster11.5T: T1 = 958 ± 16 ms (n = 1280); Cluster21.5T: T1 = 1027 ± 19 ms (n = 386). For 3 T, three clusters emerged: Cluster13T: T1 = 1160 ± 21 ms (n = 330); Cluster23T: T1 = 1067 ± 18 ms (n = 178); Cluster33T: T1 = 1227 ± 19 ms (n = 41). We then propose the concept of an online calculator for assigning local norms to a known MOLLI-T1 cluster, allowing benchmarking against published norms. CONCLUSION: Clustered structuring allows T1 standardization of widely-divergent MOLLI variants, benchmarking local norms (usually based on smaller samples) against published norms (larger samples). This may increase confidence and quality control in method implementation, facilitating wider clinical adoption of T1-mapping.

MOCOnet: Robust Motion Correction of Cardiovascular Magnetic Resonance T1 Mapping Using Convolutional Neural Networks.

Background: Quantitative cardiovascular magnetic resonance (CMR) T1 mapping has shown promise for advanced tissue characterisation in routine clinical practise. However, T1 mapping is prone to motion artefacts, which affects its robustness and clinical interpretation. Current methods for motion correction on T1 mapping are model-driven with no guarantee on generalisability, limiting its widespread use. In contrast, emerging data-driven deep learning approaches have shown good performance in general image registration tasks. We propose MOCOnet, a convolutional neural network solution, for generalisable motion artefact correction in T1 maps. Methods: The network architecture employs U-Net for producing distance vector fields and utilises warping layers to apply deformation to the feature maps in a coarse-to-fine manner. Using the UK Biobank imaging dataset scanned at 1.5T, MOCOnet was trained on 1,536 mid-ventricular T1 maps (acquired using the ShMOLLI method) with motion artefacts, generated by a customised deformation procedure, and tested on a different set of 200 samples with a diverse range of motion. MOCOnet was compared to a well-validated baseline multi-modal image registration method. Motion reduction was visually assessed by 3 human experts, with motion scores ranging from 0% (strictly no motion) to 100% (very severe motion). Results: MOCOnet achieved fast image registration (<1 second per T1 map) and successfully suppressed a wide range of motion artefacts. MOCOnet significantly reduced motion scores from 37.1±21.5 to 13.3±10.5 (p < 0.001), whereas the baseline method reduced it to 15.8±15.6 (p < 0.001). MOCOnet was significantly better than the baseline method in suppressing motion artefacts and more consistently (p = 0.007). Conclusion: MOCOnet demonstrated significantly better motion correction performance compared to a traditional image registration approach. Salvaging data affected by motion with robustness and in a time-efficient manner may enable better image quality and reliable images for immediate clinical interpretation.

Poor Bone Quality is Associated With Greater Arterial Stiffness: Insights From the UK Biobank.

Osteoporosis and ischemic heart disease (IHD) represent important public health problems. Existing research suggests an association between the two conditions beyond that attributable to shared risk factors, with a potentially causal relationship. In this study, we tested the association of bone speed of sound (SOS) from quantitative heel ultrasound with (i) measures of arterial compliance from cardiovascular magnetic resonance (aortic distensibility [AD]); (ii) finger photoplethysmography (arterial stiffness index [ASI]); and (iii) incident myocardial infarction and IHD mortality in the UK Biobank cohort. We considered the potential mediating effect of a range of blood biomarkers and cardiometabolic morbidities and evaluated differential relationships by sex, menopause status, smoking, diabetes, and obesity. Furthermore, we considered whether associations with arterial compliance explained association of SOS with ischemic cardiovascular outcomes. Higher SOS was associated with lower arterial compliance by both ASI and AD for both men and women. The relationship was most consistent with ASI, likely relating to larger sample size available for this variable (n = 159,542 versus n = 18,229). There was no clear evidence of differential relationship by menopause, smoking, diabetes, or body mass index (BMI). Blood biomarkers appeared important in mediating the association for both men and women, but with different directions of effect and did not fully explain the observed effects. In fully adjusted models, higher SOS was associated with significantly lower IHD mortality in men, but less robustly in women. The association of SOS with ASI did not explain this observation. In conclusion, our findings support a positive association between bone and vascular health with consistent patterns of association in men and women. The underlying mechanisms are complex and appear to vary by sex. © 2020 The Authors. Journal of Bone and Mineral Research published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research (ASBMR).

Quality assurance of quantitative cardiac T1-mapping in multicenter clinical trials - A T1 phantom program from the hypertrophic cardiomyopathy registry (HCMR) study.

BACKGROUND: Quantitative cardiovascular magnetic resonance T1-mapping is increasingly used for myocardial tissue characterization. However, the lack of standardization limits direct comparability between centers and wider roll-out for clinical use or trials. PURPOSE: To develop a quality assurance (QA) program assuring standardized T1 measurements for clinical use. METHODS: MR phantoms manufactured in 2013 were distributed, including ShMOLLI T1-mapping and reference T1 and T2 protocols. We first studied the T1 and T2 dependency on temperature and phantom aging using phantom datasets from a single site over 4 years. Based on this, we developed a multiparametric QA model, which was then applied to 78 scans from 28 other multi-national sites. RESULTS: T1 temperature sensitivity followed a second-order polynomial to baseline T1 values (R2 > 0.996). Some phantoms showed aging effects, where T1 drifted up to 49% over 40 months. The correlation model based on reference T1 and T2, developed on 1004 dedicated phantom scans, predicted ShMOLLI-T1 with high consistency (coefficient of variation 1.54%), and was robust to temperature variations and phantom aging. Using the 95% confidence interval of the correlation model residuals as the tolerance range, we analyzed 390 ShMOLLI T1-maps and confirmed accurate sequence deployment in 90%(70/78) of QA scans across 28 multiple centers, and categorized the rest with specific remedial actions. CONCLUSIONS: The proposed phantom QA for T1-mapping can assure correct method implementation and protocol adherence, and is robust to temperature variation and phantom aging. This QA program circumvents the need of frequent phantom replacements, and can be readily deployed in multicenter trials.

Standardized image post-processing of cardiovascular magnetic resonance T1-mapping reduces variability and improves accuracy and consistency in myocardial tissue characterization.

BACKGROUND: Myocardial T1-mapping is increasingly used in multicentre studies and trials. Inconsistent image analysis introduces variability, hinders differentiation of diseases, and results in larger sample sizes. We present a systematic approach to standardize T1-map analysis by human operators to improve accuracy and consistency. METHODS: We developed a multi-step training program for T1-map post-processing. The training dataset contained 42 left ventricular (LV) short-axis T1-maps (normal and diseases; 1.5 and 3 Tesla). Contours drawn by two experienced human operators served as reference for myocardial T1 and wall thickness (WT). Trainees (n = 26) underwent training and were evaluated by: (a) qualitative review of contours; (b) quantitative comparison with reference T1 and WT. RESULTS: The mean absolute difference between reference operators was 8.4 ±â€¯6.3 ms (T1) and 1.2 ±â€¯0.7 pixels (WT). Trainees' mean discrepancy from reference in T1 improved significantly post-training (from 8.1 ±â€¯2.4 to 6.7 ±â€¯1.4 ms; p < 0.001), with a 43% reduction in standard deviation (SD) (p = 0.035). WT also improved significantly post-training (from 0.9 ±â€¯0.4 to 0.7 ±â€¯0.2 pixels, p = 0.036), with 47% reduction in SD (p = 0.04). These experimentally-derived thresholds served to guide the training process: T1 (±8 ms) and WT (±1 pixel) from reference. CONCLUSION: A standardized approach to CMR T1-map image post-processing leads to significant improvements in the accuracy and consistency of LV myocardial T1 values and wall thickness. Improving consistency between operators can translate into 33-72% reduction in clinical trial sample-sizes. This work may: (a) serve as a basis for re-certification for core-lab operators; (b) translate to sample-size reductions for clinical studies; (c) produce better-quality training datasets for machine learning.

Deep learning with attention supervision for automated motion artefact detection in quality control of cardiac T1-mapping.

Cardiac magnetic resonance quantitative T1-mapping is increasingly used for advanced myocardial tissue characterisation. However, cardiac or respiratory motion can significantly affect the diagnostic utility of T1-maps, and thus motion artefact detection is critical for quality control and clinically-robust T1 measurements. Manual quality control of T1-maps may provide reassurance, but is laborious and prone to error. We present a deep learning approach with attention supervision for automated motion artefact detection in quality control of cardiac T1-mapping. Firstly, we customised a multi-stream Convolutional Neural Network (CNN) image classifier to streamline the process of automatic motion artefact detection. Secondly, we imposed attention supervision to guide the CNN to focus on targeted myocardial segments. Thirdly, when there was disagreement between the human operator and machine, a second human validator reviewed and rescored the cases for adjudication and to identify the source of disagreement. The multi-stream neural networks demonstrated 89.8% agreement, 87.4% ROC-AUC on motion artefact detection with the human operator in the 2568 T1 maps. Trained with additional supervision on attention, agreements and AUC significantly improved to 91.5% and 89.1%, respectively (p < 0.001). Rescoring of disagreed cases by the second human validator revealed that human operator error was the primary cause of disagreement. Deep learning with attention supervision provides a quick and high-quality assurance of clinical images, and outperforms human operators.

Left ventricular hypertrophy in middle-aged endurance athletes: is it blood pressure related?

Both regular physical activity and hypertension may be related to increased myocardial thickness, but the interplay between these two factors in causing cardiac remodeling in athletes is still a matter of debate. The aim of this study was to analyze the relation between resting and peak exercise blood pressure (BP) and myocardial hypertrophy in healthy middle-aged amateur endurance athletes. The study included 30 male, long-term athletes (mean age 40.9±6.6 years) who underwent resting BP assessment, cardiopulmonary exercise testing with peak exercise BP measurement, and cardiac magnetic resonance. We found that interventricular septal diameter is increased in athletes with high-normal resting BP (n=11, 37%) - median 13 mm (interquartile range: 12-13.75 mm), but not in those with optimal or normal BP (n=19) - median 10 mm (10-11.75 mm), P=0.001. This finding is accompanied by significantly higher left and right ventricular mass index and larger left atrial area in the first group. These differences are even more pronounced in athletes in whom high-normal BP is accompanied by exaggerated blood pressure response (EBPR) to exercise, whereas isolated EBPR to exercise does not lead to hypertrophy or further left atrial enlargement. Prehypertension, isolated or combined with EBPR to exercise, affects cardiac remodeling in athletes. Identification of increased myocardial thickness in pure endurance middle-aged athletes should merit further investigation on masked hypertension.

Cardiovascular magnetic resonance with parametric mapping in long-term ultra-marathon runners.

PURPOSE: There is a direct reverse dose-effect relationship between the amount of physical activity and cardiovascular risk. It is unknown whether this is true for extreme, persistent endurance training. The aim of the study was to assess structural changes of the heart in long-time ultra-marathon runners with special focus on myocardial fibrosis using parametric mapping. METHOD: We studied a group of 30 healthy, male ultra-marathon runners (mean age 40.9 ±â€¯6.6 yrs, median 9 yrs of running with frequent competitions) and 10 matched controls not engaged in any regular activities. All of them underwent cardiovascular magnetic resonance (CMR) with 3 T scanner including T1-mapping, late gadolinium enhancement (LGE) and extracellular volume (ECV) quantification. RESULTS: Athletes demonstrated significantly larger heart chambers and left ventricular (LV) mass. LV systolic function was unchanged. 73.3% of athletes fulfilled volumetric criteria for dilated cardiomyopathy or arrhythmogenic right ventricular cardiomyopathy. Non-ischemic, small volume LGE was found in 8 athletes and in 1 control (27% vs. 10%, p = 0.40). It was localised at insertion points (5 athletes, 1 control) or in the septum or infero-lateral wall (3 athletes). Athletes with insertion point LGE had higher right ventricular end-diastolic volume index in comparison to athletes without LGE (p = 0.04), which suggests its relation to volume overload. There were no differences between athletes and non-athletes in terms of ECV values (26.1% vs. 25%, p = 0.29). CONCLUSIONS: Ultra-marathon runner's hearts demonstrate a high degree of structural remodelling, but there is no significant increase in focal or diffuse myocardial fibrosis.

Automated localization and quality control of the aorta in cine CMR can significantly accelerate processing of the UK Biobank population data.

INTRODUCTION: Aortic distensibility can be calculated using semi-automated methods to segment the aortic lumen on cine CMR (Cardiovascular Magnetic Resonance) images. However, these methods require visual quality control and manual localization of the region of interest (ROI) of ascending (AA) and proximal descending (PDA) aorta, which limit the analysis in large-scale population-based studies. Using 5100 scans from UK Biobank, this study sought to develop and validate a fully automated method to 1) detect and locate the ROIs of AA and PDA, and 2) provide a quality control mechanism. METHODS: The automated AA and PDA detection-localization algorithm followed these steps: 1) foreground segmentation; 2) detection of candidate ROIs by Circular Hough Transform (CHT); 3) spatial, histogram and shape feature extraction for candidate ROIs; 4) AA and PDA detection using Random Forest (RF); 5) quality control based on RF detection probability. To provide the ground truth, overall image quality (IQ = 0-3 from poor to good) and aortic locations were visually assessed by 13 observers. The automated algorithm was trained on 1200 scans and Dice Similarity Coefficient (DSC) was used to calculate the agreement between ground truth and automatically detected ROIs. RESULTS: The automated algorithm was tested on 3900 scans. Detection accuracy was 99.4% for AA and 99.8% for PDA. Aorta localization showed excellent agreement with the ground truth, with DSC ≥ 0.9 in 94.8% of AA (DSC = 0.97 ± 0.04) and 99.5% of PDA cases (DSC = 0.98 ± 0.03). AA×PDA detection probabilities could discriminate scans with IQ ≥ 1 from those severely corrupted by artefacts (AUC = 90.6%). If scans with detection probability < 0.75 were excluded (350 scans), the algorithm was able to correctly detect and localize AA and PDA in all the remaining 3550 scans (100% accuracy). CONCLUSION: The proposed method for automated AA and PDA localization was extremely accurate and the automatically derived detection probabilities provided a robust mechanism to detect low quality scans for further human review. Applying the proposed localization and quality control techniques promises at least a ten-fold reduction in human involvement without sacrificing any accuracy.

Biventricular mechanics in prediction of severe myocardial fibrosis in patients with dilated cardiomyopathy: CMR study.

PURPOSE: The purpose of this study was to compare the ability of various parameters of myocardial mechanics to predict large amounts of biventricular fibrosis assessed via T1 mapping in patients with dilated cardiomyopathy (DCM). MATERIAL: Cardiovascular magnetic resonance feature tracking analysis and T1 mapping were performed in 26 patients with DCM [mean age: 34.4±9.1years, 15 (57.6%) males]. The values of various parameters of myocardial mechanics at predicting advanced left-ventricle (LV) and right-ventricle (RV) fibrosis were compared using logistic regression analysis and receiver operating characteristic curve (ROC) analysis. RESULTS: There were 7 (26.9%) patients with a large amount of LV fibrosis and 9 (34.6%) patients with severe RV fibrosis. ROC curve analysis revealed that the model of combined LV strain rates (AUC=0.902) offered superb ability at predicting large amounts of LV fibrosis. The models including RV strain rates (AUC=0.974), a combination of RV strains, strain rates and clinical parameters (AUC=0.993) as well as the RV radial strain rate alone model (AUC=0.961) yielded outstanding performance in discriminating large and small amounts of RV fibrosis. In multivariate analysis, the LV circumferential strain (LVCR) and RV radial (RVR) strain rate were the only independent predictors of large amounts of LV and RV fibrosis, respectively. CONCLUSIONS: Indices of myocardial deformation, especially combined with clinical features, offered a superlative ability to differentiate high from low degrees of fibrosis in DCM patients. Among all analyzed parameters of myocardial mechanics, LVCR and RVR rate alone were the independent predictors of high degrees of LV and RV fibrosis, respectively.

Native T1-mapping for non-contrast assessment of myocardial fibrosis in patients with hypertrophic cardiomyopathy--comparison with late enhancement quantification.

BACKGROUND: Myocardial fibrosis was shown to influence prognosis in hypertrophic cardiomyopathy (HCM). It is typically assessed by late gadolinium enhancement (LGE) on cardiac magnetic resonance. Native T1-mapping has been proposed, as a contrast-free method of fibrosis assessment. The aim of the study was to define a cut-off value for native T1 relaxation time that best reflects LGE quantification of myocardial fibrosis. METHODS: In 25 patients with HCM and 20 controls we performed T1-mapping pre-contrast using ShMOLLI technique. This was followed by LGE assessment in the studied group 10 minutes after gadolinium contrast injection. Relative myocardial fibrosis size was calculated for varying T1 time thresholds (940-1100 ms) and compared with 6 standard deviations (6SD) method for LGE. RESULTS: Median fibrosis size calculated with T1-mapping was insignificantly different from LGE only for native T1 time threshold of 1060 ms (p = 0.62). Using this threshold, Bland-Altman plots demonstrated very good agreement between fibrosis sizes from the two methods (slightly better only for 1080 ms threshold). For threshold of 1060 ms we also observed good correlation (rho = 0.73) with LGE 6SD method (insignificantly better for lower thresholds, best for threshold of 980 ms-rho = 0.88). In control group with no diagnosis of HCM, fibrosis size <1% was reached for thresholds of 1040 ms and higher. CONCLUSION: Native T1-mapping can be used for non-contrast assessment of myocardial fibrosis in HCM. The 1060 ms threshold of the native T1 relaxation time is characterized by the best balance between agreement and correlation with fibrosis assessed by LGE 6SD method.