scholarly journals Automated myocardial segmentation in native t1-mapping cardiovascular magnetic resonance images based on machine learning: a validation study in the UK biobank"s covid-19 subset

2021 ◽  
Vol 22 (Supplement_2) ◽  
Author(s):  
E Rauseo ◽  
L Lockhart ◽  
JM Paiva ◽  
K Fung ◽  
MY Khanji ◽  
...  
Keyword(s):  
Machine Learning ◽  
Quality Control ◽  
Cardiovascular Magnetic Resonance ◽  
Magnetic Resonance ◽  
T1 Mapping ◽  
Automated Analysis ◽  
Uk Biobank ◽  
Native T1 ◽  
The Mean ◽  
Good Agreement

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Innovate UK Background  Regional assessment of septal native T1 values with cardiovascular magnetic resonance (CMR) is used to characterise diffuse myocardial diseases. Previous studies suggest its potential role in detecting early pathological alterations, which may help identify high-risk subjects at early disease stages. Automated analysis of myocardial native T1 images may enable faster CMR analysis and reduce inter-observer variability of manual analysis. However, the technical performance of such methodologies has not been previously reported. Purpose  We tested, in a subset of UK Biobank participants, the degree of agreement between CMR septal myocardial T1 values obtained from our machine learning (ML) algorithm and septal native T1 values computed from manual segmentations. Methods  We analysed the first 292 participants who were tested for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and had CMR imaging (1.5 Tesla, Siemens MAGNETOM Aera). T1 mapping was performed in a single mid-ventricular short axis (SAX) slice using ShMOLLI (WIP780B) sequences. Three experienced CMR readers independently measured native T1 values by manually placing a single region of interest (ROI) covering half of the anteroseptal and half of the inferoseptal wall using cvi42 post-processing software (version 5.11). A mean T1 value for each participant was then calculated. A ML algorithm developed by Circle Cardiovascular Imaging Inc. was then applied to the same images to derive the myocardium T1 values automatically. The algorithm was previously trained to segment myocardium from SAX T1 and non-T1 mapping images on two external CMR datasets. We compared the mean septal ROI T1 values to the mean myocardium T1 values predicted by the ML algorithm. Results  Two studies were excluded after quality control. The ML-derived and the manually calculated mean T1 values were significantly correlated (r = 0.82, p < 0.001). The Bland-Altman analysis between the two methods showed a mean bias of 3.64 ms, with 95% limits of agreement of −38.88 to 53.46 ms, indicating good agreement (figure 1). Conclusions  We demonstrated strong correlation and good agreement between native T1 values obtained from our automated analysis method and manual T1 septal analysis in a subset of UK Biobank participants. This algorithm may represent a valuable tool for clinicians allowing for fast and potentially less operator-dependent myocardial tissue characterisation. However, validation of more extensive datasets and quality control processes are needed.

scholarly journals Global, segmental and layer specific analysis of myocardial involvement in Duchenne muscular dystrophy by cardiovascular magnetic resonance native T1 mapping

2021 ◽  
Vol 23 (1) ◽  
Author(s):  
Ke Xu ◽  
Hua-yan Xu ◽  
Rong Xu ◽  
Lin-jun Xie ◽  
Zhi-gang Yang ◽  
...  
Keyword(s):  
Cardiovascular Magnetic Resonance ◽  
Duchenne Muscular Dystrophy ◽  
Magnetic Resonance ◽  
Muscular Dystrophy ◽  
T1 Mapping ◽  
Healthy Controls ◽  
Native T1 ◽  
Specific Analysis ◽  
Epicardial Layer ◽  
Myocardial Involvement

Abstract Background Progressive cardiomyopathy accounts for almost all mortality among Duchenne muscular dystrophy (DMD) patients.‍ Thus, our aim was to comprehensively characterize myocardial involvement by investigating the heterogeneity of native T1 mapping in DMD patients using global and regional (including segmental and layer-specific) analysis across a large cohort. Methods We prospectively enrolled 99 DMD patients (8.8 ± 2.5 years) and 25 matched male healthy controls (9.5 ± 2.5 years). All subjects underwent cardiovascular magnetic resonance (CMR) with cine, T1 mapping and late gadolinium enhancement (LGE) sequences. Native T1 values based on the global and regional myocardium were measured, and LGE was defined. Results LGE was present in 49 (49%) DMD patients. Global native T1 values were significantly longer in LGE-positive (LGE +) patients than in healthy controls, both in basal slices (1304 ± 55 vs. 1246 ± 27 ms, p < 0.001) and in mid-level slices (1305 ± 57 vs. 1245 ± 37 ms, p < 0.001). No significant difference in global native T1 was found between healthy controls and LGE-negative (LGE−) patients. In segmental analysis, LGE + patients had significantly increased native T1 in all analyzed segments compared to the healthy control group. Meanwhile, the comparison between LGE− patients and healthy controls showed significantly elevated values only in the basal anterolateral segment (1273 ± 62 vs. 1234 ± 40 ms, p = 0.034). Interestingly, the epicardial layer had a significantly higher native T1 in LGE− patients than in healthy controls (p < 0.05), whereas no such pattern was noticed in the global myocardium. Epicardial layer native T1 resulted in the highest diagnostic performance for distinguishing between healthy controls and DMD patients in receiver operating curve analyses (area under the curve [AUC] 0.84 for basal level and 0.85 for middle level) when compared to global native T1 and endocardial layer native T1. Conclusions Myocardial regional native T1, particularly epicardial native T1, seems to have potential as a novel robust marker of very early cardiac involvement in DMD patients. Trial registration: Chinese Clinical Trial Registry (http://www.chictr.org.cn/index.aspx) ChiCTR1800018340, 09/12/2018, Retrospectively registered.

scholarly journals The reliability and feasibility of non-contrast adenosine stress cardiovascular magnetic resonance T1 mapping in patients on haemodialysis

2020 ◽  
Vol 22 (1) ◽  
Author(s):  
Federica E Poli ◽  
Gaurav S Gulsin ◽  
Daniel S March ◽  
Ahmed MSEK Abdelaty ◽  
Kelly S Parke ◽  
...  
Keyword(s):  
Cardiovascular Magnetic Resonance ◽  
Magnetic Resonance ◽  
T1 Mapping ◽  
Intraclass Correlation ◽  
Adenosine Stress ◽  
Single Shot ◽  
Retest Reliability ◽  
Native T1 ◽  
Dialysis Vintage ◽  
Test Retest Reliability

Abstract Background Identifying coronary artery disease (CAD) in patients with end-stage renal disease (ESRD) is challenging. Adenosine stress native T1 mapping with cardiovascular magnetic resonance (CMR) may accurately detect obstructive CAD and microvascular dysfunction in the general population. This study assessed the feasibility and reliability of adenosine stress native T1 mapping in patients on haemodialysis. Methods The feasibility of undertaking rest and adenosine stress native T1 mapping using the single-shot Modified Look-Locker inversion recovery (MOLLI) sequence was assessed in 58 patients on maintenance haemodialysis using 3 T CMR. Ten patients underwent repeat stress CMR within 2 weeks for assessment of test-retest reliability of native T1, stress T1 and delta T1 (ΔT1). Interrater and intrarater agreement were assessed in 10 patients. Exploratory analyses were undertaken to assess associations between clinical variables and native T1 values in 51 patients on haemodialysis. Results Mean age of participants was 55 ± 15 years, 46 (79%) were male, and median dialysis vintage was 21 (8; 48) months. All patients completed the scan without complications. Mean native T1 rest, stress and ΔT1 were 1261 ± 57 ms, 1297 ± 50 ms and 2.9 ± 2.5%, respectively. Interrater and intrarater agreement of rest T1, stress T1 and ΔT1 were excellent, with intraclass correlation coefficients (ICC) > 0.9 for all. Test-retest reliability of rest and stress native T1 were excellent or good (CoV 1.2 and 1.5%; ICC, 0.79 and 0.69, respectively). Test-retest reliability of ΔT1 was moderate to poor (CoV 27.4%, ICC 0.55). On multivariate analysis, CAD, diabetes mellitus and resting native T1 time were independent determinants of ΔT1 (β = − 0.275, p = 0.019; β = − 0.297, p = 0.013; β = − 0.455; p < 0.001, respectively). Conclusions Rest and adenosine stress native T1 mapping is feasible and well-tolerated amongst patients with ESRD on haemodialysis. Although rater agreement of the technique is excellent, test-retest reliability of ΔT1 is moderate to poor. Prospective studies should evaluate the relationship between this technique and established methods of CAD assessment and association with outcomes.

Abstract 14877: Native T1-based Cardiovascular Magnetic Resonance for Characterizing Chronic Myocardial Infarctions: Proof-of-Concept Study in Canines

Circulation ◽  
2014 ◽  
Vol 130 (suppl_2) ◽  
Author(s):  
Avinash Kali ◽  
Ivan Cokic ◽  
Richard L Tang ◽  
Hsin J Yang ◽  
Behzad Sharif ◽  
...  
Keyword(s):  
Cardiovascular Magnetic Resonance ◽  
Magnetic Resonance ◽  
Infarct Size ◽  
T1 Mapping ◽  
Ischemia Reperfusion Injury ◽  
Ex Vivo ◽  
Ischemia Reperfusion ◽  
Infarcted Myocardium ◽  
Native T1 ◽  
Replacement Fibrosis

Introduction: Gadolinium infusion required for Late Gadolinium Enhancement (LGE) Cardiovascular Magnetic Resonance (CMR) imaging is contraindicated in nearly 20% of myocardial infarction (MI) patients due to chronic end-stage kidney disease. Hypothesis: Using a canine model of MI, we investigated whether native T1 mapping at 3T could be an alternative to LGE CMR for characterizing chronic MIs (CMIs). Methods: Canines (n=29) were subjected to ischemia-reperfusion injury. Native T1 maps, native T2 maps and LGE images were acquired at 7 days (acute, AMI) and 4 months (CMI) post-MI at 1.5T and 3T. Infarct location, size and transmurality, measured using Mean + 5 Standard Deviations criterion, were compared between T1 maps and LGE images. Native T2 maps were used to examine the resolution of edema between AMI and CMI. Following the CMR studies, animals were euthanized and ex-vivo histology was performed. Results: T1 maps and LGE images were not different for measuring infarct size (p=0.61) and transmurality (p=0.81) in CMI at 3T. In AMI at 3T, T1 maps overestimated both infarct size (p=0.007) and transmurality (p=0.007) relative to LGE images. At 1.5T, T1 maps underestimated both infarct size and transmurality relative to LGE images in both AMI and CMI (p<0.001 for all cases). Relative to the remote territories, T1 of the infarcted myocardium was elevated in AMI (3T: p<0.001; 1.5T: p<0.001) and CMI (3T: p<0.001; 1.5T: p=0.037). T2 of the infarcted myocardium was elevated in AMI (p<0.001 at both 3T and 1.5T), but not in CMI (3T: p=0.19, 1.5T: p=0.55) indicating that myocardial edema resolved by 4 months post-MI. Masson’s trichrome staining showed extensive replacement fibrosis within CMIs. Sensitivity and specificity of T1 maps to detect CMI were 95% and 97% respectively at 3T, and 58% and 78% respectively at 1.5T. Conclusions: Native T1 mapping at 3T can characterize CMIs with high diagnostic accuracy. T1 elongations in CMI appear to arise predominantly from replacement fibrosis.

scholarly journals Non-contrast cardiovascular magnetic resonance detection of myocardial fibrosis in Duchenne muscular dystrophy

2021 ◽  
Vol 23 (1) ◽  
Author(s):  
Frank J. Raucci ◽  
Meng Xu ◽  
Kristen George-Durrett ◽  
Kimberly Crum ◽  
James C. Slaughter ◽  
...  
Keyword(s):  
Cardiovascular Magnetic Resonance ◽  
Duchenne Muscular Dystrophy ◽  
Magnetic Resonance ◽  
Muscular Dystrophy ◽  
Myocardial Fibrosis ◽  
Severity Score ◽  
T1 Mapping ◽  
Left Ventricular ◽  
Left Ventricular Ejection ◽  
Native T1

Abstract Background Duchenne muscular dystrophy (DMD) leads to progressive cardiomyopathy. Detection of myocardial fibrosis with late gadolinium enhancement (LGE) by cardiovascular magnetic resonance (CMR) is critical for clinical management. Due to concerns of brain deposition of gadolinium, non-contrast methods for detecting and monitoring myocardial fibrosis would be beneficial. Objectives We hypothesized that native T1 mapping and/or circumferential (εcc) and longitudinal (εls) strain can detect myocardial fibrosis. Methods 156 CMRs with gadolinium were performed in 66 DMD boys and included: (1) left ventricular ejection fraction (LVEF), (2) LGE, (3) native T1 mapping and myocardial tagging (εcc-tag measured using harmonic phase analysis). LGE was graded as: (1) presence/absence by segment, slice, and globally; (2) global severity from 0 (no LGE) to 4 (severe); (3) percent LGE using full width half maximum (FWHM). εls and εcc measured using feature tracking. Regression models to predict LGE included native T1 and either εcc-tag or εls and εcc measured at each segment, slice, and globally. Results Mean age and LVEF at first CMR were 14 years and 54%, respectively. Global εls and εcc strongly predicted presence or absence of LGE (OR 2.6 [1.1, 6.0], p = 0.029, and OR 2.3 [1.0, 5.1], p = 0.049, respectively) while global native T1 did not. Global εcc, εls, and native T1 predicted global severity score (OR 2.6 [1.4, 4.8], p = 0.002, OR 2.6 [1.4, 6.0], p = 0.002, and OR 1.8 [1.1, 3.1], p = 0.025, respectively). εls correlated with change in LGE by severity score (n = 33, 3.8 [1.0, 14.2], p = 0.048) and εcc-tag correlated with change in percent LGE by FWHM (n = 34, OR 0.2 [0.1, 0.9], p = 0.01). Conclusions Pre-contrast sequences predict presence and severity of LGE, with εls and εcc being more predictive in most models, but there was not an observable advantage over using LVEF as a predictor. Change in LGE was predicted by εls (global severity score) and εcc-tag (FWHM). While statistically significant, our results suggest these sequences are currently not a replacement for LGE and may only have utility in a very limited subset of DMD patients.

scholarly journals The dynamics of extracellular gadolinium-based contrast agent excretion into pleural and pericardial effusions quantified by T1 mapping cardiovascular magnetic resonance

2019 ◽  
Vol 21 (1) ◽  
Author(s):  
Simon Thalén ◽  
Maren Maanja ◽  
Andreas Sigfridsson ◽  
Eva Maret ◽  
Peder Sörensson ◽  
...  
Keyword(s):  
Cardiovascular Magnetic Resonance ◽  
Magnetic Resonance ◽  
T1 Mapping ◽  
Rare Occurrence ◽  
Pleural Effusions ◽  
Contrast Injection ◽  
Fluid Content ◽  
Native T1 ◽  
Pericardial Effusions ◽  
Vicarious Excretion

Abstract Introduction Excretion of cardiovascular magnetic resonance (CMR) extracellular gadolinium-based contrast agents (GBCA) into pleural and pericardial effusions, sometimes referred to as vicarious excretion, has been described as a rare occurrence using T1-weighted imaging. However, the T1 mapping characteristics as well as presence, magnitude and dynamics of contrast excretion into these effusions is not known. Aims To investigate and compare the differences in T1 mapping characteristics and extracellular GBCA excretion dynamics in pleural and pericardial effusions. Methods Clinically referred patients with a pericardial and/or pleural effusion underwent CMR T1 mapping at 1.5 T before, and at 3 (early) and at 27 (late) minutes after administration of an extracellular GBCA (0.2 mmol/kg, gadoteric acid). Analyzed effusion characteristics were native T1, ΔR1 early and late after contrast injection, and the effusion-volume-independent early-to-late contrast concentration ratio ΔR1early/ΔR1late, where ΔR1 = 1/T1post-contrast - 1/T1native. Results Native T1 was lower in pericardial effusions (n = 69) than in pleural effusions (n = 54) (median [interquartile range], 2912 [2567–3152] vs 3148 [2692–3494] ms, p = 0.005). Pericardial and pleural effusions did not differ with regards to ΔR1early (0.05 [0.03–0.10] vs 0.07 [0.03–0.12] s− 1, p = 0.38). Compared to pleural effusions, pericardial effusions had a higher ΔR1late (0.8 [0.6–1.2] vs 0.4 [0.2–0.6] s− 1, p < 0.001) and ΔR1early/ΔR1late (0.19 [0.08–0.30] vs 0.12 [0.04–0.19], p < 0.001). Conclusions T1 mapping shows that extracellular GBCA is excreted into pericardial and pleural effusions. Consequently, the previously used term vicarious excretion is misleading. Compared to pleural effusions, pericardial effusions had both a lower native T1, consistent with lesser relative fluid content in relation to other components such as proteins, and more prominent early excretion dynamics, which could be related to inflammation. The clinical diagnostic utility of T1 mapping to determine quantitative contrast dynamics in pericardial and pleural effusions merits further investigation.

scholarly journals Aortic elasticity after aortic coarctation relief: comparison of surgical and interventional therapy by cardiovascular magnetic resonance imaging

2019 ◽  
Vol 19 (1) ◽  
Author(s):  
Theresa Pieper ◽  
Heiner Latus ◽  
Dietmar Schranz ◽  
Joachim Kreuder ◽  
Bettina Reich ◽  
...  
Keyword(s):  
Cardiovascular Magnetic Resonance ◽  
Magnetic Resonance ◽  
Stent Implantation ◽  
Aortic Coarctation ◽  
T1 Mapping ◽  
Aortic Distensibility ◽  
Post Contrast ◽  
Aortic Elasticity ◽  
Native T1 ◽  
Significant Difference

Abstract Background Patients after aortic coarctation (CoA) repair show impaired aortic bioelasticity and altered left ventricular (LV) mechanics, predisposing diastolic dysfunction. Our purpose was to assess aortic bioelasticity and LV properties in CoA patients who underwent endovascular stenting or surgery using cardiovascular magnetic resonance (CMR) imaging. Methods Fifty CoA patients (20.5 ± 9.5 years) were examined by 3-Tesla CMR. Eighteen patients had previous stent implantation and 32 had surgical repair. We performed volumetric analysis of both ventricles (LV, RV) and left atrium (LA) to measure biventricular volumes, ejection fractions, left atrial (LA) volumes, and functional parameters (LAEFPassive, LAEFContractile, LAEFReservoir). Aortic distensibility and pulse wave velocity (PWV) were assessed. Native T1 mapping was applied to examine LV tissue properties. In twelve patients post-contrast T1 mapping was performed. Results LV, RV and LA parameters did not differ between the surgical and stent group. There was also no significant difference for aortic distensibility, PWV and T1 relaxation times. Aortic root distensibility correlated negatively with age, BMI, BSA and weight (p < 0.001). Native T1 values correlated negatively with age, weight, BSA and BMI (p < 0.001). Lower post-contrast T1 values were associated with lower aortic arch distensibility and higher aortic arch PWV (p < 0.001). Conclusions CoA patients after surgery or stent implantation did not show significant difference of aortic elasticity. Thus, presumably other factors like intrinsic aortic abnormalities might have a greater impact on aortic elasticity than the approach of repair. Interestingly, our data suggest that native T1 values are influenced by demographic characteristics.

scholarly journals Automated Quality-Controlled Cardiovascular Magnetic Resonance Pericardial Fat Quantification Using a Convolutional Neural Network in the UK Biobank

2021 ◽  
Vol 8 ◽  
Author(s):  
Andrew Bard ◽  
Zahra Raisi-Estabragh ◽  
Maddalena Ardissino ◽  
Aaron Mark Lee ◽  
Francesca Pugliese ◽  
...  
Keyword(s):  
Neural Network ◽  
Quality Control ◽  
Cardiovascular Magnetic Resonance ◽  
Magnetic Resonance ◽  
Model Performance ◽  
Uk Biobank ◽  
Test Set ◽  
Automated Method ◽  
Pericardial Adipose Tissue ◽  
The Uk

Background: Pericardial adipose tissue (PAT) may represent a novel risk marker for cardiovascular disease. However, absence of rapid radiation-free PAT quantification methods has precluded its examination in large cohorts.Objectives: We developed a fully automated quality-controlled tool for cardiovascular magnetic resonance (CMR) PAT quantification in the UK Biobank (UKB).Methods: Image analysis comprised contouring an en-bloc PAT area on four-chamber cine images. We created a ground truth manual analysis dataset randomly split into training and test sets. We built a neural network for automated segmentation using a Multi-residual U-net architecture with incorporation of permanently active dropout layers to facilitate quality control of the model's output using Monte Carlo sampling. We developed an in-built quality control feature, which presents predicted Dice scores. We evaluated model performance against the test set (n = 87), the whole UKB Imaging cohort (n = 45,519), and an external dataset (n = 103). In an independent dataset, we compared automated CMR and cardiac computed tomography (CCT) PAT quantification. Finally, we tested association of CMR PAT with diabetes in the UKB (n = 42,928).Results: Agreement between automated and manual segmentations in the test set was almost identical to inter-observer variability (mean Dice score = 0.8). The quality control method predicted individual Dice scores with Pearson r = 0.75. Model performance remained high in the whole UKB Imaging cohort and in the external dataset, with medium–good quality segmentation in 94.3% (mean Dice score = 0.77) and 94.4% (mean Dice score = 0.78), respectively. There was high correlation between CMR and CCT PAT measures (Pearson r = 0.72, p-value 5.3 ×10−18). Larger CMR PAT area was associated with significantly greater odds of diabetes independent of age, sex, and body mass index.Conclusions: We present a novel fully automated method for CMR PAT quantification with good model performance on independent and external datasets, high correlation with reference standard CCT PAT measurement, and expected clinical associations with diabetes.

Sign in / Sign up

Export Citation Format

Share Document