Deep Learning Analysis of Cardiac MRI in Legacy Datasets: Multi-Ethnic Study of Atherosclerosis

An artificial intelligence tool for automated analysis of large-scale unstructured clinical cine cardiac magnetic resonance databases

Aims

Artificial intelligence (AI) techniques have been proposed for automating analysis of short-axis (SAX) cine cardiac magnetic resonance (CMR), but no CMR analysis tool exists to automatically analyse large (unstructured) clinical CMR datasets. We develop and validate a robust AI tool for start-to-end automatic quantification of cardiac function from SAX cine CMR in large clinical databases.

Methods and results

Our pipeline for processing and analysing CMR databases includes automated steps to identify the correct data, robust image pre-processing, an AI algorithm for biventricular segmentation of SAX CMR and estimation of functional biomarkers, and automated post-analysis quality control to detect and correct errors. The segmentation algorithm was trained on 2793 CMR scans from two NHS hospitals and validated on additional cases from this dataset (n = 414) and five external datasets (n = 6888), including scans of patients with a range of diseases acquired at 12 different centres using CMR scanners from all major vendors. Median absolute errors in cardiac biomarkers were within the range of inter-observer variability: <8.4 mL (left ventricle volume), <9.2 mL (right ventricle volume), <13.3 g (left ventricular mass), and <5.9% (ejection fraction) across all datasets. Stratification of cases according to phenotypes of cardiac disease and scanner vendors showed good performance across all groups.

Conclusion

We show that our proposed tool, which combines image pre-processing steps, a domain-generalizable AI algorithm trained on a large-scale multi-domain CMR dataset and quality control steps, allows robust analysis of (clinical or research) databases from multiple centres, vendors, and cardiac diseases. This enables translation of our tool for use in fully automated processing of large multi-centre databases.

Multilevel comparison of deep learning models for function quantification in cardiovascular magnetic resonance: On the redundancy of architectural variations

Background

Cardiac function quantification in cardiovascular magnetic resonance requires precise contouring of the heart chambers. This time-consuming task is increasingly being addressed by a plethora of ever more complex deep learning methods. However, only a small fraction of these have made their way from academia into clinical practice. In the quality assessment and control of medical artificial intelligence, the opaque reasoning and associated distinctive errors of neural networks meet an extraordinarily low tolerance for failure.

Aim

The aim of this study is a multilevel analysis and comparison of the performance of three popular convolutional neural network (CNN) models for cardiac function quantification.

Methods

U-Net, FCN, and MultiResUNet were trained for the segmentation of the left and right ventricles on short-axis cine images of 119 patients from clinical routine. The training pipeline and hyperparameters were kept constant to isolate the influence of network architecture. CNN performance was evaluated against expert segmentations for 29 test cases on contour level and in terms of quantitative clinical parameters. Multilevel analysis included breakdown of results by slice position, as well as visualization of segmentation deviations and linkage of volume differences to segmentation metrics via correlation plots for qualitative analysis.

Results

All models showed strong correlation to the expert with respect to quantitative clinical parameters (rz ' = 0.978, 0.977, 0.978 for U-Net, FCN, MultiResUNet respectively). The MultiResUNet significantly underestimated ventricular volumes and left ventricular myocardial mass. Segmentation difficulties and failures clustered in basal and apical slices for all CNNs, with the largest volume differences in the basal slices (mean absolute error per slice: 4.2 ± 4.5 ml for basal, 0.9 ± 1.3 ml for midventricular, 0.9 ± 0.9 ml for apical slices). Results for the right ventricle had higher variance and more outliers compared to the left ventricle. Intraclass correlation for clinical parameters was excellent (≥0.91) among the CNNs.

Conclusion

Modifications to CNN architecture were not critical to the quality of error for our dataset. Despite good overall agreement with the expert, errors accumulated in basal and apical slices for all models.

A deep learning approach for fully automated cardiac shape modeling in tetralogy of Fallot

BACKGROUND

Cardiac shape modeling is a useful computational tool that has provided quantitative insights into the mechanisms underlying dysfunction in heart disease. The manual input and time required to make cardiac shape models, however, limits their clinical utility. Here we present an end-to-end pipeline that uses deep learning for automated view classification, slice selection, phase selection, anatomical landmark localization, and myocardial image segmentation for the automated generation of three-dimensional, biventricular shape models. With this approach, we aim to make cardiac shape modeling a more robust and broadly applicable tool that has processing times consistent with clinical workflows.

METHODS

Cardiovascular magnetic resonance (CMR) images from a cohort of 123 patients with repaired tetralogy of Fallot (rTOF) from two internal sites were used to train and validate each step in the automated pipeline. The complete automated pipeline was tested using CMR images from a cohort of 12 rTOF patients from an internal site and 18 rTOF patients from an external site. Manually and automatically generated shape models from the test set were compared using Euclidean projection distances, global ventricular measurements, and atlas-based shape mode scores.

RESULTS

The mean absolute error (MAE) between manually and automatically generated shape models in the test set was similar to the voxel resolution of the original CMR images for end-diastolic models (MAE = 1.9 ± 0.5 mm) and end-systolic models (MAE = 2.1 ± 0.7 mm). Global ventricular measurements computed from automated models were in good agreement with those computed from manual models. The average mean absolute difference in shape mode Z-score between manually and automatically generated models was 0.5 standard deviations for the first 20 modes of a reference statistical shape atlas.

CONCLUSIONS

Using deep learning, accurate three-dimensional, biventricular shape models can be reliably created. This fully automated end-to-end approach dramatically reduces the manual input required to create shape models, thereby enabling the rapid analysis of large-scale datasets and the potential to deploy statistical atlas-based analyses in point-of-care clinical settings. Training data and networks are available from cardiacatlas.org.

Multi-Ethnic Study of Atherosclerosis: Relationship between Left Ventricular Shape at Cardiac MRI and 10-year Outcomes

Background Left ventricular (LV) subclinical remodeling is associated with adverse outcomes and indicates mechanisms of disease development. Standard metrics such as LV mass and volumes may not capture the full range of remodeling. Purpose To quantify the relationship between LV three-dimensional shape at MRI and incident cardiovascular events over 10 years. Materials and Methods In this retrospective study, 5098 participants from the Multi-Ethnic Study of Atherosclerosis who were free of clinical cardiovascular disease underwent cardiac MRI from 2000 to 2002. LV shape models were automatically generated using a machine learning workflow. Event-specific remodeling signatures were computed using partial least squares regression, and random survival forests were used to determine which features were most associated with incident heart failure (HF), coronary heart disease (CHD), and cardiovascular disease (CVD) events over a 10-year follow-up period. The discrimination improvement of adding LV shape to traditional cardiovascular risk factors, coronary artery calcium scores, and N-terminal pro-brain natriuretic peptide levels was assessed using the index of prediction accuracy and time-dependent area under the receiver operating characteristic curve (AUC). Kaplan-Meier survival curves were used to illustrate the ability of remodeling signatures to predict the end points. Results Overall, 4618 participants had sufficient three-dimensional MRI information to generate patient-specific LV models (mean age, 60.6 years ± 9.9 [SD]; 2540 women). Among these participants, 147 had HF, 317 had CHD, and 455 had CVD events. The addition of LV remodeling signatures to traditional cardiovascular risk factors improved the mean AUC for 10-year survival prediction and achieved better performance than LV mass and volumes; HF (AUC, 0.83 ± 0.01 and 0.81 ± 0.01, respectively; P < .05), CHD (AUC, 0.77 ± 0.01 and 0.75 ± 0.01, respectively; P < .05), and CVD (AUC, 0.78 ± 0.0 and 0.76 ± 0.0, respectively; P < .05). Kaplan-Meier analysis demonstrated that participants with high-risk HF remodeling signatures had a 10-year survival rate of 56% compared with 95% for those with low-risk scores. Conclusion Left ventricular event-specific remodeling signatures were more predictive of heart failure, coronary heart disease, and cardiovascular disease events over 10 years than standard mass and volume measures and enable an automatic personalized medicine approach to tracking remodeling. © RSNA, 2022 Online supplemental material is available for this article.

Le Cœur en Sabot: shape associations with adverse events in repaired tetralogy of Fallot

BACKGROUND

Maladaptive remodelling mechanisms occur in patients with repaired tetralogy of Fallot (rToF) resulting in a cycle of metabolic and structural changes. Biventricular shape analysis may indicate mechanisms associated with adverse events independent of pulmonary regurgitant volume index (PRVI). We aimed to determine novel remodelling patterns associated with adverse events in patients with rToF using shape and function analysis.

METHODS

Biventricular shape and function were studied in 192 patients with rToF (median time from TOF repair to baseline evaluation 13.5 years). Linear discriminant analysis (LDA) and principal component analysis (PCA) were used to identify shape differences between patients with and without adverse events. Adverse events included death, arrhythmias, and cardiac arrest with median follow-up of 10 years.

RESULTS

LDA and PCA showed that shape characteristics pertaining to adverse events included a more circular left ventricle (LV) (decreased eccentricity), dilated (increased sphericity) LV base, increased right ventricular (RV) apical sphericity, and decreased RV basal sphericity. Multivariate LDA showed that the optimal discriminative model included only RV apical ejection fraction and one PCA mode associated with a more circular and dilated LV base (AUC = 0.77). PRVI did not add value, and shape changes associated with increased PRVI were not predictive of adverse outcomes.

CONCLUSION

Pathological remodelling patterns in patients with rToF are significantly associated with adverse events, independent of PRVI. Mechanisms related to incident events include LV basal dilation with a reduced RV apical ejection fraction.