Regional left ventricle scar detection from routine cardiac computed tomography angiograms using latent space classification

• Published in: Computers in biology and medicine 2022-11, Vol.150, p.106191-106191, Article 106191
• Main Authors: O’Brien, Hugh, Whitaker, John, O’Neill, Mark D., Grigoryan, Karine, Gill, Harminder, Mehta, Vishal, Elliot, Mark K., Rinaldi, Christopher Aldo, Morgan, Holly, Perera, Divaka, Taylor, Jonathan, Rajani, Ronak, Rhode, Kawal, Niederer, Steven
• Format: Article
• Language: English
• Subjects: Angiography; Cardiology; Cicatrix - diagnostic imaging; Classification; Coders; Computed tomography; Contraindications; Contrast Media; Gadolinium; Heart; Heart Ventricles - diagnostic imaging; Humans; Machine learning; Magnetic resonance imaging; Magnetic Resonance Imaging - methods; Medical imaging; Networks; Patients; Regional development; Scar detection; Tomography; Ventricle
• ISSN: 0010-4825; 1879-0534; 1879-0534
• DOI: 10.1016/j.compbiomed.2022.106191

The aim of this study is to develop an automated method of regional scar detection on clinically standard computed tomography angiography (CTA) using encoder–decoder networks with latent space classification. Localising scar in cardiac patients can assist in diagnosis and guide interventions. Magnetic resonance imaging (MRI) with late gadolinium enhancement (LGE) is the clinical gold standard for scar imaging; however, it is commonly contraindicated. CTA is an alternative imaging modality that has fewer contraindications and is widely used as a first-line imaging modality of cardiac applications. A dataset of 79 patients with both clinically indicated MRI LGE and subsequent CTA scans was used to train and validate networks to classify septal and lateral scar presence within short axis left ventricle slices. Two designs of encoder–decoder networks were compared, with one encoding anatomical shape in the latent space. Ground truth was established by segmenting scar in MRI LGE and registering this to the CTA images. Short axis slices were taken from the CTA, which served as the input to the networks. An independent external set of 22 cases (27% the size of the cross-validation set) was used to test the best network. A network classifying lateral scar only achieved an area under ROC curve of 0.75, with a sensitivity of 0.79 and specificity of 0.62 on the independent test set. The results of septal scar classification were poor (AUC < 0.6) for all networks. This was likely due to a high class imbalance. The highest AUC network encoded anatomical shape information in the network latent space, indicating it was important for the successful classification of lateral scar. Automatic lateral wall scar detection can be performed from a routine cardiac CTA with reasonable accuracy, without any scar specific imaging. This requires only a single acquisition in the cardiac cycle. In a clinical setting, this could be useful for pre-procedure planning, especially where MRI is contraindicated. Further work with more septal scar present is warranted to improve the usefulness of this approach. •Localised scar detection in cardiac CT angiograms using convolutional neural networks.•No additional sequences or contrast beyond standard clinical scans.•Ground truth established with paired contrast enhanced MRI.•Lateral wall scar was classified well in an independent test set.