Using Deep Learning to Predict Cardiovascular Magnetic Resonance Findings From Echocardiographic Videos.

BACKGROUND

Echocardiography is the most common modality for assessing cardiac structure and function. Although cardiac magnetic resonance (CMR) imaging is less accessible, it can provide unique tissue characterization, including late gadolinium enhancement (LGE), T1 and T2 mapping, and extracellular volume (ECV), which are associated with tissue fibrosis, infiltration, and inflammation. Deep learning has been shown to uncover findings not recognized by clinicians, but it is unknown whether CMR-based tissue characteristics can be derived from echocardiographic videos using deep learning. The aim of this study was to assess the performance of a deep learning model applied to echocardiography to detect CMR-specific parameters, including LGE presence and abnormal T1, T2, or ECV.

METHODS

In a retrospective single-center study, adult patients with CMR and echocardiographic studies within 30 days were included. A video-based convolutional neural network was trained on echocardiographic videos to predict CMR-derived labels, including LGE presence and abnormal T1, T2, or ECV across echocardiographic views. The model was also trained to predict the presence or absence of wall motion abnormality (WMA) as a positive control for model function. The model performance was evaluated in a held-out test data set not used for training.

RESULTS

The study population included 1,453 adult patients (mean age, 56 ± 18 years; 42% women) with 2,556 paired echocardiographic studies occurring at a median of 2 days after CMR (interquartile range, 2 days before to 6 days after). The model had high predictive capability for the presence of WMA (area under the curve [AUC] = 0.873; 95% CI, 0.816-0.922), which was used for positive control. However, the model was unable to reliably detect the presence of LGE (AUC = 0.699; 95% CI, 0.613-0.780) and abnormal native T1 (AUC = 0.614; 95% CI, 0.500-0.715), T2 (AUC = 0.553; 95% CI, 0.420-0.692), or ECV (AUC = 0.564; 95% CI, 0.455-0.691).

CONCLUSIONS

Deep learning applied to echocardiography accurately identified CMR-based WMA but was unable to predict tissue characteristics, suggesting that signal for these tissue characteristics may not be present within ultrasound videos and that the use of CMR for tissue characterization remains essential within cardiology.