Assessment of Artificial Intelligence in Echocardiography Diagnostics in Differentiating Takotsubo Syndrome From Myocardial Infarction

Abstract

Importance: Machine learning algorithms enable the automatic classification of cardiovascular diseases based on raw cardiac ultrasound imaging data. However, the utility of machine learning in distinguishing between takotsubo syndrome (TTS) and acute myocardial infarction (AMI) has not been studied.

Objectives: To assess the utility of machine learning systems for automatic discrimination of TTS and AMI.

Design, settings, and participants: This cohort study included clinical data and transthoracic echocardiogram results of patients with AMI from the Zurich Acute Coronary Syndrome Registry and patients with TTS obtained from 7 cardiovascular centers in the International Takotsubo Registry. Data from the validation cohort were obtained from April 2011 to February 2017. Data from the training cohort were obtained from March 2017 to May 2019. Data were analyzed from September 2019 to June 2021.

Exposure: Transthoracic echocardiograms of 224 patients with TTS and 224 patients with AMI were analyzed.

Main outcomes and measures: Area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity of the machine learning system evaluated on an independent data set and 4 practicing cardiologists for comparison. Echocardiography videos of 228 patients were used in the development and training of a deep learning model. The performance of the automated echocardiogram video analysis method was evaluated on an independent data set consisting of 220 patients. Data were matched according to age, sex, and ST-segment elevation/non-ST-segment elevation (1 patient with AMI for each patient with TTS). Predictions were compared with echocardiographic-based interpretations from 4 practicing cardiologists in terms of sensitivity, specificity, and AUC calculated from confidence scores concerning their binary diagnosis.

Results: In this cohort study, apical 2-chamber and 4-chamber echocardiographic views of 110 patients with TTS (mean [SD] age, 68.4 [12.1] years; 103 [90.4%] were female) and 110 patients with AMI (mean [SD] age, 69.1 [12.2] years; 103 [90.4%] were female) from an independent data set were evaluated. This approach achieved a mean (SD) AUC of 0.79 (0.01) with an overall accuracy of 74.8 (0.7%). In comparison, cardiologists achieved a mean (SD) AUC of 0.71 (0.03) and accuracy of 64.4 (3.5%) on the same data set. In a subanalysis based on 61 patients with apical TTS and 56 patients with AMI due to occlusion of the left anterior descending coronary artery, the model achieved a mean (SD) AUC score of 0.84 (0.01) and an accuracy of 78.6 (1.6%), outperforming the 4 practicing cardiologists (mean [SD] AUC, 0.72 [0.02]) and accuracy of 66.9 (2.8%).

Conclusions and relevance: In this cohort study, a real-time system for fully automated interpretation of echocardiogram videos was established and trained to differentiate TTS from AMI. While this system was more accurate than cardiologists in echocardiography-based disease classification, further studies are warranted for clinical application.

Figure 1.. Overview of the Fully Automated Echocardiographic Video Interpretation

The overall method consists of 3 main components: input processing, feature extraction, and sequence classification. First, each frame of echocardiogram video is semantically segmented (A) and the resulting videos are artificially augmented (B). In the next step, an autoencoder model is trained to reconstruct the segmented masks (C). The trained autoencoder is then used to extract a lower dimensional representation of each frame. A multivariate sequence originating from concatenating such representations is depicted along the time axis (D). In the final step, the sequences are artificially augmented and resampled (E), and a temporal neural network based on 1-dimensional time convolutional architecture is trained to classify the sequences according to takotsubo syndrome and acute myocardial infarction (F). CNN indicates convolutional neural network; 2-ch, 2 chamber; 4-ch, 4 chamber.

Figure 2.. Performance of the Machine Learning Algorithm Compared With Cardiologists

Receiver operating characteristic (ROC) curve (TTS [takotsubo syndrome] considered as the positive class) of the algorithm plotted against the ROC of the 4 different readers. The left of each panel compares the algorithm performance against individual readers. The right of each panel compares the performance with an expert voting committee consisting of 3 of 4 readers. In panel A, the performance of the machine learning algorithm is evaluated on the whole test data set. In panel B, the algorithm is evaluated on a subset of patients with common subtypes, ie, apical TTS and LAD-AMI. AMI indicates acute myocardial infarction, AUC, area under the curve; LAD-AMI, left anterior descending AMI; R, reader.

Figure 3.. Interpretability Analysis (Apical TTS vs LAD-AMI)

A, The individual latent sequences extracted from a video track the movement of different parts of the myocardium. The colored parts in the ultrasound image correspond to the area on which the latent sequences that performed best in terms of area under the curve (AUC) predominantly focused. The analysis is done separately for the 2-chamber and 4-chamber view. B, The ROC curve based on the 2 different chamber views using all sequences is presented. The SD is calculated based on 5 different training and classification runs of the temporal neural network. LAD-AMI indicates left anterior descending acute myocardial infarction; TTS, takotsubo syndrome; 2-ch, 2 chamber; 4-ch, 4 chamber.