Detection of atrial septal aneurysm on ECG based on Deep Learning algorithm (ANN).

An Artificial Neural Network trained on ECG data achieved a sensitivity of 73% and specificity of 84% for detecting Atrial Septal Aneurysm.

• The model was trained on 80% of the dataset and tested on the remaining 20%.
• Sensitivity was 73% and specificity was 84%.
• Positive Predictive Value (PPV) was 80% and Negative Predictive Value (NPV) was 73%.
• The F-1 score was 0.79.

The ROC curve exhibited an Area Under the Curve (AUC) of 0.8, described as indicative of excellent diagnostic test performance.

• AUC was 0.8 on the held-out test set (20% of the dataset).
• The authors characterized this AUC as 'indicative of excellent diagnostic test performance.'
• The model used key ECG parameters as input features.

The study population consisted of 233 individuals, with ASA diagnosis confirmed by trans-thoracic Echocardiography as the reference standard.

• 123 participants had ASA confirmed by trans-thoracic Echocardiography (TTE).
• 110 participants did not have ASA.
• The study was a retrospective analysis.
• TTE served as the gold standard for ASA confirmation.

Atrial Septal Aneurysm diagnosis is frequently incidental due to the absence of specific symptoms or electrocardiogram criteria.

• ASA is described as 'a real clinical challenge due to its possible association with other relevant conditions.'
• The absence of specific symptoms or ECG criteria explains why its diagnosis is 'very often qualified as incidental.'
• The study motivation was to assess whether ML applied to ECG data could improve detection.