AI Accuracy in Pulmonary Embolism Diagnosis

AI models exhibited high accuracy in detecting pulmonary embolism in internal validation studies, achieving a pooled sensitivity of 0.91 and a specificity of 0.94.
External validation showed a minor decrease in diagnostic accuracy, with a pooled sensitivity of 0.89 and specificity of 0.88, indicating potential limits in model generalizability.
The study found substantial heterogeneity across AI models assessed in different validation settings, as reflected by high I² values.
Despite the overall strong pooled results, the model’s performance was slightly diminished when tested in external datasets, signaling a need for caution in real-world applicability.
The researchers emphasized the importance of considering the observed performance variance and heterogeneity for practical deployment of AI diagnostics in pulmonary embolism detection.

You may have missed