Objective:

To propose an automatic recognition method that combines lesion feature enhancement and boundary-related structural cue fusion for pulmonary embolism in CT images, addressing challenges of complex lesion representation, blurred vascular boundaries, and insufficient model generalization ability.

Approach:

Key Findings:

• Achieved an Accuracy of 0.956, Precision of 0.961, Recall of 0.951, and AUC of 0.971 on 523 single-center chest CT cases, outperforming multiple comparison models.
• In external validation, achieved Accuracy values of 0.779 and 0.672 on the FUMPE and RSNA datasets, respectively, indicating variability in performance across different datasets.

Interpretation:

The proposed method effectively improves the discriminative performance of binary pulmonary embolism recognition from CT images.

Limitations:

• The generalization ability of the method under cross-center and cross-protocol conditions remains limited, which may affect its clinical applicability.

Conclusion:

The method may provide a useful technical reference for computer-aided screening of pulmonary embolism.

Attribution Notice

This content is an AI-generated, fully rewritten summary based on a published scholarly article. It does not reproduce the original text and is not a substitute for the original publication. Readers are encouraged to consult the source for full context, data, and methodology.