Objective:

To validate an MRI-based deep learning algorithm for predicting lymphatic metastasis in rectal cancer and to construct an integrated fusion model combining imaging and clinicopathological factors, enhancing preoperative diagnostic performance.

Key Findings:

• The prediction algorithm achieved an AUC of 0.760, outperforming two radiologists (AUC: 0.665 and 0.676).
• Independent risk factors for lymph node metastasis included CEA level >5 µg/L and poor differentiation.
• The integrated fusion model demonstrated an AUC of 0.873 in the primary cohort and 0.838 in the external validation cohort, indicating strong predictive capability.
• The fusion model provided higher clinical net benefit than default treatment strategies across various threshold probabilities.

Interpretation:

The preoperative fusion model significantly improves the accuracy of N-staging in rectal cancer, suggesting its potential as a clinical aid for identifying patients who may benefit from neoadjuvant therapy.

Limitations:

• The study is retrospective and may be subject to selection bias.
• The generalizability of the findings needs further validation in larger, diverse populations, and potential biases in external validation should be considered.

Conclusion:

The study presents a promising MRI-based fusion model for preoperative assessment of lymphatic metastasis in rectal cancer, enhancing diagnostic performance and reducing interobserver variability.