This machine learning model improves detection of perineural invasion (PNI) in colorectal cancer, crucial for surgical outcomes.

• LightGBM model achieved an AUC of 0.996 in training, 0.935 in testing, and 0.918 in external validation across 430 samples.
• PNI correlates with higher T stage, increased lymph node metastasis, and more lymphovascular invasion.

Adopting this model could refine patient selection and enhance prognosis understanding in colorectal surgery.