Spectral CT and Machine Learning Improve Preoperative Prediction of Gastric Cancer Invasion

Accurate preoperative prediction of lymphovascular and perineural invasion status in gastric cancer patients is crucial for identifying high-risk individuals. This study developed and tested a machine learning model that successfully combines spectral CT parameters and clinical indicators to accurately predict LVI/PNI status. The integration of portal venous and ep spectral CT parameters significantly enhances the detection of LVI/PNI, with CT stage, N status, and EMVI status also playing important roles in predicting invasion status.