Graphic abstract for machine learning pipeline for predicting mesenteric lymph node metastasis using clinicopathologic data and imaging (IMAGE)
Caption
Clinical data and CT images were obtained from the hospital’s electronic system. The attenuation values of the tumor and external iliac artery within the same slice were independently obtained by two radiologists, who were blinded to all clinical and pathological information. The LASSO algorithm was employed for feature selection. Five widely used and current machine learning techniques were used to construct the predictive models. SHAP values were utilized for model interpretation, highlighting key predictors.
Credit
Fusheng Ouyang
Usage Restrictions
Not applicable
License
Original content