Abstract： Objective:To explore the value of shapley additional explanations(SHAP)value in differentiating malignant tu-mor(MT)and Warthin tumor(WT)of parotid gland in extreme gradient boosting(XGBoost)MRI model.Methods:MRI data of 22 MT in 22 cases and 51 WT in 38 cases confirmed by surgery and pathology were retrospectively analyzed.The tumors'shape,location and enhancement mode,as well as T1WI,T2WI,FS-T1WI and FS-T2WI imaging signs were analyzed.The sta-tistically significant signs were screened by univariate analysis and included in XGBoost model,using receiver operating char-acteristic curve area under the curve(AUC),sensitivity and specificity to evaluate the diagnostic efficiency of the model.The model was analyzed by SHAP value,and the weight of MRI signs in the model was made clear.Results:Among 22 MT and 51 WT,there were statistical differences in progressive reinforcement(P＜0.05),irregular shape(x2=20.707,P＜0.05),non-caudal lobe(x2=8.911,P＜0.05),T2WI high signal(x2=7.581,P＜0.05),FS-T1WI equal or low signal(P＜0.05)and FS-T2WI high signal(x2=9.016,P＜0.05).The AUC was 0.847,the sensitivity was 77.3％,and the specificity was 92.2％when six single factors were included in XGBoost model.The absolute average SHAP value of six MRI signs was 0.21～0.99,among which the irregular shape had the largest weight.Conclusion:Irregular shape,T2WI high signal,progressive reinforcement,FS-T2WI high signal,non-caudal lobe and FS-T1WI equal or low signal have different weights in the differential diagnosis of MT and WT.Accu-rate identification of MRI signs is conducive to accurate diagnosis of MT and WT.