Making an accurate preoperative diagnosis of endometrial lesions is critical to avoid unnecessary surgical procedures and protect the patients’ fertility. Noninvasive pre-surgical evaluation of abnormalities in the uterine cavity remains to be challenging. The volumetric ADC histogram analysis provides a mechanism for quantifying the overall heterogeneity and comprehensive assessment of a given abnormality. We demonstrated that whole-lesion ADC histogram-based texture analysis could serve as a quantitative technique of differentiating benign endometrial lesions (BELs) from stage IA endometrial cancer (EC) preoperatively. The combined ADC histogram model showed high classification ability for both premenopausal and postmenopausal patients.