Four-compartment Restriction Spectrum Imaging model for Preoperative Grading in Endometrial Cancer

Motivation: Inaccurate preoperative grading of endometrial cancer (EC) can lead to either over-treatment or under-treatment, affecting survival and life quality of patients. Goal(s): To explore the application of Restriction Spectrum Imaging (RSI) model in the preoperative grading of EC. Approach: Multi-b-value MR images were collected, regions of interest (ROIs) were manually delineated to perform model fitting for further Mann-Whitney U test and receiver operating curve (ROC) analysis. Results: RSI4-derived parameters could differentiate between all grades of EC lesions. Further combination of histogram ADC parameters could improve diagnostic performance. Only RSI model derived parameters could discriminate grade-2 from grade-3 lesions. Impact: Four-compartments RSI model demonstrated a potential for preoperative discrimination across all grades of EC lesions, especially between grade-2 and grade-3. Our findings underscore the potential of the RSI model for non-invasive preoperative imaging grading of lesions in EC patients.