Risk Stratification Prediction of Endometrial Cancer Using Microstructural Mapping Based on Time-Dependent Diffusion MRI

Abstract

Time-dependent diffusion MRI (t_d-dMRI) has potential in characterizing microstructural features; however, its value in imaging endometrioid endometrial adenocarcinoma (EEA) remains uncertain. Patients surgically confirmed with EEA were finally enrolled in our study. The t_d-dMRI data were acquired using pulsed gradient spin echo sequence and oscillating gradient spin echo sequences. The microstructural markers, including cell diameter, intracellular volume fraction (V_in), cellularity, and extracellular diffusivity (D_ex), were fitted with the imaging microstructural parameters using a limited spectrally edited diffusion (IMPULSED) model. The parameters were compared between low- and high-risk groups and between low- and high-proliferation groups. The diagnostic performance was evaluated using receiver-operating characteristic curve and logistic regression analysis. Diameter, D_ex, ADC_PGSE, ADC_N1, and ADC_N2 were significantly low, whereas cellularity, ΔADC1 and ΔADC2 were significantly high in the high-risk and high-proliferation groups. Cellularity, ΔADC1, and ΔADC2 demonstrated excellent diagnostic efficacy in predicting both risk stratification and proliferation status. Cellularity was the only independent predictor for risk stratification, which exhibited a satisfactory positive correlation with cell density in histopathologic examination. The diagnostic potential of t_d-dMRI-based microstructural mapping was demonstrated to noninvasively probe the pathologic characteristics of patients with EEA in a clinical setting, which provided a valuable contribution to surgical guidance.