APPLICATION OF DIFFUSION KURTOSIS IMAGING BASED ON MULTIMODAL MRI IN MICROSATELLITE INSTABILITY OF ENDOMETRIAL CANCER

Background: Microsatellite instability (MSI) based on multimodal MRI is a critical molecular marker in endometrial cancer (EC), associated with distinct biological behaviors and therapeutic responses. Diffusion kurtosis imaging (DKI), an advanced diffusion-weighted imaging technique, offers insights into tissue microstructural complexity and has shown potential for non-invasive cancer characterization. Objective: This study aims to evaluate the application and diagnostic performance of DKI in assessing MSI status in endometrial cancer. Methods: A retrospective analysis was conducted on patients with confirmed endometrial cancer who underwent preoperative DKI and MSI testing. Quantitative DKI parameters, including mean kurtosis (MK), mean diffusion(MD), and ADC were extracted and compared between MSI-high (MSI-H) and microsatellite stable (MSS) groups. Statistical analyses were performed to assess the diagnostic accuracy. Results: MSI-H tumors demonstrated significantly higher MK,MD and ADC values compared to MSS tumors (p < 0.01), reflecting increased tissue heterogeneity and microstructural disorganization. Receiver operating characteristic (ROC) analysis indicated that MK had the highest diagnostic performance, with an area under the curve (AUC) of 0.731, sensitivity of 52.2%, and specificity of 95.1%. Conclusion: DKI provides a non-invasive, quantitative method for differentiating MSI-H from MSS endometrial cancers. Its excellent diagnostic accuracy highlight its potential as a biomarker for MSI status of endometrial cancer.