Diffusion weighted MRI for the detection of myometrial invasion in carcinoma endometrium

Doi: https://doi.org/10.55522/ijti.v4i2.0142

Abstract

Depth of myometrial invasion (<50% vs. >50%) is a critical prognostic factor in endometrial carcinoma, influencing the need for lymphadenectomy. MRI, particularly diffusion-weighted imaging (DWI), is increasingly playing a role in preoperative assessment. To determine the diagnostic accuracy of DWI in detecting myometrial invasion compared to dynamic contrast-enhanced (DCE) and T2-weighted (T2W) sequences, and to assess interobserver agreement. This retrospective study included 40 patients with surgically confirmed grade 1–2 endometroid adenocarcinoma who underwent preoperative MRI (T2W, DCE, and DWI). Two blinded radiologists independently assessed myometrial invasion. Sensitivity, specificity, accuracy, predictive values, and interobserver agreement (kappa statistics) were calculated using histopathology as the reference standard. DWI demonstrated the highest diagnostic accuracy for both observers (92.5% and 90%), followed by DCE (90% and 82.5%) and T2W (72.5% and 70%). DWI showed the highest sensitivity (100% and 94.7%) and negative predictive value (100% and 94.7%). Accuracy of DWI was significantly better than T2W for both observers (p=0.004). In the subset with coexisting fibroids/adenomyosis (n=10), DWI maintained high accuracy (90%) and sensitivity (100%). Interobserver agreement was near-perfect for DWI (κ=0.849) versus moderate for DCE (κ=0.554) and T2W (κ=0.529). DWI outperforms T2W and is at least equivalent to DCE in assessing myometrial invasion, with excellent interobserver agreement. Given its advantages—no contrast requirement, shorter scan time, and utility in renal failure or contrast allergy—DWI has the potential to replace DCE in routine MRI protocols for endometrial carcinoma staging.