Comparison of FLASH 3D and VIBE 3D dynamic MRI sequences for evaluation of myometrial invasion and cervical involvement in endometrial cancer

Objective: The aim of this study is, to assess the efficacy of dynamic contrast enhanced MRI (DCE-MRI) techniques (FLASH 3D and VIBE 3D), in determining the presence of myometrial invasion, depth of myometrial invasion, differentiation of stage IA and IB tumor and cervical stromal invasion in endometrial cancer patients preoperativelyand to compare these two dynamic techniques with each other to find out utility of the different techniques for each parameter. Materials and Methods: 72 patients with histologically proven endometrial cancer were included in this prospective study. T2-weighted (T2W) images were obtained for all patients, then half of them (n=36) had FLASH 3D while the other half (n=36) had VIBE 3D DCE-MRI. T2-weighted and dynamic images were assessed independently by two experienced radiologists for the parameters: presence or absence of myometrial invasion, if present the depth of myometrial invasion and cervical involvement. Also confidence levels were given for each parameter. The images that the observers had conflict were re-evaluated together and had a consensus.Sensitivity, specificity, positive predictive value, negative predictive value and accuracyvalues were calculated from the consensus results for every parameter for T2-weighted and dynamic sequences separately. Fisher's Exact test, Mc Nemar test, Mann-Whitney U test, Wilcoxon test and Kappa statistics were used for statistical evaluation. Results: The accuracy of T2-weighted and DCE-MRI sequences were 92%, 96% for presence or absence of myometrial invasion, 76%, 92% for the depth of myometrial invasion, 78%, 92% for differentiation of stage IA and IB tumorand 89%, 90% for cervical involvement, respectively. There was statistically significant difference between two groups in evaluation of depth of myometrial invasion and in differentiation of stage IA and IB tumor. For cervical involvement, both T2-weighted and DCE-MRI sequences had low sensitivity, 64%, 57% respectively, but the specificity of the techniques were higher (95%, 98% respectively).The accuracy of VIBE 3D and FLASH 3D sequences were 97%, 94% for presence or absence of myometrial invasion, 87%, 94% for the depth of myometrial invasion, 89%, 94% for differentiation of stage IA and IB tumorand 92%, 89% for cervical involvement, respectively. There was no statistically significant difference between VIBE 3D and FLASH 3D sequences.There was significant increase for the confidence level of dynamic sequences with respect to T2W sequence. Confidence levels of VIBE 3D and FLASH 3D were not significantly different while the FLASH 3D levels were slightly higher for both of the observers. Kappa agreement between two observers for all parameters was statistically significant and highest agreement values were found in FLASH 3D sequence. Conclusion: The accuracies of dynamic sequences were higher in the evaluation of presence of myometrial invasion which is important for premenopausal fertility sparing treatments. VIBE 3D sequence with higher sensitivity and accuracy percentages is recommended in these patients. Depth of myometrial invasion which is especially important for the prognosis and decision of treatment choice could be assessed more accurately with dynamic images. For the staging (stage IA and stage IB) DCE-MRI should be used since there is significant difference between T2-weighted and DCE-MRI. With the use FLASH 3D technique, the accuracy and the confidence levels of the observers could be increased. With higher spatial resolution taken into account the use of FLASH 3D technique is recommended for preoperative evaluation. VIBE 3D with shorter acquisition times could be preferred for the patients who cannot tolerate long MRI examination times. For evaluation of cervical stromal invasion both T2 and dynamic images can be of value.