3-T MRI in the preoperative evaluation of depth of myometrial infiltration in endometrial cancer.

Objective: The objective of our study was to evaluate the diagnostic accuracy of 3-T MRI in determining the depth of myometrial infiltration in patients with endometrial cancer.

Subjects And Methods: Fifty-two patients (43 postmenopausal) with histopathologically proven endometrial carcinoma underwent preoperative 3-T MRI. The following sequences were performed: axial T1 fast spin-echo (FSE); axial, parasagittal, and paracoronal T2 FSE; paracoronal 3D T1 inversion recovery gradient-echo after contrast administration; and parasagittal fat-suppressed T1 FSE. All patients underwent a hysterectomy. The MRI findings were compared with histopathology results. The quantity and degree of artifacts were evaluated.

Result: . MRI performed on a 3-T unit was in agreement with histopathology in assessing the depth of invasion in 86.4% (44/52) of the patients with a mean sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of 83.5%, 93.9%, 77.8%, 92.2%, and 89.7%, respectively. Performance values were also assessed for single stages of myometrial infiltration. For the detection of an intramucosal lesion (MRI, 12/52; histopathology, 6/52), sensitivity was 100%; specificity, 86.9%; PPV, 50%; NPV, 100%; and accuracy, 88.5%. For the detection of myometrial infiltration that was less than 50% (MRI, 12/52; histopathology, 16/52), sensitivity was 62.5%; specificity, 94.4%; PPV, 83.3%; NPV, 85%; and accuracy, 84.6%. For the detection of myometrial infiltration that was greater than 50% (MRI, 28/52; histopathology, 30/52), sensitivity was 93.3%; specificity, 100%; PPV, 100%; NPV, 91.7%; and accuracy, 96.2%. The following artifacts were found: abdominal wall movement, nine patients (not affecting image quality); peristalsis, 16 patients (two deeply affecting, one affecting, and 13 scarcely affecting); magnetic susceptibility artifact, four patients (not affecting); chemical shift, 20 patients (four scarcely affecting and 16 not affecting); and dielectric effect, six patients (four deeply affecting and two affecting).

Conclusion: In evaluating the depth of myometrial infiltration in patients with endometrial cancer, 3-T MRI showed high diagnostic accuracy-equivalent to that of 1.5-T MRI reported in the literature. Artifacts did not significantly affect image quality.