Proposed MRI sequences for local staging of endometrial cancer: a single center experience

Abstract

Purpose: To present our experience with preoperative staging of endometrial cancer (EC) using dedicated MRI protocol at a tertiary referral center for female pelvic imaging. Material and Methods: Between April 2013 and June 2018, 117 women (mean age: 59.9 years ± SD: 12.7) with newly diagnosed, biopsy-confirmed EC, underwent pelvic MRI for staging. In all patients, T2-weighted (T2W), T1- weighted (T1W), dynamic T1W gradient echo contrast-enhanced images (DCE) and diffusion weighted (DWI) images were obtained. Two expert radiologists prospectively recorded the following tumour prognostic factors: size, depth (< or ≥50%) of myometrial involvement, cervical extension, extrauterine spread and metastatic nodes; MRI predictive ability for each of the above characteristics was statistically tested with ROC curve analysis, using surgicopathological results as the standard of reference. Chi square and Student’s t-test were used for possible association of maximal tumour/myometrial contrast on DCE with prognostic tumour histological features including type, grade, size and depth of myometrial invasion. Qualitative assessment regarding usefulness of applied MRI sequences for tumour local staging was also addressed. Results: MRI exhibited high diagnostic accuracy for prediction of myometrial invasion (AUC=0.86, 95% CI: 0.78-0.93, p<0.001), cervical invasion (AUC=0.86, 95% CI: 0.70-1.00, p<0.001) and pelvic nodal involvement (AUC=0.90, 95% CI: 0.69-1.00, p=0.003). No association was found between maximal tumour/myometrial contrast on DCE and known histologic prognostic factors. For evaluation of myometrial invasion, DCE received significantly better rating compared to T2W or DWI alone; the combination of T2W and DCE (T2W+DCE) exhibited significantly better results for assessment of myometrial invasion compared to T2W+DWI and it was equally effective compared to T2W+DWI+DCE. For tumour cervical extension, DCE received significantly better rating compared to T2W or DWI alone; T2W+DCE and T2W+DCE+DWI significantly improved diagnosis of cervical invasion. For evaluation of metastatic lymphadenopathy, T2W was significantly more helpful than DWI; T2W+DWI were not superior to T2W alone for the evaluation of nodal metastases. Conclusions: MRI is accurate for preoperative local staging of EC. DCE-MRI optimises assessment of myometrial and cervical invasion.