Comparison of the determination of the local tumor extent of primary endometrial cancer using clinical examination and 3 Tesla magnetic resonance imaging compared to histopathology.

Purpose: The aim of this study is to analyze the correct staging of primary endometrial cancer (EC) using clinical examination and 3 Tesla (T) magnetic resonance imaging (MRI) results compared to histopathology.

Methods: In this prospective, non-randomized, single-center study, 26 women with biopsy-proven EC were evaluated. All women underwent clinical examination including transvaginal ultrasound (CE/US) and 3T MRI (T2-weighted, diffusion-weighted and dynamic contrast-enhanced sequences) prior to surgery. Spearman's correlation coefficient was employed to analyze the correlation between both staging methods and histopathology and generalized estimation equation analysis to compare their staging results. Main outcome measures are determinations of local tumor extent for EC on CE/US and 3T MRI compared to histopathology (gold standard).

Results: Sixteen women had an early-stage pT1a tumor, 10 a locally advanced ≥ pT1b tumor. The early stage was correctly diagnosed at CE/US in 100%, by MRI in 81%. Spearman's correlation coefficient was r = 1.0 (p < 0.001) for correlation of CE/US and histopathology, r = 0.93 (p < 0.001) for correlation of MRI and pathology. A locally advanced tumor stage was exactly diagnosed by MRI in 70% and at CE/US in 50%.

Conclusions: CE/US is sufficient for staging T1a endometrial cancer, while MRI provides higher sensitivity in detecting locally advanced tumors. Based on our results, combining CE/US and 3T MRI in patients with at least suspected deep myometrial invasion offers a more reliable workflow for individual treatment planning.