Borderline and malignant ovarian epithelial tumors: differentiating using multiparameter MRI

Abstract Objectives To evaluate the value of multiparameter MRI (mp-MRI) including conventional MRI image features, diffusion-weighted imaging (DWI) and dynamic contrast-enhanced MRI (DCE-MRI) in differentiating borderline from malignant ovarian epithelial tumors (OETs). Methods Forty-three patients with borderline OETs (BOETs) and 119 patients with malignant OETs (MOETs) who underwent mp-MRI examinations for pre-treatment assessment were respectively enrolled. Conventional MRI features (eg, tumor shape, configuration, signal intensity SI on T1WI and T2WI) were retrospectively analyzed. Apparent diffusion coefficient (ADC), and DCE-MRI derived parameters (rel-enhancement RELENH, time to peak TTP, wash-in-rate WIR, and wash-out-rate WOR) were evaluated. Independent samples t-test, Mann-Whitney U test, χ2 test, multivariate logistic regression analysis, binary logistic regression, and receiver operating characteristic (ROC) curve analyses were employed as appropriate. Results BOETs group showed significantly lower age than MOETs group (P .001). Multivariate logistic regression analysis indicated that tumor configuration was the independent imaging features associated with MOETs (P .001). BOETs group showed significantly higher ADC value than MOETs group (P .001). Among DCE-MRI derived parameters, MOETs group showed significantly shorter TTP than BOETs (P = .014). ROC analyses indicated that a combination of age ≥ 43.5 years old + non-cystic predominant type + ADC ≤ 1.05 × 10−3 mm2/s + TTP ≤ 238.87 s showed the highest efficiency (AUC, 0.930; sensitivity, 84.9%; specificity, 86.0%) in diagnosing MOETs, which was significantly higher than that of age (P = .002), configuration (P .001), ADC (P = .027), and TTP (P .001) alone. Conclusions mp-MRI might be effective in differentiating MOETs from BOETs. Advances in knowledge The study which combining conventional MRI, DWI, and DCE-MRI for differentiating BOETs from MOETs is still lacked until now. What is more, the study may be more accurate for the differentiation of borderline malignancy or low-grade malignant potential tumors.