Benign and malignant orbital lymphoproliferative disorders: Differentiating using multiparametric MRI at 3.0T.

Abstract

To determine the optimal combination of parameters derived from 3T multiparametric (conventional magnetic resonance imaging [MRI], diffusion-weighted [DW] and dynamic contrast-enhanced [DCE]) MRI for differentiating malignant from benign orbital lymphoproliferative disorders (OLPDs). Forty patients with OLPDs (18 benign and 22 malignant) underwent conventional 3.0T MR, DW, and DCE-MRI examination for presurgery evaluation. Conventional MRI features (including tumor laterality, shape, number of involved quadrants, signal intensity on T1 -weighted imaging (WI) and T2 WI, flow void sign on T2 WI, and findings suggestive of sinusitis) were reviewed, and multivariate logistic regression analysis was used to identify the most significant conventional MRI features. Apparent diffusion coefficient (ADC) and DCE-MRI derived parameters (area under curve [AUC], time to peak [TTP], maximum rise slope [Slopemax ]) were measured and compared between two groups. Receiver operating characteristic (ROC) curve analyses were used to determine the diagnostic ability of each combination that was established based on identified qualitative and quantitative parameters. Multivariate logistic regression analysis showed that the presence of flow void sign on T2 WI significantly associated with benign OLPDs (P = 0.034). Malignant OLPDs demonstrated significantly lower ADC (P = 0.001) and AUC (P = 0.002) than benign mimics. ROC analyses indicted that, ADC alone showed the optimal sensitivity (threshold value, 0.886 × 10-3 mm2 /s; sensitivity, 90.9%), while a combination of no presence of flow void sign on T2 WI + ADC ≤ 0.886 × 10-3 mm2 /s + AUC ≤ 7.366 showed optimal specificity (88.9%) in differentiating benign from malignant OLPDs. Multiparametric MRI can help to differentiate malignant from benign OLPDs. DWI offers optimal sensitivity, while the combination of conventional MRI, DWI, and DCE-MRI offers optimal specificity. 3 J. Magn. Reson. Imaging 2017;45:167-176.