Diagnostic Performance of Multiparametric MRI in Characterizing Adrenal Lesions

Abstract
Background: Adrenal nodules are common incidental find-ings in radiological daily routine with lipid-rich adenomas are
the most frequent. Sometimes lipid-poor adenomas may not be fully characterized on CT or chemical shift imaging (CSI).
These indeterminate lesions pose management dilemmas par-ticularly in patients with known primary malignancy.
Aim of Study: The aim of this study was to evaluate the di-agnostic performance of mutliparametric MRI in characteriza-
tion of adrenal lesions, with particular regard to the distinction between adenomas and malignant tumors.
Material and Methods: Forty patients with 42 adrenal le-sions underwent multiparametric MRI including conventional
MRI, CSI, diffusion weighted imaging (DWI), diffusion tensor imaging (DTI, in 12 patients) and dynamic contrast enhanced (DCE) MRI. Imaging features on conventional MRI were eval-uated, CSI was assessed both qualitatively and quantitatively by calculating the adrenal signal intensity index (ASII), appar-ent diffusion coefficient (ADC) and fractional anisotropy (FA) values of the lesions were measured and enhancement pattern on DCE-MRI was assessed both qualitatively and quantitative-ly. Statistical analysis was performed. Results: Most adenomas displayed iso- to low signal inten-sity (SI) on T2-weighted images (WI), while most non-adeno-mas exhibited heterogeneous high T2-SI. A significant signal drop on CSI was seen in most of adenomas. ASII exhibited highstatistically significant difference (p=0.01) in  differentiating ad-enomas from non-adenomas, at a cut off value of 23% the sen-sitivity was 92% and the specificity was 78%. On DWI, ADC values could not show significant statistical difference between adenomas and non-adenomas. On the other hand, FA values measured on DTI revealed high statistically significant differ-ence in differentiating adenomas from non-adenomas (p=0.01). In receiver operating characteristics (ROC) curve analysis, FA