A multiparametric analysis combining DCE-MRI- and IVIM -derived parameters to improve differentiation of parotid tumors: a pilot study

Abstract

Aim: To evaluate dynamic contrast-enhanced (DCE)-MRI and diffusion weighted (DW)-MRI diagnostic value to differentiate Warthin tumors (WT) by pleomorphic adenomas (PA). Materials & methods: Seven WT and seven PA were examined. DCE- and  DW-MRI parameters were extracted from volumes of interest; volume of interest-based averages and standard deviations were calculated. Statistical analysis included: linear discriminant analysis, receiver operating characteristic curves, sensitivity and specificity. Results: No single feature was able to differentiate WT by PA (p > 0.05); linear discriminant analysis analysis showed that a combination of all features or combinations of feature pairs (namely: Ktrans(std) & f(std), Ktrans(std) & D(std), kep(std) & D(std), MRE(av) & TTP(av)) might achieve sensitivity (SENS), specificity (SPEC) = 100%, with a slight reduction after cross-validation analysis (SENS = 0.875; SPEC = 1). Conclusion: Although preliminary and not conclusive, our results suggest that differentiation between WT and PA is possible through a multiparametric approach based on combination of DCE- and DW-MRI parameters.