Diagnostic performance of diffusion‐weighted (DWI) and dynamic contrast‐enhanced (DCE) MRI for the differentiation of benign from malignant soft‐tissue tumors

Abstract

Background A wide range of specificity values for the differentiation of benign and malignant soft‐tissue tumors show the limitations of conventional MRI features. The data obtained by quantitative analysis of diffusion‐weighted image (DWI) and dynamic contrast‐enhanced (DCE) MRIs would provide more objective results, especially in terms of cellularity and perfusion. Purpose To evaluate the diagnostic efficacies of DWI and DCE MRI for the differentiation of malignant and benign soft‐tissue tumors. Study Type Retrospective. Subjects In all, 136 patients (68 females, 68 males; age range 18–86 years, mean age 57.2 years) with soft‐tissue tumors. Field Strength/Sequence 3 T, DWI, DCE. Assessment Tumor sizes, margins, locations, the presence of involvement in bone or neurovascular bundle, peritumoral edema, heterogeneity, and tumor necrosis were investigated on conventional MR images. On DWIs, visual signal drops were assessed and ADC (apparent diffusion coefficient) values were measured. K trans , K ep , V e , and iAUC values, and time–concentration curve (TCC) types were determined using DCE images. Statistical Tests The data were statistically analyzed to determine the abilities to differentiate benign and malignant tumors using the chi‐square test, two‐sample t ‐test, and receiver operating characteristic (ROC) analysis. Results Seventy‐three cases were malignant and 63 benign. Age (mean ages of benign/malignant tumors, 51.75/61.86 years; P = 0.0002) and gender (F:M = 40:23 [benign], F:M = 28:45 [malignant], P = 0.003) influenced the distinction between benign and malignant. Sizes, margins, neurovascular bundle involvement, peritumoral edema, and heterogeneity of the tumors on conventional MR images and DCE parameters (K trans , K ep , V e , and iAUC, and TCC plots) obtained from focal region of interest within a narrow volume of interest significantly differentiated benign and malignant lesions (all P < 0.0001, except V e [ P = 0.0004]). For DWI with ADC mapping, all ADC values and visually signal drops were also significant ( P < 0.0001). Data Conclusion DWI and DCE‐MRI and derived variables were significantly helpful in discriminating benign and malignant soft‐tissue tumors complementary to conventional MRI. Level of Evidence: 3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2019;50:798–809.