Comparison of the diagnostic performances of diffusion parameters in diffusion weighted imaging and diffusion tensor imaging of breast lesions

Abstract

PurposeTo evaluate the diagnostic efficiency of the diffusion parameters measured by conventional diffusion-weighted imaging (DWI) and diffusion tensor imaging (DTI) for discrimination of malignant breast lesions from benign lesions and the normal breast. Materials and methodsThe study included 52 women with 55 breast lesions (30 malignant, 25 benign). DTI and DWI were performed complementary to dynamic contrast MRI at 3T. Apparent diffusion coefficient (ADC) of DWI, mean diffusivity (MD) and fractional anisotropy (FA) values of DTI were measured for lesions and contralateral breast parenchyma in each patient. We used b factors of 0, 50, 850, 1000 and 1500s/mm2 for DWI and b 0 and 1000s/mm2 for DTI. ADC, MD and FA values were compared between malignant and benign lesions, and the normal parenchyma by univariate and multivariate analyses. ResultsDiffusion parameters showed no difference according to menopausal status in the normal breast. ADC and MD values of the malignant lesions were significantly lower than benign lesions and normal parenchyma (p=0.001). The FA showed no statistical significance. With the cut-off values of ≤1.23×10−3mm2/s (b 0–1000s/mm2) and ≤1.12×10−3mm2/s (b 0–1500s/mm2), ADC showed 92.85% and 96.15% sensitivity; 72.22% and 73.52% PPV, respectively. With a cut-off value of ≤1.27×10−3mm2/s (b 1000s/mm2), MD was 100% sensitive with a PPV of 65.90%. Comparing the diagnostic performance of the parameters in DTI with DWI, we obtained similar efficiency of ADC with b values of 0,1000 and 0,1500s/mm2 and MD with a b value of 0, 1000s/mm2 (AUC=0.82±0.07). ConclusionADC of DWI and MD of DTI values provide significant discriminative factors for benign and malignant breast lesions. FA measurement was not discriminative. Supported with clinical and dynamic contrast MRI findings, DWI and DTI findings provide significant contribution to the final radiologic decision.