PURPOSE

To evaluate the diagnostic performances of the diffusion tensor imaging (DTI) parameters in the diagnosis of breast cancer and to investigate the variations in DTI parameters according to the breast cancer biomarkers.

MATERIALS AND METHODS

At 3.0 Tesla (T), DTI was performed in 85 patients with 92 enhancing breast lesions. λ , λ , λ , mean diffusivity (MD), radial diffusivity (RD), fractional anisotropy (FA), relative anisotropy (RA), and geodesic anisotropy (GA) were studied and compared with diffusion-weighted imaging-derived apparent diffusion coefficient. Lesions were analyzed according to BIRADS lexicon. Logistic regression models were constructed to determine the contribution of DTI to the specificity and the accuracy of DCE-MRI. Breast cancer biomarkers; estrogen receptor (ER), HER-2 status, and Ki-67 were correlated with DTI in malignant cases.

RESULTS

Malignant lesions exhibited significantly lower MD, RD, λ , λ , λ and higher FA, RA, GA values (P <  0.001). Logistic regression models showed that MD, RD, λ , λ , λ , FA, and RA increase the specificity of the DCE-MRI (from 83.0% to 89.4-93.6%; P <  0.05). Higher RD, λ , λ and lower FA, RA, and GA values were observed in ER-negative breast cancer (P <  0.05). Ki-67 showed significant, negative correlation with FA, RA, GA, λ -λ and λ -λ (r = -0.336 to -0.435; P <  0.05).

CONCLUSION

Besides its ability to differentiate malignant breast lesions, DTI improves the specificity of conventional 3.0T breast MRI and shows correlation with biomarkers ER and Ki-67.

LEVEL OF EVIDENCE

1 J. Magn. Reson. Imaging 2017;45:660-672.