Magnetic Resonance Imaging Features in Triple-Negative Breast Cancer: Comparison With Luminal and HER2-Overexpressing Tumors

Early recognition of triple-negative (TN) breast cancer is useful for optimal management. We analyzed the clinicopathologic data and magnetic resonance imaging (MRI) features in 225 women with breast cancer to identify the criteria of early detection before pathologic results are available. The association of several MRI findings may be used to reliably predict the TN status of breast lesions. As independent predictors of TN lesions, our findings may assist clinicians in building a diagnostic algorithm for early detection of TN lesions. Background: It has been ascertained that triple-negative (TN) breast cancer is characterized by an aggressive clinical course and a poor prognosis. The purpose of our study was to compare the magnetic resonance imaging (MRI) features of the 3 major different breast cancer subtypes (TN, luminal, and human epidermal growth factor receptor 2 [HER2]-overexpressing) and to suggest the criteria that might predict TN phenotype. Materials and Methods: From October 2007 to April 2011, we studied 77 patients with histologically confirmed TN breast cancer who underwent breast MRI. We randomly included 148 patients with non-TN breast cancer (110 luminal and 38 HER-overexpressing) as a control group. We evaluated the clinicopathologic data, the MRI morphologic and kinetic features, the signal intensity on T2-weighted images, and the apparent diffusion coefficient (ADC). Results: Our results confirmed that TN tumors are more aggressive, are usually diagnosed at a younger age compared with the other study groups, and show benign morphologic features with MRI. Backward stepwise logistic regression identified some parameters as independent predictors of TN-type lesions: age, size, shape, presence of edema, and infiltrative characteristics. The receiver operating characteristic (ROC) curve, built with 4 of 5 these factors as criteria to predict TN status, showed a 0.664 area under the curve (AUC) value (sensitivity 58.4%, specificity 73.2%). The inclusion of the fifth criterion showed a 0.699 AUC value (sensitivity, 49.4%; specificity, 89.4%). Conclusion: We identified the clinicoradiologic parameters that are independent predictors of TN breast lesions, which might be helpful for earlier prediction of the TN status of a breast lesion.