Differentiation of Triple-Negative Phenotypes of Breast Cancer from Others Using Multiparametric Magnetic Resonance Imaging: A Prospective Imaging Study

Abstract To predict and differentiate triple-negative (TNBC) from non-triple-negative breast cancer (NTNBC) phenotypes using multiparametric magnetic resonance imaging (MpMRI). One hundred and two patients with BI-RADS 5/6 lesions who underwent MpMRI (3 Tesla) and were diagnosed with invasive ductal carcinoma on histopathological analysis after ultrasound-guided biopsy were categorized into TNBC and NTNBC groups based on immunohistochemistry. MRI included T1-weighted imaging, T2-weighted imaging, dynamic contrast-enhanced MRI (DCE-MRI), diffusion-weighted imaging, and magnetic resonance spectroscopy (MRS). Morphological, semiquantitative, and quantitative DCE parameters, apparent diffusion coefficient (ADC) values, and MRS (tCho) were studied. Nonparametric categorical variables were analyzed using the chi-square test, continuous nonparametric variables using the Kruskal–Wallis, and continuous parametric variables using the ANOVA test to determine the association of parameters with NTNBC/TNBC subtypes. Receiver operating characteristic and logistic regression analyses were done to determine the diagnostic performance and independent predictors. Seventy-two cases were NTNBCs, and 30 cases were TNBCs. TNBCs showed more circumscribed margins (51.7%), rim enhancement (65.5%), perilesional edema (80%), unifocal lesions (76.7%), and axillary lymphadenopathy (73.3%). NTNBCs showed more irregular shapes (81.8%) and spiculated margins (56.1%). Choline peaks on MRS were more frequent in TNBC (60%). TNBCs had higher mean Ktrans (>0.53 min−1), higher signal enhancement ratio (>72.95), lower mean Ve (