Magnetic resonance imaging reveals functional diversity of the vasculature in benign and malignant breast lesions

BACKGROUND. Tumor perfusion through the microvascular network can be imaged noninvasively by dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). The objective of the current study was to quantity the microvascular perfusion parameters in various human breast lesions and to determine whether they varied between benign lesions and malignancy and whether they were altered with increased invasiveness. METHODS. Perfusion parameters in 22 benign fibrocystic changes, 15 ductal carcinomas in situ (DCIS), 30 infiltrating ductal carcinomas (IDC), and 22 fibroadenomas were measured using high-resolution DCE-MRI. Pixel-by-pixel image analysis yielded parametric images of two perfusion indicators: the influx transcapillary transfer constant (k(trans)) and the efflux transcapillary rate constant (k(ep)). Correlations of lesion type and perfusion parameters were calculated using Spearman correlation. Logistic regression analysis evaluated the best predictors of the kinetic parameters that differentiate between IDC and benign lesions. RESULTS. The perfusion parameters exhibited a progressive increase from benign fibrocystic changes to DCIS and IDC, with a significant correlation between lesion type and the parameters' values (range of correlation coefficients, 0.56-0.76; P < 0.0001). In addition, k(trans) increased from low-grade DCIS to high-grade DCIS. Fibroadenomas were characterized uniquely by high k(trans) but low k(ep). Stepwise logistic regression selected k(trans). as the best predictor for distinguishing benign fibrocystic changes from IDC, yielding 93% sensitivity and 96% specificity. ONCLUSIONS. The microvascular perfusion parameters in breast lesions were elevated with invasiveness. Quantification of these parameters using high-resolution DCE-MRI was helpful for differentiating between breast lesions and should improve breast carcinoma diagnosis.