Pharmacokinetic analysis using dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) directly reflects the physiological properties; including vessel permeability, perfusion and the volume of the extravascular/extracellular spaces. In breast lesions, it has been shown that pharmacokinetic analysis can potentially improve the differentiation of malignant from benign lesion, and is valuable in monitoring the responses to neoadjuvant chemotherapy^1-3.

Pharmacokinetic parameters are calculated using the tissue T1 values. In the breast DCE MRI, fat-suppression techniques are routinely applied to improve the delineation of tumors. Two flip angles (FA) images with fat-suppression are commonly used to create T1 mapping. However, the breast fibroglandular tissue contains various amount of fat tissue, and T1 values of both the breast tissue and tumor are expected to change from with and without fat-suppression. The difference of T1 values may influence the calculated permeability parameters.

Prototype Dixon-TWIST-VIBE (DT-VIBE) technique provides high temporal and spatial resolution and makes it possible to obtain both fat suppression (water image) and non-fat suppression (in-phase) images from the same data set. In this study, we evaluated the influence of fat-suppression on T1 values and pharmacokinetic parameters in DCE-MRI of breast lesions using the prototype DT-VIBE technique.

MRI was performed using a 3.0-T MRI system (MAGNETOM Skyra; Siemens Healthcare, Erlangen, Germany) with a 16-channel phased-array bilateral breast coil. A 2-point variable FAs (5°, 15°) with Dixon images were applied for T1 mapping. Prototype DT-VIBE (temporal resolution: 5 s, spatial resolution: 1 mm x 1mm x 3 mm) was performed in both early (25 flames within 2 min 22 s, started from 22 s before contrast material injection) and delayed phases (9 flames within 30 s, started from 5 min after injection). The total number of scan frames was 34 with a scan time of 5 min 30 s.

T1 values and pharmacokinetic parameters (K^trans) were calculated for five female patients (mean age, 59.8 years; range, 49-75 years) with locally advanced breast cancers who were scheduled for neoadjuvant chemotherapy (NAC) including bevacizumab. Four of the five patients underwent MRI before and after a cycle of NAC. We obtained total of nine MR examinations (before treatment, n=5; after NAC, n=4). On DT-VIBE, regions of interest (ROIs) were delineated by an experienced radiologist, who free-handedly traced the whole tumor as carefully as possible (Fig. 1). These ROIs were copied on both T1 and Ktrans maps generated from the DT-VIBE datasets.

Pharmacokinetic evaluation was performed based on the two-compartment Tofts model using Tissue 4D® software (Siemens Healthcare). T1 and K^trans values were compared between in-phase and Dixon-water images on each examination using the Wilcoxon rank-sum test (Matlab®; MathWorks, Natick, MA). Values of p<0.05 were treated as significant.

T1 values for breast tumors showed no significant difference between in-phase and Dixon-water images (Fig. 2; p = 1.00), but some cases showed large differences of more than 50% (Fig. 3a, P4). In terms of K^trans values, some cases also showed large differences of more than 30% (Fig. 4). The change in intra-patient K^trans values between before and after NAC showed the same decreasing tendency for both in-phase and Dixon-water images, with no significant difference in decreasing ratio (Fig. 4; p=0.875).

In our study, we have shown that some of the cases may exhibit large differences in T1 values between in-phase and Dixon-water images, although there was no statistical significant differences. One of the causes of this variety in T1values might be attributed to the fat fraction within the placed ROI. In long T1 value tissues, such as breast tumors on pre-contrast images, slight measured signal difference can yield large difference in the calculated T1 value using 2-point valuable flip angle method. Thus, the reliability of K^trans measurement using different image data sets seemed inappropriate. Assessment of the intra-patient K^trans change ratio using the same dataset might be feasible.Our results suggest that K^trans can be used as a relative value, but not an absolute value that allows comparison between patients.

There were several limitations in our study. First, the sample size was small. Second, we evaluated only one ROI in each lesion. Third, influence of fat suppression only using Dixon method was evaluated. Phantom study was needed to evaluate the influence of other fat suppression method.

In breast cancers, measured T1 values and K^trans may change according to with or without fat-suppression. Assessments of intra-patient K^trans changes would be feasible as relative values, but not feasible as an absolute value to compare among patients, owing to the variability in T1 values observed in fat-suppressed images.