In some previous studies, the effectiveness of initial enhancement for differentiating between benign and malignant breast lesions using high temporal resolution has been reported^1,2. However, these protocols were less suitable in clinical usage because of the limitation in spatial resolution. A recently available prototype sequence, time-resolved angiography with interleaved stochastic trajectories (TWIST)-volume interpolated breath-hold examination (VIBE) (work in progress supported by Siemens Healthcare, Erlangen, Germany) might be able to improve the temporal resolution of breast dynamic contrast-enhanced (DCE)-MRI with preserved spatial resolution. The TWIST-VIBE sequence makes it possible to assess both the parameters of initial enhancement using high temporal resolution and routine morphological evaluation using high spatial resolution at a single session³.To assess the additional value of parameters derived from the initial enhancement on TWIST-VIBE in differentiating between benign and malignant breast lesions on MRI.

Patients: This study included 136 consecutive patients referred for breast MRI from October 2014 to October 2015 because of suspicious findings on ultrasonography or mammography or for presurgical staging. These 136 patients (median age 44 years, range 17-83 years; all women) had 149 enhanced lesions (49 benign, 100 malignant) on DCE MRI.

MRI: All examinations were performed on a 3.0-T MR system (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany) with a 16-channel phased-array breast coil. Dynamic imaging was performed by a combination of two different protocols (Figure 1): prototype TWIST-VIBE (temporal resolution: 4.8 s x 22 flames, total scan time 107 s, spatial resolution: 1 mm x 1mm x 3 mm) was obtained through precontrast to the initial contrast phase, and routine DCE sequences (temporal resolution: 60 s, spatial resolution: 1 mm x 1 mm x 1 mm) were measured before and 2 times after TWIST VIBE.

Data analysis: Morphologic and kinetic analyses were performed on routine DCE MRI. The lesions were classified according to the BI-RADS category4. In this study, BI-RADS category 4 was divided into two subcategories: 4a was a lesion with lower suspicion; 4b was a lesion with higher suspicion5. Initial enhancement analysis was performed using TWIST-VIBE. Two new parameters were obtained using non-product TWIST Breast Viewer application (Siemens Healthcare, Enlangen, Germany) : time to enhancement (TTE) and maximum slope (MS). TTE was derived from the time lapse between the beginning of enhancement of the aorta and the lesion. MS was the maximum change of relative enhancement among three time points. The TTEs and MSs of the benign and malignant lesions were compared by the Wilcoxon rank sum test. For the combined assessment, morphologic kinetic features and TTE and/or MS were evaluated together in each lesion type: mass and non-mass enhancement (NME). Diagnostic values of BI-RADS only and combined initial enhancement analysis were compared by the McNemar test.

Mean TTEs of the benign and malignant lesions were 15.08±5.98 s (standard deviation) and 11.42±3.97 s (Figure 2), and MSs were 7.89%/s±5.92%/s and 10.56%/s±5.07%/s (Figure 3), respectively. There were significant differences in both TTE (p < 0.0001) and MS (p = 0.0003) (Figures 2,3). For mass lesions, only TTE showed a significant difference between benign and malignant lesions (p = 0.0024). Ten of 15 (66.7%) BI-RADS category 4a lesions were changed to category 3 in combination with TTE using a cutoff value of 10.63 s, without decreasing sensitivity (100%) (Figure 4). For NME lesions, only MS showed a significant difference between benign and malignant lesions (p = 0.0149). Seven of 8 (87.5%) BI-RADS category 4a lesions were changed to category 3 in combination with MS using a cutoff value of 7.62%/s, without significantly decreasing sensitivity (p = 0.8013, McNemar test) (Figure 5).In this study, the parameters derived from TWIST-VIBE during initial enhancement were found to be valuable for differentiating between benign and malignant breast lesions on MRI. On both mass and NME, the combination of BI-RADS and the new parameters (MS and TTE) has the potential to increase the specificity of breast MRI and may be useful as additional information for determining the need for biopsy.The value of adding initial enhancement analysis using TWIST-VIBE was shown, and we believe that it can further improve the diagnostic accuracy of breast MRI.