High spatial and temporal resolution breast DCE-MRI: A feasibility study.

Abstract

74 Background: The typical 1-2 min low temporal resolution (tRes) of conventional gradient-echo (GRE) breast dynamic contrast-enhanced (DCE) MRI precludes accurate quantitative pharmacokinetic (PK) analysis. The commercially available GRE-based TWIST (time-resolved angiography with stochastic trajectories) sequence employs an alternative k-space sampling scheme to provide high spatial and temporal resolution for dynamic imaging, originally intended for time-resolved MR angiography. Here we sought to evaluate the feasibility of breast DCE-MRI with the TWIST sequence, using both tumor morphology and quantitative PK analyses. Methods: 22 patients with 25 mammography-detected suspicious lesions underwent TWIST DCE-MRI on a 3T Siemens system prior to biopsy. The axial bilateral images were acquired with 20 s tRes and 1.0x1.0x1.4 mm3 spatial resolution (sRes). Immediately after TWIST DCE-MRI, a single DCE image set was acquired in 68 s using a conventional GRE sequence with the same sRes. Blinded to pathology, 3 radiologists compared lesion morphology assessments from the last TWIST DCE and the conventional images, and gave MRI BIRADS scores based on morphology and qualitative kinetic curve analyses. The TWIST DCE lesion signal time-courses were also subjected to Shutter-Speed Model PK analysis. Results: Normal parenchyma S/N ratios were not statistically different between TWIST and conventional GRE images. All readers indicated 88% [95% CI: (68%, 97%)] agreement in morphology evaluation between the TWIST and conventional DCE images. The degree of inter-reader agreement was substantial (κ = 0.75), as was the agreement in BIRADS scores (κ = 0.76). Pathology revealed 8 malignant and 17 benign lesions. Based on MRI BIRADS scores, all 3 readers attained 100% sensitivity and 82% specificity. With PK analysis, a Ktrans (contrast agent plasma/interstitium transfer rate constant) cut-off value of 0.1 min-1 further improved specificity to 94%. Conclusions: TWIST breast DCE-MRI provides both high tRes and sRes for accurate quantitative PK analysis and morphology evaluation, respectively. Utilizing TWIST for clinical breast DCE-MRI with quantitative PK analysis is feasible and may improve MRI interpretation.