Target Audience

Basic and clinical scientists interested in quantitative imaging and breast MRI.

Introduction

3D-Magnetic Resonance Fingerprinting (MRF) is a recently developed technique which provides volumetric breast coverage and allows simultaneous mapping of T₁ and T₂ relaxation times of breast tissue in a clinically feasible scan¹. The T₁ and T₂ relaxation times for normal breast tissue obtained from 3D MRF have been comparable to literature values at 3T². Clinically, measurements obtained from 3D MRF may be useful in lesion characterization or assessment of treatment response. However, in order to interpret the T₁ and T₂ in tumors and change after treatment, it is important to understand the repeatability of this technique in normal breast tissue. The purpose of this work is to evaluate 3D breast MRF test-retest repeatability in the same visit and across two different visits in normal breast tissue.

Methods

In this IRB approved study, 12 healthy pre-menopausal volunteers were scanned with 3D MRF twice within 7-15 days. In addition, 5 volunteers also consented to a test-retest scan in the same sitting, in which 3D MRF was done before and after repositioning with a 10 minute scan interval. All scans were performed on a 3T scanner (Verio, Siemens) using an 8-channel breast coil. The 3D MRF sequence was based on a FISP acquisition with a fat-saturation module¹. Other imaging parameters included: FOV=40×40 cm; matrix size 256×256; slice thickness 3 mm; number of partitions, 48; partial Fourier in the partition direction, 6/8. The overall acquisition time for 48 partitions was 5.5 minutes. The 3D MRF partition with the largest area of normal breast tissue in that partition was selected. Region of Interests (ROIs) were manually drawn by one radiologist (7 years radiology experience). Separate ROIs were drawn for right and left breast (Figure 1). For retest and repeatability analysis, ROIs were drawn on the partition showing anatomy of breast tissue as the reference scan. The mean T₁ and T₂ values from all voxels within each ROI were obtained. For both test-retest (n = 5) and two-visit (n = 12) repeatability, the between-subject coefficient of variance (bCV), and within-subject coefficient of variance (wCV) were computed for both breasts. Intraclass correlation coefficient (ICC) and coefficient of repeatability (CR) were also estimated for both breasts.

Results

The mean age was 28 years, median age 29 years, range 22-33 years. Table 1 summarizes all results for two-visit as well as test-retest repeatability analysis. Figure 2 shows relationship of T₁ and T₂ values between visit 1 and visit 2. Figure 3 shows Bland-Altman plots for two-visit repeatability for all ROIs in both breasts. While bCV was in the range of 10-15% (Table 1), the wCV was 5.5-5.7% for T₁ and 4.3-4.8% for T₂. Agreement was good-to-excellent for both T₁ and T₂ (ICC > 0.75). Figure 3 shows test-retest relationship for T₁ and T₂ for both breasts. Again the bCV was higher but the wCV was within 3-7% for T₁ and T₂ (Table 1).

Discussion

This work shows the repeatability of 3D MRF measurements in a single-institution setting. The wCV for breast MRF is comparable or smaller than wCV of ADC (5-11%) reported previously^3,4,5. The higher bCV indicates a larger inter-subject variation either due to difference in breast tissue composition or scanning in different phases of hormonal cycles across different subjects.. While repeat scans were not confined to a particular phase of menstrual cycle, the high ICC and low wCV suggests that the values are repeatable across different cycles, though further work is needed. Thus 3D MRF may be useful in longitudinal follow-up of individual subjects and affect size greater than 7% ought to be measurable in longitudinal follow-ups.

Conclusion

3D MRF measurements in normal breast tissue are repeatable across various scan timings and may be useful in clinical applications in breast imaging.

Acknowledgements

Research Support: NIH grants 1R01DK098503, 1R00EB016728 and Siemens Healthineers

References

1. Chen Y, et al. 3D Magnetic Resonance Fingerprinting for Quantitative Breast Imaging,, Abstract 0399, 24th ISMRM, 2016. 2. Rakow-Penner R, et al. Relaxation times of breast tissue at 1.5T and 3T measured using IDEAL. JMRI. 2006;23:87-91. 3. Aliu S, et al. Repeatability of Quantitative MRI Measurements in Normal Breast Tissue. Trans Oncol. 2014:7:130-137. 4. O’Flynn EAM, et al. Diffusion weighted imaging of the normal breast: reproducibility of apparent diffusion coefficient measurements and variation with menstrual cycle and menopausal status. Eur Radiol. 2012. 22; 1512-1518. 5 Newitt DC, et al. Reproducibility of ADC measurements by Breast DWI: Results of the ACRIN 6698 Trial. Abstract 0949, 25th ISMRM 2017.