Cross-field strength and multi-vendor reliability of MagDensity for MRI-based quantitative breast density analysis

Abstract

Purpose: Breast density (BD) is a significant risk factor for breast cancer, yet current assessment methods lack automation, quantification, and cross-platform consistency. This study aims to evaluate the reliability and cross-platform consistency of MagDensity, a novel magnetic resonance imaging (MRI)-based quantitative BD measure, across different imaging platforms.

Methods: Ten healthy volunteers participated in this prospective study, undergoing fat-water MRI scans on three scanners: 3T Siemens Prisma, 3T Siemens Biograph mMR, and 1.5T GE Signa. Great effort was made to schedule all scans within a narrow three-hour window on the same day to minimize any potential intra- or inter-day variations, requiring substantial logistical coordination. BD was assessed using the MagDensity technique, which included combining magnitude and phase images, applying a fat-water separation technique, employing an automated whole-breast segmentation algorithm, and quantifying the volumetric water fraction. Agreement between measures across scanners was analyzed using mean differences, two-tailed t-tests, Pearson's correlation, and Bland-Altman analysis.

Results: MagDensity measures obtained from the two 3T Siemens scanners demonstrated no statistically significant differences, with high correlation (Pearson's r > 0.99) and negligible mean differences (< 0.2%). Cross-platform comparison between the 3T Siemens and the 1.5T GE scanners showed larger mean differences (< 4.2%). However, after applying linear calibration, these variations were reduced to within ±0.2%, with strong inter-scanner correlation maintained (Pearson's r > 0.97).

Conclusion: MagDensity showed strong intra-vendor consistency and promising cross-platform reliability after leave-one-out calibration. While full standardization remains a long-term goal, these findings provide clear evidence that scanner-related variability can be effectively mitigated through calibration. This technique offers a step further toward more consistent MRI-based BD quantification and may help enable broader clinical implementation.

Fig 1. Multi-echo magnitude images from different scanners.

Multi-echo magnitude images acquired from a representative participant, using 3T Siemens Prisma (top row), 3T Siemens Biograph mMR (middle row), and 1.5T GE Signa (bottom row).

Fig 2. Reconstructed images using the multi-echo data.

Reconstructed images from a representative participant, including fat-only images (left column), water-only images (middle column), and regions of interest on the fat fraction maps (right column) used for MagDensity calculations.

Fig 3. Pearson’s correlation and Bland–Altman analyses between the MagDensity measures of 3T Siemens Prisma and 3T Siemens Biograph mMR.

Pairwise t-test showed no statistically significant difference between the MagDensity measures (p > 0.05) with a Pearson coefficient larger than 0.99 (p < 0.001). The Bland-Altman analysis showed a mean bias of −0.16% with 95% limits of agreement between −4.8% and 4.4%.

Fig 4. Pearson’s correlation and Bland-Altman analyses between the MagDensity measures of 3T Siemens Biograph mMR/Prisma vs. 1.5T GE Signa (before and after calibration).

After calibration, pairwise t-test showed no statistically significant difference between the MagDensity measures between 3T Siemens Biograph mMR/Prisma and 1.5T GE Signa (p > 0.05) with Pearson coefficients larger than 0.97 (p < 0.001). The calibration was performed in a leave-one-out manner. For measures between 3T Siemens Biograph mMR and 1.5T GE Signa, the Bland-Altman analysis showed a mean bias of −0.01% with 95% limits of agreement between −5.5% and 5.5%. For measures between 3T Siemens Prisma and 1.5T GE Signa, the Bland-Altman analysis showed a mean bias of 0.13% with 95% limits of agreement between −7.9% and 8.1%.