Objectives: The purposes of this study were to introduce and assess an automated user-independent quantitative volumetric (AUQV) breast density (BD) measurement system on the basis of magnetic resonance imaging (MRI) using the Dixon technique as well as to compare it with qualitative and quantitative mammographic (MG) BD measurements.
Materials and Methods: Forty-three women with normal mammogram results (Breast Imaging Reporting and Data System 1) were included in this institutional review board–approved prospective study. All participants were subjected to BD assessment with MRI using the following sequence with the Dixon technique (echo time/echo time, 6 milliseconds/2.45 milliseconds/2.67 milliseconds; 1-mm isotropic; 3 minutes 38 seconds). To test the reproducibility, a second MRI after patient repositioning was performed. The AUQV magnetic resonance (MR) BD measurement system automatically calculated percentage (%) BD. The qualitative BD assessment was performed using the American College of Radiology Breast Imaging Reporting and Data System BD categories. Quantitative BD was estimated semiautomatically using the thresholding technique Cumulus4. Appropriate statistical tests were used to assess the agreement between the AUQV MR measurements and to compare them with qualitative and quantitative MG BD estimations.
Results: The AUQV MR BD measurements were successfully performed in all 43 women. There was a nearly perfect agreement of AUQV MR BD measurements between the 2 MR examinations for % BD (P < 0.001; intraclass correlation coefficient, 0.998) with no significant differences (P = 0.384). The AUQV MR BD measurements were significantly lower than quantitative and qualitative MG BD assessment (P < 0.001).
Conclusions: The AUQV MR BD measurement system allows a fully automated, user-independent, robust, reproducible, as well as radiation- and compression-free volumetric quantitative BD assessment through different levels of BD. The AUQV MR BD measurements were significantly lower than the currently used qualitative and quantitative MG-based approaches, implying that the current assessment might overestimate breast density with MG.
From the *Division of Molecular and Gender Imaging, †Computational Imaging Research Lab, ‡MR Centre of Excellence, Department of Biomedical Imaging and Image-guided Therapy, Medical University of Vienna, Vienna, Austria.
Received for publication April 10, 2014; and accepted for publication, after revision, August 6, 2014.
Preliminary results of the study were presented at RSNA 2012.
Conflicts of interest and sources of funding: none declared.
Reprints: Thomas H. Helbich, MD, Division of Molecular and Gender Imaging, Department of Biomedical Imaging and Image-guided Therapy, Medical University of Vienna/Vienna General Hospital, Waehringer Guertel 18-20, 1090 Vienna, Austria. E-mail: firstname.lastname@example.org.