Abstract P3-02-03: Accurate and reliable automated breast density measurements with no ionizing radiation using fat-water decomposition MRI

• Published in: Cancer research (Chicago, Ill.) Vol. 77; no. 4_Supplement; pp. P3 - P3-02-03
• Main Authors: Ding, J, Thompson, PA, Gao, Y, Marron, MT, Wertheim, BC, Altbach, MI, Galons, J-P, Roe, DJ, Wang, F, Maskarinec, G, Thomson, CA, Stopeck, A, Huang, C
• Format: Journal Article
• Language: English
• Published: 15.02.2017
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.SABCS16-P3-02-03

Abstract Objective Breast density(BD) is a measure of the distribution of variable tissue types within the breast and higher BD has been shown to positively correlate with breast cancer risk. As such, the accurate measurement of BD has become a priority for risk assessment and for evaluating the effects of prevention strategies aimed at reducing BD. Mammography(MG) is the most common method of BD determination but is limited by the exposure to ionizing radiation, particularly for studies requiring repeated measures. BD derived from fat-water decomposition magnetic resonance imaging(FWMRI-BD) has been proposed as an alternative, safe, and quantitative method for BD. To optimize its use, we developed a new FWMRI-BD that is automated, more accurate and reliable. In this study, we compare our automated method to digital MG and a previous reported algorithm for MRI derived BD. Methods From a completed prevention trial, 42 pre- and post-menopausal patients receiving tamoxifen therapy for early stage breast cancer or as primary chemoprevention were identified. Patients had undergone prior digital MG within 6 months from the date of MRI scan and MG-BD was calculated using a well-established method(Cumulus). MRI scans were performed on a 1.5T GE Signa NV-CV/i scanner using an axial radial IDEAL-GRASE sequence to generate quantitative fat fraction maps of the entire breast. Total acquisition time was < 5 min and automated breast segmentation was applied to all scans. Only the contralateral, unaffected breast was analyzed. Pearson correlation analysis compared BD as measured by MG(range 0-100%) and FWMRI based methods. BD by FWMRI was initially calculated as the ratio of breast voxels with<80% apparent fat fraction(Fra80). Fra80 had been previously shown by our group to correlate with MG-BD(Spearman ρ=0.86, p<0.001). Here, BD was calculated using a new algorithm(FraG+W) that accounts for the total amount of fibroglandular tissue and water content in the breast after correction for fat-water signal intensity bias and fat-water signal shine-through. Reliability of FWMRI measurements was tested in 24 repeated scans from 9 patients and evaluated using intra-class correlation(ICC) analysis. Results Table 1 shows the correlation and reliability analysis results between MG-BD and FWMRI-BD. Both FWMRI-BD measures(Fra80 and FraG+W) were strongly correlated with MG-BD. More importantly, they exhibit superior test-retest reliability(ICC>0.98) compared to MG-BD values from the literature(reported ICC range 0.91-0.95). FraG+W showed improvement over Fra80 in all measures tested including correlation to MG-BD, dynamic range, standard errors and ICC. Table 1. Accuracy and Reliability of the FWMRI-BD measuresFWMRI-BDFra80FraG+WPearson correlation coefficient* with MG-BDR=0.86R=0.94Test-retest reliabilitystandard error0.02300.0134dynamic range0.0902 – 0.65370.0736 – 0.6588standard error/ dynamic range4.1%2.3%ICC [95% confidence interval]0.985 [0.966,0.993]0.990 [0.976,0.995]* All P-values < 1e-10 Conclusion The refined and automated FWMRI-BD that quantifies the entire fibroglandular and water content of the breast(FraG+W) strongly correlates with MG-BD and is more accurate and reliable than previous FWMRI-BD method. Acknowledgement NIH grants CA149417, CA161534. Citation Format: Ding J, Thompson PA, Gao Y, Marron MT, Wertheim BC, Altbach MI, Galons J-P, Roe DJ, Wang F, Maskarinec G, Thomson CA, Stopeck A, Huang C. Accurate and reliable automated breast density measurements with no ionizing radiation using fat-water decomposition MRI [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P3-02-03.