KMID : 1114620150120020149
|
|
Journal of the Korean Society for Breast Screening 2015 Volume.12 No. 2 p.149 ~ p.155
|
|
Quantification of Volumetric Breast Density Using MR Images
|
|
Jeyanth Josef Xavier
Kim Won-Hwa Seo Mi-Ri-Nae Moon Woo-Kyung
|
|
Abstract
|
|
|
Objective: To determine whether volumetric breast density (VBD) measured from 3 representative slices correlate with the VBD measured from the all slices of breast MR imaging.
Materials and Methods: Precontrast, T1-weighted, sagittal images of the normal breast obtained in 151 women (29-78 years; mean, 45 years) were used for VBD measurement. The borders of the breasts were outlined manually and fibroglandular tissue was segmented from the fat tissue using fuzzy c-means algorithm. VBD (fibroglandular volume/ total breast volume ¡¿ 100) was measured from the all slices as well as from 3 representative slices [the middle (n), (n+15)th and (n-15)th slices] for each patient. The results were compared using t-test, intraclass correlation coefficient (ICC) and coefficient of variation (CV).
Results: VBDs measured by the two methods showed excellent agreement (ICC = 0.96; 95% CI = 0.95-0.97) and the CV was 9.74%. The mean absolute difference in VBD between the two methods was not different for fatty (< 20% VBD) and dense (> 40% VBD) breasts (2.37% vs. 2.62%; p = 0.67). The 3-slice method showed excellent intra- and inter-observer reproducibility (ICC > 0.9). The mean measurement time for the 3 slice method was significantly lower than that for whole breast measurement (133 sec vs. 572 sec; p < 0.0001).
Conclusion: The VBD measured from the 3 representative slices is in excellent agreement with the VBD measured from the all slices and can serve as an efficient tool for measuring VBD in breast MR imaging.
|
|
KEYWORD
|
|
Breast density, MR imaging, fuzzy c-means algorithm
|
|
FullTexts / Linksout information
|
|
|
|
Listed journal information
|
|
|