유방밀도평가: 영상의학과의사의육안평가와 Volpara유방밀도의비교

Abstract

Purpose : To retrospectively compare Volpara mammographic density estimation with the density by radiologists based on the Breast Imaging Reporting and Data System (BI-RADS) density classification.

Materials and Methods : Materials and methods: From twentieth December 2011 to twenty fourth December 2011, 116 patients aged 25-74 years who had mammography with Volpara density grade were included in this study. We reviewed the patient’s breast density grade of BI-RADS categories by radiologists and Volpara. Weighted kappa statistic was used in measuring agreement between radiologists and Volpara imaging software. Mann Whitney test was used for comparison total breast volume (mL), fibroglandular tissue volume (mL), and degree of breast density (%) in group with technical error of Volpara imaging software and group without error.

Results : Overall agreement of breast density between the radiologist’s BI-RADS and Volpara imaging software was moderate (k=0.428). Five (4.3%) patients had technical error of Volpara imaging software in breast density measurements. Statistical analysis showed a significant difference in total breast tissue volume (mL) and degree of breast density (%) when comparing group with technical error of Volpara software and group without technical error of that method (p=0.0301, p=0.0067).

Conclusion : The agreement of radiologists and Volpara imaging software was moderate (kappa value = 0.428). Small total breast volume (mL) and high degree of breast density (%) were causes of the Volpara softwear error.