Eun-Kyung Kim, Presenter: Nothing to Disclose

Hye-Jeong Lee MD, Abstract Co-Author: Nothing to Disclose

Min Jung Kim, Abstract Co-Author: Nothing to Disclose

Ji Hyun Youk MD, Abstract Co-Author: Nothing to Disclose

Ji Young Lee, Abstract Co-Author: Nothing to Disclose

Dae Ryong Kang PhD, Abstract Co-Author: Nothing to Disclose

Chung Mo Nam PhD, Abstract Co-Author: Nothing to Disclose

Ki Keun Oh, Abstract Co-Author: Nothing to Disclose

et al, Abstract Co-Author: Nothing to Disclose

To retrospectively evaluate inter- and intraobserver variability between breast radiologists in sonographic feature analysis and management according to the fourth edition of the Breast Imaging Reporting and Data System(BI-RADS).

One hundred thirty-six patients with 150 breast lesions who underwent breast US and core needle biopsy were included. A pathologic diagnosis was available for all 150 lesions and 77(51%) were malignant and 73(49%) were benign. The size of the lesions ranged from 3mm to 45mm (mean, 14.1mm). Four radiologists retrospectively reviewed these lesions twice with 8 weeks interval. Each observer described each lesion with BI-RADS descriptors and final assessment. Inter- and intraobserver variability was assessed with the cohen κ static. Positive predictive value, negative predictive value, sensitivity and specificity for each observer was also calculated.

Interobserver agreements for ultrasonographic descriptors were as follows : substantial agreement was obtained for lesion calcification and final assessment(κ=0.61 for both). Moderate agreement was obtained for lesion shape, orientation, boundary and posterior acoustic features (κ=0.49, 0.56, 0.59 and 0.49, respectively). Fair agreement was achieved for lesion margin and echo pattern (κ=0.33 and 0.37, respectively). About intraobserver agreement, substantial to perfect agreement was found for almost lesion descriptors and final assessment. We also evaluated inter- and intraobserver agreement for each descriptors, according to the mass size(less than 1cm versus more than 1cm). However, there was no significant difference in inter- and intraobserver agreement, according to the mass size. With the results of all observers combined, positive predictive value, negative predictive value, sensitivity and specificity were 60%, 95%, 98% and 33%, respectively in this study.

Inter- and intraobserver agreements with the BI-RADS descriptors and final assessment category for sonography are good and validates the US lexicon.

Inter- and intraobserver agreements with the BI-RADS lexicon for sonography are good, so the use of BI-RADS lexicon can make accurate and consistent breast sonography.

Kim, E, Lee, H, Kim, M, Youk, J, Lee, J, Kang, D, Nam, C, Oh, K, et al, , BI-RADS (Breast Imaging Reporting and Data System) for Breast Ultrasound: Inter- and Intraobserver Variability.  Radiological Society of North America 2006 Scientific Assembly and Annual Meeting, November 26 - December 1, 2006 ,Chicago IL. http://archive.rsna.org/2006/4439860.html