arXiv.itsnick.co

an independent field guide to 23 million papers from arXiv & OpenAlex — not affiliated with arXiv.org

Computer Science · Artificial Intelligence

Interobserver variability for characteristics of breast lesions between junior and senior radiologists with BI-RADS

HE Ping-qing

See on the map

Abstract

Objective To retrospectively evaluate interobserver variability for lesion features on mammogram(MMG) and ultrasonography(US) between junior and senior radiologists using terminology of the breast imaging reporting and data System(BI-RADS) . Methods The MMG and US images were obtained by an experienced breast doctor,respectively. Four breast radiologists independently reviewed all the 441 lesions on US and MMG images,respectively. Each observer described each lesion with BI-RADS terminology and recorded lesion features,respectively. Interobserver variability for lesion features on US and MMG were obtained and analyzed with the Cohen's statistic. Results ①For MMG descriptors,substantial agreement was obtained for lesion presence or absence(K=0.71) ,moderate agreement for lesion shape and margin(K=0.45,0.47,respectively) ,fair agreement for lesion density(K=0.26) . Substantial agreement was obtained for calcification presence or absence and calcification morphology(K=0.73,0.78,respectively) ;②For US descriptors,substantial agreement was obtained for lesion shape,moderate agreement for lesion pseudocapsule and echo patter(K=0.57,0.41,respectively) ,Fair agreement lesion margin and posterior acoustic feature(K=0.35,0.40,respectively) . Conclusion Interobserver variability are substantial and moderate for most lesion features on MMG,and moderate for most lesion features on US;thus indicates that BI-RADS descriptors can express the lesion characteristics in details and is less influenced by radiologists’ experience.

More like this