M Rominger 1 , C Wisgickl , N Timmesfeld . Show Affiliations »
Abstract
PURPOSE: To use a systematic review and meta-analysis to determine the predictive value of five suspicious type descriptors of microcalcifications, with special focus on round/punctate microcalcifications because of controversy. MATERIALS AND METHODS: We performed a literature search using three databases and one search engine. We screened citations for malignancy rates of type descriptors. We regarded malignancy rates below 2% (BI-RADS 3) and above 95% (BI-RADS 5) as clinically significant threshold values for suspicion (BI-RADS 4). RESULTS: Forty studies from 14 countries with a total of 10,665 microcalcification lesions were included. The pooled malignancy rates were: coarse heterogeneous, 13% (95% confidence interval (95%-CI): 7 - 20%); amorphous or indistinct, 27% (95%-CI: 21-33%); pleomorphic, 50% (95%-CI: 43-58%); linear, 78% (95%-CI: 68 - 86%). The pooled malignancy rate of all round/punctate microcalcifications was 9% (95%-CI: 6-13%), for the subgroup follow-up it was 0,5% (95%-CI: 0.08-2.57%), and with histological verification it was 14% (95%-CI: 11-19%). Woman selection and consideration of additional suspicious image findings were reasons for between-study heterogeneity. Addition of ACR distribution descriptors diversified risk stratification, but did not alter BI-RADS assessment category. CONCLUSION: All suspicious type descriptors including round/punctate microcalcifications as well as combinations of type with suspicious distribution descriptors fell into BI-RADS assessment category 4. Exclusion of suspicious adjunct factors can direct clustered, round/punctate microcalcifications to a lower BI-RADS assessment category. © Georg Thieme Verlag KG Stuttgart · New York.
PURPOSE: To use a systematic review and meta-analysis to determine the predictive value of five suspicious type descriptors of microcalcifications, with special focus on round/punctate microcalcifications because of controversy. MATERIALS AND METHODS: We performed a literature search using three databases and one search engine. We screened citations for malignancy rates of type descriptors. We regarded malignancy rates below 2% (BI-RADS 3) and above 95% (BI-RADS 5) as clinically significant threshold values for suspicion (BI-RADS 4). RESULTS: Forty studies from 14 countries with a total of 10,665 microcalcification lesions were included. The pooled malignancy rates were: coarse heterogeneous, 13% (95% confidence interval (95%-CI): 7 - 20%); amorphous or indistinct, 27% (95%-CI: 21-33%); pleomorphic, 50% (95%-CI: 43-58%); linear, 78% (95%-CI: 68 - 86%). The pooled malignancy rate of all round/punctate microcalcifications was 9% (95%-CI: 6-13%), for the subgroup follow-up it was 0,5% (95%-CI: 0.08-2.57%), and with histological verification it was 14% (95%-CI: 11-19%). Woman selection and consideration of additional suspicious image findings were reasons for between-study heterogeneity. Addition of ACR distribution descriptors diversified risk stratification, but did not alter BI-RADS assessment category. CONCLUSION: All suspicious type descriptors including round/punctate microcalcifications as well as combinations of type with suspicious distribution descriptors fell into BI-RADS assessment category 4. Exclusion of suspicious adjunct factors can direct clustered, round/punctate microcalcifications to a lower BI-RADS assessment category. © Georg Thieme Verlag KG Stuttgart · New York.
Entities: Disease
Species
Mesh: See more »
Year: 2012
PMID: 22923222 DOI: 10.1055/s-0032-1313102
Source DB: PubMed Journal: Rofo ISSN: 1438-9010