Format

Send to

Choose Destination
Radiology. 2018 Dec;289(3):630-638. doi: 10.1148/radiol.2018181180. Epub 2018 Oct 2.

Microcalcifications Detected at Screening Mammography: Synthetic Mammography and Digital Breast Tomosynthesis versus Digital Mammography.

Author information

1
From the Department of Radiology, Taipei Veterans General Hospital, Taipei, Taiwan (Y.C.L.); School of Medicine, National Yang-Ming University, Taipei, Taiwan (Y.C.L.); Department of Radiology and Biomedical Imaging, University of California San Francisco, 1600 Divisadero St, Box 1667, Room C250, San Francisco, CA 94115 (A.Y.L., J.H.H., R.I.F., I.V.L., B.N.J.); and Department of Radiology, The Permanente Medical Group, 3600 Broadway, Oakland, CA 94611 (K.M.R.).

Abstract

Purpose To compare the performance of two-dimensional synthetic mammography (SM) plus digital breast tomosynthesis (DBT) versus conventional full-field digital mammography (FFDM) in the detection of microcalcifications on screening mammograms. Materials and Methods In this retrospective multireader observer study, 72 consecutive screening mammograms recalled for microcalcifications from June 2015 through August 2016 were evaluated with both FFDM and DBT. The data set included 54 mammograms with benign microcalcifications and 18 mammograms with malignant microcalcifications, and 20 additional screening mammograms without microcalcifications used as controls. FFDM alone was compared to synthetic mammography plus DBT. Four readers independently reviewed each data set and microcalcification recalls were tabulated. Sensitivity and specificity for microcalcification detection were calculated for SM plus DBT and for FFDM alone. Interreader agreement was calculated with Fleiss kappa values. Results Reader agreement was kappa value of 0.66 (P < .001) for FFDM and 0.63 (P < .001) for SM plus DBT. For FFDM, the combined reader sensitivity for all microcalcifications was 80% (229 of 288; 95% confidence interval [CI]: 74%, 84%) and for malignant microcalcifications was 92% (66 of 72; 95% CI: 83%, 97%). For SM plus DBT, the combined reader sensitivity for all microcalcifications was 75% (215 of 288; 95% CI: 69%, 80%) and for malignant microcalcifications was 94% (68 of 72; 95% CI: 86%, 98%). For FFDM, the combined reader specificity for all microcalcifications was 98% (78 of 80; 95% CI: 91%, 100%) and for malignant microcalcifications was 98% (78 of 80; 95% CI: 91%, 100%). For SM plus DBT, combined reader specificity for all microcalcifications was 95% (76 of 80; 95% CI: 88%, 99%) and for malignant microcalcifications was 95% (76 of 80; 95% CI: 88%, 99%). Mixed-effects model concluded no differences between modalities (‒0.03; 95% CI: ‒0.08, 0.01; P = .13). Conclusion Relative to full-field digital mammography, synthetic mammography plus digital breast tomosynthesis had similar sensitivity and specificity for the detection of microcalcifications previously identified for recall at screening mammography.

PMID:
30277445
DOI:
10.1148/radiol.2018181180
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Atypon
Loading ...
Support Center