Characterization of True and False Positive Findings on Contrast-Enhanced Mammography.
Acad Radiol
; 2022 Feb 18.
Article
em En
| MEDLINE
| ID: mdl-35190261
ABSTRACT
RATIONALE AND OBJECTIVES:
The purpose of this paper is to characterize true and false positive findings on contrast-enhanced mammography (CEM) and correlate enhancement pattern and method of detection with pathology outcomes. MATERIALS ANDMETHODS:
This was an IRB-approved retrospective review of diagnostic CEM performed from December 2015 through December 2019 for which biopsy was recommended. Background parenchymal enhancement, tissue density, finding features, pathologic/clinical outcomes, and method of detection were captured. CEM includes low-energy images (LE), similar to standard 2D mammography, and recombined images (RI) that show enhancement. 'MG-detected' findings were identified on mammography or LE. 'RI-detected' findings were identified due to enhancement on RI. The positive predictive value (PPV2) was calculated on a per-case and a per-finding level. Comparisons were performed using Pearson chi-square and Fisher exact tests.RESULTS:
One hundred sixty CEM cases with 220 findings were evaluated with a case PPV2 of 58.1%. 32.3% (71/220) of lesions were RI-detected. The PPV2 of RI-detected enhancement was 40.8% with subanalysis revealing PPV2 of 22.2%, 32%, and 51.4% for foci, NME, and masses, respectively. The PPV2 of MG-detected enhancement was 73.5% with subanalysis revealing PPV2 of 50%, 54.1%, and 83.8% for foci, NME, and masses, respectively. There were 100 false positives findings, 42 of which were RI-detected.CONCLUSION:
PPV2 of diagnostic CEM is within the range of other diagnostic breast imaging exams. However false positives remain a challenge, especially for RI-detected findings. Additional efforts to improve specificity of RI-detected findings are worthwhile.
Texto completo:
1
Coleções:
01-internacional
Base de dados:
MEDLINE
Tipo de estudo:
Diagnostic_studies
/
Prognostic_studies
Idioma:
En
Revista:
Acad Radiol
Assunto da revista:
RADIOLOGIA
Ano de publicação:
2022
Tipo de documento:
Article