Contrast-Enhanced Mammography in the Work-up of Asymmetries: A Systematic Review and Meta-analysis

Contrast-enhanced mammography (CEM) is increasingly used to evaluate mammographic or digital breast tomosynthesis (DBT) detected asymmetries. We performed a systematic review and meta-analysis to estimate the diagnostic performance of CEM in this setting, focusing on the negative predictive value (NPV) of non-enhancing asymmetries.

MEDLINE and Embase were searched through October 2025 for studies reporting outcomes of CEM in mammographic/DBT-detected asymmetries with histopathology and/or imaging follow-up. Study selection and data extraction were performed by two readers. Random-effects models with Hartung-Knapp adjustment were used for pooled estimates. Exact aggregated binomial confidence intervals were calculated for zero-event analyses. Risk of bias was assessed using QUADAS-2.

Four retrospective single-center studies met inclusion criteria. Across 147 non-enhancing asymmetries (NEAs) with definitive outcomes, no false-negative cancers were observed. Study-level NPVs were 100%. Random-effects pooling yielded an NPV of 0.98 (95% CI, 0.85-0.997; I² = 0%). Exact aggregated binomial analysis demonstrated an NPV of 100% with a conservative 95% lower bound of 97.5%. Enhancing asymmetries, particularly mass-type and suspicious non-mass enhancement, were strongly associated with malignancy. Accuracy and positive predictive value (PPV) varied by study setting and prevalence.

In appropriately selected mammographic/DBT asymmetries, absence of enhancement on CEM is associated with an extremely low malignancy risk and may support surveillance strategies, whereas enhancement markedly increases malignant probability and refines risk stratification.