Purpose: To compare performance of contrast-enhanced mammography (CEM) in patients with marked background parenchymal enhancement (BPE) versus minimal BPE.
Materials and Methods: Our IRB-approved HIPAA-compliant study included consecutive screening CEMs performed at our institution from January 2015-December 2021. CEM, with dual-energy mammography, was acquired 2-2.5 min following contrast injection. BPE categories (i.e. Marked, Moderate, Mild, Minimal) were extracted from radiology reports. Abnormal interpretation rate (AIR) was calculated as the proportion of exams assigned BI-RADS 0/3/4/5. Exams with suspicious/indeterminate contrast-only findings (i.e. no suspicious correlate on mammography and ultrasound) were referred to MRI. Positive predictive value (PPV3) was calculated, on a per exam basis, as the number of cancers diagnosed on MRI-guided biopsy over all CEM patients who underwent MR-guided biopsy (i.e. contrast-only CEM findings). Metrics were calculated across all exams, and for the Marked and Minimal BPE subgroups. Chi-squared tests were performed to evaluate for differences in AIR across BPE subgroups, and for differences in PPV3 across the BPE subgroups.
Results: Of the 8,015 CEMs in 3,216 patients included in the study, 4.0% (322) exams had Marked BPE, 19.3% (1549) had Moderate BPE, 45.5% (3646) had Mild BPE and 31.3% (2498) had Minimal BPE. The AIR was 14.8% (95% CI: 14.1-15.6) across all exams, 29.2% (95% CI: 24.2-34.2) in Marked BPE exams, and 8.4% (95% CI: 7.4-9.5) in Minimal BPE exams. A chi-squared test shows that the AIR was increased in Marked versus Minimal BPE exams (p < 0.001). See Figure 1a.
The PPV3 of CEM patients who underwent MR-guided biopsy was 26.3% (95% CI: 19.6-33.0) across all exams, 7.7% (95% CI: -6.7-22.1) across Marked BPE exams, and 46.7% (95% CI: 28.8-64.5) across Minimal BPE exams. A chi-squared test shows that the PPV3 was decreased in Marked versus Minimal BPE exams (p = 0.03). See Figure 1b.
Conclusion: CEM exams with marked BPE are more likely to be interpreted as abnormal. When MRI is performed for further evaluation of an indeterminate or suspicious enhancing abnormality, patients with marked BPE are also more likely to undergo an MR-guided biopsy with a benign result. Further work is needed to determine if CEM acquisition techniques can be modified to improve performance in women with marked BPE.
Clinical Relevance Statement: Our study demonstrates that contrast-enhanced mammography performs better in patients with minimal versus marked BPE, an important point as the clinical implementation of CEM continues to evolve and expand. The improvement of CEM acquisition techniques in marked BPE patients should also be explored.
