Return To Abstract Listing

Scientific Session 08 — Breast - Ultrasound

Tuesday, May 2, 2017

Abstracts 2395-3218

2815. Impact of Age and Density on the Performance of Mammography and Ultrasound in Diagnostic Imaging

Blaschke E1*,  Hughes K1,  Roark A2,  Smith B1,  Lehman C1 1. Massachusetts General Hospital, Boston, MA; 2. Baylor College of Medicine, Houston, TX

Address correspondence to E. Blaschke (

Objective: Although the impact of breast density on screening mammography performance is well documented, the American College of Radiology (ACR) Appropriateness Criteria recommend differential imaging evaluation of patients experiencing symptoms on the basis of their age. We measured the impact of age and density on diagnostic mammography and ultrasound (US) performance in patients presenting with a focal area of clinical concern.

Materials and Methods: After institutional review board approval, review of medical records from March 2006 to March 2015 identified all patients presenting for diagnostic mammography and US for an area of focal clinical concern. Each breast with the area or areas of clinical concern was designated as a case. Outcomes were determined by imaging, biopsy, or any pathology findings in our hospital tumor registry within a minimum 12-month follow-up. Performance measures were defined according to the 5th edition of BI-RADS. Modality-specific outcomes of diagnostic mammography and US were determined for lesions marked “seen” or “not seen” by the radiologist. Breast density was grouped as dense (heterogeneously dense or extremely dense) or nondense (fatty tissue or scattered fibroglandular tissue). Logistic regression and chi-square analyses were used to assess impact of age and density on sensitivity and specificity of diagnostic mammography, US, and the combination of both techniques.

Results: In all, 607 cancers out of 18,135 cases in 14,500 women (mean age, 46.3 years; range, 20–102 years) were diagnosed for a cancer yield of 41.9 per 1000 women. Sensitivity, specificity, and negative predictive values were 80.9%, 96.0%, and 99.3% for diagnostic mammography; 84.7%, 95.1%, and 99.4% for US, and 93.2%, 94.4%, and 99.8% for the two combined. After adjusting for density, sensitivities of diagnostic mammography, US, and the combined techniques were not significantly impacted by age (p > 0.35 for all). After adjusting for age, density was associated with significantly decreased sensitivity for diagnostic mammography (p = 0.0003); however, sensitivities of US and the combined techniques were not impacted by density after adjusting for age (p > 0.10 for both). Age and density were independent predictors of specificity of diagnostic mammography, US, and the combined techniques (p < 0.0001 for all).

Conclusion: In patients presenting for evaluation of a focal area of clinical concern in the breast, diagnostic mammography and US have complementary roles in cancer detection. Breast density, not age, is the significant predictor of sensitivity of mammography. Our findings support the ACR Appropriateness Criteria for management of patients with symptoms of breast cancer and highlight the importance of emphasis on breast density in future guideline development.