Abstract A48: Pitfalls of using breast magnetic resonance imaging (MRI) in an early-phase breast cancer chemoprevention trial of vitamin D among high-risk postmenopausal women

Abstract

Abstract Background: Mammographic density (MD), which refers to the relative proportions of radiodense fibroglandular tissue and radiolucent fat seen on a mammogram, is one of the strongest predictors of breast cancer risk. A decrease in MD in high-risk women within a year of starting tamoxifen has been correlated with a reduction in breast cancer incidence. Therefore, a number of early phase breast cancer chemoprevention trials are using MD as an intermediate endpoint for clinical benefit. Breast MRI can assess the volume of fibroglandular tissue within the breast as a 3-dimensional object. Although less well-established compared to MD, breast MRI volume may be a more sensitive image-based biomarker for detecting changes over time. As a screening tool for high-risk women, MRI has a higher sensitivity for detecting breast cancer compared to mammography, but with the potential for more false positive results which may lead to additional scans and biopsies. We conducted an evaluation of changes in breast density on mammography and MRI in a 1-year intervention trial of high-dose vitamin D in high-risk postmenopausal women. Methods: This pilot study included 20 high-risk postmenopausal women, defined as 5-year Gail risk score ≥1.67%, lobular or ductal carcinoma in situ (LCIS/DCIS), BRCA1/BRCA2 mutation carrier, or history of stage I-II breast cancer in remission for >5 years. Other eligibility criteria included baseline MD ≥25%, serum 25-hydroxyvitamin D (25-OHD) ≤32 ng/ml and no history of kidney stones. Women were assigned to a 1-year intervention of vitamin D3 20,000 IU or 30,000 IU PO weekly. Participants underwent a mammogram and breast MRI at baseline and 12 months, as well as serial blood collections every 3 months. We assessed the results of mammograms and breast MRIs during the trial and the need for follow-up breast imaging and biopsy. Results: From January 2009 to July 2011, 152 women were screened, 86 were eligible (most common reason for ineligibility was 25-OHD >32 ng/ml), 26 consented and underwent baseline breast imaging (including 2 screen failures and 4 who withdrew consent). Median age: 59 (49–73); White/Hispanic/Black/Asian: 14/10/1/1; median body mass index: 28.8 kg/m2 (19.4–38.9); breast cancer risk category, Gail/LCIS/DCIS/stage I-II breast cancer: 12/4/8/2; median serum 25-OHD: 23 ng/ml (8–31); baseline MD, scattered fibroglandular densities/heterogeneously dense: 16/10. For 85% of women, the baseline study MRI represented their first breast MRI. A total of 46 standard-of-care or study mammograms (BIRADS 0/1/2/3/4: 1/4/28/9/4) and 45 breast MRIs (BIRADS 0/1/2/3/4: 3/1/22/13/6) were performed. Ten women required follow-up breast imaging. A total of 12 biopsies were performed on 7 women, including 2 fine needle aspirations, 8 core biopsies, and 2 excisional biopsies for atypia and an intraductal papilloma. No biopsies showed evidence of malignancy. Discussion: Among high-risk postmenopausal women participating in a vitamin D intervention trial with serial breast imaging, there was a high rate of false positive results on mammograms and breast MRIs and need for additional breast procedures. Use of breast MRI volume as a modifiable biomarker of breast cancer risk in early phase chemoprevention trials may be limited by cost and the high frequency of incidental findings. Citation Information: Cancer Prev Res 2011;4(10 Suppl):A48.