Return To Abstract Listing

Scientific Session 17 — Breast -MRI/New Technologies

Wednesday, May 3, 2017

Abstracts 1031-2861

2376. Suspicious Breast MRI Lesions Assigned to MRI Follow-Up: Was Outcome Negatively Affected?

Lavista F1*,  Milans S1,  Leung J2 1. Republic University School of Medicine, Montevideo, Uruguay; 2. The University of Texas MD Anderson Cancer Center, Houston, TX

Address correspondence to J. Leung (

Objective: In Uruguay, a country with limited MRI resources, MRI-guided biopsy is not available. This study was performed to determine whether the delay in MRI-guided biopsy of breast lesions that were assessed as suspicious or highly suggestive of malignancy at MRI negatively impacts outcome.

Materials and Methods: This is an observational cohort study, performed with institutional review board approval. From 2010 to 2014, 257 patients (median age 49 years; range, 16–86 years) underwent breast MRI (screening, n = 100; problem-solving, n = 88; local disease extent assessment, n = 45; neoadjuvant therapy monitoring, n = 14; and occult primary tumor, n = 10). The clinical, imaging, and pathology records were retrospectively analyzed, with identification of 118 lesions on MRI for which biopsy was prospectively determined to be indicated according to MRI criteria. These 118 lesions constituted our study cohort.

Results: Among these 118 lesions, there were 65 enhancing masses and 53 cases of nonmass enhancement. All 118 lesions were evaluated with second-look ultrasound; a correlate was identified in 50 (42%) cases (30 mass, 17 architectural distortion, three solitary duct). Ultrasound-guided core biopsy using a 14-gauge spring-loaded needle was performed in all 50 cases, yielding three fibroadenoma, five ductal carcinoma in situ, 30 invasive ductal carcinoma, and 12 invasive lobular carcinoma. Positive predictive value (PPV) was 94% (47/50). The remaining 68 (58%) cases were assigned to MRI follow-up (every 6 months for 2 years). We have follow-up data on 40 cases; the remaining 28 were lost to follow-up. Of the 40 cases, 20 (50%) were stable in size at 2-year MRI follow-up, 12 (30%) decreased in size, six (15%) resolved, and two (5%) increased in size. The two cases that increased were again subjected to second-look ultrasound; a sonographic correlate was now identified in these two cases. Subsequent ultrasound-guided biopsy yielded invasive ductal carcinoma in both cases. Hence, PPV of lesions assigned to follow-up (with available follow-up data) was 5% (2/40). One of these two cancers increased from 6 mm to 9 mm, and the second from 5 mm to 10 mm. Both cancers were initially assessed as BI-RADS category 4 and were diagnosed at the first 6-month MRI follow-up. Of the 47 cancers diagnosed at initial ultrasound-guided core biopsy, median size was 7 mm (range, 5–10 mm). There were 18 low-grade lesions, 28 intermediate-grade lesions, and one high-grade lesion among these 47 cancers. The two cancers diagnosed upon interval increase in size during MRI follow-up were 9 mm and 10 mm, and both were of intermediate grade.

Conclusion: PPV of 5% among lesions assigned to MRI follow-up exceeded the usually accepted PPV of 2% for short-interval imaging follow-up. This result supports the consensus that MRI-detected lesions deemed suspicious or highly suggestive of malignancy on the basis of MRI criteria should be biopsied even in the absence of a sonographic correlate. Nevertheless, the absence of MRI-guided biopsy capability and associated delay in breast cancer diagnosis did not appear to have significant negative effect on outcome in terms of cancer size or grade as long as follow-up was obtained.