The effect on detection rate of second cancers after the addition of MRI to conventional screening mammography in patients with a history of breast cancer.

10572 Background: Mammography is currently used in the surveillance of breast cancer survivors, who are at increased risk of developing ipsilateral and contralateral breast cancers regardless of age at diagnosis or time since diagnosis. Several prospective studies have shown the utility of breast MRI in other high risk populations; however, little data exists on the use of MRI for surveillance of breast cancer survivors. We aimed to compare the outcome of surveillance breast MRI vs. mammography in this population. Methods: We identified women <65 with non-metastatic breast cancer or DCIS with at least one MRI performed at our center >11 months after initial diagnosis, along with a mammogram done within 6 months of the MRI. We compared the outcome of MRI and mammography in terms of biopsies performed as well as in detection of new cancers. Results: Of 512 consecutive charts reviewed, 204 patients met inclusion criteria, 105 (51.4%) of whom were African-American. The average number of MRIs per patient was 2.3 (range 2-7), with a total of 474 MRIs performed between 2005 and 2011. MRI resulted in BIRADS scores of 1 or 2 in 73.5% of studies vs. 84.4% for mammography. There were 19 biopsies performed due to MRI findings alone, 7 done due to mammographic findings alone, and 6 biopsies done based on abnormalities seen on both MRI and mammography. There were 7 malignancies identified based on an abnormal MRI, 3 seen on both MRI and mammography, and none identified via mammography alone. The malignancies identified via MRI alone included 1 patient with DCIS, 5 with stage I disease, and 1 with isolated lung metastases. Of the 10 recurrences detected, 5 (50%) were in African Americans. Two patients developed interval cancers within 6 months of normal screening MRI and mammography. Sensitivity and specificity for MRI were 83.3% (95% CI 0.51-0.97) and 92.2% (95% CI 0.87-0.95), vs. 25% (95% CI 0.05-0.57) and 94.8% (95% CI 0.90-0.97) for mammography. Positive and negative predictive values were 40% and 98.9% for MRI vs. 25% and 95.2% for mammography. Conclusions: Gadolinium-enhanced breast MRI is a useful surveillance modality in breast cancer survivors < age 65. Prospective studies are needed in this population.