Discordance between pathologic and radiologic tumor size on breast MRI may contribute to increased re-excision rates

Preoperative breast MRI does not decrease re-excision rates in patients who undergo lumpectomy. We evaluated concordance of tumor size on MRI and pathologic size in patients who underwent re-excision of margins after lumpectomy. A retrospective review of patients at the Cedars-Sinai Breast Center who received breast MRI was performed. We found that MRI was performed before lumpectomy in 136 patients. Mean age was 55.2 years (standard deviation ± 12.6). Re-excision occurred in 34 per cent (n = 46). Of those undergoing re-excision, 35 per cent (16/46) were re-excised for ductal carcinoma in situ (DCIS) at the lumpectomy margin. There was no significant difference between radiologic and pathologic size of the tumor (1.94 vs 2.12 cm; P = 0.159). In those who underwent re-excision, the radiologic size was underestimated compared with the pathologic size (2.01 vs 2.66 cm; P = 0.032). Patients with pure DCIS lesions (n = 9) also had smaller radiologic tumor size compared with pathologic (0.64 vs 2.88 cm; P = 0.039), and this difference trended toward significance in those who underwent re-excision (0.55 vs 3.50 cm; P = 0.059). Discordance between tumor size on MRI and pathologic size may contribute to re-excisions in patients who undergo lumpectomy. The limitations of breast MRI to evaluate the extent of DCIS surrounding many breast cancers, and the impact on re-excision rates, should be further evaluated.