Genetic counseling of medullary breast cancer patients

To the Editor: Germline mutations in the BRCA1 and BRCA2 genes account for approximately 40% of families with evidence of inherited susceptibility to breast cancer but for only 2–3% of all breast cancer cases. However, these genes are highly penetrant, and mutations in either of these genes confer a 60–85% lifetime risk of breast cancer and a 15–40% lifetime risk of ovarian cancer (1). Mutation analysis in the BRCA1 and BRCA2 genes is cumbersome and expensive, because both genes are extremely large, with a total cDNA of 16.7 kb, and mutations are scattered throughout the entire coding region. It is therefore imperative for genetic counselors to be able to efficiently identify those patients in real need of the complete genetic analysis of both genes. An overrepresentation of the rare subtype of medullary breast carcinoma among patients with mutations in the BRCA1 gene has recently been reported, while this is not occurring for the BRCA2 gene (2–4). The inclusion of the histology of medullary breast cancer as an independent selection criterion for referral to BRCA1 mutation analysis, besides family history and age, has been proposed as a possibility for genetic counseling (2–4). However, in a recent investigation of 42 cases of typical and atypical medullary breast cancer, only three BRCA1 mutations (7%) were detected, and these were among young patients with a significant family history (5). It has also been proposed that besides family history, other morphological and biological features of breast cancer could improve the efficacy of BRCA1 mutation screening (6, 7). It is now realized that the tumor phenotype of BRCA1 mutation carriers differs significantly from the phenotype of sporadic breast cancer, while the phenotype of BRCA2 tumors lies somewhere in between. BRCA1 tumors tend to be of a higher grade and higher mitotic index than the sporadic cases (8). They are often negative for both estrogen and progesterone receptors (ER and PgR) and are associated with aneuploidy, high S-phase, and presence of p53 somatic mutations that confer a far more aggressive pattern in the BRCA1 tumor compared with a sporadic one (8). In the most recent and largest of all cohorts of cases with a BRCA1 germline mutation (n1⁄4 165), these observations were substantiated and three immunohistochemical markers emerged as the ones showing significant difference compared to sporadic breast cancer: 90% and 79% of the BRCA1 tumors were negative for the presence of the hormone receptors ER and PgR, respectively, and 97% of them were negative for the overexpression of the c-erbB-2 oncogene (9). Our data agree with the above remarks and we also suggest to include a fourth parameter: an estimate of the tumor proliferation with either the Ki-67 proliferation marker assessed immunohistochemically or the proliferation profile obtained from cytometry which in our BRCA1 tumors is always extensive (10, 11). The goal of our study was to test for the above two proposals and examine their validity alone or in combination in a group of medullary carcinoma patients. Medullary carcinoma of the breast is rather scarce (<3% of breast cancer cases in most populations), but we managed to collect phenotypic and genetic data from 17 Greek patients suffering from this rare subtype. Patients were asked about their family history in counseling sessions. After informed consent, mutation analysis was performed on genomic DNA isolated from blood leukocytes employing a combination of the Protein Truncation Test (PTT) for the large exon 11 of the BRCA1 gene and DNA sequencing on an ABI 310 Genetic Analyzer (Perkin-Elmer, Applied Biosystems, Foster City, CA, USA) of the other small exons and all exon/intron boundaries of the BRCA1 gene (including exon 11) (10). Complex rearrangements constituting about 10% of BRCA1 mutations were not covered by our mutation detection methods. Histological diagnosis and classification in the corresponding tumor specimens on hematoxylin and eosin slides were based on criteria previously described (12, 13). We performed immunohistochemistry (IHC) with the LSAB methodology by using the following monoclonal antibodies for the four phenotypic parameters mentioned above: 1D5 (ER), 10A9 (PgR), MIB-1 (Ki-67), and 3B5 (c-erbB-2), all purchased from Clin Genet 2004: 65: 343–344 Copyright # Blackwell Munksgaard 2004 Printed inDenmark. All rights reserved CLINICALGENETICS doi: 10.1111/j.1399-0004.2003.00231.x