Triple-Negative Phenotype and/or High Mitotic Rate May Improve Selection of High-Risk Patients for BRCA Testing
TR Lester, BK Arun, A Gutierrez Barrera, L Huo, CT Albarracin, MZ Gilcrease. MD Anderson Cancer Center, Houston, TX
Background: Current selection of patients for BRCA testing relies heavily on family history and early onset of breast and ovarian cancer, but improved selection methods are needed because of the high cost of testing.
Design: Two pathologists independently examined multiple histomorphologic features of invasive breast carcinomas in 260 specimens (188 surgicals and 72 core biopsies) from 216 high-risk patients who had undergone BRCA testing. Each feature was scored on a scale of 0 to 2, and associations between individual and combined scores and BRCA mutation status were evaluated.
Results: Morphologic features found by both pathologist to be associated with BRCA1 mutations in 129 untreated surgical specimens were tumor circumscription (Fisher's exact test, P=0.005 and P=0.001), tumor necrosis (P=0.026 and P=0.007), and a high mitotic rate (P<0.001 for both pathologists). Triple-negative phenotype and tumor grade were also associated with BRCA1 mutations (P<0.001 and P=0.003, respectively). None of the patients with low-grade tumors had BRCA1 mutations. A total morphology score of ≥4 was associated with a BRCA1 mutation (P<0.001 for both pathologists), but a high mitotic rate alone was better able to predict a BRCA1 mutation. A mitotic rate of ≥25 per 10 HPF had a sensitivity and specificity for predicting a BRCA1 mutation of 81% and 83%, respectively, and 79% and 91%, respectively, by each pathologist. The sensitivity could be increased at the expense of specificity using triple-negative phenotype and/or high mitotic rate. The sensitivity and specificity of this combination was 94% and 74%, respectively, and 87% and 81%, respectively, by each pathologist. This combination did not maintain significance in 59 additional surgical specimens after neoadjuvant chemotherapy, but the combination had a high sensitivity but decreased specificity when only core biopsy specimens were evaluated. The sensitivity and specificity for these were 95% and 57%, respectively, and 95% and 65%, respectively, by each pathologist. No good predictors of BRCA2 mutations were observed.
Conclusions: The combination of triple-negative phenotype and/or mitotic rate ≥25 per 10 HPF in an untreated surgical or core biopsy specimen may improve the selection of high-risk patients for BRCA testing.
Tuesday, March 23, 2010 9:30 AM
Poster Session III # 52, Tuesday Morning