Association between BRCA2 status and histologic variants (intraductal [IDC] and cribriform [CRIB] histology) in prostate cancer (PC).

Abstract

5579 Background: IDC histology in PC has been suggested to associate with germline BRCA2 mutations (gBRCA2) in small series, leading to the potential recommendation of genetic testing for all PC patients with IDC in the primary tumor. Methods: We conducted a case-control study in which primary PC from 58 germline BRCA2 mutation carriers ( gBRCA2) and 116 from non-carriers (NC) were matched 1:2 by Gleason score and specimen type (core biopsy/prostatectomy). Samples were independently reviewed by two expert pathologists blinded to genetic status who established the presence of IDC and/or CRIB morphology with supportive immunohistochemical stains in a subset of cases. Next-generation sequencing, aCGH and/or FISH were used to assess for somatic mono-/bi-allelic BRCA2 alterations. PTEN protein loss was determined by IHC, and TMPRSS2-ERG was detected by FISH/qRT-PCR. Chi-square tests were used to compare the frequency of IDC and cribriform histology in gBRCA2 vs NC, as well as to assess the associations with other variables. Logistic regression models were built to identify independent factors associated with IDC and CRIB histology. Results: gBRCA2 cases were younger at diagnosis (median 61.3 vs 64, p < 0.01) and had T3-4 disease more often than NC cases (31% vs 10.5%, p < 0.01), but the two groups did not differ in any other clinical-pathologic characteristics. After independent histopathological review, 79 cases demonstrated IDC histology and 81 had CRIB histology. No differences in the prevalence of IDC (50% NC vs 36% gBRCA2, p = 0.09) or CRIB (43% NC vs 53% gBRCA2, p = 0.20) patterns were observed. The probability of IDC was higher in PC with bi-allelic BRCA2 alterations (OR 5.1, 95%CI 2.1-12.6), PTEN loss (OR 5.1, 95%CI 1.9-13.5) or both (OR 23.0, 95%CI 4.9-107.2) compared to those without these alterations. Bi-allelic BRCA2 alteration was also associated with higher probability of CRIB histology (OR 7.2, 95%CI 3.1-16.4). TMPRSS2-ERG fusions were not associated with IDC or CRIB histology. MVA confirmed the independent association of bi-allelic BRCA2 alteration (p < 0.01) and PTEN loss (p < 0.01) with IDC histology. Bi-allelic BRCA2 alteration (p < 0.01) and Gleason >8 (p < 0.01) were independent risk factors for CRIB histology. Conclusions: Primary PC with bi-allelic BRCA2 alterations was significantly associated with IDC and CRIB histology, independent of other clinical-pathologic factors (while gBRCA2 status alone was not). PTEN loss in primary PC was also independently associated with IDC, but not CRIB, histology.