ETS gene aberrations in atypical cribriform lesions of the prostate: Implications for the distinction between intraductal carcinoma of the prostate and cribriform high-grade prostatic intraepithelial neoplasia.

Background: Atypical cribriform lesions (ACLs) of the prostate consist of cribriform glands lined with cytologically malignant cells with partial or complete basal cell lining. It may represent cribriform "high-grade prostatic intraepithelial neoplasia" (HGPIN) or "intraductal carcinoma of the prostate" (IDC-P), which is almost always associated with clinically aggressive prostate carcinoma (PCa). Distinction between these 2 lesions has profound clinical significance, especially on needle biopsies. However, there are lesions that do not fully satisfy the criteria for IDC-P yet are worse than typical HGPIN and are difficult to distinguish based on morphologic criteria alone.

Methods: To better understand the biologic and molecular basis of distinction between cribriform HGPIN and IDC, we used break-apart fluorescence in-situ hybridization assay to assess ETS gene aberrations, a specific and commonest molecular alteration involving PCa, in a cohort of 16 isolated ACL, presumed to be an isolated cribriform HGPIN, and 45 carcinoma-associated ACL (ACL-PCa) on radical prostatectomy specimens, presumed to be spectrum of IDC-P. The latter was further divided into 2 groups: group A with marked nuclear atypia (nuclear size 6xnormal or larger) and/or comedonecrosis (n=21) and group B that did not fulfill these criteria (n=24).

Results: Overall, ERG rearrangement was absent (0 of 16) in isolated cribriform HGPIN, whereas present in 75% (36 of 48) of IDC-P, of which 65% (23 of 36) were through deletion and 35% (13 of 36) through insertion. Notably, 17% (6 of 36) of the IDC-P showed duplication of ERG rearrangement in combination with deletion of 5'-ERG. Hundred percent (34 of 34) of the IDC-P showed concordance of ERG rearrangement status with adjacent invasive carcinoma. There was no difference between the 2 groups of IDC-P lesions regarding prevalence of ERG rearrangement (group A 79% vs. group B 74%) and EDel2+ (20% vs. 15%). No case with ETV1, ETV4, or ETV5 rearrangement was identified.

Conclusions: Our molecular data suggest that isolated cribriform HGPIN and IDC-P are biologically distinct lesions. Majority of ACL-PCa most likely represent intraductal spread of PCa. There is a significant overlap between IDC-P and HGPIN at the lower grade morphologic spectrum. ERG break-apart fluorescence in-situ hybridization assay provides insight into understanding the molecular basis of cribriform HGPIN and IDC-P and has potential clinical implications in their distinction on needle biopsies.