Columnar cell lesions of the breast: terminology, histologic classification, molecular pathology and clinical significance

Columnar cell lesions (CCLs) are one of the most common abnormalities in the adult female human breast and comprise a spectrum of histologic changes characterized by the presence of one to several layers of columnar-shaped epithelial cells in enlarged terminal duct lobular units. CCLs are being seen increasingly in core biopsies taken for non-palpable calcifications identified by screening mammography. The increased incidence may reflect improved recognition of CCLs by pathologists or a true increase in incidence related to biological and/or environmental factors. We identified CCLs in 17 % of normal breast tissue samples obtained from a forensic autopsy series, 4.2 %> of reduction mammoplasty specimens and 44 % of breast cancer patients.

CCLs may have been first recognized by Warren in 1905 as an abnormal pattern of involution. They have since been described under a variety of names such as blunt duct adenosis, clinging carcinoma, columnar alteration with prominent apical snouts and secretions (CAPSS), enlarged lobular units with columnar alteration (ELUCA), hyperplastic enlarged lobular units (HELUs), monomorphic epithelial proliferation (MEP), columnar cell change (CCC), columnar cell hyperplasia (CCH) with or without atypia, and flat epithelial atypia, originally designated as ductal intraepithelial neoplasia, type 1 b (DINlb), and later as DINla. The lack of standardized terminology has been a serious obstacle to wide recognition of CCLs by practicing pathologists and clinicians.

The cellular origin of CCLs, and their possible relationship to either expansion or metaplasia of a pre-existing normal cell phenotype remains unclear. Although columnar cells lack mature luminal or basal/myoepithelial and apocrine markers, they are often positive for estrogen receptor-a, progesterone receptors, Bcl-2 and FASN. It is likely that the development of CCLs is influenced by steroid hormones and/or growth factors. CCLs are frequently associated with lobular and ductal in-situ tumors, and invasive lobular and tubular carcinomas. The genetic alterations in CCLs are not well characterized. A few studies have reported changes in ERBB and amphi-regulin genes by DNA microarray, allelic imbalance targeting loci at 9q, 10q, 17p and 17q by PCR, and recurrent 16q loss by comparative genomic hybridization. After age and BRCA1/2 genetic predisposition, high mammographic density is the third largest population risk factor
for subsequent development of in-situ or invasive carcinoma. We reported an association between CCLs and breast tissue composition, including mammographic density.

Both histological and molecular genetic studies suggest that CCLs represent the earliest histologically identifiable, non-obligate precursor of low-grade breast cancer. Although the follow-up
data show that the risk of local recurrence and progression of most CCLs to invasive cancer is low, we believe that there is sufficient evidence to recommend regular follow-up when CCLs are diagnosed in women participating in mammographic screening programs. Whether further excision should be recommended for CCLs detected in core biopsies remains controversial. As the natural history of CCLs is currently uncertain,further detailed phenotype and genotype studies across the spectrum of CCLs and in the context of normal and other pre-neoplastic breast lesions will be required to delineate the biology and better define the clinical implications of CCLs.