Encyclopedia of Pathology

Living Edition
| Editors: J.H.J.M. van Krieken

Columnar Cell Lesions

  • Margaret C. Cummings
  • Peter T. SimpsonEmail author
  • Sunil R. Lakhani
Living reference work entry
DOI: https://doi.org/10.1007/978-3-319-28845-1_4700-1



Columnar cell lesions refers to a group of proliferative lesions in the breast which have a characteristic cystic dilatation of acini of terminal duct lobular units such that they appear enlarged and unfolded, and which are lined by cells showing a columnar morphology. Columnar cell lesions that display cytological atypia are referred to as flat epithelial atypia (FEA).

Clinical Features

  • Incidence

    As a typical feature of columnar cell lesions is the presence of calcification, they have been increasingly identified through the use of screening mammography and have been described as occurring in up to 50% of core needle biopsies that are performed for microcalcification. A frequency of 1.9% has been reported for FEA in reduction mammoplasty specimens (Desouki et al. 2013) and between 2.4% and 4.5% in core needle biopsies derived from breast screening programs (Said et al. 2015; Yu et al. 2015). For FEA, excluding those cases which coexist with higher grade lesions, such as atypical hyperplasia, in situ or invasive carcinoma, an incidence of up to 12.2% has been reported (Berry et al. 2016).

  • Age

    Reflecting the general age distribution of women undergoing screening mammography, an age range in the 40s–60s has been described for women with columnar cell lesions including FEA.

  • Sex

    Columnar cell lesions, including FEA, have only been reported in females, no doubt because of the lack of lobular units in male breast tissue.

  • Site

    These lesions occur only in breast tissue.

  • Clinical Presentation

    Columnar cell lesions are most commonly identified on core needle biopsy taken because of microcalcification which has been identified on screening mammography. The microcalcification has been variably described as indistinct, amorphous, fine, irregular, and pleomorphic. Calcification such as this can also be seen in more significant breast diseases, hence the need for core needle biopsy. There are no mammographic features that can be used to distinguish columnar cell lesions showing atypia from those without atypia. Columnar cell lesions may also be identified histologically when a core needle biopsy has been taken for another reason, such as a nonspecific density and they may also be found incidentally in surgical excision specimens.

  • Treatment

    For columnar cell change and columnar cell hyperplasia diagnosed on core needle biopsy, no further treatment is required.

    The situation with regards to the management of patients in whom FEA is diagnosed on core needle biopsy is not so straightforward. A contributing factor to this uncertainty is the not insignificant frequency of FEA in the screening population. A related factor is the difficulty of making the diagnosis of FEA histologically. Pathologists concerned about the possibility of underdiagnosis and the implications of missing an atypical lesion tend to overdiagnose the condition, which of course makes the most appropriate management difficult to determine. And for the instances where FEA is accurately diagnosed, another confounding factor is the possible pathology upgrade on surgical excision, and this is of the most immediate concern, considerations of long-term risk reduction being less relevant. The most commonly described treatment for FEA identified on core needle biopsy is surgical excision. If no higher grade lesion, such as in situ or invasive carcinoma is identified initially in the excision specimen, all the material should be blocked thoroughly and levels performed as required. However, the presence of FEA at the margin of a surgical excision specimen is not an indication for further surgery.

    If FEA is found to coexist with atypical ductal hyperplasia, in situ or invasive carcinoma, then naturally the treatment recommendation is dictated by the higher grade lesion.

    Other treatment recommendations for FEA found on core needle biopsy include radiology and pathology correlation with multidisciplinary team discussion. Some may recommend using vacuum-assisted biopsy to excise any residual tissue showing microcalcification or other areas of radiological concern, instead of advocating surgical excision. If no further microcalcification is identified radiologically, if the extent of involvement of the terminal duct lobular units by FEA is small and the patient does not have a personal history of breast cancer, ongoing radiological surveillance instead of surgical excision has been suggested as the preferred management (Berry et al. 2016; Acott and Mancino 2016). However, long-term follow-up studies of women with FEA diagnosed on core needle biopsy for whom the management has been prolonged surveillance are not yet available, and some use this lack of definitive information to support the use of surgical excision as the primary treatment recommendation.

  • Outcome

    The clinical outcomes of the various columnar cell lesions have not been precisely defined, and while some generalizations can be made, there is still a lack of clarity with respect to the long-term significance of these lesions. A lower level of risk has been attributed to columnar cell change and columnar cell hyperplasia compared with that for FEA.

    The biological importance of columnar cell lesions is underscored by their association with other breast lesions such as atypical ductal hyperplasia and lobular neoplasia, including atypical lobular hyperplasia and lobular carcinoma in situ, which have been noted to coexist with them, sometimes in the same terminal duct lobular unit (Schnitt and Collins 2009). An association with tubular carcinoma, which together with columnar cell lesions and lobular neoplasia forms the eponymously named Rosen’s triad, was first described in 1999 (Rosen 1999). Whether or not columnar cell lesions constitute a direct precursor lesion or are an indicator of a more general increased risk of subsequent breast cancer has not been established. The time course of possible progression to in situ or invasive carcinoma and actual upgrade rates are also contentious areas.

    There is probably sufficient evidence to support columnar cell lesions being at least one early feature of the low-grade breast neoplasia pathway, but the data are difficult to compare and analyze because of the range of methodologies the various studies have used. In some follow-up studies it is not clear if the subsequent carcinoma has been in the same area as the initial columnar cell lesion. When columnar cell lesions are removed at core biopsy it is generally not known how much, if any, lesional tissue remains. It is difficult to then interpret the significance of subsequent cancers as it would not be known from precisely what tissue they may have arisen.

    An association between columnar cell lesions and carcinoma has also been investigated by assessing whether columnar cell lesions are present in breast tissue adjacent to in situ and invasive carcinomas. Another way to explore this association is to retrospectively study biopsies taken previously from women with invasive carcinomas to see if columnar cell lesions, including FEA, might have been present earlier. Follow-up studies have investigated whether cancers arising subsequent to columnar cell lesions have occurred in the same area or somewhere else in the ipsilateral breast or indeed in the contralateral breast, the latter, of course, not being used to support any direct precursor role (Boulos et al. 2008). Another confounding factor is that many of the studies have not been able to provide information as to the size or possible multifocality of the columnar cell lesions although pathology-radiology correlation as to the extent of residual calcification has been very helpful in this regard.

    Overall, the long-term relative risk of carcinoma for those women with nonatypical columnar cell lesions diagnosed in isolation appears to be very low, similar to that seen for nonatypical proliferative breast disease in general, around the order of a 1.5-fold increase (Verschuur-Maes et al. 2012b).

    A closely related issue with columnar cell lesion is that of the potential upgrade to more significant lesions, particularly in situ and invasive carcinoma, which occurs when there is an immediate surgical excision taken on the basis of a columnar cell lesion identified on core biopsy. A very wide range of upgrade rates (between 0% and 67%) has been described in the literature, and this considerable variation partly reflects the very different ways in which the upgrade rate has been calculated. The numerator is generally those who have had a malignancy in the subsequent excision specimen, while the denominator can include all women who have had a columnar cell lesion on core biopsy, or only those who have gone on to surgical excision. This of course would give a higher upgrade rate. In this latter group too, other clinical and radiological features may have also contributed to the decision to recommend surgical excision, thus confounding the analysis of columnar cell lesions. Some studies have reviewed columnar cell lesions when there is also atypical ductal or lobular hyperplasia present, while others have specifically excluded cases with concomitant atypical lesions. Many more studies have addressed upgrade rates from FEA than from pure columnar cell change or columnar cell hyperplasia. The problem of correctly diagnosing these lesions histologically only adds to the difficulty of obtaining meaningful outcome data. Verschuur-Maes in a comprehensive meta-analysis of 24 papers concerning the risk of carcinoma in women with columnar cell lesions derived pooled underestimation risks of 1.5% for columnar cell lesions without atypia, 9% for columnar cell lesions with atypia compared with 20% for those who also had atypical ductal hyperplasia (Verschuur-Maes et al. 2012b).


Columnar cell lesions, including FEA are diagnosed microscopically, and except for occasional floridly cystic examples, are not able to be identified macroscopically.


A distinctive feature of columnar cell lesions on low power examination is the cystic dilatation of enlarged terminal duct lobular units, and for columnar cell change and columnar cell hyperplasia, these gland spaces often have an irregular contour (Fig. 1). The lesions may also be conspicuous on low power examination because of a basophilic appearance, partly due to the regular arrangement of the cells, which are generally uniform, columnar to cuboidal in shape, and mostly orientated perpendicular to the basement membrane. Quite frequently, small apical cytoplasmic blebs or snouts are seen, projecting into the lumen of some glandular spaces, but these are not usually prominent. Thin, pale proteinaceous secretions which may be pink or blue are commonly seen in the gland lumina (Fig. 2). Luminal microcalcification, often the reason for these lesions being detected radiologically, may be seen and these range from being fine to granular or coarse and even sometimes psammomatous (Fig. 3). The nuclei of columnar cell change or hyperplasia are generally oval or elongated, with smooth nuclear membranes and fine chromatin. Nucleoli usually are not seen and mitoses are rarely identified. Occasionally the underlying myoepithelial cells become cuboidal and even relatively prominent. For columnar cell change, 1–2 layers of cells are seen above the basement membrane, with greater numbers seen for columnar cell hyperplasia (Fig. 4), where stratification of cells can occur within dilated acini, often giving an impression of hyperchromasia. Luminal tufting of cells, blunt micropapillae, and luminal secretions can occur (Fig. 5).
Fig. 1

Columnar cell lesion. Cystic dilatation of terminal duct lobular unit. (H & E)

Fig. 2

Columnar cell lesion. dilated gland lumina containing watery secretions. (H & E)

Fig. 3

Columnar cell lesion and columnar cell hyperplasia. Luminal microcalcification. (H & E)

Fig. 4

Columnar cell hyperplasia. irregularly dilated gland spaces with some overlapping of cells. (H & E)

Fig. 5

Columnar cell hyperplasia. Luminal tufting of cells and small micropapillae. (H & E)

Enlarged or dilated acini of terminal duct lobular units, often appearing rounded or distended (Fig. 6) rather than the more convoluted contours of columnar cell change, characterize lesions of FEA and these are lined by 3–5 layers of cells. With a loss of polarity, the cells have a more haphazard overlapping arrangement. “Flat” refers to the absence of complex architectural features such as Roman bridges or well-developed rigid micropapillae. Apical cytoplasmic projections, characteristic of columnar cell change, may on occasion be more pronounced. The lining epithelial cells are usually rounded, with a low-grade monomorphic appearance (Fig. 7) unlike that seen in columnar cell change or hyperplasia. The nuclei are enlarged, with an increased nuclear to cytoplasmic ratio and hyperchromasia, making the lesions appear distinctive on low-power examination. The cellular crowding, loss of polarity, and increased nuclear to cytoplasmic ratio all contribute to a basophilic appearance. Nuclear membranes may be irregular, the chromatin relatively dense, and nucleoli can generally be seen. Stromal changes, such as a chronic inflammatory infiltrate and a fibrinous or myxoid change can be found in the intralobular stroma, surrounding glandular spaces involved by FEA. Microcalcification, as occurs in columnar cell change and hyperplasia, can also be seen in gland lumina and this may be granular or irregular (Schnitt and Collins 2009).
Fig. 6

Flat epithelial atypia. Rounded dilated acini of terminal duct lobular unit. (H & E)

Fig. 7

Flat epithelial atypia. Overlapping cells with monomorphic atypia and occasional nucleoli visible. (H & E)


Columnar cell lesions, including columnar cell change, columnar cell hyperplasia, and FEA, show an immunoprofile similar to that seen in atypical ductal hyperplasia and low-grade ductal carcinoma in situ, reflecting the clonality of these entities. Diffuse expression of the low molecular weight cytokeratins CK7, CK8/18, and CK19 can be seen in the luminal cells, with complete absence of expression of high molecular weight cytokeratins such as CK14, CK5/6, and 34βE12. Unlike the cells of the normal terminal duct lobular units which show a heterogeneous pattern of staining, columnar cell lesions show diffuse, intense staining with antibodies to both estrogen and progesterone receptor proteins. This is also in contrast with the cells of apocrine metaplasia which are usually uniformly negative for hormone receptor proteins and sometimes can be confused with those of columnar cell lesions. The proliferation rate of columnar cell lesions as indicated by Ki67 staining is generally low, with less than 10% of cells staining, and while it may be increased compared with that of the adjacent noninvolved breast tissue, it is not a reliable discriminating feature and nor can it be used to distinguish between FEA and nonatypical columnar cell lesions. The immunoprofile of columnar cell lesions also includes Bcl2 positivity, with strong diffuse cytoplasmic staining, positive E-cadherin, and absence of staining with HER2 immunohistochemistry.

Molecular Features

Prompted by morphological similarities with more advanced lesions, potential molecular changes in columnar cell lesions has been investigated to try to determine the clinical significance of these entities as possible (even nonobligate) precursors of atypical ductal hyperplasia or in situ and invasive carcinoma, as part of the low-grade breast neoplasia pathway. A variety of methods of analysis have been used, and while the findings so far are suggestive of a causal relationship, rather than just co-occurrence, the evidence currently available is relatively scanty. A loss of heterozygosity study of clinging ductal carcinoma in situ, also called ductal intraepithelial neoplasia (DIN) flat type, one of the earlier names for what is now referred to as FEA, demonstrated changes at chromosomes 11q21-23.2, 16q23.1-24.2, and 3p14.2 in 50%, 45%, and 41% of informative cases, respectively, with corresponding changes identified in the in situ and invasive carcinoma also present in these cases (Moinfar 2009). These findings were later supported by the work of Aulman in a study of FEA, tubular carcinoma, and low-grade ductal carcinoma in situ where loss of heterozygosity was noted for 16q in up to 80% of cases, 8p21 (26% of cases) 3p14 (20% of cases), 1p35 (20% of cases), and 11q14 (24% of cases), with many comparable losses seen in the adjacent low-grade ductal carcinoma in situ and tubular carcinoma (Aulmann et al. 2009).

Using comparative genomic hybridization Simpson was able to demonstrate chromosomal losses (6q, 11q, 12q, 16q, 17p, 19q, and 22) and gains (3p, 3q, 7, 8q, 12p, 15q, 16p, 19, and 20) in columnar cell lesions, with increasing chromosomal changes seemingly occurring in parallel with increasing cytological and architectural atypia, particularly the recurrent loss of 16q (Simpson et al. 2005).

Gains and losses of 17 breast cancer-related genes were documented for columnar cell lesions with accompanying ductal carcinoma in situ and invasive carcinoma. Copy number gains were identified for C11orf30, MYC, CPD, MTDH, CCND1, CCNE1, ESR1, and TOP2A, while losses were seen for CDH1 and TOP2A in a small number of cases, the copy number changes increasing from columnar cell lesions through to invasive carcinomas. The relative infrequency of the changes in the columnar cell lesions compared with those for the in situ and invasive carcinomas suggested that they may represent relatively late events in the presumed low-grade pathway (Verschuur-Maes et al. 2014).

Twenty-five percent of invasive breast carcinomas have been demonstrated to contain an activating mutation in the catalytic subunit of phosphatidylinositol-3-kinase (PIK3CA), and the presence of these mutations have been investigated in columnar cell lesions, with one study showing them to be present in 13/24 (54%) cases (Troxell et al. 2012) and in 31/62 (50%) of columnar cell/usual ductal hyperplasia lesions in another (Ang et al. 2014). However, the same point mutations were not consistently identified in the corresponding carcinoma, sometimes with different mutations identified, or wild-type PIK3CA in the carcinoma with mutations in the columnar cell lesions, or vice versa, wild-type in the columnar cell lesions and point mutations identified in the carcinomas making the association less certain.

Promoter methylation of the tumor suppressor genes ID4, CCND2, and CDH13 has been demonstrated in columnar cell lesions with increasing levels shown to occur in parallel with progression to invasive carcinoma (Verschuur-Maes et al. 2012a) supporting a precursor role for these lesions, although atypia was not associated with increased levels. Other changes identified as linked to the early phase of the low-grade pathway include, in columnar cell hyperplasia, downregulation of the let-7c microRNA in the epithelial compartment and upregulation of miR-132 in the stroma surrounding columnar cell hyperplasia (Bjorner et al. 2014) possibly reducing some controls on anti-proliferative activity.

Differential Diagnosis

The diagnosis of columnar cell change, columnar cell hyperplasia, and FEA can be difficult to be certain about, particularly on a core needle biopsy where there may be only limited material for evaluation; however, distinguishing FEA is particularly important as it is associated with a different management outcome for the patient.

Similar to columnar cell lesions, apocrine metaplasia can also be seen in cystically dilated acini and the lining cells may also show apical snouts. The cytoplasm of apocrine cells is more eosinophilic and granular than that of columnar cell lesions, the apical snouts are usually quite inconspicuous, and the nuclei of apocrine cells tend to be more round and to have a central prominent nucleolus. Subtle, refractile calcium oxalate crystals can be associated with apocrine metaplasia rather than the darker blue calcium phosphate seen with columnar cell lesions. Immunostaining for both hormone receptors and Bcl-2 is usually negative in apocrine lesions.

Cystic columnar cell lesions may contain luminal mucin, and this may calcify and may be fine, granular, or more coarse. Rupture of such cysts with mucin extravasation into the stroma creates a differential diagnosis with mucocele-like lesions. The distinction may be difficult to make on a core biopsy.

Usual ductal hyperplasia may be a consideration when a columnar cell lesion becomes more proliferative, but this is not commonly a difficulty because of the characteristic columnar shape of the cells on the one hand and the more heterogeneous architectural features associated with usual ductal hyperplasia on the other. As they are both benign lesions, differentiating the two does not have practical consequences.

Distinguishing columnar cell change and columnar cell hyperplasia from FEA is the key consideration. Columnar cell hyperplasia shows a relatively orderly multilayered stratification of cells; the lining cells in FEA tend to have fewer layers and show a loss of polarization, including a loss of the perpendicular orientation that characterizes both columnar cell change and columnar cell hyperplasia. The affected gland spaces of FEA appear more rounded and rigidly distended, and with a more basophilic appearance overall. Monomorphic cytological atypia, with rounder nuclei and perceptible nucleoli, is seen. Architectural atypia, however, in the form of cellular bridges, arcades, punched out cribriform spaces, or more structured micropapillae, together with polarization of cells around these various formations, indicates either atypical ductal hyperplasia or low-grade ductal carcinoma in situ. Though, of course, both atypical ductal hyperplasia and FEA may coexist.


  1. Acott, A. A., & Mancino, A. T. (2016). Flat epithelial atypia on core needle biopsy, must we surgically excise? American Journal of Surgery, 212, 1211–1213.CrossRefPubMedGoogle Scholar
  2. Ang, D. C., Warrick, A. L., Shilling, A., Beadling, C., Corless, C. L., & Troxell, M. L. (2014). Frequent phosphatidylinositol-3-kinase mutations in proliferative breast lesions. Modern Pathology, 27, 740–750.CrossRefPubMedGoogle Scholar
  3. Aulmann, S., Elsawaf, Z., Penzel, R., Schirmacher, P., & Sinn, H. P. (2009). Invasive tubular carcinoma of the breast frequently is clonally related to flat epithelial atypia and low-grade ductal carcinoma in situ. The American Journal of Surgical Pathology, 33, 1646–1653.CrossRefPubMedGoogle Scholar
  4. Berry, J. S., Trappey, A. F., Vreeland, T. J., Pattyn, A. R., Clifton, G. T., Berry, E. A., et al. (2016). Analysis of clinical and pathologic factors of pure, flat epithelial atypia on core needle biopsy to aid in the decision of excision or observation. Journal of Cancer, 7, 1–6.CrossRefPubMedPubMedCentralGoogle Scholar
  5. Bjorner, S., Fitzpatrick, P. A., Li, Y., Allred, C., Howell, A., Ringberg, A., et al. (2014). Epithelial and stromal microRNA signatures of columnar cell hyperplasia linking Let-7c to precancerous and cancerous breast cancer cell proliferation. PLoS One, 9, e105099.CrossRefPubMedPubMedCentralGoogle Scholar
  6. Boulos, F. I., Dupont, W. D., Simpson, J. F., Schuyler, P. A., Sanders, M. E., Freudenthal, M. E., et al. (2008). Histologic associations and long-term cancer risk in columnar cell lesions of the breast: A retrospective cohort and a nested case-control study. Cancer, 113, 2415–2421.CrossRefPubMedPubMedCentralGoogle Scholar
  7. Desouki, M. M., Li, Z., Hameed, O., Fadare, O., & Zhao, C. (2013). Incidental atypical proliferative lesions in reduction mammoplasty specimens: Analysis of 2498 cases from 2 tertiary women’s health centers. Human Pathology, 44, 1877–1881.CrossRefPubMedGoogle Scholar
  8. Moinfar, F. (2009). Flat ductal intraepithelial neoplasia of the breast: A review of diagnostic criteria, differential diagnoses, molecular-genetic findings, and clinical relevance – it is time to appreciate the Azzopardi concept! Archives of Pathology & Laboratory Medicine, 133, 879–892.Google Scholar
  9. Rosen, P. P. (1999). Columnar cell hyperplasia is associated with lobular carcinoma in situ and tubular carcinoma. American Journal of Surgical Pathology, 23, 1561.CrossRefPubMedGoogle Scholar
  10. Said, S. M., Visscher, D. W., Nassar, A., Frank, R. D., Vierkant R. A., Frost, M. H. et al (2015). Flat epithelial atypia and risk of breast cancer: A Mayo cohort study. Cancer, 121, 1548–1555.Google Scholar
  11. Schnitt, S., & Collins, L. (2009). Biopsy interpretation of the breast (1st ed., pp. 96–122). Philadelphia: Lippincott Williams and Wilkins.Google Scholar
  12. Simpson, P. T., Gale, T., Reis-Filho, J. S., Jones, C., Parry, S., Sloane, J. P., et al. (2005). Columnar cell lesions of the breast: The missing link in breast cancer progression? A morphological and molecular analysis. The American Journal of Surgical Pathology, 29, 734–746.CrossRefPubMedGoogle Scholar
  13. Troxell, M. L., Brunner, A. L., Neff, T., Warrick, A., Beadling, C., Montgomery, K., et al. (2012). Phosphatidylinositol-3-kinase pathway mutations are common in breast columnar cell lesions. Modern Pathology, 25, 930–937.CrossRefPubMedGoogle Scholar
  14. Verschuur-Maes, A. H., de Bruin, P. C., & van Diest, P. J. (2012a). Epigenetic progression of columnar cell lesions of the breast to invasive breast cancer. Breast Cancer Research and Treatment, 136, 705–715.CrossRefPubMedGoogle Scholar
  15. Verschuur-Maes, A. H., van Deurzen, C. H., Monninkhof, E. M., & van Diest, P. J. (2012b). Columnar cell lesions on breast needle biopsies: Is surgical excision necessary? A systematic review. Annals of Surgery, 255, 259–265.CrossRefPubMedGoogle Scholar
  16. Verschuur-Maes, A. H., Moelans, C. B., de Bruin, P. C., & van Diest, P. J. (2014). Analysis of gene copy number alterations by multiplex ligation-dependent probe amplification in columnar cell lesions of the breast. Cellular Oncology (Dordrecht), 37, 147–154.CrossRefGoogle Scholar
  17. Yu, C. C., Ueng, S. H., Cheung, Y. C., Shen, S. C., Kuo, W. L., Tsai, H. P., et al. (2015). Predictors of underestimation of malignancy after image-guided core needle biopsy diagnosis of flat epithelial atypia or atypical ductal hyperplasia. The Breast Journal, 21, 224–232.CrossRefPubMedGoogle Scholar

Copyright information

© Springer International Publishing AG 2018

Authors and Affiliations

  • Margaret C. Cummings
    • 1
    • 2
  • Peter T. Simpson
    • 1
    Email author
  • Sunil R. Lakhani
    • 1
    • 2
  1. 1.Centre for Clinical Research, Faculty of Medicine, The University of QueenslandBrisbaneAustralia
  2. 2.Pathology Queensland: The Royal Brisbane & Women’s HospitalBrisbaneAustralia