Case Histories
Case 1 - Figure 1 - Breast FNA (air-dried, Giemsa). Woman aged 43. Symptomatic mass, imaging probably benign.
|
Case 2 - Figure 1 - Core biopsy. Woman aged 52. Screen-detected, indeterminate calcification.
|
Case 3 - Figure 1 - Core biopsy. Woman aged 50. Screen-detected, suspicious microcalcification.
|
Case 4 - Figure 1 - Core biopsy. Woman aged 49. Indeterminate mass lesion on screening mammography.
|
Introduction
Definitive pre-operative diagnosis of breast lesions avoids patient uncertainty and anxiety, in part
by obviating the need for frozen section assessment. It allows planning of operating lists and bed
occupancy. In some cases open biopsy is avoided and benign lesions can be safely left in situ. Both
fine needle aspiration cytology (FNAC) and core biopsy are simple and cheap techniques for the diagnosis
of breast lesions which can be performed in the outpatient setting and have a low complication rate.
Both modes of diagnosis have disadvantages; in particular FNAC requires highly skilled and trained
personnel in both aspiration and assessment
[1,
2]
; as early as 1933 it was recorded regarding
aspiration cytology that "until the pathologist has familiarized himself with the various pitfalls,
errors are certain to occur" [3]. Errors in diagnosis of breast lesions by either technique, may lead to
over-treatment or delay in diagnosis. In the first part of this course we will cover some of the most
commonly encountered and serious potential pitfalls in pre-operative diagnosis of breast lesions.
Fine Needle Aspiration Cytology (FNAC)
Reporting Categories for FNAC
In ideal circumstances one should aim for a definitive diagnosis of malignancy or benignity. The
proportion where this is possible will increase with experience of both the pathologist and aspirator.
Problems arise where paucity of the sample or interpretation of cell morphology, make such a clear
distinction impossible.
Inadequate
The designation of an aspirate as "inadequate" is to a certain extent a subjective matter and may
depend on the experience of the aspirator and/or the interpreter. It is generally based on the presence
of sufficient numbers of epithelial cells to provide a sample adequate for confident assessment. There
are a number of reasons for labelling a smear as inadequate. These fall into three main groups: (a)
hypocellularity, (b) poor aspiration, spreading or staining or (c) excessive blood. In some cases the
sample may be representative of the lesion, for example, adipose tissue fragments could support a
clinical diagnosis of lipoma. Aspirates from certain lesions, such as cysts, abscesses, fat necrosis and
nipple discharge specimens may not contain epithelial cells but should clearly not be classified as
inadequate.
Benign
Indicates an adequate sample showing no evidence of atypia or malignancy. The aspirate is usually
poorly to moderately cellular and consists mainly of regular duct epithelial cells. These are arranged
as monolayers and have characteristic benign cytological features. The background is often composed of
dispersed individual and paired naked nuclei. A mixture of foamy macrophages and regular apocrine cells
may also be part of the picture. Fragments of fibro-fatty and/or fatty tissue are common.
Atypia probably benign
The aspirate can have all the characteristics of a benign aspirate as described in the previous
paragraph. There are however, in addition, certain features not commonly seen in benign aspirates.
These could be any, or a combination of the following:
 | Nuclear pleomorphism |
| Some loss of cellular cohesiveness |
| Nuclear and cytoplasmic changes resulting from e.g. hormonal (pregnancy, pill, HRT) or treatment influences |
| Increased cellularity. |
|
Suspicious of malignancy
This category should be used for those aspirates where there are highly atypical features in the
smear, such that the pathologist is almost certain that they come from a malignant lesion although a
confident diagnosis cannot be made. This may be for three main reasons:-
- The specimen is scanty, poorly preserved or poorly prepared, but some cells with features of malignancy are present.
- The sample may show some malignant features without overt malignant cells present
- The sample has an overall benign pattern with large numbers of naked nuclei and/or cohesive sheets of cells, but with occasional cells showing distinct malignant features.
Malignant
Indicates an adequate sample containing cells characteristic of carcinoma, or other malignancy.
Malignancy should not be diagnosed on the basis of a single criterion.
Table 1: General Diagnostic Criteria for Benign and Malignant FNAC
| Criterion | Benign | Malignant |
| Cellularity | Usually poor or moderate | Usually high |
| Cell to cell cohesion | Good with large defined clusters | Poor with cell separation resulting in dissociated cells with cytoplasm or small groups |
| Cell arrangement | Even, usually in flat sheets (monolayers) | Irregular with overlapping and three-dimensional arrangements |
| Cell types | Mixture of epithelial, myoepithelial and other cells with fragments of stroma | Usually a uniform cell population |
| Bipolar bare nuclei | Present, often in high numbers | Not conspicuous |
| Background | Generally clean, except in inflammatory conditions | Occasionally necrotic debris, inflammatory cells and macrophages |
| Nuclear characteristics |
| Size (in relation to RBC diameter) | Small | Variable, often large |
| Pleomorphism | Rare | Common |
| Nuclear membranes | Smooth | Irregular with indentations |
| Nucleoli | Indistinct or small and single | Variable - may be prominent, large and multiple |
| Chromatin | Smooth or fine | Clumped and may be irregular |
| Additional features | Apocrine metaplasia, foamy macrophages | Mucin, intracytoplasmic lumina |
General Principles to Avoid Pitfalls in FNAC Interpretation
In order to avoid the misdiagnosis of breast FNAC as a general rule smears should be of high quality;
they should be thin, contain no blood clots and show no preparative artefacts. They should be prepared
without undue force causing disruption of membranes. This may also produce apparent discohesion,
although completely disrupted nuclei and fragments of cells will also be present indicating that the
appearances are artefactual. Air-dried preparations must be rapidly dried or the nuclear outlines may
appear blurred and the cells increased in size, whereas wet-fixed specimens must be placed into fixative
immediately. The staining quality should be good. If any of the stages of aspiration, smear preparation
or subsequent fixation and staining is of poor quality it is advisable to use one of the non-diagnostic
categories of diagnosis (atypia or suspicious) or to report the specimen as inadequate unless the
appearances are incontrovertible. Indeed in a review of benign lesions misdiagnosed on FNAC, Kline noted
that most errors were based on insufficiently cellular specimens or poorly preserved cells lacking the
majority of the criteria of malignancy [4].
It is vital that sufficient clinical and, if appropriate, mammographic details are provided to the
pathologist. The age and sex of the patient must be supplied with full information regarding previous
treatment and relevant history. Some pathologists believe that any sections or smears should be examined
"blind" without knowledge of the patient's history or symptoms in order to avoid bias from the clinical
impression. We, conversely, strongly recommend that at some stage prior to definitively reporting a
breast FNAC sample the patient's age and sex is noted; in this way at least some errors in the diagnosis
of cellular fibroadenomas in young patients and pregnancy and lactational change may be avoided.
Causes of False Positive Diagnoses in Breast FNAC
Fibroadenoma
Often smears from fibroadenoma may give very worrisome appearances with marked pleomorphism and some
dissociation. This usually happens in actively growing lesions in teenage women. The clue to the
diagnosis is the presence of 'stripped' bipolar nuclei. Smears containing these in significant numbers
should not be diagnosed as malignant unless there are benign as well as a distinct separate population of
dissociated malignant cells. These smears, where the needle has passed through both a benign and a
malignant lesion may be very difficult but the two distinct populations of epithelial cells should aid
their recognition. Smears from some malignant tumours contain bare nuclei. These are not bipolar and
have malignant features identical to co-existing intact tumour cells. Often in fibroadenomas two cell
types can be recognised in the cell clumps, even in the rather pleomorphic examples.
Apocrine cells
Apocrine cells in smears may appear rather pleomorphic and may dissociate. Degenerate apocrine cells
in cyst fluids may also have a rather worrisome appearance. Recognition of the dusty blue cytoplasm,
with or without cytoplasmic granules with Giemsa stains or pink cytoplasm on Papanicolaou or Haematoxylin
& Eosin stains coupled with the prominent central nucleolus is the key to identifying cells as
apocrine. Awareness of the marked pleomorphism which may occur in degenerate benign apocrine cells and
careful assessment of the cellularity and chromatin pattern should allow the distinction from the rare
apocrine carcinoma.
Spreading artefacts
Excessive pressure during spreading of slides may produce dissociation of cells from benign clumps.
If the cells within these clumps are also somewhat pleomorphic due to degenerative or atypical changes,
then the dissociation may cause the cells to resemble dissociated malignant cells. The clue to this is
often the finding of nuclear lysis and trails of chromatin due to the over-spreading artefact.
Fibroadenomata are the most likely lesions to produce these problems when over-spread.
Papilloma
Aspiration of papillomas usually produces cellular aspirates with "staghorn" or "antler horn"
clusters of cells similar on low power appearance to those seen in fibroadenomas although they may appear
three-dimensional [5]. In some cases connective tissue cores may be seen within these clusters. These
may be diagnostic of papilloma but are not common. While it is important clinically to distinguish
papilloma from intracystic papillary carcinoma, this may not be possible on cytological grounds. Some
features of malignancy such as nuclear pleomorphism, increased nuclear cytoplasmic ratio, and cellular
crowding or overlapping may occur with some benign forms of papilloma. No single feature can
differentiate the two conditions.
Atypical lobular hyperplasia (ALH) and Lobular Carcinoma in Situ (LCIS)
It is not possible to distinguish atypical lobular hyperplasia, lobular carcinoma in situ and even
invasive lobular carcinoma reliably on fine needle aspiration smears alone. The difference between LCIS
and ALH is one of extent of lobule involvement seen in histological sections and is not based on the
cytological appearances of the cell. The cells are identical. The cytological features of ALH have been
described by Salhany and Page [6]. Cytologically dissociated small epithelial cells with rounded or
squared-off nuclei are seen. These are present singly or in small groups with nuclear moulding. The
cells may contain intracytoplasmic lumina (private acini) seen best on mucin staining where they appear
like a "bulls-eye" with an alcian blue stained microvillous membrane and a periodic acid Schiff (PAS)
stained mucin droplet in the centre. ALH and LCIS are usually seen as a chance finding in association
with another lesion, which can result in complex appearances in FNAC.
Atypical ductal hyperplasia (ADH)
ADH is another lesion for which the diagnosis depends on the architectural features and extent of the
lesion seen on histology. As it can be difficult to distinguish ADH from ductal carcinoma in situ (DCIS)
on histological grounds it is not surprising that it may be difficult or impossible cytologically.
Columnar cell change
This may produce dissociation and some authors have noted that the cells may resemble lobular
carcinoma cells. Some of the cells are columnar in nature resembling bronchial epithelial cells.
Lactational change
Even in the screening age group focal lactational changes can occur. This is uncommon but can
produce occasional dissociated cells within an otherwise benign appearing smear. The dissociated cells
may possess nucleoli and have larger nuclei than the surrounding benign cells. They do however have a
moderate quantity of pale blue cytoplasm on Giemsa staining with lipid droplets in the cytoplasm.
Caution in interpreting occasional dissociated cells in an otherwise benign pattern should be exercised
even in the screening age range and the question "could these be lactational/secretory cells" can be
specifically asked in these cases. Outside the screening age a history of pregnancy/lactation should
always be sought and clinicians should always tell the pathologist of lactation or pregnancy.
Radiotherapy changes
These can lead to a false positive cytological diagnosis especially when the history of previous
irradiation is not provided. The aspirate, however, is usually not very cellular and the interpretation
of poorly cellular smears especially with a history of irradiation should be undertaken with caution.
Irradiation can cause marked nuclear pleomorphism and dissociation.
Organising Haematoma / Previous Recent Aspiration or Biopsy
Smears are not very cellular and the haemosiderin may reportedly be interpreted as melanin leading to
an erroneous diagnosis of metastatic melanoma. Problems can also be encountered in aspirates following a
previous aspiration shortly before. This is due to activated macrophages and fibroblasts involved in the
repair process. Re-aspiration should not be performed until 2-3 weeks after a previous aspirate in order
to let this reaction settle.
Intra-mammary lymph nodes
These should not cause a problem if the pathologist recognises the cells as lymphoid. Awareness that
these can occur and can be aspirated should be enough to avoid an error. Lymphomas may be more difficult
to distinguish from carcinoma, but the lack of clumps should suggest the possibility. Careful assessment
including immunocytochemistry should distinguish the occasional carcinoma which shows almost complete
dissociation with a rather plasmacytoid appearance. Examples of bone marrow in aspirates of lesions
stated to be in the breast are rarely seen; the origin of these are assumed to be rib or myelo-lipoma.
Degenerate cells in cyst fluids
Degeneration of cells within cysts or nipple discharge specimens can give pleomorphic appearances
especially when these are larger apocrine cells. Cautious interpretation of cells within degenerate
cysts is advised.
Ultrasound gel
The amorphous appearance of the gel in the background of the smear may suggest necrosis and if the
preparation is rather cellular, with some cellular pleomorphism, this may lead the unwary to a false
positive diagnosis. The problem may be compounded if the gel produces cell lysis.
Granulomatous mastitis
Epithelioid macrophages in granulomatous mastitis can mimic carcinoma cells. They are associated
with other inflammatory cells. The smear is also very cellular. In the presence of inflammation and a
cellular smear the finding of multinucleate macrophages should alert the observer to the possibility of
granulomatous mastitis. The rare cribriform carcinomas with multinucleate giant cells do not usually
contain other inflammatory cells and are therefore distinguishable by their dimorphic picture of small
malignant cells in clumps and singly and more basophilic "osteoclast-like" giant cells with larger nuclei
and prominent nucleoli. Mononuclear forms of the multinucleate cells may also be present.
Other unusual lesions
A range of other
lesions have rarely been described as causing difficulties in diagnosis on FNAC including granular cell
tumours,
adenomyoepithelial lesions and collagenous spherulosis [7], silicone, soya oil or paraffin granuloma, benign stromal
lesions, phyllodes tumours, metastatic tumours and lymphoma and malignant stromal tumours.
Potential False Negative Diagnosis in Breast FNAC
The most common cause of false negative cytological diagnosis is an aspiration miss. There are,
however, types of carcinoma [8] which, by their nature, may produce a false negative diagnosis. The most
common of these are:
Tubular carcinoma [9]
Tubular carcinoma cells often have much in common with benign breast epithelial cells, including
uniformity, nuclear size and often absence of immediately obvious nuclear abnormalities. Knowledge of
the mammographic findings, a lack of bare nuclei, individual cells with cytoplasm and occasional tubular
profiles are pointers to the diagnosis. Paradoxically the nuclei are often more regular and orderly than
benign ductal epithelium and there is a single cell population in the clumps. Often it is not possible
to give an unequivocal diagnosis but care should always be taken in interpreting smears from stellate
opacities to avoid false negative results from this type of tumour. It should be noted that tubules can
occasionally be obtained from benign lesions including radial scars and fibroadenomas.
Infiltrating lobular carcinoma
[6,
8]
Aspirates from this type of carcinoma are often difficult to interpret. The cellularity of these
specimens is usually less than that seen in other forms of invasive carcinoma. A number of patterns can
be observed, ranging in cytological appearance from benign looking uniform cells to atypical cells not
dissimilar to those seen in invasive of no special type (NST). The presence of small three-dimensional
collections of cells with only slightly enlarged nuclei is helpful. A large number of cells with
intracytoplasmic lumina (private acini), in association with the above features, are an indication of
lobular carcinoma, although not specific. Nuclear irregularities and small protrusions from the nucleus
("noses") may also be seen.
Apocrine carcinoma
This rare type of carcinoma produces cellular smears. Difficulty in interpretation is related to the
subtle appearance of the neoplastic apocrine cells and their resemblance to benign apocrine cells with
degenerative changes. Clustering of cells and papillary formations are seen in benign as well as
malignant lesions and are of little help. The key features of a malignant aspirate are the uniform cell
population with nuclear atypia, which one should not confuse with degenerative changes. Necrosis is also
a helpful feature. The diagnosis on FNAC of apocrine carcinoma should always be approached with a great
deal of caution.
DCIS
It should be noted that DCIS and invasive carcinoma cannot be distinguished accurately by FNAC. Some
features have been identified which have been described as being helpful but are not uniformly present [5]
While some of the cases of DCIS are overtly malignant, the small cell type may be missed. The
cellularity of these samples is only moderate. One should be guided by the increased nuclear/cytoplasmic
ratio in the presence of normal size cells and an abnormal nuclear chromatin pattern. The presence of
some necrotic debris in the background should alert the interpreter to the possible malignant nature of
the lesion.
Core Biopsy
With the development of core biopsy "guns" the use of this technique for obtaining a pre-operative
diagnosis of breast lesions has increased significantly [10]. Core biopsy is particularly useful for the
diagnosis of microcalcifications; the calcification can be directly visualised in the histological
section after X-ray of the core has confirmed the presence of the lesion. Benign calcifications can thus
be left in situ in the breast with greater certainty that the radiological abnormality has been sampled
and histologically classified. Malignant calcification can also been identified in foci of DCIS and for
the definitive diagnosis of mammographically suspicious calcification core biopsy is particularly useful,
especially as FNAC is relatively insensitive in the diagnosis of in situ carcinoma in the breast, as
described above.
Reporting Categories for Core Biopsy
Histological examination of core biopsy samples is performed to fulfil the assessment process role by
giving a pathology category classification and not designed to give a definitive diagnosis, although this
is possible in the majority of cases. Thus whilst most core biopsy samples can be readily categorised as
normal, benign or malignant, it must be recognised that a small proportion (probably less than 10%) of
samples cannot. It is also important to remember that although there are five reporting categories
similar to those used in fine needle aspiration cytology (FNAC), these are not equivalent. These
categories are designed to take account purely of the histological nature of the specimen and not the
clinical or imaging characteristics. Similarly it is not feasible for pathology interpretation to judge
independently whether a sample is adequate and from the mammographic lesion. This judgement requires
multidisciplinary discussion. For these reasons there is no inadequate biopsy category for core biopsy
specimens.
Normal Tissue (UK/European category = B1)
This indicates a core of normal tissue whether or not breast parenchymal structures are present; thus
this category is equally appropriate for a core including normal breast ducts and lobules or mature
adipose tissue or stroma only. A normal report should include a description of the components present
and comment should be made regarding the presence of breast epithelial structures. Normal histology may
indicate that the lesion has not been sampled but this is not necessarily so; in the case of certain
benign lesions such as hamartomas and lipomas normal histological features would be expected on core
biopsy. Minor architectural distortions seen mammographically may also result in minimal changes such as
a slight increase in stromal fibrosis on biopsy.
Cores with a normal histological appearance may contain microcalcification, for example within
involutional lobules. It is important in these cases that discussion between pathology and radiology
colleagues is undertaken to confirm the appropriateness of the microcalcification in the histological
specimen. Small foci of calcification within involuted lobules are common and frequently too small to be
visible mammographically, thus a report that merely records the presence of this calcification without
additional comment on its nature, size and site may be misleading and lead to false reassurance. It is
evident that microcalcification, either singly or in clusters, less than 100mm in diameter is not visible
radiologically [11].
Exceptionally some specimens may be classified as uninterpretable, for example due to excessive crush
artefact or composition of blood clot only. Such samples should also be classified as normal.
Benign Lesion (UK/European category = B2)
A core is classified as benign when it contains a benign abnormality. This category is appropriate
for a range of benign lesions including fibroadenomas, fibrocystic changes, sclerosing adenosis and duct
ectasia and extends to include other non-parenchymal lesions such as abscesses and fat necrosis. In some
cases it may be difficult to determine whether a specific lesion is present, for example if minor
fibrocystic changes are seen. The multi-disciplinary approach is once again vital in these cases to
determine whether the histopathological features are in keeping with the radiological and clinical
findings. It may be appropriate and prudent to classify the lesion as normal, rather than benign if only
very minor changes are present, such histopathological features would clearly be insufficient to explain
a well-defined mass lesion and classification as benign would be inappropriate.
Lesion of Uncertain Malignant Potential (UK/European category = B3)
This category mainly consists of lesions which may provide benign histology on core biopsy, but
either are known to show heterogeneity or to have an increased risk (albeit low) of associated
malignancy.
1. Atypical intraductal epithelial proliferations and ADH
There is a range of severity of atypical intraductal epithelial proliferations that fall under this
category, from those which are insufficient for a definite diagnosis of DCIS but highly suspicious to
those which only show a minor degree of atypia, normally architectural which requires further assessment.
Judgement of appropriate categorisation as of uncertain malignant potential or suspicious is required
(see Table 2).
The definition of ADH is derived from surgical resection specimens and relies on a combination of
histological, morphological and size extent criteria. For this reason it is our opinion that accurate
diagnosis of ADH is not possible on core biopsy. It has however been shown that core biopsy samples
which include atypical intraductal epithelial proliferative foci, of insufficient extent for
classification as DCIS, on subsequent surgical resection may form part of an established in situ
neoplastic lesion with or without associated invasion. This view is based on several studies, which
describe the subsequent surgical diagnoses in cases described as ADH in non-operative core biopsy. In
over 50% of cores surgical excision biopsy has shown either in situ or invasive carcinoma [12]. Clearly
this will depend on the number of cores taken and also the size of the cores. In larger, vacuum-assisted
biopsies this "understaging" is less of a problem. This is not surprising as ADH is basically defined as
an intraductal epithelial proliferation showing the features of low grade DCIS, but in less than two duct
spaces or less than 2 mm in diameter. The limited tissue sampling which can be undertaken by core biopsy
guns (often by stereotactic methods for foci of microcalcification) may thus provide insufficient
material for definitive diagnosis of low grade DCIS if only a few duct spaces are obtained. In these
cases a diagnosis of atypical intraductal epithelial proliferation and a classification of uncertain
malignant potential or suspicious of malignancy should be made dependant on the severity and extent of
the lesion.
2. Papillary lesions (Table 3)
Papillary lesions may show significant intra-lesional heterogeneity and the limited sampling achieved
with core biopsy may miss areas of in situ cancer. The majority of these lesions should, therefore, also
be designated being of uncertain malignant potential. On rare occasions when a small lesion has been
very widely sampled and submitted for pathological examination a benign classification may be considered.
Conversely, when a sample of a papillary lesion in a core biopsy shows atypia, for example strongly
suspicious of papillary carcinoma in situ, a suspicious designation may be more appropriate.
3. Radial scar/complex sclerosing lesion
Biopsies which show features of a radial scar/complex sclerosing lesion such as areas of
hyalinisation, elastosis and tubular entrapment with epithelial proliferation should be categorised as of
uncertain malignant potential. Although still a matter of debate, many authorities believe that a
proportion of these lesions are associated with malignancy. Thus we believe at the present that, unless
the sclerosing lesion is very widely sampled, the process should be designated as uncertain malignant
potential as the presence of an associated area of DCIS or an invasive carcinoma cannot be excluded.
In several centres, trials are underway regarding the possibility and efficacy of excising some of
these lesions of uncertain malignant potential with vacuum-assisted biopsy devices. Thus both papillomas
and radial scars may potential be completely excised by these techniques rather than by surgical
diagnostic excision following core biopsy. These lesions can then be diagnosed unequivocally and a core
biopsy diagnostic category is not used.
Table 2: Epithelial Proliferative Lesions in Core
Biopsies
| | |
| I - Usual hyperplasia | | Benign |
| II - Microfocal epithelial atypia in lobules | Minimal
Moderate
High grade | Benign
Uncertain malignant potential (UMP)
Suspicious |
| III - Lobular neoplasia | Typical
Indistinguishable from low grade DCIS (rare) | UMP
Malignant |
| IV - Low & intermediate grade atypical intraductal epithelial proliferation |
| 1. One or a small number of ducts involved | Sufficiently worrisome not to ignore, but lacking in extent or degree of duct / lobule involvement to classify as suspicious of DCIS. Similar features in a surgical excision would be classified as ADH | UMP |
| 2. Greater extent / multiple ducts | Features of low grade DCIS in architecture and epithelial character but insufficient for confident diagnosis of DCIS | Suspicious |
| 3. Multiple ducts | Complete involvement of at least 2 spaces with definite features of DCIS | Malignant |
| V - High grade atypical epithelial proliferation | Part of one duct
One or more complete duct profiles involved. Caution is advised when a single profile only is present. Additional features such as necrosis may be useful | Suspicious
Malignant |
Table 3: Papillary lesions in core biopsies
4. Lobular neoplasia
A small cell regular epithelial proliferation within lobules which is considered by the pathologist
to represent intralobular neoplasia (ALH/LCIS) should be classified as uncertain malignant potential.
This process does not have the same management implications as a diagnosis of DCIS or invasive malignancy
and does not per se require therapeutic excision. Lobular neoplasia is most frequently a co-incidental
finding in a core biopsy from a screen-detected lesion however and multidisciplinary discussion is
essential, as the abnormality identified radiologically may not be represented. These cases must be
managed cautiously.
On occasions it may be impossible to classify a small cell epithelial proliferation in lobules and/or
ducts as either lobular neoplasia or low grade DCIS and in these circumstances a higher category
(suspicious or malignant) is prudent and should be considered.
5. Phyllodes tumour
Fibroadenomatoid lesions with cellular stroma, stromal overgrowth and possibly some mitotic activity
suggesting a phyllodes tumour should be designated uncertain malignant potential. Thus the presence of a
cellular stroma within a fibroepithelial lesion should prompt a search for other features that may aid in
discrimination from a fibroadenoma. In practice, however, this distinction is often impossible and
careful appraisal of the entire clinical picture will usually allow appropriate management to be
undertaken.
Suspicious (UK/European category = B4)
Technical problems such as crushed or poorly fixed cores, which contain probable carcinoma but cannot
provide the definitive diagnosis, are best included as suspicious. Similarly, apparently neoplastic
cells contained within blood clot or adherent to the outer aspect of the sample should be classified as
suspicious.
A complete single duct space bearing an unequivocal high-grade atypical epithelial proliferative
process can be classified as malignant. However care must be taken if one or only part of a duct space
is seen containing a highly atypical epithelial process (particularly if no necrosis is present); this
may be regarded as suspicious rather than definitively malignant. In particular great care should be
taken if the epithelial cells show any features of an apocrine phenotype, which may represent an atypical
apocrine proliferation rather than DCIS.
Another lesion that can be allocated to this category is a non-high grade intraductal proliferation
with a significant degree of atypia probably representing intermediate or low grade DCIS, where
relatively few involved duct spaces are represented in the biopsy. A pragmatic approach is usually
required by reporting an atypical intraductal proliferation and qualifying this according to the degree
of suspicion, i.e. 'at least ADH, probably low grade DCIS'. The case will then be allocated to either
to the uncertain or to suspicious category on the basis of the extent and severity of atypia.
The management of cases classified as suspicious will usually be either diagnostic excision biopsy of
the area or repeat core biopsy sampling to obtain definitive diagnosis. We believe
that definitive therapeutic surgery should not be undertaken as a result of an uncertain malignant
potential or suspicious core biopsy diagnosis.
Malignant (UK/European category = B5)
This category is appropriate for cases of unequivocal malignancy on core biopsy. Further
categorisation into in situ and invasive malignancy should be undertaken whenever possible. Other forms
of malignancy such as malignant lymphoma should also be classified as malignant.
1. Ductal carcinoma in situ
One of the benefits of core biopsy is that it can allow distinction between in situ and invasive
carcinoma. It should however be borne in mind that, due to sampling error, presence of DCIS alone in the
core does not exclude the possibility of an invasive focus being present. In approximately 20% of cases
sampled by standard methods co-existing invasive carcinoma will be identified in the subsequent surgical
excision specimen [13]. The nuclear grade, architecture and the presence of necrosis of the DCIS can be
indicated on the core biopsy report. In particular, the presence of associated calcification should be
recorded.
As noted above in exceptional circumstances lobular neoplasia may be impossible to distinguish from
small cell solid DCIS and may be classified as malignant.
2. Invasive carcinoma
An advantage of core biopsy over FNAC is the ability to diagnose invasion positively. Invasive
carcinoma can be unequivocally identified in core biopsy with a positive predictive value of 98% [13].
As noted above, however, the negative predictive value for invasion is only 80% when only DCIS is
identified.
General Principles to Avoid Pitfalls in Core Biopsy Interpretation
Proper interpretation of core biopsies requires knowledge of details of both clinical and
mammographic findings and this information should be provided on the request form. The completed request
form should include clinical details, specifying the radiographic sign and the site of biopsies.
| Biopsies performed from microcalcifications should be x-rayed to determine the presence of calcium. Whenever possible a radiological comment regarding the presence of representative microcalcification of the mammographic lesion in the sample should be provided along with the specimen x-ray. Examination of further levels should be performed if the calcification is not immediately apparent on histological examination. |
| Optimal fixation is paramount. Biopsies should be placed in fixative solution immediately and sent promptly to the laboratory. Ideally biopsies should be fixed routinely for a minimum of 6 hours although specimens may be fixed rapidly with the aid of microwave techniques. |
| After processing haematoxylin and eosin stained sections from one level are usually sufficient for core biopsies from mass lesions but core biopsies taken for the investigation of microcalcification should have a minimum of three levels examined. In practice most laboratories choose to examine all core biopsies from screen detected lesions at least 3 levels initially. In problematic cases further levels may be helpful. |
| Immunohistochemical studies may be invaluable in difficult cases. |
|
Calcifications
If calcifications are present on the specimen X-rays, microcalcification can be seen histologically
in 78% of cases and diagnosis made in 81% [14]. However if calcifications are not present on the core
biopsy radiographs they may nevertheless be identified histologically in 13% of specimens and a specific
diagnosis can still be made in 38% of cases [14]. Care must, however, be taken to review the
histological calcifications and mammographic films at a multidisciplinary meeting so that a decision can
be reached regarding the correlation of the appearances. If this is not done, small foci of benign
histological calcification, which are in fact too small to be those mammographically visible, may be
believed to be the source of the X-ray features.
Problems and Pitfalls in Core Biopsy Diagnosis
Table 4: Common causes of false positive and negative core biopsy diagnoses.
| COMMON CAUSES OF FALSE POSITIVE DIAGNOSIS |
| 1. | Sclerosing adenosis or radial scar/CSL mistakenly diagnosed as tubular carcinoma. |
| 2. | Apocrine atypia in lobules, ducts or sclerosing lesions mistakenly diagnosed as DCIS. |
| 3. | Chronic inflammation mistakenly diagnosed as infiltrating lobular carcinoma. |
| 4. | Invasion mistakenly diagnosed in DCIS. |
| 5. | Radiotherapy changes mistakenly diagnosed as carcinoma. |
COMMON CAUSES OF FALSE NEGATIVE DIAGNOSIS |
| 1. | Tubular carcinoma mistakenly diagnosed as sclerosing adenosis or radial scar/CSL. |
| 2. | Infiltrating lobular carcinoma mistakenly interpreted as chronic inflammation or missed. |
| 3. | Radiotherapy effect with missed foci of carcinoma. |
| 4. | Metaplastic carcinoma mistakenly diagnosed as a stromal proliferation/fibroblastic scar. |
|
Diagnostic pitfalls and problems in diagnosis in core biopsy include many of the lesions that cause
difficulties in FNAC diagnosis. Other lesions however may present particular diagnostic problems in core
samples.
Minor degrees of epithelial atypia
Mild atypia of epithelium within lobular units is one of the commonest problems encountered in core
biopsy samples. Care must be taken not to overdiagnose such minimal degrees of atypia, which may
represent usual epithelial hyperplasia, apocrine change or reactive changes (for example adjacent to
previous sampling procedure). Conversely more severe degrees of atypia must be sought which may reflect
cancerisation of lobules by high grade DCIS. The degree of atypia should be helpful in distinguishing
the process and the nuclear chromatin and presence of mitoses (although rarely seen) may aid in the
diagnosis. Similarly, usual epithelial hyperplasia (UEH) is commonly seen in cores from benign
fibroadenomas. This often shows apparent discohesion due to the trauma of the core biopsy sampling
process and "telescoping" of epithelium is seen within the duct spaces thus resembling a hyperplastic
process. As with UEH in surgical excision specimens, the lack of uniformity and distribution/ streaming
of the epithelial cells with bland nuclear features and paucity of mitoses is of assistance in reaching a
diagnosis. ADH should not be diagnosed in these cases unless uniformity of nuclear size and shape and
regular, evenly placed nuclei are seen. Usual epithelial hyperplasia of gynaecomastoid type with a
micropapillary architecture should not be mistaken for micropapillary ADH/DCIS.
Apocrine atypia and apocrine DCIS
Apocrine atypia, particularly in association with a sclerosing lesion such as sclerosing adenosis
(so-called "apocrine adenosis") may be especially difficult to identify correctly in non-operative
diagnostic samples. In core biopsy large nuclei, often with prominent nucleoli may be mistaken for DCIS
if pleomorphism is also present. The typical granular eosinophilic cytoplasmic appearance of apocrine
cells should be sought. Pure apocrine DCIS is relatively rare and when an apocrine proliferation is seen
within ducts in a core biopsy, additional features of malignancy such as significant atypia, intraluminal
necrosis and the presence of mitoses as well as multiple duct involvement should be sought for
confirmatory evidence. In addition multiple duct involvement indicating a more extensive lesion may
provide further supportive evidence. Mild or moderate degrees of apocrine proliferation with atypical
features in a duct space should be assessed with caution and it may be prudent not to record a definite
diagnosis but to classify such a process as being of uncertain malignant potential. Conversely papillary
apocrine change should not be mistakenly classified as other than benign.
Lactational change
As described above, focal lactational change may be seen in women who are neither lactating, nor
pregnant and indeed are nulliparous and/or post menopausal. The involved acini are usually lined by
plump vacuolated cells with a "hobnail" architecture but may, less frequently, appear atypical with
irregular, large or pyknotic nuclei. The epithelial cells may appear degenerative and rarely the benign
nature of the process may be mistaken for cancerisation of lobules by DCIS. The recognition of the
vacuolation of the cytoplasm and the typical hobnail architecture will enable the correct diagnosis to be
established.
Sclerosing lesions/tubular carcinoma
There is a risk of overdiagnosis of invasive carcinoma when confronted by sclerosing adenosis in a
core biopsy, particularly as the normal lobular arrangement may be less apparent than on an excision
biopsy specimen. Immunohistochemical staining with collagen IV, laminin and/or smooth muscle actin to
demonstrate the presence of a basement membrane and a dual epithelial/myoepithelial layer respectively
can be extremely useful in this situation. The stromal appearances may be helpful in achieving a correct
diagnosis; sclerosing lesions do not induce the fibroblastic/desmoplastic reaction generally seen in an
invasive carcinoma. Radial scars/complex sclerosing lesions have an eosinophilic fibrotic and elastotic
central focus with entrapped tubules. These latter elements have, however, a surrounding myoepithelial
layer. In difficult cases immunohistochemistry for smooth muscle actin may be invaluable; the absence of
the myoepithelial component in tubular carcinomas can be confirmed.
Stromal proliferations and spindle cell lesions
Occasionally a second biopsy sample will be taken with a fibroblastic proliferation which may
represent the target lesion, but which may reflect tissue reaction and repair at a previous core biopsy
site. If the lesion represents the core site, an associated histiocyte reaction or fat necrosis may be
present and haemosiderin-laden macrophages seen. Sometimes a fibroblastic stroma may be identified in a
sample from a patient who has not undergone previous FNAC or core biopsy and which may represent a
spindle cell proliferation such as a fibromatosis or part of a spindle cell tumour such as a nerve sheath
tumour or myofibroblastoma. A stromal proliferation may also be seen as the only evidence of a phyllodes
tumours and an epithelial component should be sought, for example by performing additional levels.
Metaplastic carcinomas may also mimic stromal proliferations and a high index of suspicion may enable
confirmatory diagnosis by immunohistochemical examination with a range of anti-cytokeratin antibodies (at
least one broad spectrum and a high molecular weight cytokeratin). Immunohistochemistry may prove
unhelpful in stromal proliferations and the multidisciplinary approach must be applied to the clinical,
radiological and histopathological features. When a definitive histological diagnosis cannot be made the
abnormality should be reported as a spindle cell lesion of uncertain histogenesis or nature and
classified as uncertain malignant potential.
Fibroepithelial tumours
As noted above, phyllodes tumours may rarely be difficult to distinguish from other stromal lesions.
More commonly the differential diagnosis lies between a cellular benign fibroadenoma and a phyllodes
tumour. Features including stromal atypia, if present, can be useful, but the degree of cellularity of
the stroma is the most valuable feature to assess. In rare cases it is not possible to distinguish the
two lesions and the sample should be reported as a "fibroepithelial lesion" and classified as uncertain
malignant potential, to avoid underdiagnosis of a phyllodes tumour.
Radiation induced changes
Radiotherapy changes to the breast may be difficult to differentiate from foci of recurrent or
residual carcinoma, both in situ and invasive. The radiation induces a degree of atypia of the breast
epithelium but also in the histiocyte population, which is prominent as a result of the radiotherapy and
also recent surgery. The macrophages may also show degenerative features. Carcinoma cells can
conversely mimic macrophages. Immunocytochemistry can be helpful in difficult cases as irradiated
neoplastic cells retain cytokeratin expression whilst macrophages demonstrate a histiocytic phenotype
(e.g. CD68 reactivity).
Infiltrating lobular carcinoma
Small foci of invasive lobular carcinoma can be missed in histological sections and be dismissed as
chronic inflammation or stromal cells. The targetoid infiltrative pattern of classical lobular carcinoma
may be of assistance but a reactive lymphocyte process can also have a peri-ductal or peri-lobular
distribution. Cytokeratin immunohistochemistry, to demonstrate the neoplastic cells is of value in
difficult cases but recognition of the abnormal cell proliferation requires vigilance, as the features
can be subtle.
Rare Potential Pitfalls in Core Biopsy Diagnosis
Lymphoma
Malignant lymphoma may rarely be identified in core biopsies and should be classified as malignant.
The majority of these lesions are of high grade B cell morphology and may mimic carcinoma. The cells
frequently show less cohesion, a higher nuclear to cytoplasmic ratio and do not demonstrate the
architectural features of carcinoma. To avoid misclassification as carcinoma, however, the correct
diagnosis must be considered. Low-grade lymphomas may be more difficult to distinguish, mimicking a
chronic inflammatory process. Infiltration of the lobular epithelium should be sought and the degree of
lymphoid infiltrate, if high, should raise the possibility of a neoplastic process. A panel of lymphoid
markers (CD45, CD20, CD3, CD30 etc.) will demonstrate the immunophenotype of the cells present and to
allow correct diagnosis.
Metastasis to the breast
Metastasis to the breast from malignancies derived elsewhere is well recognised although rarely
biopsied. Lesions that are recognised as metastasising to the breast include lung, ovarian, renal and
prostatic carcinomas but non-epithelial malignancies such as melanoma, myelomas and rhabdomyosarcomas may
also be seen. A full clinical history is essential. A panel of antibodies frequently allows
identification of the likely site of a metastatic adenocarcinoma and enables appropriate clinical
investigation/management. Breast carcinoma usually express Cytokeratin 7 and 18 (and not Cytokeratin
20), epithelial membrane antigen, CEA/NCA and approximately 80% will express oestrogen receptor.
Sarcomas
Primary breast sarcomas are rare. They most commonly originate in association with phyllodes tumours
but in core biopsy specimens an epithelial component is often not present. The most common phyllodes
associated sarcomas seen are liposarcoma and fibrosarcoma although other differentiation including
osteosarcoma, chondrosarcoma and rhabdomyosarcomas can be identified. Angiosarcomas may be a cause of
false negative diagnosis as they may be relatively subtle and bland and may be mistaken for radiotherapy
changes, particularly when they occur in this situation in the treated breast. Primary leiomyosarcoma
(and leiomyoma) may be found in the breast; the latter most commonly seen in a retroareolar site. All
these lesions can be difficult to diagnose definitively in core samples. A high index of suspicion and
judicious use of immunohistochemistry can facilitate or support a diagnosis.
Conclusion
Breast FNAC required experience and training in both aspiration and assessment. False positive
diagnoses can be minimised by good smearing and staining of samples with abundant clinical information
and a multidisciplinary approach by the pathologist, radiologist and surgeon and the other members of the
breast "team". However pitfalls in the diagnosis of core biopsies can also be made without this same
approach and whilst some clinicians and pathologists believe that core biopsy interpretation is
"histology" and therefore easy and straightforward, this is not always the case.
References
- Cytology Sub-Group of the National Coordinating Committee for Breast Screening Pathology. Guidelines for cytology procedures and reporting in breast cancer screening. Sheffield: NHSBSP Publications, 1992
- Zakowski MF. Fine-needle aspiration cytology of tumors: diagnostic accuracy and potential pitfalls. Cancer Invest 1994;12:505-515.
- Stewart FW. The diagnosis of tumours by aspiration. Am J Pathol 1933;9:801-812.
- Kline TS. Masquerades of malignancy. A review of 4,241 aspirates from the breast. Acta Cytologica 1981;25:263-266.
- Zakhour H, Wells C. Diagnostic Cytopathology of the Breast. London: Churchill Livingstone; 1999.
- Salhany KE, Page DL. Fine needle aspiration of mammary lobular carcinoma in situ and atypical lobular hyperplasia. Am J Clin Pathol 1989; 92: 22-6.
- Wells CA, Wells CW, Yeomans P, Vina M, Jordan S, d'Ardenne AJ. Spherical connective tissue inclusions in epithelial hyperplasia of the breast ("collagenous spherulosis"). J Clin Pathol 1990; 43:905-8.
- Lamb J, Anderson TJ. Influence of cancer histology on the success of fine needle aspiration of the breast. J Clin Path 1989; 42: 733-5.
- 28. Bondesan L, Lindholm K. Aspiration cytology of tubular breast carcinoma. Acta Cytologica 1990; 34: 15-20.
- Parker SH, Lovin JD, Jobe WE, Burke BJ, Hopper KD, Yakes WF. Nonpalpable breast lesions: stereotactic automated large-core biopsies. Radiology 1991;180:403-407
- Dahlstrom JE, Sutton S, Jain S. Histologic-radiologic correlation of mammographically detected microcalcification in stereotactic core biopsies. Am J Surg Pathol 1998;22:256-9
- Liberman L, Cohen MA, Dershaw DD, et al. Atypical ductal hyperplasia diagnosed at stereotaxic core biopsy of breast lesions: an indication for surgical biopsy. AJR 1995; 164: 1111-3.
- Liberman L, Dershaw DD, Rosen PP, Giess CS, Cohen MA, Abramson AF, et al. Stereotaxic core biopsy of breast carcinoma: accuracy at predicting invasion. Radiology 1995;194:379-81.
- Liberman L, Evans Wr, Dershaw DD, et al. Radiography of microcalcifications in stereotaxic mammary core biopsy specimens. Radiology 1994;190:223-225.
|
