Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
The technique of fine needle aspiration biopsy (FNAB) of breast has developed over the past 60 years to be an extremely useful and accurate method of diagnosing palpable and impalpable lesions.
Breast FNAB can attain sensitivity in the range of 90% to 95% and a positive predictive value (PPV) of malignancy of 99%, with a very low false-positive rate, usually related to FNAB of fibroadenomas, papillomas, and papillary lesions, and a low false-negative rate related to low-grade ductal and lobular carcinoma.
There has been a long and highly successful practice of using FNAB for palpable lesions of the breast. More recently, the use of FNAB in assessment of mammographically detected lesions, including masses and clustered calcifications, has allowed mammographic screening programs and their assessment clinics to function effectively, by providing an accurate and inexpensive method of diagnosing detected lesions. The use of FNAB in the multidisciplinary setting, especially with rapid on-site assessment to triage palpable and impalpables cases needled under ultrasound direction, has been proven to be a highly cost-effective management of breast lesions.
But the use of FNAB for mammographically detected lesions has highlighted difficulties presented by the workup of calcificied lesions and the subclassification of proliferative breast lesions, and the role of breast FNAB in mammographically detected lesions has been debated. Proliferative lesions include epithelial hyperplasia with and without atypia, fibroadenomas, papillomas with epithelial hyperplasia, radial scars, and columnar cell change and its variants. Their distinction from atypical ductal hyperplasia, low- and intermediate-grade ductal carcinoma in situ, papillary carcinomas, and low-grade carcinomas can be difficult on FNAB. Often the technical limitations of stereotactic and ultrasound localization may limit the sample, but cytopathologists have to be aware of the diagnostic criteria for these proliferative lesions and have clear criteria for the unequivocal diagnosis of malignancy. It is essential to avoid false-positive diagnoses with their inherent patient distress, risk of overtreatment, and medicolegal problems.
FNAB can drain and confirm cysts and confirm nonspecific and specific benign lesions such as fibroadenomas, thereby completing the all-negative “triple test.” The triple test consists of clinical, imaging, and cytologic assessment and has a negative predictive value (NPV) of virtually 100%. FNAB of breast can also achieve high accuracy in diagnosing proliferative lesions and carcinoma, confirm inoperable carcinomas for neoadjuvant chemotherapy, diagnose axillary lymph node metastases and chest wall recurrence, and, most importantly, provide rapid on-site evaluation (ROSE) for cost-effective immediate triage of the patient in the multidisciplinary breast clinic.
In cases where this provisional report is indeterminate due to inadequate sampling or the presence of atypical cells, or the findings do not correlate with the triple test, the patient can be triaged for repeat FNAB or immediate core biopsy. Both FNAB and core biopsy can provide material for ancillary testing of malignant lesions: FNAB-derived cell blocks can be used for immunohistochemistry (IHC) for estrogen and progesterone receptors, cytokeratin 5/6 or 14 for basal-type carcinomas, and E-cadherin to help confirm lobular carcinomas and for HER2, although fluorescence in situ hybridization (FISH) is preferred for HER2. FNAB can also provide material for RNA extraction for molecular typing. Similarly, if carcinoma is diagnosed, core biopsy can be performed to allow precise carcinoma typing and material for prognostic indicators, as well as Oncotype DX (Genomic Health) and MammaPrint (Agendia).
Breast FNAB has difficulties distinguishing intraductal from invasive carcinoma, although criteria favoring invasion and high-grade intraductal carcinoma have been published and are presented in this chapter. This difficulty is far less of a problem when FNAB is used correctly within the triple test and with immediate provisional reports, especially in the era of preoperative axillary lymph node FNAB and sentinel lymph node biopsy, which now commonly determine whether axillary dissection is performed.
FNAB and core biopsy in many institutions are regarded as complementary, with FNAB offering major benefits in the speed of initial diagnosis, low cost, a high rate of acceptance by patients, lower rate of complications such as bleeding, and the provision of an excellent screening test, especially for palpable masses, with a sensitivity in the region of 90% to 95% for diagnosing carcinoma. This is especially so when FNAB, using ultrasound when necessary, is performed by experienced cytopathologists with immediate assessment of Giemsa-stained slides. Combining FNAB with immediate assessment as a screening test decreases costs significantly by triaging each case, decreasing inadequacy rates, increasing sensitivity, and decreasing patient waiting time, anxiety, clinic recalls, and costs. The majority of women attending a symptomatic clinic for a multidisciplinary assessment with rapid on-site evaluation of FNAB smears are managed with a same-day diagnosis obviating the need for recall. At the same clinic visit, core biopsy can be then used only where necessary, reducing the need for the more expensive and invasive test.
Core biopsies have similar sampling error and crush artifact problems as FNAB; cause greater bruising and pain; are more expensive; take longer to process, requiring a histopathology laboratory; can remove marker calcifications and smaller lesions; and in general cannot provide an immediate report. As with FNAB, core biopsies have difficulty distinguishing lobular neoplasia, papillomas versus intraductal papillary carcinoma, cellular fibroadenomas versus phyllodes tumors, and atypical ductal proliferations versus low-grade intraductal carcinomas. Core biopsies do offer greater specificity in the diagnosis of microcalcifications, some proliferative lesions, low-grade intraductal carcinoma, and possibly in the subtyping of invasive carcinomas. Some clinical trials demand core biopsies.
In the multidisciplinary breast clinic setting, where committed radiologists and cytopathologists perform the FNAB and correlate findings in the triple test, sensitivity and specificity rates for FNAB and core biopsies are comparable. Of course, in the 80% of the world where medical resources and funds are severely limited, FNAB of the breast is a practical, inexpensive, and rapid diagnostic tool for breast lesions in an environment where core biopsies, surgical excision, and well-resourced histopathology laboratories are not readily available.
The risk of FNAB causing damage to a lesion, making subsequent histologic interpretation difficult, is low and less likely than with core biopsy, which may actually remove a lesion. Repeat FNAB may yield considerable hemosiderin and debris related to hemorrhage. Infarcted lesions are occasionally sampled by FNAB and are most commonly fibroadenomas, phyllodes tumors, or papillomas in younger women and may be associated with lactation.
Liquid-based preparations largely destroy the patterns used in the cytologic diagnosis of breast lesions in this chapter and are more expensive. The key to FNAB of the breast, as with all procedures in diagnostic pathology and surgery, is the expertise of the operator performing the FNAB and making the direct smears, who may be a clinician trained to perform excellent FNAB or ideally a cytopathologist, with or without ultrasound direction.
In young women mammography has little role in screening or diagnosing breast carcinoma, and reliance is placed on ultrasound and particularly FNAB to avoid a delay in diagnosis due to a low index of suspicion. It is probable that when grade and stage are taken into account, breast carcinoma in younger women has a slightly worse prognosis, but when delays are avoided the usual triple test workup, using ultrasound, can provide similar diagnostic accuracy to that achieved in older women. Similarly, FNAB is preferred in pregnant or breast-feeding women because FNAB reduces complications inherent in core or excision biopsy of the lactating breast.
Sentinel lymph node biopsy has largely replaced routine axillary dissection, with its significant lymphedema, pain, joint problems, and numbness, but it is expensive and time consuming and still has its own complications and false negatives, leading to the need for repeat surgery. Preoperative FNAB of axillary lymph nodes to detect metastatic disease in women with breast lesions using ultrasound guidance and ROSE has a low inadequate rate, high specificity and PPV approaching 100%, and moderate sensitivity and NPV around 75% to 80%. The FNAB provides information to help in deciding the need for neoadjuvant chemotherapy and allows for assessment of the effectiveness of various preoperative chemotherapy protocols when surgical excision subsequently is carried out.
Because ultrasound sensitivity is not that high, it has been suggested that FNAB should be carried out on any lymph node found clinically or on ultrasound rather than just on lymph nodes thought to be abnormal at ultrasound. False positives are rare, and patients with positive FNAB are spared sentinel node biopsy, while false negatives due to sampling are picked up on subsequent sentinel lymph node biopsy. Recently, several large studies have suggested that the complications of axillary lymph node dissection may outweigh its diagnostic value, even when the patient is found to have positive sentinel nodes.
In the developing world, FNAB of axillary lymph nodes is a cost-effective and efficient alternative to sentinel node biopsy and routine axillary dissection. Core biopsy has similar false-negative rates and is more costly and invasive in the axilla but may offer a slightly better NPV.
A breast FNAB report should include a statement of the adequacy of the specimen, the degree of cellularity, a clear cytologic description, a specific diagnosis wherever possible, a statement of whether malignancy is present, and a code to allow overall categorization and management of the lesion. The report and conclusion should always provide a clear conclusion to facilitate communication between the cytopathologist and clinician, and a code should never be the sole conclusion but should always be accompanied by the descriptive conclusion. The code can be useful for clarity and quality assurance measures. Five codes are recommended:
Code 1: Insufficient material
Code 2: Benign
Code 3: Atypical, probably benign
Code 4: Suspicious, probably in situ carcinoma or malignant
Code 5: Malignant
The atypical (Code 3) category for “probably benign, malignancy cannot be excluded” should be kept to a minimum. The “suspicious, probably malignant” category should yield a majority, that is, more than 80% of carcinomas including undersampled or low-grade invasive carcinomas and ductal and lobular in situ lesions. Lesions that may yield an atypical or suspicious result include radial scars, epithelial hyperplasias with or without atypia, papillomas, fibroadenomas, low- and high-grade intraductal carcinomas, and low-grade invasive ductal carcinoma and lobular carcinoma. Categories 3 and 4 decrease with increased operator and cytopathologist experience, as do insufficient rates that include false-negative cases.
Optimal diagnostic accuracy is achieved when the FNAB is performed by an experienced cytopathologist or radiologist using ultrasound where necessary, and good smear-making techniques are applied with immediate staining and assessment for adequacy and provisional diagnosis. The results are ideally correlated with imaging and clinical findings in a multidisciplinary clinic setting, to triage the patient for immediate further FNAB or core biopsy if required. The FNAB should, as much as possible, correlate with the clinical findings and other findings in the triple test. If there are discrepancies, further biopsy is required.
The cellularity is determined by the operator experience, number of passes, and size and nature of the lesion. In terms of adequacy, the cellularity that is adequate to confirm the imaging findings for fibrocystic change, a cyst, entrapped fat, or fat necrosis is not adequate for a well-defined lesion found on imaging, which requires a specific diagnosis. In all such cases, including ill-defined thickenings, it is recommended that a minimum of seven distinct epithelial tissue fragments should be present, each of more than 20 cells to allow for assessment of fragment architecture. A case is categorized Code 1 if the amount of cellular material is less than this, unless the imaging shows either a cyst, which is subsequently completely drained, or probably fat tissue. Code 1 is not used to categorize cases in which the cytologic findings do not correlate with the imaging findings. In such cases where the cytologic findings are discrepant in the triple test, the cytology findings should be clearly reported, the imaging should be reviewed, and further biopsy is usually required.
Ideally, each FNAB should have both alcohol-fixed Papanicolaou (Pap)-stained smears and air-dried Giemsa (DiffQuick or similar)–stained smears because the two stains are complementary and allow assessment of different features. Sample splitting of the material by quickly transferring part of the sample to a second slide before smearing of both slides is recommended. Routine cell block preparation from the normal saline washings of needle and syringe is also recommended, and dedicated whole FNAB passes can be placed in the cellblock material when IHC for prognostic markers or other antigens is required.
The FNAB needle size used includes 22G, 23G, 25G, or 27G, but the 23G or 22G is recommended. The finer-gauge needles will tend not to provide the larger tissue fragments that are diagnostic in intraductal papillomas, fibroadenomas, and low-grade intraductal carcinomas. The needle should be attached to a syringe held in a holder when first needling a cyst or simple palpated nodule. The needle by itself can be used effectively in most other situations and produces less obscuring blood on the slide. The needle is passed rapidly into the lesion 10 to 20 times in a “woodpecker” action, with swiveling of the needle if it is held between the fingers, maximizing the cutting action of the bevel. The needle should be withdrawn after 6 to 8 seconds of these rapid passes. Except in cystic lesions, when there is an immediate flash of cyst fluid into the needle hub, aspiration should not be applied till late in the FNAB and not at all once material or blood is seen in the hub of the needle, which is an indicator to cease immediately the FNAB. Keeping the needle in the lesion longer leads to fibrin clot forming in the needle, and this is particularly true of second and subsequent passes.
Rather than using 100% alcohol fixation for wet-fixed slides, we recommend placing the slides immediately into modified Carnoys, which lyses the blood (modified Carnoys fixative: 1 liter, consisting of 665 mL methanol, 250 mL chloroform, 35 mL distilled water, and 50 mL acetic acid). We also recommend using a warm hair dryer to rapidly fix air-dried slides to avoid “slow air-drying artifact.”
Interpreting breast FNAB Pap alcohol-fixed and Giemsa air-dried slides requires a consistent stepwise approach based on pattern recognition and high-power assessment.
First, low-power assessment is made of the cellularity and pattern of tissue fragments and dispersed cells, the architecture of the tissue fragments, the presence or absence of other features such as stroma, and the degree of smearing artifact, which impacts on the pattern.
Second, high power is used to confirm features seen at low power, such as myoepithelial cells on epithelial tissue fragments and bare bipolar myoepithelial cell nuclei in the background, and then to assess the types of dispersed cells and tissue fragments. Finally, high power determines the degree of epithelial nuclear atypia in the tissue fragments and dispersed cells.
Finally, the low- and high-power features are synthesized in a final diagnosis.
The diagnostic process requires constant switching from the crucial low-power assessment of pattern to high-power confirmation, while avoiding a rush to high-power assessment of individual cell nuclear atypia. Assessment of nuclear atypia is the final mandatory step in diagnosing breast malignancy, but atypia can be marked in benign breast lesions such as fibroadenomas, and nuclear atypia should always be assessed in light of the overall pattern and other features. If there are discrepancies or noncorrelating cytologic features between low- and high-power assessments, reassessment of the cytologic diagnosis is required, and this is best performed fully aware of the clinical and imaging findings in the triple test.
In an ideal situation, the cytopathologist performs the FNAB, gaining valuable information about the lesion from the texture as the needle enters the lesion. In a multidisciplinary clinic the cytopathologist may work closely with an experienced radiologist or clinician who performs the FNAB and will often know radiologic and clinical details of the case when reporting. However, in the setting where the FNAB is performed elsewhere or by a radiologist or clinician under ultrasound direction, the cytopathologist should assess the breast FNAB and reach a provisional diagnosis knowing only the patient’s age and hormone status unbiased by any imaging impressions. The cytopathologist should then reassess after reading any clinical and imaging information available and correlate his diagnosis with the other findings in the triple test. If the cytologic finding is discrepant, the cytopathologist should reassess the cytology and the cytology and radiology findings should be discussed in a multidisciplinary setting. If a discrepancy remains, repeat FNAB or core biopsy is usually warranted ( Table 5-1 ).
Step 1: Low-power assessment (×2-×20)
Step 2: High-power assessment (×40)
|
First, the epithelial cellularity is assessed as adequate or inadequate and categorized as scant or mildly, moderately, or highly cellular. Scant cellularity is defined as 7 to 10 tissue fragments each of more than 20 cells, which provides tissue fragments large enough to allow assessment of architecture. The cellularity is an indication of the diagnostic potential of a particular slide or set of slides and can also be a measure of the confidence that can be put in the diagnosis; that is, adequately smeared slides with more material provide a better opportunity to make a diagnosis. For this reason a minimum of at least two and up to four passes are recommended in all breast FNAB.
The expected cellularity varies with the nature of the lesion and patient’s age, menstrual cycle, menopausal status, and medications, especially hormone replacement therapy. A postmenopausal patient on estrogen supplementation may have marked cellularity in her smears, making the age and hormone replacement status of a patient essential information.
Low epithelial cellularity may reflect normal breast tissue, in which small cohesive ductal epithelial tissue fragments with myoepithelial cells are present with a small number of bare bipolar nuclei, or it may be the result of a sclerosed lesion such as a sclerosed fibroadenoma, or a lowly cellular lesion, such as some lobular carcinomas, in which widely spaced single cells or strands of carcinoma infiltrate desmoplastic stroma.
If the cytologic findings correlate with the clinical history and imaging, then a lack of epithelium, as in the contents of a fully aspirated cyst, does not preclude adequacy. But if there is discrepancy with imaging and clinical findings, then the triple test must be applied and repeat FNAB or core biopsy recommended.
It is acceptable for FNAB smears of lesions that are thought on imaging and clinical findings to be fat, “entrapped fat,” or a lipoma to have no epithelium as long as there is considerable fibroadipose tissue. But the specific diagnosis of lipoma cannot be made. Similarly, smears containing a proteinaceous background with histiocytes can be diagnosed as “cyst contents” or as a “cyst” if apocrine cells are also present, if this correlates with the imaging findings and drainage of the cyst. But if the clinical and imaging findings are of a nodule or other lesion, then smears just containing fat or cyst contents are inadequate.
Other causes of low epithelial cellularity include galactoceles and lactational change with a milky background, histiocytes, and a variable number of scattered acinar cells; extravasated mucin or mucoceles; mucinous carcinomas with scant malignant epithelium; or fat necrosis where histiocytes and multinucleated giant cells are present with necrotic fat and debris.
Importantly, low cellularity may also reflect poor sampling or slide smearing technique, most often related to the operator’s inexperience.
In breast FNAB, in contrast to other body sites, high cellularity can be associated with benign lesions, such as fibroadenomas and papillomas, as well as intraductal or invasive ductal, lobular, and other carcinomas.
Recognition at low power of the pattern is a powerful tool in assessing breast FNAB slides. The pattern includes the presence and the varying prominence of small and large epithelial tissue fragments and the degree of single-cell dispersal.
There will be variations in the pattern of specific lesions due to variations in the skill and technique of the operator performing the FNAB and smear making, as well as due to the range of epithelial and stromal cellularity in lesions (e.g., cellular to sclerosed fibroadenomas). However, in establishing a routine approach to reporting FNAB of the breast, particular lesions produce characteristic patterns in most cases. It is recognition of these patterns that allows an experienced breast cytopathologist to rapidly diagnose lesions and which is the starting point in assessing all breast FNAB slides.
The pattern is best assessed in the center of a well-made direct FNAB smear, but the entire slide should be assessed, because the fluid hemodynamics of smearing lead to various degrees of cell dispersal and breakup of epithelial and stromal tissue fragments in varying areas of the slide. The degree of cellular dispersal and tissue fragment discohesion should always be assessed with reference to the degree of crush artifact and smearing of nuclear chromatin. Smearing, particularly when forceful, can produce dispersal and lack of cohesion in any cellular breast lesion whether benign or malignant, especially in the distal third or “tail” of the smear ( Fig. 5-1 ), and inexperienced smearing can ruin good material from the FNAB. The classic cause of a false-positive breast FNAB is the undersampled fibroadenoma, where forceful smearing can produce dispersal and smearing artifact has been ignored ( Figs. 5-2 and 5-3 ). Stripped nuclei may represent smearing artifact or fragile malignant cells and can be assessed at high power, but a diagnosis should never be made purely on the basis of stripped nuclei.
In general terms, as the pattern of large cohesive tissue fragments gives way to smaller discohesive tissue fragments and dispersed cells, and as the degree of nuclear enlargement and atypia increase, and as the number of bare bipolar nuclei and discernible myoepithelial nuclei on epithelial tissue fragments decrease, a breast lesion becomes atypical and more likely to be malignant. Some benign lesions such as fibroadenomas can yield high cellularity and marked dispersal, particularly at the tail of the smear, and, if assessment at low power for the characteristic pattern of large epithelial fragments, stromal fragments, and bare bipolar nuclei is bypassed in favor of a high-power assessment of the dispersed cells and nuclear atypia, there is a risk of a false-positive diagnosis of carcinoma.
At low power other features can contribute to the pattern: calcifications are either granular and fine throughout a cystic proteinaceous background or large psammomatous calcifications; mucin is fibrillary as in mucinous carcinomas but can be mimicked by finely granular proteinaceous material in the background of a myxoid fibroadenoma; stromal fragments can be large and rounded, clover leaf in shape from fibroadenomas, or stellate from papillomas or small, irregular, tufted, and fibroblastic as in invasive carcinoma; and necrosis can suggest high-grade intraductal carcinoma or high-grade invasive carcinomas or fat necrosis.
There are eight main pattern types.
This pattern is typical of benign breast tissue, with low cellularity and usually some fat fragments and intact lobules. The tissue fragments represent terminal ductules or small ducts and are small and 3-D, consisting of small epithelial cells with some overlapping of small regular nuclei and overlying myoepithelial cells, or they are tiny flat epithelial sheets with myoepithelial cells. Undersampled or sclerotic fibroadenomas should be considered, and the slides screened for their distinctive stromal tissue fragments and plentiful bare bipolar. FLOAT NOT FOUND
This is a granular thin to thick background with a variable number of histiocytes and multinucleated histiocytes, siderophages, and cholesterol crystals and debris, consistent with cyst contents. The addition of apocrine cells in sheets confirms a cyst, and if small tissue fragments of ductal cells are also present, then a diagnosis of fibrocystic change is made. Galactoceles resemble cysts, while lactating breasts have a milky, proteinaceous background containing microglobules of fat with acinar cells and occasional lactating lobules. A proteinaceous background is also seen in epithelial hyperplasia with fibrocystic change, radial scars, and papillomas. A granular background can also be seen in granular cell tumors. FLOAT NOT FOUND
This is the typical benign pattern where large cohesive ductal epithelial tissue fragments represent epithelial hyperplasia and can be monolayered or folded large sheets with myoepithelial cells or dense 3-D, irregular, branched, or rounded tissue fragments with nuclear overlapping and haphazard arrangement. Typically, the hyperplastic epithelial tissue fragments consist of cells with nuclei arranged in a chaotic unorientated pattern with irregular or slitlike secondary lumina around which the small, mildly pleomorphic nuclei stream. High power will demonstrate the benign, bimodal pattern of myoepithelial cells in a different focal plane on the tissue fragments and confirm the bare bipolar nuclei in the background, which are actually stripped myoepithelial cell nuclei. Rounded or balloon-shaped epithelial tissue fragments mixed with irregular hyperplastic ductal epithelial tissue fragments with orientated cells at their edges and myoepithelial cells and bare bipolar nuclei suggest columnar cell change. FLOAT NOT FOUND
This pattern can be produced by a number of lesions, where additional features provide a specific diagnosis:
When histiocytes, a proteinaceous background, and apocrine sheets, which may also be hyperplastic, are added to the pattern, fibrocystic change with epithelial hyperplasia can be diagnosed ( Plate 5-3 [ii] ). If cellularity is marked, the epithelial hyperplasia is prominent and ragged, and hyaline stromal fragments and occasional benign tubules are also present, a radial scar can be suggested.
If meshwork or stellate papillary tissue fragments or true papillary tissue fragments are found in addition to the large hyperplastic ductal epithelial tissue fragments with apocrine sheets, siderophages, histiocytes, and bare bipolar nuclei in a proteinaceous background, intraductal papilloma ( Plate 5-3 [iii] ) can be diagnosed.
If the hyperplastic ductal epithelial tissue fragments are characteristically branched or staghorn in shape and associated with plentiful bare bipolar nuclei and rounded or scalloped stromal fragments, then a fibroadenoma ( Plate 5-3 [iv] ) can be diagnosed. Fibroadenomas in some cases can produce more irregular, fibrillary stromal fragments with less distinctive epithelial tissue fragments, but the diagnosis can usually be made when plentiful bare bipolar nuclei contribute to the pattern. Benign and borderline phyllodes tumors show hypercellularity and nuclear atypia of the stroma, as well as increased spindle cells in the background.
If the epithelial fragments are more complex and 3-D and show micropapillary, cribriform, or more rigid cell arrangement architecture, associated with mild to moderate nuclear atypia and variable single cell dispersal and occasional psammomatous calcifications, then low- to intermediate-grade intraductal carcinoma should be suspected. FLOAT NOT FOUND
If the large tissue fragments consist of crowded cells with overlapped atypical enlarged nuclei and lack myoepithelial cells, and smaller discohesive tissue fragments and plentiful dispersed atypical cells are also present, invasive carcinoma, particularly low to intermediate grade, must be considered.
The absence of benign elements such as myoepithelial cells and bare bipolar nuclei supports the diagnosis of intraductal or invasive carcinoma.
This pattern suggests carcinoma. The small epithelial fragments are crowded with overlapping enlarged nuclei accompanied by dispersed intact epithelial cells, showing nuclear atypia varying with the grade of carcinoma, and there is absence of benign features, such as myoepithelial cells and bare bipolar nuclei. FLOAT NOT FOUND
Rarely, a benign lesion that has been heavily smeared or includes apocrine cells may show this small tissue fragment with dispersed cell pattern, but usually the degree of nuclear atypia is mild.
When nuclear grade is low to moderate and rigid, angulated tubules are prominent, tubular carcinoma is suggested, and when nuclear grade is high and lymphocytes are plentiful and may infiltrate small syncytial epithelial sheets, carcinoma with medullary features should be considered. Metastatic carcinomas are suggested by the lack of typical breast carcinoma cytologic features, presence of multiple lesions, young age of the patient, and a history of a primary lesion elsewhere in the patient. Metaplastic carcinomas have a variable mix of squamous cell carcinoma, high-grade poorly differentiated carcinoma, and sarcomatous components, often with necrosis. Invasive micropapillary carcinoma is suggested by a pattern of crowded small tissue fragments fitting together like a jigsaw, with plentiful dispersed cells and high nuclear grade.
The classic pattern of breast carcinoma is one of dispersed single epithelial cells with scattered, minute, loosely cohesive tissue fragments. This is most characteristically seen in lobular carcinoma, associated with low to moderate nuclear enlargement, angulation, and atypia and often plentiful intracytoplasmic lumina, and in high-grade infiltrating ductal carcinoma, where nuclear enlargement and atypia are marked. Lymphomas and metastatic melanoma can produce a similar pattern, as can benign intramammary lymph nodes. Cases in which cells with apocrine type cytoplasm dominate suggest apocrine carcinomas. FLOAT NOT FOUND
High-grade malignant epithelial cells in small to large crowded tissue fragments and as dispersed single cells, when associated with prominent necrotic debris and irregular calcifications, suggest high-grade intraductal carcinoma. High-grade “basal-like” carcinomas show high cellularity, marked nuclear atypia, and prominent nucleoli and frequently have necrosis without calcifications. Metaplastic carcinomas frequently have necrosis along with poorly differentiated, spindle, or squamous cell carcinoma and spindle or chondroid fragments. FLOAT NOT FOUND
Fat necrosis with scattered histiocytes, necrotic fat, and acellular pyknotic and eosinophilic debris usually lacks epithelium.
Mucinous carcinoma has stringy fibrillary mucin with variable epithelial cellularity including rounded and discohesive tissue fragments, strands of cells, and dispersed single cells, which may have intracytoplasmic vacuoles or resemble signet rings cells. The degree of nuclear atypia is usually mild to moderate but on occasion can be high. The rare secretory carcinoma has a thin mucinous background with prominent intracytoplasmic vacuoles in tumor cells and signet ring cells. Fibroadenomas can have rounded myxoid stromal fragments in a faintly granular magenta background mimicking mucin, but the epithelial fragments have myoepithelial cells and bare bipolar nuclei will be present. Mucocele-like lesions and inspissated duct or cystic mucin have low epithelial cellularity and an absence of cytologic atypia. FLOAT NOT FOUND
High-power examination of the slides is used to confirm features noted at the low-power pattern assessment, determine whether myoepithelial cells and bare bipolar nuclei are present, assess the type of epithelial cells seen dispersed singly or in tissue fragments, assess the architecture of tissue fragments, and then analyze epithelial nuclear features. It is essential not to go to high power immediately, because nuclear atypia can be pronounced in some benign breast lesions. The low-power assessment of cellularity, pattern of tissue fragments and cell dispersal, tissue architectural features, and presence of other components such as stroma should be emphasized because it provides powerful diagnostic information as to the nature of the lesion and prevents false-positive diagnoses.
First, assess and confirm in every case the presence or absence of myoepithelial cell nuclei on epithelial tissue fragments and bare bipolar nuclei in the background (see Figs. 5-6, 5-7, and 5-15 ). Bare bipolar nuclei represent dispersed myoepithelial cell nuclei. Their presence supports a benign diagnosis and precludes an outright diagnosis of malignancy in most cases. The myoepithelial cells overlying ductal epithelial tissue fragments are best seen in the alcohol-fixed, Pap-stained smears and generally should be in the correct focal plane above the ductal cells, not like apoptotic debris in carcinoma, admixed with the epithelial cells. Because such diagnostic reliance is placed on bare bipolar nuclei and myoepithelial cells, strict criteria must be adhered to in deciding whether they are present: the nuclei must be small, perfectly oval with smooth outlines without indentations or irregularity, and with fine uniform dark chromatin and no nucleoli. Bloody and proteinaceous backgrounds can dilute bare bipolar nuclei, while plentiful bare bipolar nuclei are seen in fibroadenomas.
Second, the tissue fragments need to be assessed at high power. Are they discohesive, with cells pulling away at the edges, irregular and crowded with overlapping of the enlarged nuclei and poor delineation of individual cells, as seen in carcinoma (see Figs. 5-17 and 5-20 )? Are they flat apocrine or ductal sheets showing folding and myoepithelial cells (see Plate 5-2 ; Figs. 5-14 and 5-15 )? In fibroadenomas the epithelial tissue fragments are typically staghorn, branched, or drumstick shaped (see Plate 5-3 [iv] ). Small, slitlike irregular secondary lumina with streaming or irregular nuclear orientation are commonly seen in benign epithelial tissue fragments from epithelial hyperplasia, where they are associated with myoepithelial cells and bare bipolar nuclei (see Plate 5-3 [i] ). Meshwork or stellate papillary fragments with epithelial fragments attached and true papillary fragments with fibrovascular cores strongly suggest an intraductal papilloma (see Plate 5-3 [iii] ). Micropapillary fragments associated with nuclear multilayering and atypia, few or absent myoepithelial cells, and some dispersed cells suggest papillary, low-grade intraductal carcinoma (see Plate 5-4 ). Tubular carcinoma has crowded, angulated, small tubules, lacking myoepithelial cells, admixed with larger, sometimes cribriform fragments of an intraduct or invasive cribriform component and dispersed single cells.
Third, the dispersed cells are assessed at high power. Are they indeed intact? Are most of the “cells” actually stripped hyperchromatic malignant nuclei, or are they just bare bipolar nuclei? Do the cells have relatively dense, well-defined cytoplasm with nuclear atypia suggesting malignancy (see Plate 5-5 ; Fig. 5-20 ), or is the nuclear-to-cytoplasmic (N:C) ratio low and the nuclei regular with bland, uniform chromatin as seen in benign dispersed cells (see Figs. 5-6, 5-7, 5-14, and 5-15 )? Are they lymphoid cells from an intramammary lymph node? Histiocytes and apocrine cells can be extremely hard to distinguish, but both can be dispersed in fibrocystic change with or without epithelial hyperplasia as seen most typically in radial scars (see Plate 5-3 [ii] ).
What is the differentiation of the intact epithelial cells? Do they show columnar cytoplasmic differentiation with a low N:C ratio as found in cellular dispersed smears from some benign intraductal papillomas (see Plate 5-3 [iii] ), or is the N:C ratio higher with nuclear atypia in the columnar cells as found in intraductal papillary carcinomas (see Plate 5-4 )? Are there intracytoplasmic lumina or vacuoles, commonly seen in lobular carcinoma (see Plate 5-6 and Figs. 5-21 to 5-22 ), but occurring also in ductal carcinomas and occasionally in apocrine cells and benign ductal cells? Are the dispersed cells mixed in with necrotic material (see Plate 5-7 ) or mucin (see Plate 5-8 ) as seen in carcinoma?
Fourth, assess at high power other features including stromal fragments for their degree of cellularity and the presence or absence of atypia in stromal nuclei. High cellularity and atypia are seen in phyllodes tumor. Any mucin present should be carefully analyzed looking for dispersed single cells or atypical tissue fragments and whether it is fibrillary as seen in mucinous carcinoma (see Plate 5-8 ; Figs. 5-25 to 5-26 ) or thin and finely granular as seen in myxoid change in fibroadenomas. Microcalcifications may be finely granular, debris-like calcifications from cysts ( Figs. 5-27 and 5-28 ); clear, jagged, angulated, calcium oxalate crystals from fibrocystic or columnar cell change ( Figs. 5-29 and 5-30 ); larger psammomatous bodies as seen in epithelial hyperplasia, columnar cell change, or intraductal carcinoma ( Figs. 5-31 and 5-32 ); or coarsely granular coalescing masses as seen in high-grade ductal carcinoma in situ. Calcifications in the presence of necrosis and high-grade nuclear atypia in epithelial tissue fragments suggest high-grade intraductal carcinoma (see Plate 5-7 ).
Finally, the degree of epithelial nuclear atypia in the dispersed cells and tissue fragments is assessed. Nuclear enlargement is the most significant feature, along with variation in nuclear size and shape; the degree of chromatin clumping and hyperchromasia; nuclear indentations, particularly acute clefts; the size, shape, irregularity, and pleomorphism of nucleoli; and the presence of mitoses (see Fig. 5-20 ). These are the classic nuclear features of malignancy assessed in all areas of cytopathology and are required to make the diagnosis of malignancy in breast FNAB. However, it should always be remembered that many benign lesions of the breast such as fibroadenomas can have marked though usually focal nuclear atypia, possibly related to partial apocrine metaplasia, cystic degeneration, or the proliferative stage of the menstrual cycle and that some low-grade intraductal carcinomas show a more rigid placement of minimally atypical nuclei (see Plate 5-4 ). Pattern recognition in conjunction with high-power assessment of nuclear features can make the correct diagnosis.
Reporting breast FNAB requires a cytopathologist to reach as specific a diagnosis as possible so that correlation with clinical and particularly imaging findings using mammography, ultrasound, and magnetic resonance imaging (MRI) can be as accurate and effective as possible in the triple test. Checklists of groups of cytologic criteria that are diagnostic for specific lesions assist in making specific diagnoses, foster uniformity and reproducibility of the diagnosis, and also highlight discrepant findings that alert the cytopathologist to the possibility of an alternative diagnosis.
Smear Pattern: Proteinaceous Background
Cysts can present as palpable, sometimes painful breast masses and are rounded on mammography and ultrasound, where they can be mimicked by fibroadenomas, papillomas, medullary carcinomas, or mucinous carcinomas. Typically, cysts completely disappear on ultrasound and mammography after aspiration, but attempted aspiration may leave residual fluid or a residual mass may be palpable. These lesions should be re-needled. Some “complex” cysts on the ultrasound have a heterogeneous pattern with possible internal structures or wall thickenings, possibly related to the density of the cyst contents, a large number of siderophages, inflammation, scarring, or a proliferative focus and may yield thick material due to inspissation or aseptic inflammation. Similar aspirated fluid that is thick, resembling pus and reflecting its high cellularity, or bloodstained or dark brown due to old hemorrhage, can be obtained from carcinomas that are necrotic or partially cystic, including the uncommon encysted papillary carcinoma. On the other hand, the commonest cause of freshly bloodstained cyst fluid is incidental hemorrhage due to the FNAB. For all these reasons, in our experience, the fluid from ALL cystic lesions, not just bloodstained fluids, should be assessed by cytology.
Cellularity low.
Pattern: proteinaceous background .
Foamy histiocytes and siderophages in variable number.
Apocrine cells in sheets and singly.
“Cyst contents” is reported when only a proteinaceous background with histiocytes is present (see Figs. 5-11 and 5-12 ). The proteinaceous background is best seen in Pap smears because it often washes off the air-dried Giemsa smears, and it can be thin or thick and obscuring, finely granular, or colloidal. It may contain cholesterol crystals, laminated round bodies (Liesegang rings) (see Fig. 5-34 ), and calcific debris. The calcifications can be granular and blue in the Giemsa and grayish black in the Pap smears, clear calcium oxalate birefringent angulated fragments, or occasional small psammomatous calcifications (see Figs. 5-27 to 5-32 ).
Histiocytes are usually seen singly but sometimes in aggregates. They have small, round, kidney-shaped or irregular, indented nuclei and copious, finely vacuolated or granular cytoplasm (see Fig. 5-13 ), which may contain bluish black (Giemsa) or yellow-brown (Pap) hemosiderin pigment (“hemosiderophage”) representing previous hemorrhage (see Fig. 5-40 ). Multinucleated histiocytes and siderophages are frequently seen ( Fig. 5-33 ).
Metaplastic apocrine cells occur in flat sheets of evenly spaced polygonal cells with clearly defined cytoplasmic margins and have a large amount of finely granular cytoplasm, which is grayish blue with fine red granules in the Giemsa smears ( Fig. 5-35 ), and green with varying dark green to reddish brown granules in the Pap stain ( Fig. 5-36 ). Apocrine cells have single or binucleated round central nuclei and often quite large single nucleoli and can show columnar differentiation or spindling in small sheets, consistent with the attenuated lining of cysts ( Figs. 5-37 and 5-38 ). Apocrine cells can have intracytoplasmic lumina, containing eosinophilic material in the Pap smear or bluish purple material in the Giemsa smear, and there may be nuclear enlargement, pleomorphism, and hyperchromasia, possibly related to degeneration, when the chromatin becomes blurred ( Fig. 5-39 ).
Such apocrine atypia can be pronounced and associated with marked dispersal or hyperplastic sheets or micropapillary tissue fragments, when apocrine intraduct carcinoma enters the differential diagnosis (DD) ( Figs. 5-40 to 5-42 ). A moderate to high N:C ratio with still considerable eosinophilic granular cytoplasm and marked pleomorphism of nuclei with coarse chromatin showing perinucleolar clearing, spiculated large nucleoli, and loss of the flat sheet architecture favor apocrine intraduct carcinoma, which can be cribriform or high grade and associated with necrosis and calcifications or apocrine carcinoma.
Large numbers of neutrophils are seen in “inflamed cysts” in cases where cyst rupture has produced an inflammatory reaction that may or may not have caused pain or redness. The neutrophils and histiocytes in the proteinaceous background resemble pus and raise the possibility of a septic abscess that can be excluded by culture of the aspirate. Sheets of apocrine cells and ductal epithelial cells infiltrated by neutrophils and showing inflammatory atypia may be present ( Figs. 5-43 to 5-44 ), along with fragments of myxoid granulation tissue with branched, anastomosing capillaries ( Fig. 5-45 ). The N:C ratio usually remains low in these inflamed epithelial cells, and within the sheets there is a uniform increase in nuclear size and uniform chromatin, despite the hyperchromasia and atypia. A neutrophilic infiltrate is found uncommonly in carcinomas.
Recurrent subareolar abscesses (Suska disease) are a distinct entity associated with chronic inflammation and squamous metaplasia in the lactiferous ducts, and smears show pus with considerable keratinous debris and squamous cells, which may show inflammatory atypia. These lesions are differentiated from inflamed epithelial cysts of the skin of the breast by their location within and deep to the nipple and by their history.
Granulomatous mastitis can be seen in specific infections such as tuberculosis involving the breast but also in a nonspecific inflammatory process, where multinucleated giant cells, plentiful histiocytes occasionally in vague granulomas, and lymphocytes are seen in a proteinaceous background with mildly atypical ductal tissue fragments. Similar findings can be seen in granulomas in the breast and axillary lymph nodes reacting to silicon derived from breast prostheses, where the histocytes have foamy cytoplasm (silicon is not birefringent [ Figs. 5-46 and 5-47 ]). Culture and polymerase chain reaction to exclude mycobacterial infection are mandatory in granulomatous mastitis.
Dilated ducts with periductal sclerosis and inflammation, known as periductual mastitis or duct ectasia when deep to the nipple, produce a proteinaceous background, with histiocytes, lymphocytes, plasma cells, granular debris, multinucleated histiocytes, and usually some apocrine or ductal epithelium. The distinction from a cyst relies on ultrasound findings.
Fat necrosis can be seen in a previous biopsy or operative site, in sites of trauma, or can occur incidentally and be unsuspected on imaging. Giemsa-stained smears show multicolored yellow to orange to blue to black, punctate, and irregular granular debris and degenerate cells, with multinucleated and single histiocytes and siderophages ( Figs. 5-48 and 5-49 ), and fragments of fat, some of which will be necrotic with the reticular outline of fat cells lacking nuclei ( Fig. 5-50 ). If fat necrosis is associated with large numbers of multinucleated histiocytes with bubbly vacuolated cytoplasm, the possibility of a reaction to silicon, which is not birefringent, from implants should be considered. “Oil cysts,” seen on ultrasounds, can produce viscous yellow to white colloidal material on the unstained smears, but only a little debris is seen after fixation and staining.
In the postoperative setting where there has been adjuvant radiotherapy, scattered cohesive epithelial tissue fragments may show radiation changes characterized by occasional prominent enlarged cells, with pleomorphic, large hyperchromatic nuclei with vacuoles and smudged dark chromatin, low to moderate N:C ratio, and considerable vacuolated cytoplasm, often in a background of debris and possible fat necrosis, with plump atypical fibroblastic cells. Any increase in epithelial cellularity or dispersal of atypical cells with preserved chromatin should suggest residual carcinoma, and core biopsy should be recommended.
Galactoceles usually occur in lactating women and resemble cysts on imaging. They have a proteinaceous background that variably resembles cyst fluid or milk, with fat globules in thin proteinaceous material, with scattered histiocytes. Lactational change and lactational nodules (sometimes referred to inappropriately as “lactating adenomas”) show a milky background of microvesicular and macrovesicular fat globules in a thin casein proteinaceous background, along with small numbers of acinar cells in sheets and occasionally singly. The acinar cells have pale vacuolated fragile cytoplasm and round nuclei containing a single central nucleolus ( Fig. 5-51 ). Stripped, round, nucleolated acinar cell nuclei and large lobules in which the terminal ductules are expanded ( Fig. 5-52 ) are seen as well. Acinar cells can show marked dispersal, but the milky background and lack of nuclear atypia help in the distinction from lobular and rare secretory carcinoma.
Smear Pattern: Predominantly Small Discohesive Tissue Fragments With Plentiful Dispersed Cells
Granular cell tumor is an S100-positive tumor arising from nerve sheath Schwann cells, occurring in many sites in both males and females including the breast, where it often presents clinically and radiologically as a suspicious firm and spiculated mass sometimes attached to skin or chest wall.
Cellularity: moderate to high.
Pattern: predominantly small discohesive tissue fragments with plentiful dispersed cells .
Faint granularity in background with sheets and tissue fragments of cells with low N:C ratio and abundant granular cytoplasm resembling histiocytes.
Usually central small round regular nuclei lacking nucleoli or mitoses.
No necrosis.
Single dispersed cells in small numbers and spindle stromal fragments are present.
The DD includes fibrocystic change with apocrine cells and histiocytes, as well as apocrine carcinoma. If necrosis is present, the possibility of a malignant granular cell tumor should be considered and excision biopsy recommended.
In cell blocks, the granular cell tumor cells are S100 and vimentin positive, and cytokeratin is negative.
Smear Pattern: Proteinaceous Background
Fibrocystic change is common in FNAB of women of all ages and can produce palpable masses and mammographic and ultrasound abnormalities. Histologically, small and large cysts arising from the terminal duct lobular unit and usually associated with apocrine metaplasia are seen with intralobular and interlobular fibrosis. Fibrocystic change and proliferative breast disease such as epithelial hyperplasia are often lumped together in histopathology reports, but an attempt should be made in both cytology and histology to distinguish these entities, so as to improve correlation with imaging, maximize the role of the triple test, and avoid false-positive diagnoses.
Cellularity: low.
Pattern: proteinaceous background .
Apocrine sheets and small cohesive tissue fragments of ductal epithelial cells with myoepithelial cells.
Histiocytes, siderophages, and multinucleated histiocytes and a variable number of bare bipolar nuclei in the background.
The diagnosis of “cyst contests” is made when there is only a proteinaceous background with histiocytes and there is correlation with imaging findings, particularly complete drainage under ultrasound real-time imaging or when FNAB of palpable lesions leaves no residual mass (see Figs. 5-11 and 5-12 ). When apocrine sheets are present in a proteinaceous background a diagnosis of “cyst with apocrine cells” is appropriate (see Plate 5-2 ; Figs. 5-56 and 5-57 ). When ductal epithelial tissue fragments with myoepithelial cells are also present with apocrine cells and histiocytes in a proteinaceous background, the term fibrocystic change is used ( Fig. 5-58 ). The aim is to allow close correlation with imaging, especially ultrasound findings.
Ductal epithelial cells can show a transition to apocrine metaplasia that can be associated with nuclear enlargement, atypia, and dispersal ( Fig. 5-59 ). Apocrine cells have well-defined cell margins, a large amount of granular cytoplasm that is magenta to gray in the Giemsa stain and khaki to light green in the Pap stain with red granules, and a central round nucleolus and single prominent nucleolus (see Figs. 5-35 and 5-36 ).
A small number of bare bipolar nuclei are usually present but often diluted by the proteinaceous background and may be scant. They must be distinguished from stripped apocrine cell nuclei, which are round with a single nucleolus. Microcalcifications can be seen in the proteinaceous background as in cysts (see Figs. 5-27 to 5-29 ).
Apocrine sheets have scant to no myoepithelial cells. In some cases dispersal of apocrine cells can be marked and, when associated with nuclear degeneration, may cause some concern. But the overall cellularity is low, the N:C ratio is low, and the marked anisonucleosis, increase in nuclear size, and hyperchromasia and large and irregular nucleoli of apocrine carcinoma are not seen.
Histiocytes have granular copious cytoplasm and at times may resemble apocrine cells, particularly degenerate desquamated apocrine cells but the distinction is of little importance.
At times apocrine metaplasia can be associated with marked papillary hyperplasia, and features of an intraductal papilloma such as meshwork or stellate “papillary” or true papillary tissue fragments with fibrovascular cores can be seen, matching the apocrine papillary hyperplasia seen histologically ( Fig. 5-60 ). Apocrine dispersal can be present in these cases, but the individual apocrine cells are usually bland and the lesions clearly benign.
Smear Pattern: Predominantly Large Epithelial Tissue Fragments With Myoepithelial Cells and Bare Bipolar Nuclei, a Variable Number of Cohesive Smaller Epithelial Tissue Fragments, and Few Dispersed Epithelial Cells
Epithelial hyperplasia is commonly seen in the context of fibrocystic change and can be associated with palpable lumps, mammographic asymmetry and architectural disturbances and calcifications, and ultrasound abnormalities that are usually regarded as “fibrocystic change.” Epithelial hyperplasia can also be seen in intraductal papillomas and in fibroadenomas and phyllodes tumors, and it is the presence of other features such as stromal fragments that allow a more specific diagnosis to be made.
Cellularity: moderate to high.
Pattern: predominantly large epithelial tissue fragments with myoepithelial cells and bare bipolar nuclei, a variable number of cohesive smaller epithelial tissue fragments, and few dispersed epithelial cells .
Large hyperplastic ductal cell sheets with well-ordered cells and regular nuclear arrangement and myoepithelial cells; some sheets may fold.
Large 3-D tissue fragments with chaotically arranged cells and nuclei that lack orientation, stream around irregular slitlike secondary lumina, and show variable mild nuclear enlargement and pleomorphism; myoepithelial cells are usually numerous.
A variable number of smaller cohesive ductal epithelial tissue fragments with irregularly arranged, small, mildly pleomorphic nuclei, and myoepithelial cells.
A small number of bare bipolar nuclei and variable dispersal of single epithelial cells, especially at the tail of the smears.
Once large tissue fragments are seen in a breast FNAB smear, the pattern is regarded as a large tissue fragment pattern, even though there may be a mix of small tissue fragments and dispersed cells, as this aids the pattern recognition approach and DD.
The myoepithelial cells have small, perfectly oval nuclei with uniform fine chromatin and no nucleoli and sit at a focal plane above that of the epithelial cells in the bimodal sheets. They are more obvious at the edge of the tissue fragments. These features are perhaps more easily seen in the Pap-stained slides (see Fig. 5-62 ). Myoepithelial cells must be distinguished from irregular apoptotic debris found in the same focal plane as the nuclei in large sheets and tissue fragments from intraductal carcinomas and some carcinomas. The bare bipolar nuclei or myoepithelial nuclei in the background similarly must be perfectly oval in shape with fine, even chromatin and no nucleoli to separate them from stripped malignant nuclei (see Figs. 5-5 and 5-6 ).
The ductal cells show small, round to oval to mildly pleomorphic indented nuclei with fine chromatin, single nucleoli, and occasional nuclear pseudoinclusions and grooves, arranged in a relatively uniform pattern when seen in a flat sheet (see Fig. 5-62 ), but more haphazardly in a multilayered overlapping pattern in the thicker 3-D tissue fragments (see Figs. 5-63 to 5-65 ). At the margins of the cohesive epithelial sheets, the cytoplasm of cells can be pulled out or tufted as the cells remain firmly attached to the tissue fragment but respond to the forces of smearing. The nuclei can be large, with prominent single nucleoli, but the N:C ratio remains low, and these “switched-on” nuclear features are usually uniform or predictable throughout a particular tissue fragment or sheet. Nuclear hyperchromasia or irregular chromatinic clearing such as perinucleolar clearing and prominent nuclear envelope abnormalities are not usually seen. The nucleoli can be quite large and prominent, particularly in younger women, possibly related to the proliferative phase of the menstrual cycle or to the oral contraceptive pill ( Fig. 5-66 ).
If a proteinaceous background, histiocytes, and apocrine sheets are also present, the appropriate diagnosis is “fibrocystic change with epithelial hyperplasia,” noting that the proteinaceous background may dilute the bare bipolar nuclei and apocrine hyperplasia may also be present ( Fig. 5-67 ). If large or scalloped or irregular or myxoid stromal fragments are present, then fibroadenoma or perhaps phyllodes tumor can be diagnosed, and if meshwork or stellate papillary tissue fragments are present, intraductal papilloma is diagnosed. Correlation with imaging findings is required.
Partial apocrine metaplasia (see Fig. 5-59 ) can be associated with nuclear enlargement and hyperchromasia in ductal cells that are larger, show increased but not typical copious granular apocrine cytoplasm, and can disperse singly. Seeing the transition from ductal to apocrine cells in sheets is often a clue to this benign process, and myoepithelial cells and bare bipolar nuclei are usually present. Cytologic atypia should be correlated with imaging and clinical findings, but if there is significant atypia (Code 3), then repeat FNAB or core biopsy is required, even if imaging is negative.
Columnar cell change can present incidentally or in FNAB for mass lesions or calcifications. Histologically, the terminal ductules of the lobules are grossly dilated and lined by columnar cells, which often feature a luminal apical bleb of cytoplasm and show a range of simple columnar cell change to columnar cell hyperplasia where the columnar cells are multilayered but still orientated. There is a spectrum with flat epithelial atypia, where the nuclei of the thickened epithelium lining the ballooned terminal duct lack orientation to the lumen and show nuclear atypia. Calcifications including calcium oxalate crystals are frequently present, and the columnar cell change is also often mixed with fibrocystic change and epithelial hyperplasia in adjacent lobules. The FNAB usually produces a cohesive large epithelial tissue fragment pattern with myoepithelial cells and bare bipolar nuclei in a proteinaceous background, but it is variable and can include large ballooned tissue fragments of ductal epithelial cells ( Figs. 5-68 to 5-70 ). These tissue fragments are not papillary and represent the ballooned terminal ductules of the involved lobules. They have an orderly array of nuclei but fewer myoepithelial cells and a central lumen, which can be demonstrated by focusing up and down on the tissue fragment. Large flat sheets and smaller 3-D ductal tissue fragments may have an ordered columnar array with apical snouts at their luminal edge. A small number of dispersed columnar cells, bare bipolar nuclei, and in some cases calcifications are present in a proteinaceous background ( Figs. 5-71 and 5-72 ). Micropapillary and cribriform tissue fragments are not seen. The degree of epithelial hyperplasia and nuclear atypia varies with the spectrum of columnar cell lesions, and in some cases the tissue fragment architecture may suggest cribriform or micropapillary low-grade intraduct carcinoma or the nuclear atypia may be more marked leading to an atypical diagnosis. Core biopsy is required.
The distinction of epithelial hyperplasia from atypical ductal hyperplasia and low-grade intraductal carcinoma is difficult in histopathology, where there is considerable interobserver variability. Ducts showing epithelial hyperplasia, atypical ductal hyperplasia, and low-grade intraductal carcinoma are often intermingled and vary as a lesion is cut through. Pure atypical ductal hyperplasia is seldom seen in histology, making any study that attempts to correlate it with a cytologic diagnosis of “atypical epithelial hyperplasia” problematic. It is not surprising therefore that the diagnosis of “atypical ductal hyperplasia” on cytology is not possible and that the cytologic diagnosis of “epithelial hyperplasia with atypia” correlates poorly with histologic “atypical ductal hyperplasia,” although it is arguable that most cases of atypical ductal hyperplasia probably produce such a cytologic finding. The cytopathologist should decide if the lesion’s features are within the range acceptable for epithelial hyperplasia or whether they are atypical and suggestive of low-grade intraductal carcinoma. In other words, there is no rationale for attempting to make a cytologic diagnosis of “atypical ductal hyperplasia.”
“Atypia” in a breast FNAB report should refer to a pattern of increased dispersed cells or increased architectural complexity of tissue fragments including rigid nuclear arrays and cribriform and micropapillary architecture, or it may refer to nuclear features such as increased size, anisonucleosis, irregularity or hyperchromasia of chromatin, and enlargement or spiculation of nucleoli. However, there are no standardized criteria among cytopathologists for “atypia.” Despite this, lesions with these features require biopsy and are associated in most institutions with increased risks of carcinoma.
Awareness of the characteristic cytologic features of the histologic spectrum from fibrocystic change to proliferative breast disease, which includes epithelial hyperplasia, fibroadenomas, intraductal papillomas, columnar cell change, adenosis, sclerosing adenosis, radial scars, florid epithelial hyperplasia, and so-called “papillomatosis,” allows the cytopathologist in many cases to make a specific diagnosis and at least assess where a particular case should be placed in the spectrum. Each FNAB should be clearly described and, if possible, a specific diagnosis made. If the conclusion raises a DD, including the possibility of low-grade intraductal carcinoma, a core biopsy or, in some cases, excision biopsy is required.
Proliferative breast lesions in the spectrum from epithelial hyperplasia to low-grade intraductal carcinoma produce highly cellular smears. The smear pattern shifts from Pattern 5-3 with large cohesive 3-D tissue fragments with myoepithelial cells and ductal cells with small nuclei and minimal nuclear atypia, which stream around irregular slitlike secondary lumenal holes with bare bipolar nuclei and only focally dispersed cells in the background (see Plate 5-3 [i] ; Figs. 5-61 to 5-67 ), to Pattern 4 where there is a mix of predominantly large but increasing number of small epithelial tissue fragments showing varying degrees of nuclear enlargement and atypia (see Plate 5-4 ). In low-grade intraduct carcinoma the nuclei may show a more rigid organization and cribriform holes and micropapillary extensions may be present with few if any myoepithelial cells (see Plate 5-4 ). Pattern 5-4 can be associated with prominent cell dispersal showing mild to moderate nuclear atypia and scant or no bare bipolar nuclei. The cribriform pattern of intraductal carcinoma should have nuclear orientation to the punched-out lumina, rather than the streaming of nuclei as seen in epithelial hyperplasia ( Figs. 5-73 and 5-74 ), but this can be difficult to assess, particularly in Giemsa smears, where cribriform tissue fragments appear cratered or pockmarked, rather than having punched-out holes (see further discussions and Figs. 5-119 and 5-120 ) in section on low-grade intraductal carcinoma).
Entities in the Differential Diagnosis
|
Entities in the Differential Diagnosis
|
Entities in the Differential Diagnosis
|
Entities in the Differential Diagnosis
|
Microcalcifications can be seen in epithelial hyperplasia, columnar cell change (see Figs. 5-71 and 5-72 ), and low-grade intraductal carcinoma (see Fig. 5-31 ).
Hyperplastic ductal epithelial tissue fragments may contain small, rounded magenta (Giemsa) or pale green (Pap) globules known as collagenous spherulosis ( Fig. 5-75 ), whose only significance is their differentiation from the generally larger and more numerous hyaline globules seen in adenoid cystic carcinoma. In some cases there may be plentiful globules, usually more variable in size ringed by myoepithelial cells within hyperplastic tissue fragments of bland epithelial cells with scattered bare bipolar nuclei.
Adenoid cystic carcinoma occurring in the breast can show similar variation in grade to those seen in the salivary glands but tends to show more dispersal, nuclear crowding and atypia, and larger, rounded, more plentiful globules than collagenous spherulosis, but the DD can be problematic. If a small number of spherules are seen in a background typical of fibrocystic change or papilloma or fibroadenoma, then the diagnosis of collagenous spherulosis can be made. If such a background is not present and the globules are frequent, or associated with nuclear atypia or prominent dispersal, core biopsy should be recommended to exclude adenoid cystic carcinoma (see section on adenoid cystic carcinoma for further discussion and Figs. 5-200 to 202 ).
Smear Pattern: Predominantly Large Epithelial Tissue Fragments With Myoepithelial Cells and Bare Bipolar Nuclei, a Variable Number of Cohesive Smaller Epithelial Tissue Fragments, and Few Dispersed Epithelial Cells
These mammographically detected stellate proliferative lesions are difficult on ultrasound and mammography to distinguish from small carcinomas, leading to frequent excision biopsy. Histologically, radial scars show central and radiating sclerosis, containing distorted tubules, ductal epithelial hyperplasia with or without atypia, apocrine change, adenosis, sclerosing adenosis, and in some cases, atypical ductal hyperplasia, low-grade intraduct carcinoma, or tubular carcinomas.
It is essential to recognize their cytologic features to avoid false-positive diagnoses because they often present to the cytopathologist with a malignant diagnosis on imaging. Alternatively, a mammographic “radial scar” may on FNAB reveal cancer, advancing the management of a particular patient.
The cytologic features resemble florid fibrocystic change with epithelial hyperplasia.
Cellularity: moderate to high.
Pattern: predominantly large epithelial tissue fragments with myoepithelial cells and bare bipolar nuclei, a variable number of cohesive smaller epithelial tissue fragments, and few dispersed epithelial cells .
Plentiful large epithelial ductal tissue fragments, including monolayered sheets of apocrine cells and 3-D ductal epithelial cell tissue fragments with myoepithelial cells.
Variable number of smaller epithelial fragments, some of which may be tubular tissue fragments with myoepithelial cells.
Minor component of dispersed epithelial cells.
Bare bipolar nuclei present in small numbers.
Mild nuclear atypia in some cases.
Foamy macrophages and proteinaceous background are usually present.
Small sclerotic stromal tufts.
The DD includes fibrocystic change with epithelial hyperplasia; intraductal papillomas with epithelial hyperplasia, which can coexist in radial scars, and have distinctive meshwork, stellate “papillary,” and true papillary tissue fragments (see Plate 5-3 [iii] ); fibroadenomas with branching hyperplastic epithelium, rounded stromal fragments, and plentiful bare bipolar nuclei (see Plate 5-3 [iv] ); low-grade intraductal carcinoma, which may have micropapillary or cribriform tissue fragments and increased dispersal of single cells with nuclear atypia (see Plate 5-4 ); and tubular carcinoma, which is a major imaging DD but has distinctive cytology of mild to moderate cellularity with predominantly small, angulated crowded tubules showing moderate nuclear atypia, a lack of myoepithelial cells and bare bipolar nuclei, and more frequent dispersed atypical cells with nuclear atypia in the background.
Smear Pattern: Predominantly Large Epithelial Tissue Fragments With Myoepithelial Cells and Bare Bipolar Nuclei, a Variable Number of Cohesive Smaller Epithelial Tissue Fragments, and Few Dispersed Epithelial Cells
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here