Fine-needle aspiration biopsy may be performed with a number of variations. First, the indications for the procedure are discussed with the patient. In addition, potential risks or complications or both are discussed, including bleeding and infection in the area after the procedure. The risk of a significant complication is minimal, and although such incidents have been reported, the chances of such an occurrence may be equated to the risk incurred when undergoing simple venipuncture. Indeed, drawing similarities to venipuncture often alleviates the patient's anxiety toward the procedure.10 As with venipuncture, local anesthetic is generally not required for FNA of palpable masses, with the rare exception of certain anatomic locations (i.e., periareolar breast aspirations). After consent is obtained, slides are labeled and a suitable fixative (e.g., 95% alcohol or spray fixative) is made accessible. The area is then prepared with alcohol, and after stabilization of the mass with the fingers of one hand, the needle is placed into the lesion.
Two sampling techniques may be used. An open-ended needle, without attached negative suction, may be used alone for almost all biopsies (known as the “French technique”). Short, rapid strokes within the lesion cause dislodgement of cells and allow effective collection within the needle via capillary action. A syringe with the plunger removed may be attached for the collection of excess fluid if a cystic lesion is suspected. Regardless of the attachment used, it is critical that the end of the apparatus be open to the atmosphere to allow proper collection of the specimen; care must be taken not to cover this opening with a fingertip during the procedure, particularly when using a needle alone. Rapid strokes are made within the lesion, and care is taken not to extract the tip of the needle completely at any time. After 15–20 seconds, the needle is removed, pressure is applied to the area with sterile gauze, and slides are prepared immediately. If blood appears in the needle hub at any time, the needle should be removed immediately and slides prepared. Any delay results in clotting of the blood within the hub, making retrieval of the specimen difficult.
A second technique involves application of negative pressure during the procedure, typically with the use of a syringe (10–20 mL) placed in a syringe holder (aspiration gun/handle; Fig. 1A). When using an aspiration gun or handle, one must remember to pull the plunger back only after the needle has been placed into the lesion; the plunger should remain pulled while rapid, short strokes are made. More important, it is imperative that suction be released before the needle is removed, because continued negative pressure results in suction of the material back into the syringe preventing preparation of direct smears. Although rinsing with saline solution may allow retrieval of this material, cell recovery is poor and preparations are generally inferior to direct smears.
Fig. 1. A. An aspiration gun (Cameco) with attached syringe may be used to easily generate negative pressure at needle tip. B. Franzen needle guide, which is useful in the aspiration of paravaginal and pararectal lesions. (Photograph courtesy of Wesley W. Simms, MD.)
After removal of the needle, the material collected must be expelled onto glass slides using a syringe. If using the open-ended needle technique, a ready clean syringe with the plunger pulled back should be attached to the needle. If using the negative pressure aspiration technique, the same syringe can be used to expel the material onto the slides. If material is not expelled on the initial plunging of air through the syringe, remove the needle from the syringe, pull the plunger back fully and reconnect to the needle and expel quickly again. Direct smears should be made without delay. The technique for preparing smears is similar to that of preparing peripheral blood smears, with gentle pressure applied as two slides are passed over one another. One slide of the pair should be left to air-dry, while the other is immediately fixed, either by immersion in 95% alcohol or by spraying. It is imperative that fixed slides not be allowed to air-dry even partially before being fixed. Air exposure can cause marked cellular degeneration and distortion, hindering pathologic interpretation and resulting in an equivocal diagnosis. Air-dried smears should be labeled clearly as such, as these may then undergo Giemsa staining. The number of slides prepared obviously depends on the volume of material obtained. Any remaining material within the needle may be rinsed into a saline solution for further processing.
There are essentially no contraindications to FNA of a superficially located mass. FNA may be performed in patients with thrombocytopenia and in those taking anticoagulant medication, assuming their laboratory parameters are not beyond the therapeutic range. Although firm pressure should be applied for at least 10 minutes after each needle stick, the incidence of bleeding complications does not seem to be increased significantly in this population. In addition, patients with mitral valve prolapse generally do not require prophylactic antibiotics before this procedure, provided proper technique is followed.
With the use of radiologic guidance (e.g., ultrasound, computed tomography [CT], fluoroscopy), similar techniques may be used to sample deep-seated and other nonpalpable lesions.11, 12 Generally, a needle guide is first placed percutaneously, followed by passage of a small-gauge (generally 21–22-gauge) needle. After radiologic confirmation of accurate needle placement within the mass, material is extracted either with or without negative suction. The collected material is then expelled onto slides, direct smears prepared as described earlier, and the needle and syringe are rinsed into saline solution for future processing. The diagnostic yield for such procedures, as with FNA of superficial lesions, is optimized by immediate, onsite evaluation of the material by pathology personnel. Although such an approach adds to the total time of the procedure, nondiagnostic results are minimized, and cases requiring special studies (e.g., lymphoma, poorly differentiated malignancies) may be managed appropriately with the collection and proper distribution of additional material. A bleeding diathesis or abnormal clotting/coagulation parameters are considered a contraindication to FNA of deep-seated lesions. It is recommended that such deficiencies (i.e., administration of fresh frozen plasma or interruption of anticoagulant medication) be corrected before the procedure.
Finally, transvaginal and transrectal FNAs may be performed in patients with both palpable and nonpalpable pelvic, paravaginal, and pararectal masses. Such aspirates may be approached with the use of a Franzen needle guide (Fig. 1B), which maximizes accurate needle placement while minimizing the risk of self-puncture.13, 14, 15 Similar needle guides may also be attached to ultrasound probes for image-guided aspiration.16 To minimize the risk of abscess formation, it is generally recommended that prophylactic antibiotics be administered to those undergoing a transrectal approach, as proper sterilization in this area is difficult.17
A brief review of some common misconceptions regarding the performance of FNA biopsies is in order. It is believed by some that the use of larger needles results in more cellular specimens. Larger needles, however, generally cause a greater amount of bleeding during the procedure, resulting in dilution of the specimen and the presence of large amounts of obscuring blood. Indeed, lesions that typically yield hypocellular aspirates, such as those with extensive fibrosis, are best sampled with smaller gauge needles (e.g., 25 gauge). In addition, the incidence of complications can be expected to rise as needle size increases.18, 19
Another common misconception of FNA is that the application of negative suction at the time of sampling (i.e., with the aspiration gun) leads to a specimen with a greater number of cells. Experience has shown, however, that similar amounts of material are obtained from both benign and malignant lesions regardless of the presence of negative pressure. A study comparing absolute numbers of cells obtained with versus without aspiration (aspiration gun versus French technique) showed no significant differences with respect to numbers of cells obtained.20 On a practical note, the French technique makes for a much less-cumbersome procedure and allows the technician greater sensitivity and accuracy by permitting a more comfortable hand-and-wrist position (similar to that assumed when obtaining an arterial blood gas specimen). Finally, patients generally tolerate the French technique well, but many become anxious at the sight of the relatively large, aspiration handle often used.
Finally, some believe that multiple areas of a given mass can, and should, be sampled with the same needle in a single needle stick. In a large mass, the angle of the needle may be changed in small increments during sampling. Major changes in the direction of the needle should be avoided, however, because it can lead to increased bleeding and result in suboptimal specimens. Therefore, to sample three different areas of a given lesion, one should consider performing three separate passes. Multiple needle punctures are generally well tolerated, provided the rationale for such an approach is discussed with the patient before the procedure.
The specific cytomorphologic features of lesions in a variety of body sites follows; however, general cytomorphologic criteria of malignancy for epithelial lesions are presented here. The most useful diagnostic criterion applied to FNA specimens is cellularity, as nearly all neoplasms (particularly malignancies) yield cellular specimens on aspiration, whereas aspirates of nonneoplastic lesions and some benign neoplasms are generally hypocellular. Another useful diagnostic criterion is that of architectural complexity within intact tissue fragments. Although the practice of cytology is generally thought to rely entirely on evaluation of subtle cellular and nuclear details, large tissue fragments are often present in FNA biopsy specimens of epithelial lesions and contain valuable information. In general, normal parenchyma and benign epithelial lesions are characterized by architecturally “flat”, two-dimensional fragments with well-organized and uniformly spaced and predictable nuclei (Fig. 2A), whereas malignant epithelial neoplasms generally yield complex, three-dimensional fragments with significant nuclear overlap and disorganization (Fig. 2B). Obviously, all cases must be evaluated for conventional nuclear criteria of malignancy, including variation in nuclear size, shape, and chromatin pattern in adjacent cells. In addition, nuclear membrane abnormalities, altered chromatin distribution (including both hypochromasia and hyperchromasia), and elevated nucleus-to-cytoplasm ratios are important features of malignancy in individual cells (Fig. 3).22 General cytologic features of commonly encountered lesions in the breast, lymph nodes, gynecologic tract, and pelvis are presented.
Fig. 2. (A) Well-organized, 2D tissue fragment typical of benign glandular epithelium (stain, Papanicolaou) in contrast to (B) architecturally crowded, 3D tissue fragment from case of adenocarcinoma showing complete loss of cellular polarization and organization (stain, Papanicolaou).
Fig. 3. Cytologic criteria of malignancy are shown in this case of transitional cell carcinoma. Adjacent cells show significant variation in nuclear size and shape; individual nuclei display abnormal chromatin distribution and nuclear membrane irregularities (stain, Papanicolaou).
Fine-needle aspiration is a reliable and cost-effective tool in the evaluation of palpable and nonpalpable breast masses.23, 24, 25, 26 The combination of physical examination, imaging findings (mammography or ultrasound or both), and cytologic examination, known as the triple test, has a diagnostic accuracy reported to be greater than 95% when all three elements of the test are concordant.27, 28, 29, 30, 31, 32 Current recommendations for patients undergoing the triple test are as follows:
- If all three tests are compatible with a benign process, the patient may simply be followed clinically, with close observation.
- If all three tests are compatible with malignancy, the patient may undergo definitive therapy.
- If there is a mixed or inconclusive triple test result (i.e., one or two of the tests suggest malignancy), excisional biopsy of the index mass should be performed.33, 34
The breast lesions most commonly encountered in FNA specimens are nonneoplastic masses, best classified as fibrocystic change. Aspirates from these lesions generally contain foam cells (histiocytes) and proteinaceous debris, consistent with cyst contents. In addition, most aspirates contain relatively few fragments of ductal epithelium, all of which are tightly cohesive, composed of small, well-organized nuclei and a distinct second cell population (myoepithelial cells); apocrine metaplasia, characterized by abundant granular cytoplasm, is also commonly encountered (Fig. 4). Occasional cases of fibrocystic change with ductal hyperplasia may yield increased numbers of ductal epithelial tissue fragments, often with some nuclear enlargement and overlap. Such findings, in the absence of worrisome clinical or mammographic findings, warrant simple observation rather than surgical intervention.
Fig. 4. Fine-needle aspiration of fibrocystic change of breast often yields (A) few cohesive fragments of benign ductal epithelium (stain, Diff-Quik) and (B) fragments of apocrine metaplasia (stain, Diff-Quik); proteinaceous debris may also be seen in cystic lesions.
The question of the need for pathologic examination of cyst fluid from clinically benign breast lesions should be addressed. There are two patient groups in whom one might argue against routine cytologic evaluation of all cases: (1) those whose fluid is clear, and (2) those who have total resolution of the lesion with no residual mass. Patients in whom this is not the case (e.g., those who have opaque or bloody fluid or in whom a mass remains), however, should have all aspirated material sent for pathologic examination as well as additional sampling to rule out a cystic malignancy.33, 35, 36
Another benign breast lesion often encountered in FNA specimens, particularly in the younger patient population, is fibroadenoma. Considered a true neoplasm, this lesion is composed of both an epithelial and a stromal component, which should be represented in aspirate smears. The epithelial component is classically composed of numerous cohesive antler-shaped tissue fragments, often with some proliferative changes (nuclear enlargement and overlap). The stromal component is typically represented by stripped ovoid naked nuclei in the background, often referred to as bipolar cells. In addition, intact stromal fragments, often with myxoid changes, may be observed in occasional cases. A combination of these findings, in the context of the typical clinical findings of a well-circumscribed, movable, rubbery mass, is diagnostic of fibroadenoma.
Papillary lesions in the breast are a frequent source of indeterminate cytologic diagnoses when aspirated. Both papilloma (a proliferative, benign lesion) and papillary carcinoma (a malignant lesion) yield hypercellular aspirate smears, which contain numerous intact tissue fragments as well as numerous single, intact epithelial cells that are often columnar in shape.37, 38 Because of the cytomorphologic overlap of these entities, such findings generally result in a FNA diagnosis of “papillary lesion”, with surgical excision recommended for definitive histologic classification. Although some reports have attempted to establish criteria for reliably separating benign from malignant papillary lesions,39 a conservative approach should be adopted in these cases.
Carcinoma of the breast may be reliably diagnosed by FNA in the presence of three criteria: (1) hypercellularity; (2) dyscohesion, defined as significant numbers of intact single cells; and (3) lack of a second (myoepithelial) cell population in intact tissue fragments (Fig. 5). Strict adherence to these criteria results in an extremely high specificity for the diagnosis of breast carcinoma. Aspiration of some tumor subtypes, such as lobular and tubular carcinoma, may yield material that may not fulfill all criteria of malignancy, resulting in a “suspicious” diagnosis and a recommendation for surgical biopsy and histologic confirmation. Nuclear grading may be performed on FNA smears and correlates well with nuclear grading of the subsequently resected tumors.40, 41, 42, 43 Immunocytochemical stains for the presence of hormone receptors (e.g., estrogen, progesterone) may be performed on cytologic or cell-block material or both obtained from the FNA of a primary breast tumor or metastatic lesions.44, 45, 46, 47 Assessment of c-erbB2/Her-2/neu overexpression may be analyzed by immunohistochemistry or fluorescence in situ hybridization on aspirate samples, which may provide important prognostic information for presurgical treatment planning.48, 49, 50, 51
Fig. 5. Fine-needle aspiration of adenocarcinoma of breast typically yields hypercellular smears containing a single population of dyscohesive epithelial cells (stain, Papanicolaou).
There are some limitations of FNA of the breast that should be recognized. Although many breast tumor subtypes display classic histologic features, the accurate subclassification of tumor subtype (i.e., ductal, lobular, or medullary carcinoma) cannot be performed consistently based on simple cytomorphologic criteria alone. In addition, the determination of tumor invasion cannot be made based on cytologic smears and requires definitive histologic demonstration of stromal invasion; thus, the FNA diagnosis of “adenocarcinoma” in the breast encompasses in situ and invasive carcinoma.52 In some cases, particularly image-guided FNA of nonpalpable breast lesions, the addition of core needle biopsy may result in improved diagnostic accuracy.53, 54, 55 Finally, as with FNA in any site, clinical findings must be taken into account. Thus, a positive FNA diagnosis in the breast in the context of suspicious clinical or mammographic findings or both is sufficient for definitive therapy, whereas a negative FNA diagnosis in the presence of such findings requires further evaluation to rule out an unsampled or low-grade malignancy.
Soft tissue and subcutaneous lesions
Fine-needle aspiration represents an extremely useful diagnostic technique for patients with a soft tissue or subcutaneous nodule, particularly in the presence of a history of malignancy. Exclusion of metastatic disease is obviously of paramount importance to the patient and clinician. Positive cases are easily recognized but require comparison with previous histologic or cytologic material for confirmation of metastasis (Fig. 6A and B). In contrast, multiple samples by an experienced aspirator without evidence of malignancy are generally sufficient to exclude metastasis. In addition to the absence of malignancy, however, an explanation for the presence of a mass lesion is generally reassuring for both patient and clinician. Simple keratinous or epidermal inclusion cysts and fat necrosis are two such entities that are often seen in aspirate specimens in this clinical setting. Keratinous cysts generally yield benign and anucleated squamous cells admixed with debris and variable numbers of acute inflammatory cells (Fig. 7). Aspiration of fat necrosis, which may appear clinically suspicious, yields only histiocytes and granular debris, sometimes admixed with fragments of mature adipose tissue. Identification of the diagnostic features of either of these entities allows conservative management and simple observation.
Fig. 7. Fine-needle aspiration of keratinous cyst presenting as firm subcutaneous nodule in patient with a history of malignancy yields mature and anucleated squamous cells admixed with inflammatory cells (stain, Papanicolaou).
The FNA of lymph nodes, regardless of their location, is generally performed to determine whether their enlargement is because of (1) a benign, reactive process that can be treated conservatively; (2) a neoplastic hematopoietic process (lymphoma) that may require excisional biopsy for further subclassification and subsequent treatment; or (3) a neoplastic nonhematopoietic process (e.g., metastatic carcinoma, melanoma, sarcoma) that may require an extensive workup to identify a potential primary site. Both palpable and deep-seated lymph nodes may harbor any of these processes, and FNA represents a cost-effective approach to their diagnostic evaluation.56, 57, 58
Benign, reactive conditions generally yield cellular specimens composed of a dyscohesive (i.e., single cell) mixed population of lymphoid cells. Lymphocytes in a number of different states of activation are present, including small mature lymphocytes, large activated cells, immunoblasts, and plasma cells, resulting in the classic polymorphous population of lymphocytes indicative of a reactive condition (Fig. 8A).59 Histiocytes may also be present, often with ingested intracytoplasmic cellular debris (“tingible body” macrophages) and also suggest a reactive process. In contrast, aspirates of lymphoma yield a dyscohesive population of lymphoid cells in an abnormal distribution, with one subpopulation in dominance, resulting in the classic monotonous population typical of malignant lymphoma (Fig. 8B). Such a monotonous population of small, intermediate, or large lymphoid cells without the accompanying cell types just described is highly suggestive of lymphoma.59 Material from a FNA may be submitted for immunophenotyping by flow cytometry, in the hopes of objectively showing a monotypic cell population, consistent with malignancy. Although surgical biopsy may be required for further subclassification of some types of lymphoma, FNA with flow cytometry may be used for diagnosing and subclassifying cases of primary and recurrent non-Hodgkin lymphoma.60, 61, 62, 63 Such an approach is particularly appropriate for deep-seated masses in which open biopsy is technically difficult and perhaps even contraindicated in debilitated patients.
Fig. 8. (A) Fine-needle aspiration (FNA) of reactive lymph node shows polymorphous population of small, mature lymphocytes and larger activated lymphoid cells, consistent with lymphoid hyperplasia (stain, Diff-Quik); compare with (B) FNA of malignant lymphoma displaying monotonous population of large lymphoid cells, consistent with large cell lymphoma (stain, Diff-Quik).
Metastatic neoplasms are generally easily recognized in aspirates of enlarged lymph nodes. A foreign population of nonlymphoid cells is usually obvious, often with an admixture of lymphoid cells in the background. Metastatic carcinoma is by far the most common entity encountered and yields both single tumor cells and cohesive tissue fragments of various sizes. Nuclear and architectural atypia, as described earlier, should be present and may be profound in poorly differentiated malignancies (Fig. 6C). The distinction between squamous cell carcinoma and adenocarcinoma can often be made via cytologic preparations, particularly in well-differentiated tumors. Cells of squamous carcinoma generally possess homogeneous, dense cytoplasm with distinct cell boundaries and occasional intercellular bridges. In addition, streaming or spindling of tumor cells is often seen in squamous lesions. Finally, keratin pearls may also be seen in some cases and are diagnostic of squamous differentiation. Cells of adenocarcinoma, in contrast, typically display granular or vacuolated cytoplasm, sometimes with intracytoplasmic mucin secretions. Cell boundaries are generally indistinct, resulting in tissue fragments with a syncytial appearance (Fig. 9A), and glandular structures with distinct lumen formation may be seen in some tissue fragments. It should be noted, however, that a significant percentage of cases do not show definitive features of either squamous or glandular differentiation, resulting in a diagnosis of simply “poorly differentiated carcinoma”.
Fig. 9. (A) Fine-needle aspiration of enlarged inguinal lymph node in patient with history of endometrial adenocarcinoma shows syncytial fragment of malignant epithelium with vesicular nuclei and prominent nucleoli, consistent with metastatic adenocarcinoma (stain ,Papanicolaou), morphologically similar to (B) histologic section of previously resected primary tumor of endometrium (stain, hematoxylin-eosin).
Other metastatic neoplasms (e.g., melanoma, sarcoma) are less frequently encountered in lymph node aspirates. Although special stains (i.e., immunocytochemistry) may be helpful in this setting, clinical history is often most useful in these less common cases.
Two sites in which lymphadenopathy must be evaluated in patients with a history of gynecologic malignancy are the inguinal and supraclavicular areas. Palpable lymph nodes are present in the inguinal region of most persons and simply represent chronically stimulated nodes with a variable degree of fibrosis and fatty replacement. In the presence of a history of malignancy, however, the possibility of metastatic disease must be excluded for any mass larger than the typical shotty node; FNA is perfectly suited to such a clinical setting (Fig. 9).64
Enlarged supraclavicular lymph nodes should always be viewed with high clinical suspicion, particularly in the oncologic patient population. FNA in this location is easily performed and can reliably establish the presence of metastatic disease.65, 66, 67, 68
Deep-seated lymph nodes aspirated under radiologic guidance in the gynecologic patient population include those in the pelvis and retroperitoneum. Although ultrasonography69, 70 and lymphangiography71, 72 have been used to guide aspirates of these lesions, CT guidance has become increasingly popular because of a number of advantages.73 These include a high degree of accuracy in detecting pathologic changes in these nodal groups (often superior to lymphangiography)74 and the ability to detect heterogenous tissue densities in pathologic nodes; such information allows selective sampling of viable, potentially malignant tissue rather than necrotic debris.
The volume of gynecologic, cytopathologic, and radiologic literature on the subject of needle aspiration of the ovary has markedly increased over the past two decades. Despite the expanding literature and increasing use of the technique, the role of aspiration of the ovaries is still controversial, with some authors suggesting that aspiration biopsy of the ovary (except for the purpose of oocyte retrieval) is potentially dangerous and should not be regarded as a routinely acceptable clinical practice.75 The arguments against the use of FNA include the possible spillage of malignant cells into the abdominal cavity, leading to the potential dissemination of tumor, as well as misdiagnosis related to sampling errors.76 Having noted this opinion, it is important to point out that some of the literature also indicates that the use of needle aspiration of the ovary is valuable, safe, and even the standard of care in certain clinical settings, notably in young women who wish to preserve their ovarian function.77
The technique of FNA is increasingly being used by gynecologists because of its valuable diagnostic capabilities, its lack of significant morbidity for the patient, its simplicity of performance, and its diagnostic rapidity in comparison with conventional surgical excision and processing in pathology laboratories.9, 78 Aspiration of the ovary can be performed through a transvaginal or transrectal approach with a sheathed needle guide (e.g., Franzen needle guide; see Fig. 1B). In addition, radiologically (typically ultrasonographically) guided transabdominal aspiration is possible, as is direct aspiration through either laparotomy or laparoscopy.9, 78
Most ovarian mass lesions in premenopausal women are self-limited functional cysts.76 Aspiration of such cystic lesions generally yields cytologically accurate diagnoses and reassurance of the patient while avoiding unnecessary surgery.79, 80, 81, 82 Transvaginal sonographic examination of ovarian masses has led to scoring systems that are useful for evaluating adnexal lesions before aspiration or surgical excision. These scoring systems use parameters such as cystic versus solid nature, smooth-walled versus papillated cyst lining, cyst wall thickness, presence of septations, and echogenicity. These systems can be a useful starting point in patient treatment and reportedly facilitate the distinction between benign lesions (widely accepted as aspirable) and their malignant ovarian counterparts with a specificity of 83% and a sensitivity of 100%.83, 84 FNA can also be used in young women during laparotomy when only one ovary is tumor bearing and bilateral oophorectomy is unplanned. Aspiration of the contralateral ovary may avoid an unnecessary resection and help reassure both the surgeon and patient.77
Pregnant women constitute a special group of premenopausal patients in whom cystic ovarian lesions are often noted. As with adnexal lesions in the general population, ovarian masses in pregnant women are usually simple (unilocular-to-trilocular) cysts measuring 5 cm or less in diameter. These cysts are usually functional and resolve spontaneously. The number of pregnant women reported in the literature who have undergone aspiration biopsy of the ovary is small; however, most cytologically studied cases are reported as benign cystic masses.85, 86 Pregnant patients should be selected carefully for FNA. Those with a history of pelvic surgery or pelvic inflammatory disease may be unsuitable candidates for FNA because of fibrosis, which can produce a hypocellular, nondiagnostic specimen. Aspiration is suggested when a diagnostic procedure is indicated and radiologic studies favor a benign process. Ovarian FNA of a gravid patient should be performed only by an experienced operator.85
Women presenting for in vitro fertilization (IVF) constitute another subset of young women frequently noted to have ovarian lesions. Transvaginal ultrasound is routinely performed in the course of egg retrieval in these patients, and lesions are generally noted at that time. The recent literature confirms that the nature of ovarian cysts in IVF patients can be determined in many cases by the cytologic features of the cysts.87, 88 Aspirated material may even help in defining an unsuspected cause of infertility, such as endometriosis.87 FNA cytology of ovarian cysts in IVF patients can also play a role in the detection of occult ovarian neoplasms, although detection of malignancies in this group of patients is rare. Until the exact incidence of ovarian carcinoma in this subset of patients is known, it would seem prudent to continue to examine all cyst fluids obtained by needle aspiration.87, 88
FNA may be indicated in older patient cohorts as well. For example, the use of aspiration biopsy is generally accepted (if not indicated) in that subset of patients with previously documented ovarian malignancy in whom there is a suspicion of local recurrence.89, 90 When a previous diagnosis of malignancy is known, the minimally invasive technique of aspiration can be of great value in documenting disease spread and recurrence, with a specificity of up to 100% and a sensitivity of 96% in one large review.90
Although not generally recommended, the use of aspiration biopsy for the primary diagnosis and classification of ovarian carcinoma is accurate. For most nonfollicular cystic lesions, the specificity is reported at greater than 90%, especially if strict criteria for specimen adequacy are used and on-site assessment of the specimen is made during aspiration by a cytopathologist.14, 91, 92 Because ovarian cancer rarely produces characteristic symptoms or signs in the early stages, many affected patients present with advanced disease, even carcinomatosis. In cases of diffuse abdominal involvement, FNA may be a useful first-line approach to diagnosis, as the potential for dissemination of disease associated with spillage at the time of aspiration is of much less clinical import.
Early detection of ovarian malignancies markedly improves patient outcomes, with 5-year survival rates for patients with stage I disease approaching 80–85%.93, 94 Therefore, early detection of ovarian cancer, especially in young women with strong family histories, is highly desirable. To benefit this patient group, protocols designed to screen for ovarian cancer, based on routine transvaginal sonography complemented by FNA cytology of ovarian masses, are currently under consideration.77 Only when significant numbers of early cancers are detected will the debate be resolved regarding efficacy, safety, and risk-to-benefit ratio of ovarian cyst aspiration.93 Having discussed the indications for FNA of the ovary, let us turn to the characteristic cytodiagnostic findings and clinicopathologic correlates associated with the most frequently encountered primary ovarian lesions.
The classification of ovarian tumors is primarily morphologic but is intended to reflect the current concepts of the embryogenesis of this complex organ. It is based on the premise that the female gonad is composed of four major tissue derivatives, all of which are capable of giving rise to neoplasms: (1) surface, celomic, or germinal epithelium; (2) germ cells; (3) supporting sex cords; and (4) specialized ovarian stroma.95
Nonneoplastic ovarian lesions
Appropriate patient treatment requires that all four above-mentioned neoplastic processes be recognized; however, the majority of ovarian lesions (especially in the premenopausal population) represent nonneoplastic changes in the organ. The following is a discussion of a number of benign (often physiologic or functional) ovarian mass lesions.
SIMPLE CYSTIC LESIONS
Simple ovarian, paraovarian, serous, epithelial inclusion, paramesonephric, and paratubal cysts as well as hydrosalpinx are cytologically characterized by a smear containing only a few epithelial cells in a clear fluid background.77, 96 These entities are often cytologically indistinguishable from one another.29 The cells are generally cuboidal to low columnar and on occasion may possess cilia.9, 77, 96, 97 The presence of ciliated bodies in the fluids of ovarian cysts indicates the existence of ciliated columnar epithelial cells on the wall of the cysts, which excludes the differential diagnosis of cysts of follicular origin.98 Cilia may also be noted in aspirates from malignant ovarian tumors and therefore are not pathognomonic of a benign process.98, 99 The cellularity may be exceedingly sparse and degenerated with only a few macrophages, consistent with the cystic nature of these lesions. Hematosalpinx may present as a simple paraovarian cyst, and aspirated contents may contain pleomorphic epithelial cells with degenerated erythrocytes and cellular debris, raising the possibility of a false-positive diagnosis of malignancy.91
FUNCTIONAL CYSTIC LESIONS (FOLLICULAR AND CORPUS LUTEUM CYSTS)
Physiologic (functional) cysts of the ovaries usually appear benign (i.e., thin-walled, unilocular, and not more than 5 cm in diameter) on ultrasound examination; however, they may mimic malignant tumors.77 Aspirates generally contain follicular cells, typically granulosa cells, arranged both singly and as tightly cohesive fragments (Fig. 10). These cells house ovoid nuclei with granular chromatin and occasional nuclear grooves, and they have a small rim of distinct cytoplasm; mitoses may be present.100 The background of the smear typically shows clear, proteinaceous material but may be bloody.77 Cyst fluid from functional cysts of folliculogenesis typically have a high estrogen content. Mulvany and colleagues101 reported a sensitivity of 84% for the detection of follicular cysts by estrogen assay. As a follicular cyst undergoes luteinization, cytomorphologically altered granulosa cells appear with prominent nucleoli and more ample granular cytoplasm.100 Cytoplasmic hyaline droplets may be seen in the corpus luteum of pregnancy.82 On occasion, fluid aspirated from functional cysts may be highly cellular, with cells featuring high nuclear-to-cytoplasmic ratios and hyperchromasia; papillary clusters and even small glandular structures may be noted.102, 103 Clinical and radiologic impressions, as well as the presence of well-preserved granulosa cells in the smears, are helpful for reaching the correct diagnosis.103 The granulosa cells lining follicular cysts are immunohistochemically negative for cytokeratins, whereas those epithelial neoplasms considered in the differential diagnosis would be immunoreactive with antibodies directed against cytokeratins.102
Fig. 10. A. Fluid aspirated from a unilocular ovarian cyst in a young woman shows small, cohesive aggregates of benign lining cells from the inner granulosa cell layer of a functional (follicle) cyst; the cells are intermediate in size with round, vesicular nuclei and scant cytoplasm (stain, Papanicolaou; courtesy of Bradford Tan, MD). B. Corresponding histologic section of a follicular cyst shows orderly layering of granulosa/theca cells in the cyst wall with bland nuclear features, similar to that seen in the aspirated fluid (stain, hematoxylin-eosin).
Symptomatic lesions associated with active endometriosis are composed of both cystic and solid regions and often yield a thick, chocolate-colored fluid on aspiration.9 Endometriosis primarily affects women of reproductive age, and if it is identified by FNA, the need for a diagnostic surgical procedure is eliminated.104 The cellularity in aspirations of endometriosis is typically scant. The examiner should attempt to find all three components of the classic diagnostic triad, which includes endometrial glandular cells, endometrial stromal cells, and hemosiderin-laden macrophages.9 Fresh blood may also be a prominent feature.100 The glandular cells are small, with scanty cytoplasm, round-to-ovoid nuclei, finely granular chromatin, and occasional chromocenters. When present in intact tissue fragments, glandular cells are well organized, indicating their benign nature. Stromal cells may appear as naked, isolated nuclei.77 Generally, the presence of at least two of the three classic findings is sufficient for the diagnosis of endometriosis in the appropriate clinical setting. Finally, it should be noted that endometriotic nodules may undergo malignant change and should be evaluated for such transformation.
Neoplastic Ovarian Lesions
Neoplastic ovarian tumors are grouped by the World Health Organization classification (1995) into surface epithelial-stromal tumors, sex cord-stromal tumors, germ cell tumors, tumors of uncertain histogenesis, soft tissue tumors not specific to the ovary, and secondary (metastatic) processes.95
SURFACE EPITHELIAL-STROMAL NEOPLASMS
The epithelial group of ovarian neoplasms includes benign, borderline, and malignant subsets of serous, mucinous, endometrioid, clear cell, and transitional (Brenner) tumors as well as squamous cell lesions, tumors of mixed epithelial derivation, and undifferentiated epithelial lesions that lack the characteristic findings of the more specific categories.
Serous fluid aspirated from cystic serous neoplasms generally displays a granular, eosinophilic background with low cellularity in benign and well-differentiated lesions.77 The epithelial cells are small and tightly packed, and they have a high nuclear-to-cytoplasmic ratio (sometimes resembling reactive mesothelial cells). Spindled stromal cells may be noted if the tumor has a connective tissue component (e.g., cystadenofibroma).9, 77 Histiocytes may also be a prominent component in benign and well-differentiated tumors.
Low-grade carcinomas and borderline tumors are indistinguishable on cytology. FNA of malignant serous tumors yields thick, turbid fluid with increased cellularity, including single cells.9 Papillary structures and psammoma bodies may occur.82 Cellular pleomorphism, coarsening of the chromatin pattern, and prominent nucleoli are all markers of malignancy.82 The differential diagnosis of ovarian serous malignancies includes reactive mesothelial cells (which may be papillary), endometrioid carcinoma, and metastatic processes.9 Ancillary studies such as DNA ploidy analysis (highly malignant ovarian tumors likely to be nondiploid) and CA125 measurement (markedly elevated in most serous malignancies), may prove helpful.105, 106, 107
Epithelial cells from mucinous tumors generally resemble endocervical cells (Fig. 11), particularly in benign and low-grade neoplasms. Benign cases are characterized by small, regular fragments of uniform columnar cells with bland, basally located nuclei and vacuolated cytoplasm.82 Desquamated epithelial cells may resemble foamy macrophages.77 Keratin immunohistochemical studies and cytochemical stains for mucin may be useful in distinguishing epithelial cells from macrophages.
Fig. 11. A. Fluid aspirated from a cystic ovarian mass contains cohesive fragments of epithelial cells with small, basally located nuclei without atypia and abundant vacuolated cytoplasm; cytologically, they resemble endocervical cells and are compatible with a mucinous neoplasm (stain, Papanicolaou). B. Corresponding histologic section shows a single, undulating layer of well-organized, endocervical-type glandular epithelium, consistent with mucinous cystadenoma (stain, hematoxylin-eosin).
FNA material from malignant mucinous lesions (borderline lesions as well as cystadenocarcinomas) is often highly viscous and mucoid. Single cells, irregular groupings of cells, and syncytial fragments of epithelium with cellular atypia are noted in malignant mucinous tumors.82 Elevated carcinoembryonic antigen (CEA) and CA125 levels are appreciated in the cyst fluid aspirated from primary mucinous cystadenomas and cystadenocarcinomas.106, 107 Multinucleated tumor cells may be noted, and the degree of cellular atypia tends to increase as the lesion becomes more poorly differentiated.82
Benign and borderline endometrioid tumors are uncommon. Most lesions in this category are carcinomas. Such lesions can be distinguished from their serous counterparts by their more abundant, granular, eosinophilic cytoplasm as well as focal areas of squamoid differentiation, which are not seen in serous lesions.77, 82
As with endometriod lesions, most primary clear cell tumors of the ovary are frankly malignant. Cytologically, they resemble clear cell tumors arising in the vagina, cervix, or endometrium.77 Generally, the cells of clear cell adenocarcinoma have abundant, pale, vacuolated (glycogenated) cytoplasm with indistinct cell borders and pleomorphic nuclei.9 The differential diagnosis includes metastatic renal cell carcinoma and yolk sac tumor.9, 77
FNA of benign Brenner tumors shows cohesive fragments and sheets of uniform cells with oval nuclei and moderate, eosinophilic cytoplasm. Deep nuclear grooves and small prominent nucleoli may be noted; stromal elements may also be apparent.9, 77, 82
Malignant Brenner tumors of the ovary are difficult to categorize cytologically and may be confused with metastatic transitional cell carcinoma or nonkeratinizing squamous cell carcinoma. These cells are often arranged in papillary groups and exhibit dense basophilic cytoplasm and coarsely granular chromatin.77 Many binucleate and multinucleate forms and mitotic figures may also be seen.108
SEX CORD-STROMAL NEOPLASMS
This broad category of primary ovarian lesions includes thecoma-fibroma, granulosa-stromal cell tumors, Sertoli-Leydig cell lesions, androblastoma, sex cord tumors with annular tubules, and gynandroblastomas.
Thecoma-fibroma aspirates are extremely hypocellular and may be interpreted as nondiagnostic.77 Aspirated tumor cells are ovoid with elongated, hyperchromatic nuclei and scanty, ill-defined cytoplasm.77
Aspirates of granulosa cell tumors are usually highly cellular, with relatively small tumor cells arranged both singly and in aggregates of a few to several hundred cells.109 The nuclei are round to ovoid, single, and centrally located, and they may have marked membrane indentations appearing as longitudinal grooves.109 Rosette-like structures with central, dense, eosinophilic material (Call-Exner bodies) may be found and are very useful in supporting the diagnosis.77 In the absence of such bodies and in cases with frequent mitoses, the differential diagnosis of small cell carcinoma (either primary or metastatic) must be ruled out.
Sertoli-Leydig cell tumors yield fragments of uniform, round cells with moderate amounts of cytoplasm on aspiration.82 These cells may be arranged in cords, acini, or trabeculae, and a few bare nuclei of stromal origin may be noted.77
Sex cord-stromal tumor with anular tubules is a rare primary lesion, at times seen in association with Peutz-Jeghers syndrome. Smears from such lesions have been reported as cellular; both single cells and cohesive fragments of neoplastic cells arranged in follicular, solid, and trabecular patterns.110 In classic cases, homogenous, eosinophilic, hyaline bodies surrounded by palisading nuclei are very helpful at suggesting the appropriate diagnosis.77
GERM CELL NEOPLASMS
Mature cystic teratomas are the most common germ cell tumors of the ovary. Other entities in this category include dysgerminoma, yolk sac tumor, embryonal carcinoma, choriocarcinoma, and polyembryoma.77, 95 It is not uncommon for lesions in the germ cell group to present with mixed morphologies. Malignant germ cell tumors account for less than 5% of all ovarian cancers, and such lesions are most often encountered in young females (younger than 20 years of age).111 In this age group, the tumor may be treated in such a way that reproductive function is preserved.111
Aspirates of mature cystic teratomas show numerous superficial and anucleated squamous cells; ciliated and columnar epithelial cells may also be observed.77 Amorphous material (amphophilic, extracellular sebaceous secretions) is a frequent finding in the background of smears from benign teratomas, and it may elicit a foreign body giant cell reaction.77 More important, transvaginal collection of any ovarian aspirate may yield many contaminating squamous cells. This finding alone should not be misinterpreted as suggestive of ovarian teratoma. Immature teratomas are predominantly comprised of neuroglial elements; the presence of a frankly malignant neuroglial component in a smear favors classification as a malignant neoplasm.77 Cells normally present in cerebrospinal fluid cytology, such as ependymal and choroidal cells, may be noted in ovarian aspirates.112, 113 When such cells are present in materials obtained from the ovary, malignant neuroectodermal tumors must be considered. Ovarian carcinoids may be primary to the ovary, a component of ovarian teratomas, or metastatic.77 Such tumors yield cells with an oval-to-spindled configuration and moderate cytoplasm. The classic “salt and pepper” chromatin pattern is most useful in diagnosis.
Smears from dysgerminoma are characteristically hypercellular and composed of large tumor cells with prominent nucleoli admixed with a separate population of lymphoid cells (Fig. 12).77, 114 The background is classically described as “tigroid” and is best appreciated in air-dried, Diff-Quick-stained smears.114
Fig. 12. A. Cytologic preparation of material obtained from a large ovarian mass in a young woman shows a biphasic cell population comprised of large tumor cells (solid arrow) admixed with small lymphocytes (open arrow) consistent with dysgerminoma (stain, Diff-Quik). Many of the tumor cells are stripped of their cytoplasm and contain prominent nucleoli on Papanicolaou stain. B. Corresponding histologic section displays the characteristic “fried-egg” appearance of dysgerminoma (stain, hematoxylin-eosin).
Yolk sac tumors are also characteristically hypercellular on aspiration and are composed of moderately sized oval cells, dispersed both as single elements and in fragments. Vacuolated cytoplasm and both intracellular and extracellular hyaline globules are noted.114 The serum alpha-fetoprotein is typically elevated.77
Embryonal carcinoma of the ovary is very rare, and aspirates of such lesions are hypercellular, often showing hemorrhage and necrosis. Tumor cells are bizarre with a high nucleus-to-cytoplasm ratio. The differential diagnosis is usually poorly differentiated adenocarcinoma. Immunohistochemical studies may show positivity for chorionic gonadotropin in syncytiotrophoblastic-like cells.95 Embryonal tumors largely composed of embryoid bodies are referred to as polyembryomas.95
Most choriocarcinomas of the ovary represent metastases from uterine primaries. Therefore, thorough evaluation of the uterine corpus is indicated when a diagnosis of ovarian choriocarcinoma is considered.95 Microscopically, an admixture of syncytial and cytotrophoblastic elements in a necrotic and hemorrhagic background is typical of this lesion. Immunohistochemical reactivity for human chorionic gonadotropin is the rule.95
Primary ovarian tumors of uncertain histogenesis include small cell carcinoma and oncocytoma, among others.95 In two thirds of cases, primary ovarian small cell tumors produce paraendocrine hypercalcemia.115 In ovarian small cell carcinoma, an intermediate-sized, round-to-oval cell population with scanty cytoplasm and coarse granular chromatin may predominate.115
Processes known to secondarily involve the ovary include metastatic carcinomas (from breast, colon, stomach, pancreas, and gallbladder), malignant lymphomas, and myeloid leukemias.77, 116, 117 Cases of metastasis to an ovary that already contains a primary ovarian neoplasm have been reported in the literature.118 Such cases of “double cancers” are exceedingly rare, but they show the need for careful review of obviously diagnostic aspirates to not overlook cellular populations whose recognition may dramatically alter patient treatment.
Thus, despite reported controversial views on the role of ovarian needle aspiration, the technique has been accepted by many as a relatively innocuous procedure that can be accomplished with minimal complications and risks for the patient.77 Without question, the procedure of aspiration with cytodiagnosis can be used to preserve ovarian hormonal and reproductive function. If definitive surgical treatment becomes indicated based on information gleaned from aspirated material, the surgeon and patient may plan appropriately for expected outcomes.
Fine needle aspiration of both palpable and nonpalpable pelvic masses may be readily performed in the initial evaluation and follow-up of patients with gynecologic malignancies. Although the issue of the diagnostic reliability of FNA in primary ovarian masses is the subject of some debate, as discussed previously, the utility of FNA for the evaluation of nonovarian primary tumors and potential recurrent gynecologic tumors is unquestioned. Masses noted during pelvic examination may be aspirated via either a transvaginal or transrectal approach, often with the aid of a Franzen needle guide. Lesions located above the pelvic brim and those deep within the posterior pelvis require radiologic guidance for accurate placement of the biopsy needle. Ultrasound guidance may be sufficient for a number of these cases, although CT guidance is preferred for those lesions deep to bowel or bone.73
Numerous series with histologic or long-term clinical follow-up have confirmed the extremely high specificity (99%) of the technique for pelvic lesions.15, 17, 119, 120, 121 As with FNA in other anatomic regions, the sensitivity of the procedure is slightly lower (85–95%) because of occasional false-negative results.15, 17, 119, 120 Such occurrences are generally ascribed to either technical difficulties (e.g., inexperienced aspirator) or, more frequently, to intrinsic properties of the lesion (e.g., significant fibrosis). Thus, a negative FNA result in the presence of suspicious clinical findings warrants further evaluation to exclude malignancy.
FNA is also valuable in the surveillance of patients with a history of malignancy after therapy. In a series of patients who underwent FNA biopsy for evaluation of pelvic masses (parametrial, pelvic sidewall) noted on pelvic examination or imaging studies after primary therapy (including radiation therapy and hysterectomy) for cervical carcinoma, 22 of 30 patients had a positive aspirate.122 The utility of FNA biopsy in patient follow-up after radiation therapy alone was shown in another study of 58 patients with cervical or endometrial carcinoma.17 Approximately 40% of the patients were proved to have residual or recurrent disease on aspiration biopsy, with a 7% false-negative rate. More important, one third of the patients with a positive FNA had an equivocal or negative pelvic examination and imaging study results before the biopsy. Thus, FNA is an extremely useful tool that may be valuable in the often-difficult distinction between postradiation fibrosis and recurrent tumor.
As with all specimens, proper communication with the interpreting pathologist is crucial for aspirates from this site. In addition to the mass in question, a large number of normal structures may be concurrently sampled depending on the needle approach and direction, including bladder, ureter, and colon. A history of radiation or chemotherapy or both must also be relayed to the pathologist, perhaps avoiding a falsely suspicious or positive diagnosis based on marked therapy changes.
True fine needle aspirates of endometrial lesions are rarely performed. Biopsy is the standard of care for endometrial lesions. If present, intracavitary fluid can be collected and sent for cytologic evaluation. Abnormal cells in routine cervical-endocervical smears can at times also be sufficient for a diagnosis of malignancy.