Cytopathologist-performed fine-needle aspiration (FNA), or interventional cytopathology, is a minimally invasive technique that is highly accurate at definitively diagnosing samples from palpable head and neck lesions, superficial lymph nodes, and subcutaneous nodules covering the body, to name a few.1 Furthermore, rapid onsite evaluation (ROSE) can be used at the time of FNA to limit or extend the number of passes required to make a diagnosis and to adequately triage the specimen for microbiology cultures, a liquid-based preparation, a cell block preparation, flow cytometry, chemical analysis, or molecular diagnostic tests.2

Certainly, there are many diagnostic approaches that can be utilized to interpret an FNA sample. One of these strategies is to apply a site-specific approach that relies on cytomorphologic patterns. The use of cytomorphologic patterns can aid the interventional cytopathologist by providing a reproducible, pattern-based approach to diagnostically complex lesions not only during the procedure but also during final classification. These patterns can be used to ensure adequate diagnostic material is obtained, create a workable differential diagnosis, and suggest further specimen utilization for ancillary diagnostic studies. Herein, we provide an overview of the salient cytomorphologic patterns encountered in common specimens—including major salivary glands, the thyroid gland, and superficial lymph nodes—obtained by the interventional cytopathologist.

Salivary gland FNA is most performed on a mass lesion in the parotid gland, followed next by the submandibular gland.3 Salivary gland tumors are an incredibly heterogeneous group of neoplasms with over 30 specific and distinct types currently recognized by the WHO Classification of Head and Neck Tumors.4 To further compound this diagnostic complexity, many of the tumors, including both benign and malignant neoplasms, contain diverse and overlapping morphologic features, which may require tissue architecture or ancillary studies for definitive classification. Despite these diagnostic challenges, salivary gland FNA is highly accurate at definitively diagnosing common lesions, including pleomorphic adenoma, Warthin tumor, reactive lymph nodes, and many lymphomas.5, 6, 7 Furthermore, it is accurate at differentiating between non-neoplastic and neoplastic lesions, which will guide clinical management with clinical follow-up versus surgery, respectively.8,9 Finally, it can often distinguish between low-grade and high-grade carcinomas, which when combined with radiographic characteristics and location, can guide the extent or approach to surgery.10

One of the ways to overcome the diagnostic complexity in the salivary gland is to utilize cytomorphologic patterns to create a workable differential diagnosis. These patterns can be applied at the time of ROSE by the interventional cytopathologist to ensure that adequate diagnostic material is obtained, and that the specimen is triaged appropriately. Some of the more salient patterns in salivary gland cytology include the normal salivary gland, cystic, lymphoid-rich, basaloid, oncocytic, spindle cell, and pleomorphic patterns.11 For each pattern, the defining features, differential diagnosis, ROSE strategy, and diagnostic pearls are discussed in Table 1 while morphologic details are discussed in Table 2 and shown in Figure 1.

Briefly, recognition of these patterns informs sample adequacy. For example, aspiration of bland acinar cells that are in a normal lobular configuration suggests that the lesion is not being sampled, and the needle should be targeted to a different area of the mass. In addition, aspiration of non-mucinous cystic fluid without an epithelial component is non-diagnostic according to the Milan System for Reporting Salivary Gland Cytopathology (MSRSGC), and additional samples should be targeted at a more solid component. In contrast, however, aspiration of extracellular mucin material, even in the absence of an epithelial component, is diagnostic for Atypia of Undetermined Significance (AUS) according to the MSRSGC, suggesting that lesional tissue is being obtained. Due to the paucicellular nature commonly encountered with the mucinous cystic pattern, creation of a cytospin or liquid-based preparation to enrich for the cellular components should be considered. Next, aspiration of a lymphoid-rich lesion should prompt a careful search for an epithelial component from a neoplasm that commonly contains a tumor associated lymphoid proliferation (TALP), such as a Warthin tumor or an acinic cell carcinoma. In the absence of an epithelial component, flow cytometry or cultures may be indicated to rule out a lymphoma or an infection causing sialadenitis, respectively. In the salivary gland, the presence of the basaloid pattern is typically indicative of a neoplastic lesion; and, as a result, obtaining material for cell block preparation may provide value for further classification. In contrast, the oncocytic pattern may be the result of metaplasia, hyperplasia, a benign neoplasm, or a malignant neoplasm. Based on the degree of atypia, a cell block may be useful to rule out an acinic cell carcinoma or a salivary duct carcinoma. Next, the spindle cell pattern includes both salivary primary tumors, such as a myoepithelial-rich neoplasm (i.e. cellular pleomorphic adenoma or a myoepithelioma), as well as secondary processes, such as a schwannoma or a melanoma. As a result, additional passes for cell block creation can help confirm the etiology of the spindle cell process. Finally, the pleomorphic pattern is easily recognizable as lesional tissue; however, creating a cell block can help inform a more specific definitive diagnosis, such as a salivary duct carcinoma or a metastatic squamous cell carcinoma or melanoma.

When utilizing this approach, it is important to note that many distinct neoplasms may display one or more of these patterns. In fact, this might be helpful in making the diagnosis. For example, a Warthin tumor should show the non-mucinous cystic, lymphocyte-rich, and oncocytic patterns. The possible patterns, morphology, and ancillary studies for selected salivary gland neoplasms are presented in greater detail in Table 2. In summary, utilizing patterns in salivary gland FNA is a relatively simple framework that can guide the interventional cytopathologist during the procedure and during final classification.

Thyroid gland FNA may be performed by an interventional cytopathologist, particularly when trained to utilize ultrasound guidance. Nodules can be selected for FNA based on the ACR TI-RADS criteria and the size of the lesion.29 FNA is then performed on a suspicious nodule, and the sample is categorized according to The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC).30 This categorization informs patient management, which is usually clinical follow-up, repeat FNA, lobectomy, or total thyroidectomy. For certain cases with repeat indeterminate diagnoses, molecular testing performed on dedicated passes may provide clinically useful information.31,32 In most cases, though, thyroid gland FNA is highly accurate in selecting patients to be managed conservatively versus surgically.30

For proper categorization, adequate material must be obtained. The goals during ROSE should be to establish adequacy, create a differential diagnosis, and triage the sample accordingly. Smears that contain abundant colloid, atypical follicular epithelium, numerous lymphocytes, or at least six groups of at least ten follicular cells are diagnostic according to TBSRTC 30. Next, diagnostic patterns can be used to guide additional passes, create a differential diagnosis, and determine the need for further specimen utilization. The most pertinent cytomorphologic patterns encountered on thyroid gland FNA include the paucicellular, macrofollicular, microfollicular, papillary, Hurthle cell, and discohesive patterns.11 For each major pattern, the defining features, differential diagnosis, ROSE strategy, and diagnostic pearls are described in Table 3 while morphologic details are presented in Figure 2.

Briefly, the paucicellular pattern may be comprised of predominantly cyst fluid with scant or absent follicular cells and a lack of abundant colloid. Since cystic features can be present in non-neoplastic, benign neoplastic, and malignant lesions, risk stratification is not possible. As a result, the case is considered non-diagnostic according to TBSRTC.30 In contrast, a case may be paucicellular but be comprised of abundant colloid, which is diagnostic for a benign follicular/colloid nodule; these cases are considered adequate and benign according to TBSRTC.30 For both scenarios with limited follicular cellularity, it may be helpful to create a cytospin or liquid-based preparation to enrich for the epithelial component. Next, the macrofollicular pattern is characterized by the presence of variably sized, bland and non-overlapping fragments and sheets of follicular cells. The cellularity and amount of colloid can range, but there should be a notable absence of a predominant microfollicular architecture and a lack of cytologic atypia. The macrofollicular pattern typically indicates aspiration of a benign follicular nodule or even normal thyroid tissue. If the adequacy criteria have been met, these cases are considered benign according to TBSRTC.30 In contrast, the microfollicular pattern is characterized by the presence of uniform, round groups of <15 follicular cells in a background of scant or absent colloid. This pattern is typically indicative of a follicular neoplasm, such as a follicular adenoma or a follicular carcinoma; however, other tumors with a follicular architecture, such as a non-invasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) or follicular variant of papillary thyroid carcinoma (PTC) should be in the differential if nuclear atypia is also present. For cases lacking nuclear atypia, the diagnosis will largely depend on the cellularity. Cases that have a predominant microfollicular architecture with limited but adequate cellularity should be diagnosed as atypia of undetermined significance (AUS)/follicular lesion of undetermined significance (FLUS), depending on the adopted terminology in the laboratory.30 Hypercellular cases with a predominant microfollicular architecture should be diagnosed as suspicious for a follicular neoplasm (SFN). Occasionally, parathyroid tissue can show a microfollicular pattern. As a result, it may be helpful to send the needle rinse fluid for parathyroid hormone (PTH) analysis in select cases. Next, the papillary pattern mainly refers to atypical nuclear features that raise the concern for or are diagnosed for a PTC, including nuclei that are enlarged, oval, and crowded with thick, irregular nuclear contours, nuclear grooves, pale and powdery chromatin, peripheral punctate nucleoli, and intranuclear pseudoinclusions. If any of these features are present and there is no explanation for this type of change, such as cyst lining cells in a multinodular goiter or a background of chronic lymphocytic thyroiditis (CLT), the diagnosis will depend on the degree of atypia, ranging from AUS to Malignant. For the Hurthle cell pattern, the follicular cells are enlarged with abundant and densely granular cytoplasm with round nuclei that can be central or eccentric and may show variably prominent nucleoli. The differential diagnosis for the Hurthle cell pattern is long because Hurthle cell changes can be seen in non-neoplastic conditions such as multinodular goiter and CLT or in neoplastic lesions, such as a Hurthle cell adenoma, Hurthle cell carcinoma, or variants of PTC. As a result, the diagnosis will depend on the predominance of another pattern, such as coexisting macrofollicular, microfollicular, or papillary patterns. Finally, the discohesive pattern is characterized by single individual cells in the absence of colloid, which can be seen with many lesions, including Hurthle cell adenomas and carcinomas, as well as poorly differentiated PTC, anaplastic thyroid carcinoma, and medullary carcinoma. In addition, parathyroid tissue, lymphomas, and metastases may also present with a predominant discohesive pattern. As a result, certain cases may warrant testing the aspirate fluid for calcitonin, PTH, or flow cytometry while others may benefit from preparing a cell block to diagnose a metastasis.

In summary, thyroid gland FNA is a minimally invasive, cost-effect diagnostic test for working up patients with thyroid gland nodules.35,36 Utilizing cytomorphologic patterns is a practical approach to guide the interventional cytopathologist at the time of ROSE and during final classification.

FNA of a superficial lymph node is a minimally invasive, quick, and cost-effective diagnostic test to evaluate the etiology of lymphadenopathy.36, 37, 38, 39, 40 The differential diagnosis is extensive and includes non-neoplastic lesions, such as reactive hyperplasia and infection, and neoplastic lesions such as lymphomas and metastases. The goal of the procedure is to make a diagnosis and obviate the need for an invasive diagnostic excisional biopsy. However, for some lymphomas, architectural assessment may be necessary.41 Despite this diagnostic complexity, lymph node FNA is highly accurate, especially with the aid of ancillary studies.37,38 ROSE is helpful in ensuring an adequate procedure and in triaging the specimen for culture, flow cytometry, cytogenetics, or cell block preparation based on the differential diagnosis. As previously discussed in the salivary gland and thyroid gland, utilizing cytomorphologic patterns at the time of ROSE can inform procedure adequacy, build a differential diagnosis, and indicate additional warranted studies. The major cytomorphologic patterns of superficial lymph node FNA include the infectious, polymorphous lymphocytic, small-to-medium lymphocytic, large lymphocytic, and metastatic patterns. For each pattern, the defining features, differential diagnosis, ROSE strategy, and diagnostic pearls are discussed in Table 4 while morphologic details are discussed in Table 5 and Table 6 and shown in Figure 3.

Briefly, the infectious pattern is composed of either acute suppurative or granulomatous inflammation. The Diff-Quik stain can be helpful in identifying fungal and acid-fast bacilli but, ultimately, cultures should be sent for specific organism identification and antimicrobial susceptibility testing.42 Both classic Hodgkin lymphoma and types of non-Hodgkin lymphoma can also present with mixed inflammatory or granulomatous inflammation. As a result, in certain scenarios, a cell block may be warranted. If ample material is obtained, flow cytometry may be performed to rule out the possibility of non-Hodgkin lymphoma. The polymorphous lymphocytic pattern is comprised of a mix of lymphocytes in various stages of maturation and loosely cohesive clusters of germinal center components, including tingible-body macrophages—which is typically indicative of a reactive process. The differential includes reactive lymphoid hyperplasia, as well as Hodgkin and non-Hodgkin lymphoma. Unless clinicians and/or radiologists have a strong suspicion for metastatic disease, flow cytometry should be sent to rule out a clonal population. If there is clinical concern for Hodgkin lymphoma, a cell block may be considered to highlight the rare neoplastic cells. The small-to-medium lymphocytic pattern is typically indicative of a non-Hodgkin lymphoma and displays a “clonal” low power appearance, due to a lack in the broad range in size of the lymphocytes, as seen in the polymorphous lymphocytic pattern. At a minimum, flow cytometry should be ordered. The large lymphocytic pattern is typically obviously malignant and comprised of markedly enlarged and atypical dispersed lymphocytes. The differential most commonly includes a diffuse large B-cell lymphoma, high grade B-cell lymphoma, Burkitt lymphoma, or anaplastic large T-cell lymphoma. The diagnosis can typically be made with the aid of flow cytometry; however, specific subtyping may require tissue evaluation.

Finally, the metastatic pattern is comprised of a population of abnormal, non-lymphocytic cells. Most of these tumors will be metastatic carcinomas and melanomas, and the differential will mainly depend on the anatomic site. For example, a lateral neck lymph node may contain a metastasis from a cutaneous head and neck squamous cell carcinomas (SqCC) or melanoma, oropharyngeal/oral cavity SqCC, or a papillary thyroid carcinoma, among others. If a metastatic squamous cell carcinoma with unknown primary is detected in a head and neck lymph node, an attempt should be made to obtain enough tumor for cell block preparation. A p16 immunostain with or without High-risk HPV mRNA in situ hybridization (ISH) should be performed and, if negative, Epstein Barr-virus in-situ hybridization should be considered to help further subclassify the tumor.46, 47, 48, 49 However, in the axilla, a metastatic breast primary or upper extremity cutaneous primary would be at the top of the differential. As a result, close attention to history and cell block preparation are paramount for making a definitive diagnosis. In summary, while performing ROSE, an infectious pattern should trigger cultures, a lymphocytic pattern should trigger at least flow cytometry, and a metastatic pattern should trigger a cell block.