Anaplastic Thyroid Carcinoma: Cytomorphologic Features on Fine-Needle Aspiration and Associated Diagnostic Challenges

Abstract

Objectives

Anaplastic thyroid carcinoma (ATC) is an aggressive malignancy, and early diagnosis, often aided by fine-needle aspiration (FNA), is key to improving patient prognosis. While the current literature describes some of the cytologic features (CFs) of this entity, a comprehensive examination of the CFs has not yet been performed.

Methods

We retrospectively searched our electronic database for ATC cases with available slides between January 2008 and December 2019. Cases were examined for 22 CFs and compared with a control group of differentiated thyroid carcinoma.

Results

A total of 18 ATC cases meeting our inclusion criteria were identified. Most cases showed moderate to high cellularity (83%) and epithelioid cytomorphology (83%). Architecture included either predominantly groups/clusters of tumor cells (56%) or single tumor cells (44%). The other CFs were as follows: nuclear enlargement (100%), nuclear crowding (89%), nuclear membrane irregularities (100%), multinucleated tumor cells (33%), and background acute inflammatory cells (50%). Of the CFs examined, statistically significant differences between ATC and the control groups were found in the following: nuclear pleomorphism, coarse/clumped chromatin, macronucleoli, apoptosis, and necrosis.

Conclusions

Identification of key CFs in FNA coupled with the clinical history aids in the diagnosis of ATC and helps distinguish it from other mimickers.

KEY POINTS.

• Anaplastic thyroid carcinoma (ATC) is an aggressive thyroid malignancy with poor survival outcomes.

• Identification of key characteristic cytologic features along with pertinent clinical presentation helps improve fine-needle aspiration diagnosis of ATC.

• A multidisciplinary approach should be considered, as ATC diagnosis is often challenging.

INTRODUCTION

Anaplastic thyroid carcinoma (ATC) is one of the most aggressive thyroid malignancies and is associated with low overall survival rates.¹ It typically presents as a rapidly enlarging thyroid mass most often affecting patients during the sixth and seventh decades of life.² ATC can either arise de novo or in a background of differentiated thyroid carcinoma (DTC), typically papillary thyroid carcinoma (PTC). It is essential to appropriately evaluate and diagnose these cases in a timely manner, as patients with ATC with smaller tumor size (<5 cm) and without any nodal disease have better median survival rates compared with patients with larger tumor size (>5 cm) or disease metastatic to the lymph nodes.³ Fine-needle aspiration (FNA) cytology plays an important role in the assessment and diagnosis of thyroid lesions, including ATC.⁴ Some of the cytologic features of ATC have been described in previously published studies, but a comprehensive evaluation of cytologic features has yet to be performed.^5,6 The cytologic features evaluated in this study are more comprehensive than those reported in previous studies.

MATERIALS AND METHODS

This study was conducted after getting approval from the institutional review board. We retrospectively searched our electronic data-based system from January 1, 2008, to December 31, 2019, for cases with a diagnosis of “anaplastic thyroid carcinoma.” Inclusion criteria included the following: histologically confirmed ATC cases with associated cytology and histology slides available for review and/or cytology cases with available slides and adequate material and diagnosed as “ATC” without any subsequent surgical specimen. We collected the relevant data pertaining to patient demographics, clinical features of the lesion, histologic follow-up details (if available), and patient follow-up. Each case had at least one FNA pass that was performed under ultrasound guidance by a radiologist using a 25-gauge needle. Onsite evaluation was not performed on most cases. One smear slide was air-dried and stained with Diff-Quik stain (a modified Romanowsky stain), while a second smear slide was immediately immersed in 95% ethanol and later stained with Papanicolaou stain. Needles were rinsed in CytoRich Red solution (Thermo Fisher Scientific) for a ThinPrep liquid-based preparation (Hologic). Cell blocks were prepared in seven cases from the residual specimen in CytoRich Red fluid.

Eighteen cases meeting the inclusion criteria were identified. Simultaneously, 18 cases with a histologic diagnosis of DTC were randomly selected as a control group (mean [SD] age, 54.5 [15.4] years; 17 women and 1 man). The histopathologic diagnoses of these cases were as follows: five cases of PTC, follicular variant; four cases of PTC, classic type; two cases of PTC, tall cell variant; four cases of follicular thyroid carcinoma; and three cases of Hurthle cell carcinoma. This control group was cytologically characterized as follows: eight had PTC, two were suspicious for PTC, and eight were classified as follicular/Hurthle cell neoplasms. Slides were retrieved and the following 22 cytologic features were evaluated by two cytopathologists (S.M.G. and R.A.): cellularity (low, moderate, or high), predominant architecture (single cells, loosely cohesive groups, cohesive groups, or papillary groups), predominant cell type (spindle or epithelial/epithelioid), pleomorphism (mild, moderate, or severe), nuclear changes including nuclear enlargement, nuclear overlapping and crowding, nuclear shape, nuclear membrane irregularities, nuclear grooves, macronucleoli or prominent nucleoli, small nucleoli, intranuclear pseudoinclusions, multinucleated tumor cells, quality of chromatin material (coarse clumped, fine powdery, or coarse granular), cytoplasm (scant or moderate to abundant), mitosis, apoptosis, necrosis, presence of acute inflammation in background or within tumor cells, and presence or absence of colloid and background PTC (if any). The predominant pattern of architectural features was recorded. The presence or absence of multinucleated giant cells was also documented. The cytologic features between the two groups were compared, and statistical analysis was performed by computing the odds ratio using Fisher exact test in R statistical software. To identify whether a subset of cytologic features was more helpful for ATC diagnosis, we used lasso regression (L1-regularized logistic regression) implemented via the glmnet package in R.⁷ All statistical tests were performed using R (version 4.0.3).⁸ To adjust for multiple hypothesis testing, the Bonferroni correction was applied. A P value of less than .05 was considered significant.

RESULTS

A total of 18 ATC cases were evaluated (mean [SD] age, 67.38 [11] years; 10 men and 8 women) in this study TABLE 1. Presenting clinical history was variable, but the most frequent finding was an enlarging thyroid or neck mass (present in >70% of cases), followed by locally compressive symptoms (hoarseness or shortness of breath). One case had a longstanding history of a thyroid nodule, and another case had a remote history of thyroidectomy with recurrent PTC in the lymph nodes. FNA cytology was obtained from the thyroid lesion (majority), neck lymph nodes, or a neck mass. All cytology cases had histologic follow-up (resection = 13 and biopsy = 3), except for two cases that were diagnosed as ATC on cytologic examination alone. All but one case having histologic follow-up were confirmed to be ATC. This case (neck mass) was characterized as ATC on cytology, but the concurrent biopsy specimen indicated that it was poorly differentiated thyroid carcinoma. The mean (SD) follow-up interval between cytology and histology was 33 (69.6) days.

TABLE 1.

Clinicopathologic Characteristics of Anaplastic Thyroid Carcinoma Cases

Characteristic	Value
Age, mean (range), y	67.38 (53-93)
Sex, No.
Male	10
Female	8
Clinical presentation, No.
Enlarging thyroid or neck mass	13
Othersa	5
Tumor site (based on histology follow-up: n = 16), No.
Thyroid gland	13
Left	4
Right	2
Bilateral	7
Neck mass/lymph node	3
Tumor size, mean (SD), cm	6.6 (2.7)
Cytology diagnostic category, No.
Malignant	15
Suspicious for malignancy	2
Neoplasm	1
Cell blocks available, No.	7
Histologic follow-up available, No.	16
Procedure, No.
Biopsy	3
Resection	13
Cytology-histology follow-up, mean, d	33

^aOther: Compressive symptoms (n = 5) including one case with longstanding thyroid nodule/goiter and shortness of breath.

Overall, the mean age (years) for ATC cases was higher compared with the control group (67.38 [11] vs 54.5 [15.4]; P = .0067). Clinical follow-up was available in 11 patients, and of those, 8 died within 1 year of diagnosis, and 5 had either lung nodules on imaging or a tissue diagnosis suggesting lung metastasis. On histologic evaluation, seven patients showed background DTC, most commonly PTC (6/7, 86%) TABLE 2.

TABLE 2.

Cytology-Histology Details

		Cytology		Histology
Case No.	Tumor Size, cm	Cytology FNA Site	Cytology Diagnosis	Procedure	Predominant Histologic Pattern	Differentiated Thyroid Carcinoma
1	3.4	T	AC	—	1a	—
2	7	T	AC	—	2b	—
3	12	T	PFMCc	Resection	Spindle cell	Not seen
4	2.5	NLN	MC	Resection	Epithelioid	Present
5	6	T	FN/HCN	Biopsy	Mixed spindle and epithelioid	Not seen
6	3.6	T	Suspicious for carcinoma	Resection	Pleomorphic/giant cell	Present
7	10	N	AC	Biopsy	Epithelioid	—
8	7	N	PFMCc	Resection	Epithelioid	Not seen
9	11	T	HGC with AF	Resection	Pleomorphic/giant cell	Present
10	2	NLN	PDC		Epithelioid	Not seen
11	5.7	NLN	PDMN	Biopsy	Epithelioid	—
12	7.9	T	Suspicious for SCCc	Resection	Epithelioid with SD	Present
13	7.5	T	HGMN with AF	Resection	Pleomorphic/giant cell	Present
14	4.5	T	PTC	Resection	Epithelioid	Present
15	5.8	T	PDMN with AF	Resection	Epithelioid	Not seen
16	8	T	Consistent with AC	Resection	Epithelioid with giant cells	Present
17	8	T	PDCc	Resection	Pleomorphic/giant cell	Not seen
18	6.5	NLN	HGC	Resection	Epithelioid with SD	Not seen

AC, anaplastic carcinoma; AF, anaplastic feature; FN/HCN, Hurthle cell neoplasm; HGC, high-grade carcinoma; HGMN, high-grade malignant neoplasm; MC, metastatic carcinoma; N, neck; NLN, neck lymph node; PDC, poorly differentiated carcinoma; PDMN, poorly differentiated neoplasm; PFMC, positive for malignant cells; PTC, papillary thyroid carcinoma; SCC, squamous cell carcinoma; SD, squamous differentiation; T, Thyroid; —, not applicable.

^aOn cytology: spindle cell pattern.

^bOn cytology: epithelioid pattern.

^cWith differential diagnosis.

Of these 18 cases, 15 were cytologically classified as “malignant,” 2 were characterized as “suspicious,” and 1 case was called “follicular/Hurthle cell neoplasm (FN/HCN).” Malignant diagnoses included ATC, poorly differentiated carcinoma or neoplasm, high-grade carcinoma, PTC, or others. Of these malignant cases, one case was PTC, and subsequent histologic follow-up (~10 months) showed anaplastic carcinoma with background PTC. Retrospective review of this cytology case confirmed the original findings. Two cases with a “suspicious for carcinoma” diagnosis had scant cellularity, including one with atypical/dyskeratotic squamoid cells. One case was categorized as FN/HCN with a note describing rare cells with atypia and raising the possibility of dedifferentiation.

We have evaluated cytologic features in ATC cases FIGURE 1, FIGURE 2, FIGURE 3, and FIGURE 4, and the results were as follows: increased cellularity (moderate/high) in 83%, predominant architecture (single cells, 44%; clusters or groups, 56%), predominant cell type (epithelioid, 83%; spindle cell type, 17%), pleomorphism (no or mild to moderate, 17%; moderate to severe, 83%), nuclear crowding in 89%, nuclear enlargement in 100%, irregular nuclear membranes in 100%, macronucleoli in 72%, intranuclear pseudoinclusions in 28%, multinucleated tumor cells in 33%, coarse clumped chromatin material in 78%, mitosis in 44%, necrosis in 67%, presence of an acute inflammatory response in 50%, neutrophils within tumor cells in 17%, presence of colloid in 33%, and background PTC in 17%. More than one cell type (spindle cell, epithelioid cells, or presence of osteoclast-like multinucleated giant cells) was seen in 11 (61%) cases.

FIGURE 1.

A, B, Scattered single cells along with multinucleated tumor cells (A, Diff-Quik, ×400; B, ThinPrep, ×400). C, Sheets of tumor cells and scattered multinucleated tumor cells (arrow) and osteoclast-like giant cells (arrowhead) (H&E, ×200). D, Tumor mass has replaced the majority of normal thyroid parenchyma (gross image).

FIGURE 2.

A, Tumor cells with epithelioid cytomorphology (Papanicolaou, ×400). B, Loosely cohesive group of tumor cells (Diff-Quik, ×400). C, Tumor cells with spindled cytomorphology (ThinPrep, ×400). D, Cohesive cluster of epithelioid tumor cells (Papanicolaou, ×400).

FIGURE 3.

A, Epithelioid tumor cells with intranuclear pseudoinclusion (arrow) (Papanicolaou, ×400). B, Pleomorphic tumor cells with coarse clumped chromatin (Papanicolaou, ×400). C, Tumor cells with prominent macronucleoli (Papanicolaou, ×400). D, E, Tumor cells with squamoid features (D, Papanicolaou, ×400; E, H&E, ×100).

FIGURE 4.

A-D, Tumor cells with background acute inflammatory cells and few tumor cells show phagocytized neutrophils within the cytoplasm (C, arrow) (A, Papanicolaou, ×200; B, cell block with H&E, ×400; C, Papanicolaou, ×400; D, H&E, ×200). E, Focal necrotic debris (ThinPrep, ×400).

In this study, ATC showed predominantly single cell architecture and increased nuclear pleomorphism compared with the DTC control group (P = .064 and P < .0001). Other cytologic features more pronounced in ATC were the presence of macronucleoli (P < .0001) and coarse clumped chromatin (P < .0001). Mitosis, apoptosis, and necrosis were seen in only a few ATC cases and absent in DTC cases. We also noted that nuclear pleomorphism appeared to be a necessary cytologic feature (specificity 100% and sensitivity 93% by using regularized logistic regression) compared with other features, but this finding does not imply that the other features are not significant, as our data set is limited. In our opinion, using a combination of cytologic features, including marked nuclear pleomorphism, would be superior to using one cytologic feature alone. Another interesting finding in ATC cases was the presence of background acute inflammatory cells or neutrophilic infiltrate TABLE 3. The predominant cell type was epithelioid cells (83%), followed by spindle cells (17%). Giant cells were noted in both spindle cell and epithelioid cell patterns, while squamous cells were noted in the epithelioid type only. We noticed that both Diff-Quik and Papanicolaou stains had contributed to the evaluation of cytologic features, but nuclear features were more elaborated in the Papanicolaou stain.

TABLE 3.

Comparison of Cytologic Features Between Anaplastic Thyroid Carcinoma (ATC) and Differentiated Thyroid Carcinoma (DTC)

Cytologic Features	ATC (n = 18), No.	DTC (n = 18), No.	P Value	Odds Ratio (95% CI)
Cellularity
Low	3	2	1a	1.58 (0.157-21.4)
Moderate	11	10
High	4	6
Architecture (predominant)
SC	8	0	.064b	∞ b (2.36-∞)
LCG	3	5
CG	7	9
PG	0	4
Cell type (predominant)
Spindle cells	3	0	1	∞ (0.428-∞)
Giant cells	2	0
Epithelioid cells	15	18
Giant cells	5	4
Squamous cells	2	0
Pleomorphism
None or mild to moderate	3	18	<.0001	0 (0.000-0.086)
Moderate to severe	15	0
Nuclear changes
Nuclear enlargement
Present	18	15	1	∞ (0.428-∞)
Absent	0	3
Nuclear crowding
Present	16	13	1	2.99 (0.405-36.261317)
Absent	2	5
Nuclear shape
Round or round to oval	13	18	1	0 (0.000-0.962)
Spindle/plumped spindle	5	0
Irregular nuclear membrane
Present	18	13	1	∞ (1.04-∞)
Absent	0	5
Nuclear grooves
Present	4	11	.900	0.192 (0.032-0.946)
Absent	14	7
Prominent or macronucleoli
Present	13	0	<.0001	∞(7.05-∞)
Absent	5	18
Small nucleoli
Present	11	13	1	0.613 (0.116-3.01)
Absent	7	5
Intranuclear pseudoinclusions
Present	5	7	1	0.613 (0.116-3.01)
Absent	13	11
Multinucleated tumor cells
Present	6	1	1	8 (0.814-412)
Absent	12	17
Chromatin
Coarse, clumped, irregular	14	0	<.0001	∞(8.937137-∞)
Other (powdery or coarse granular)	4	18
Cytoplasm
Moderate to abundant	15	8	.780	5.91 (1.10-43.3)
Scant	3	10
Mitosis
Present	8	0	.064	∞(2.356525-∞)
Absent	10	18
Apoptosis
Present	13	0	<.001	∞(7.051877-∞)
Absent	5	18
Necrosis
Present	12	0	<.001	∞(5.644604-∞)
Absent	6	18
Inflammation
Acute	9	1	.160	15.7 (1.72-784)
Other (no or mixed)	9	17
Neutrophils within tumor cell cytoplasm
Present	3	1	1	3.29 (0.235-189)
Absent	15	17
Colloid
Present	6	13	.960	0.202 (0.036-0.964)
Absent	12	5
Papillary thyroid carcinoma
Present	3	8	1	0.260 (0.036-1.43)
Absent	15	10

CG, cohesive group; CI, confidence interval; LCG, loosely cohesive group; PG, papillary group; SC, single cell.

^a P value is calculated between 2 groups as follows: group 1 (low) vs group 2 (moderate and high).

^b P value is calculated between 2 groups as follows: group 1 (single cells) vs group 2 (LCG, CG, and PG).

The cytologic features of cases with a definitive cytology diagnosis (ATC or malignancy/carcinoma with anaplastic features) were also compared with cases with indeterminate malignant diagnosis (includes any type of malignancy diagnosis) TABLE 4. Seven cases had available cell blocks with adequate cellularity and had contributed equally to the diagnosis along with the cytology smears. Of the seven cases with cell block material, five had a definitive diagnosis on cytology. Immunostain was performed on seven cases, including six with cell blocks and one case with additional ThinPrep slides. The most frequent immunostains performed were PAX-8 (4/5, 80%), TTF-1 (1/5, 20%), thyroglobulin (0/4), and cytokeratin AE1/AE3 (2/3, 67%).

TABLE 4.

Cases With “Malignant” Diagnosis on Cytology

Cytologic Features	Definitive Diagnosis (n = 7)	Indeterminate Diagnosis (n = 8)	P Value	Odds Ratio (95% CI)
Predominant architecture
Single cells	5	2	1	6.4 (0.520-128)
Other (LCG, CG, PG)	2	6
Cell type
Spindle	1	2	1	0.52 (0.007-12.7)
Epithelioid	6	6
Pleomorphism
Mild	0	2	1	0 (0.000-6.00)
Moderate to severe	7	6
Nuclear grooves
Present	2	2	1	1.19 (0.063-22.2)
Absent	5	6
Macronucleoli
Present	4	6	1	0.47 (0.027-6.24)
Absent	3	2
Multinucleated tumor cells
Present	4	2	1	3.62 (0.302-64.1)
Absent	3	6
Mitosis
Present	5	3	1	3.76 (0.325-65.7)
Absent	2	5
Apoptosis
Present	6	5	1	3.31 (0.191-220)
Absent	1	3
Necrosis
Present	6	4	1	5.3 (0.346-343)
Absent	1	4
Acute inflammation
Present	3	3	1	1.23 (0.101-15.4)
Absent	4	5
Cell block
Present	5	2	1	6.4 (0.520-128)
Absent	2	6
Immunostain
Performed	5	2	1	6.4 (0.520-128)
Not performed	2	6

CG, cohesive group; CI, confidence interval; LCG, loosely cohesive group; PG, papillary group.

DISCUSSION

Anaplastic thyroid carcinoma is an aggressive malignancy that usually presents as a rapidly enlarging neck mass. This rare malignancy arises from thyroid follicular cells and accounts for less than 2% of thyroid cancer.⁹ Definitive diagnosis of ATC requires either cytologic or histopathologic evaluation. Oftentimes, ATC diagnosis is challenging, especially on small and limited samples, and it is essential to differentiate this entity from its benign and malignant mimickers. In this study, we evaluated a variety of cytologic features in ATC cases to determine if there was a correlation with any cytomorphologic feature(s) that can help in establishing a definitive diagnosis on cytologic examination. We selected DTC cases as a control group. There is a spectrum of cytologic features that encompasses DTC, poorly differentiated thyroid carcinoma (PDTC), and ATC with some overlap among each group. On histologic examination, PDTC diagnosis requires the presence of certain criteria, including solid/trabecular growth pattern, lack of PTC nuclear features, and one of the following features: (1) convoluted nuclei, (2) more than 3 mitoses/10 high-power field, and (3) necrosis.¹⁰ Overall, the PDTC diagnosis on cytology has its own associated challenges due to its significant overlap with DTC and ATC. Similarly, no uniform criteria to diagnose PDTC on cytology specimens have been established.¹¹ Some authors have proposed that PDTC can be suggested on cytology based on certain features such as solid/trabecular pattern, high nuclear/cytoplasmic ratio, crowding, and single-cell architecture.¹² Due to the lack of uniformly accepted criteria, we selected DTC cases as the control group (reference group/baseline), which have more consistent and established diagnostic criteria on cytology material.¹³ ATC can share some features with PTDC and also can arise in the background of DTC, particularly in PTC components. In this study, we noticed a PTC component in three cytology cases while seven subsequent surgical cases showed background PTC components.

It is essential to separate ATC from other morphologic mimickers as prognosis and treatment approaches are different.⁹ Frequently, ATC presents as a rapidly enlarging neck mass with or without local compressive symptoms, and FNA cytology can be helpful for establishing a definitive diagnosis. It is important to recognize cytologic features of ATC to establish an accurate cytologic diagnosis. In a patient with a clinical history concerning for ATC, triaging material for cell block preparation would be helpful. In this study, 15 cases were classified as “malignant” on FNA cytology. Of those 15 cases, 7 (46%) were characterized as ATC, consistent with ATC or malignancy/carcinoma with anaplastic features. Of these seven, 5 had cell block material available, and immunostains were performed on 5 cases. The remaining 8 (54%) cases were signed out as “malignant” by using different diagnostic terminologies, and only 2 cases had cell block material. In a study of ATC cases, Rivera et al⁶ performed PAX-8 immunostain on 5 cases with available cell blocks, and all were negative. In this study, PAX-8 immunostain was positive in 4 of 5 cases. Simultaneously, other studies have reported variable PAX-8 staining in ATC ranging from 54.4% to 80%.^14-16 This staining pattern is also variable among different histopathologic patterns of ATC.¹³

We evaluated cytomorphologic features of ATC in detail to evaluate the most consistent or frequent features that could facilitate an accurate diagnosis. We noticed at least moderate nuclear pleomorphism in 15 (83%) cases, presence of macronuclei in 10 (55%) cases, coarse clumped irregular chromatin material in 15 (83%) cases, nuclear crowding in 16 (89%) cases, and nuclear enlargement in all cases (100%). Other prominent cytomorphologic features noted in ATC were the presence of mitosis, necrosis, apoptosis, and background acute inflammatory infiltrate. In comparison, Rivera et al⁶ reported marked pleomorphism, neutrophilic infiltrate, and necrosis as prominent cytologic features of ATC. In their study, most cases showed increased cellularity (75%), and macronuclei were noted on 27% of cases, but histologic follow-up was available in only 55% of cases. Others have also documented cytologic features in ATC, including increased cellularity, inflammation, and necrosis as prominent features.⁵ Most (83%) of our ATC cases were cellular (moderate/high), but our control group also showed increased cellularity (89%). In this study, one case was categorized as FN/HCN on cytology with a note describing rare cells with atypia and raising the possibility of dedifferentiation. Variable type of cytologic atypia can be seen in Hurthle cell lesions and is not a reliable criterion to differentiate a nonneoplastic lesion from a neoplastic lesion.¹⁷ Presence of isolated cytologic atypia (any type) in an endocrine organ such as thyroid should be evaluated carefully to avoid overinterpretation. Diagnosis of ATC may require identifying a group of atypical cytologic features along with clinicopathologic correlation.

In this study, we identified an epithelioid/epithelial cell–type pattern as a predominant pattern on cytology followed by a spindle cell–type pattern. Spindle cell pattern on cytology examination brings up a differential diagnosis of sarcoma or melanoma. These two entities are rare, but correlation with clinical features (ie, rapidly enlarging neck mass, clinical history) is helpful. Predominant single cell architecture was seen in eight of our cases. Melanoma and lymphoma should be ruled out in cases with only single cell architecture. In difficult cases, performing immunostains may assist in ruling out certain entities. Osteoclast-like multinucleated giant cells were noted in 7 (39%) ATC cases while neoplastic multinucleated cells were observed in 6 (33%) cases. Similar findings were reported by Jin et al,⁵ who identified multinucleated tumor cells in 33% of their cases. Other important findings are the presence of acute inflammatory infiltrate in half of our cases (50%), and 3 (17%) cases showed phagocytized neutrophils in the tumor cell cytoplasm. Similar findings are documented by Us-Krasovec et al¹⁸ in a study of 113 cases.

Presence of atypical squamous cells in thyroid FNA is rare. Two of our cases showed atypical squamous/dyskeratotic cells and were signed out in a “suspicious” diagnostic category. Other studies have also described the presence of malignant squamous cells in rare cases.⁵ Squamous cell carcinoma of the thyroid gland is very rare, and the differential diagnosis includes either ATC or PTC with squamous differentiation or metastatic carcinoma. In such cases in which the malignant squamous component is predominant, extensive sampling could be helpful for finding a DTC component. Some authors suggested that pure squamous cell carcinoma may be considered ATC.¹⁹ Cytology cases with atypical/dyskeratotic squamous cells should be characterized in a broad category (ie, “suspicious” or “malignant”) with a note stating the differential diagnosis and suggesting clinical correlation. Such cases may require further discussion with the clinical team or in multidisciplinary tumor board conferences.

Recognition of the appropriate clinical history along with identifying cytologic features is helpful for ATC diagnosis TABLE 5. We evaluated many cytologic features and found that presence of moderate to severe nuclear pleomorphism (P < .0001), macronucleoli (P < .0001), coarse clumped chromatin material (P < .0001), apoptosis (P < .001), and necrosis (P < .001) are statistically significant features. However, recognizing only isolated (one or two) cytologic features may often pose diagnostic challenges, and a cautious approach should be used. Variable types of cytologic atypia can be seen in many reactive and benign conditions.²⁰ On the other hand, inflammatory cells and multinucleated giant cells can be seen in de Quervain thyroiditis (granulomatous thyroiditis).²¹ Riedel thyroiditis can show a sparsely cellular specimen with fragments of stromal fibrous tissue, mixed inflammatory cells, and scattered bland spindles and can mimic the paucicellular variant of ATC. Similarly, fibrosis and reactive stromal spindle cells can be seen in other thyroiditis due to prior FNA-related changes and should be evaluated cautiously.²² Although cellularity is not a statistically significant feature, we noticed that all three cases with low cellularity showed macronucleoli, necrosis, and apoptosis.

TABLE 5.

Cytologic Diagnosis of Anaplastic Thyroid Carcinoma: Helpful Features and Diagnostic Considerations

Feature	Differential Diagnosis
Presence of clinical or imaging concern	Riedel thyroiditis, any other malignancy (primary vs secondary)
Cytology features
Reasonably cellular specimen (may aid in diagnosis)	Riedel thyroiditis (in cases with low cellularity and spindle cytomorphology)
Presence of the following features:
Moderate to severe pleomorphism	Reactive/inflammatory conditions, PDTC, metastatic malignancy
Macronucleoli	Hurthle cell lesions
Coarse clumped chromatin	PDTC, metastatic malignancy
Apoptosis	Lymphoid malignancy
Necrosis	PDTC, metastatic malignancy
Immunostains on cell block (cases with clinical/imaging concern for ATC)	To confirm ATC and to rule out other differential diagnosis

ATC, anaplastic thyroid carcinoma; PDTC, poorly differentiated thyroid carcinoma.

There are a few limitations pertinent to this study. The sample size of this study is small because ATC is a rare thyroid malignancy. In addition, this study was based on retrospective analysis, and cytology slides were not available for all ATC cases. Furthermore, it is debatable whether to select PDTC or DTC as an appropriate control group. There are no uniformly established cytologic criteria for PDTC, and Turin’s criteria are based on the histologic evaluation.¹⁰ On the other hand, DTC has more widely acceptable cytologic criteria, so we selected DTC as the control group/baseline reference group.¹³ Our case selection includes confirmed cases of ATC, either histologically or with definitive cytology diagnosis, which may raise a possibility of selection bias, but the rationale for this approach is to identify the cytologic features of ATC in detail for a definitive and an accurate cytology diagnosis.

In this study, we identified the following statistically significant cytologic features of ATC: moderate to severe pleomorphism, macronucleoli, coarse clumped chromatin material, apoptosis, and necrosis. We conclude that the diagnosis of ATC requires identifying these important cytologic features and then correlating with clinical and imaging features TABLE 5. Diagnosis of ATC should be established in a multidisciplinary team approach.