- Email: [email protected]
Pitfalls in cytological diagnosis of autoimmune thyroiditis
Pathology (1999) 31, pp. 1±7
PITFALLS IN CYTOLOGICAL DIAGNOSIS OF AUTOIMMUNE THYROIDITIS M ARIAN P RIYANTHI K UMARASINGHE*
AND
S HARMILA D E S ILVA²
Department of Pathology, Faculty of Medicine, University of Colombo* and National Hospital², Colombo, Sri Lanka
Summary The aims of this study are to document pitfalls in cytologic diagnosis of autoimmune thyroiditis (AT) and highlight possible ways to minimize them. One hundred consecutive thyroid aspirates with features diagnostic or suggestive of AT, performed and reported by the first author, were included in the study. Follow-up was traced and cytologic features responsible for indecisiveness were re-assessed in those reported as suggestive of AT. The features were then correlated with the results of serologic and thyroid function tests and clinical features, and an attempt was made to amend the final diagnosis using an integrated approach. Seventy eight were diagnostic and 22 were suggestive of AT. In the latter 22, features responsible for the indecisiveness were: cytologic atypia, in the form of nuclear enlargement, irregularity and grooves and altered chromatin texture, in 14 (64%); nucleoli with suspicion of a coexisting neoplasm in three (13.6%), two of which showed epithelial preponderance, crowding and discohesion; sparse inflammation in four (18%); a predominant lymphoid population without epithelial cells resembling a reactive lymphnode in one (4.5%); coexisting toxic features in two (9%); and scanty smears in one (4.5%). Eighteen of the 22 suspected of AT had follow-up. Six had been assessed histologically; three with features suspicious of a neoplasm were diagnosed respectively as a papillary carcinoma (PC), Hurthle cell carcinoma (HCC) and a multinodular goitre (MNG) with degenerate changes. The other three were confirmed as AT; one with cytologic atypia, one with sparse inflammation and the third as cytologically resembling a reactive lymphnode. In ten of the remaining 12, the final diagnosis could be revised following an integrated approach with possible reduction of the indecisiveness. Potential pitfalls are: cytologic atypia occurring in AT; abundance or scarcity of background inflammation; low cell yield; and co-existing toxicity and malignancies. Epithelial preponderance over inflammation, nuclear crowding, severe atypia and cell discohesion should raise the possibility of a neoplasm in spite of other features of AT. Awareness of possible pitfalls and adopting an integrated approach, especially in difficult situations, will minimize pitfalls. Key words: Autoimmune thyroiditis, fine needle aspiration biopsy, indecisiveness, integrated approach. Abbreviations: AT, autoimmune thyroiditis; DQ, Diff Quick stain; HCC, Hurthle cell carcinoma; MNG, multinodular goitre; PC, papillary carcinoma; S, suspicious. Accepted 7 September 1998
INTRODUCTION The value of Fine needle aspiration biopsy (FNAB) in the assessment of autoimmune thyroiditis (AT) is widely acknowledged. The term AT embraces the spectrum of a specific inflammatory process occurring as a result of form ation of auto-antibodies indicated by the presence of raised antithyroglobulin and anti-microsomal thyroid antibodies in the serum. Lymphocytic and Hashimoto’s thyroiditis are various stages of the disease within the spectrum of AT. AT m ay show a different cytologic pattern depending on the stage of the disease and age of the patient. A lymphocytic thyroiditis pattern is commonly seen in young patients and the Hashimoto’s pattern occurs in older patients. Cytologic diagnosis of AT may prevent an unnecessary thyroidectomy. There are however many pitfalls in the cytologic assessment leading to diagnostic difficulties, although a firm diagnosis is made on a combination of features rather than on one single feature. The aims of this study are to document potential pitfalls responsible for diagnostic difficulties in the cyto-diagnosis of thyroiditis and to suggest possible ways of minimizing them.
MATERIALS AND METHODS For the year 1995 to 1996, a retrospective analysis was done on 100 consecutive thyroid aspirates with diagnostic or suggestive features of AT. All FNABs were done by the first author using the classical aspiration or the non-aspiration needle jab technique 1 on two or more sites depending on the nature and the size of the thyroid enlargement. Sm ears were wet fixed in alcohol and stained with H&E or air dried and stained with Romanowsky stains. The terminology used for reporting of thyroid aspirations in our setting includes 5 categories: unsatisfactory; benign/no malignant cells; atypical; suspicious including follicular proliferations (S); and malignant. This terminology has recently been recommended by the Papanicolaou Society.2 A firm diagnosis of thyroiditis falls into the broad category of benign lesions, while som e of those suspected of AT may fall into atypical, S or inadequate categories. The features considered mandatory for a firm diagnosis of AT, as described by Kini3 and Orell et al.,4 were the presence of a mixture of sm all and large lymphocytes with or without plasma cells in a colloid deficient background and a variable number of follicular epithelial cells without a significant follicular pattern. Hurthle cells when present along with above criteria, were considered as an important feature, but the presence was not mandatory as they may not be seen in som e types of thyroiditis such as florid lymphocytic thyroiditis. Macrophages and multinucleated cells were additional but inconsistent features. The sm ears were regarded as nondiagnostic but suggestive of AT when one or more features were absent or
0031±3025/99/010001± 07 1999 Royal College of Pathologists of Australasia
2
Pathology (1999), 31, February
KUM ARASINGHE AND DE SILVA
scantily represented or occurred in the presence of unusual features. The sm ears considered as having atypical features were those with epithelial cells showing: irregular nuclear contours; nucleoli; grooves and inclusions of variable quantity and quality; cell discohesion; altered chromatin texture; and nuclear enlargement. W hen one or more of the above atypical features were prominent or wide spread with a suspicion of a neoplasm, the smears were categorised as S. Variation of the nuclear size alone, a well recognised feature of Hurthle cells, was not considered as atypia. The initial cytologic assessm ent of the aspirates was done and a report issued without the knowledge of results of biochemical and serological tests. Clinical presentation was known as all aspirates were performed by the first author. Subsequently, results of serological (anti-thyroglobulin and anti-microsomal antibody) and biochemical (total T3, T4, free thyroxine and ultra sensitive TSH) tests and histological follow-up were traced for those cytologically suspected as thyroiditis (non diagnostic cases) and correlated with the cytologic features. The initial cytologic diagnosis of those lesions which were not assessed histologically was revised, following correlation with the results of the above tests. TABLE 1
RESULTS Seventy eight aspirates were diagnostic and 22 were suggestive of AT. Classical Hurthle cells were seen in 37% (33 diagnostic cases and four suggestive cases). Follow-up was available for 18 cases; six had histologic follow-up with or without results of laboratory tests, and in the remaining 12, results of serologic and/or biochemical tests were available, the final diagnosis being am ended in ten of these 12 cases following the integrated approach. Table 1 shows the features responsible for indecisiveness in the 22 suggestive cases along with the histologic or am ended diagnosis made in a total of 16. Table 2 shows the correlation of cytologic diagnosis with clinical presentation and laboratory results in the six histologically assessed cases. Correlation of cytologic features with clinical presentation and results of serologic and biochemical tests,
Problematic cytologic feature(s) with histologic or amended diagnosis Histologic or amended diagnosis
Problematic cytologic feature(s) Cytologic atypia (n = 14) (1) Nuclear changes resembling PC and with epithelial preponderance and crowding (S) (2) Discohesive Hurthle cells with nuclear enlargement and irregularity and disproportionately low inflammation (S) (3) Marked nuclear enlargement with irregularity and nucleoli (S) (4) Nuclear enlargement with irregularity and nucleoli (5) Nuclear grooves and occasional inclusions
1
PC
1
HCC
1 5 6
MNG AT (3 of 5) AT (3 of 6)
Co-existing toxic features (n = 2)
2
AT + Toxicity
Degree of inflammation (n = 5) (1) Sparse inflammation (2) Exclusive lymphoid cells without epithelial cells
4 1
AT (3 of 4) AT (nodular)
1
AT
22
16
Overall scanty smears (n = 1) Total S = Suspicious of a neoplasm.
TABLE 2 Correlation of cytologic features with clinical presentation and laboratory results in the six histologically assessed cases Clinical features and the age (years)
Problematic cytologic feature(s)
Histologic diagnosis
Serologic results
Functional status
Male, solitary nodule (42) Female, multi-nodular (30) Female, multi-nodular (55) Female, multi-nodular (44) Female, solitary nodule (42) Female, multi-nodular (56)
Atypia, suspicious Atypia, suspicious Atypia, suspicious Atypia No epithelial cells Sparse inflammation
PC HCC MNG AT AT AT
Not done Positive Negative Positive Positive Positive
Not done Euthyroid Not done Euthyroid Euthyroid Hypothyroid
TABLE 3
Correlation of cytologic features with clinical presentation and serologic and functional status of those which were not histologically assessed
Clinical features Female, multi-nodular Female, multi-nodular Female, multi-nodular Female, multi-nodular Male, diffuse, fibrous Male, diffuse Male, nodular Female, multi-nodular
Number of cases 1 1 2 3 1 1 1 2
Problematic cytological feature(s)
Serological results
Atypia Atypia Atypia Atypia Inadequate Toxicity Toxicity Sparse inflammation
Positive Positive Negative Positive Positive Positive Positive Positive
Functional status
Am ended diagnosis
Euthyroid Not done Not done Hypothyroid Hypothyroid Hyperthyroid Hyperthyroid Hypothyroid
AT AT Not done AT AT AT + Toxicity AT + Toxicity AT
CYTOLO GICAL DIAGNO SIS OF AUTOIM M UNE THYROIDITIS
and the amended diagnosis for those which were not histologically assessed (n = 12), is given in Table 3. Smears from five of the ten (50%) cases in which the diagnosis of AT was confirmed histologically or following integrated approach, showed overrunning of epithelial cells by lymphocytes (Fig. 1); a feature which could not be appreciated in four (40%) aspirates with inadequate, overall cellularity or one component cell. Lymphocytic invasion of epithelial cells was also appreciated in one of the two cases with co-existing toxic features. Lymphocytes were intimately associated with malignant epithelial cells in the
3
smears of papillary carcinoma (PC). Problematic cytologic features with the follow-up are discussed in detail below.
Cytological atypia including those suspicious of a neoplasm (n = 14) The smears of the PC showed cellular, branching sheets and three dimensional clusters of crowded cells with sharp borders in a colloid deficient, blood stained background of abundant lymphoid cells (Fig. 2). M any cells were atypical and showed overlapping nuclei with grooves (Fig. 3).
Fig. 1 Classic appearance of AT with numerous lymphocytes overrunning the epithelial cells (H& E, original magnification 3 100).
Fig. 2 Sm ears of PC with surrounding AT, showing epithelial preponderance in a lymphoid background (DQ, original magnification 3 40).
4
Pathology (1999), 31, February
KUM ARASINGHE AND DE SILVA
The third cytologically suspicious case showed clusters of epithelial cells having enlarged, irregular nuclei with nucleoli and moderate to abundant somewhat degenerate cytoplasm (Fig. 6) together with a lymphocytic infiltrate. Lymphocytic invasion of the follicular epithelial cells was absent. This lesion was confirmed to be a multinodular goitre with degenerate changes histologically. The sections showed somewhat irregular enlarged nuclei, nucleoli and abundant foamy cytoplasm of follicular epithelial cells in some areas representing the changes observed in the smears. A non-specific lymphoid infiltrate was noted in relation to degenerate and fibrous areas. One case with atypia was confirmed as AT with atrophic follicular epithelial cells and Hurthle cell change on histology. Fig. 3 Epithelial cells of the PC showing nuclear enlargement, grooves and crowding (DQ, original magnification 3 400).
Intra-nuclear inclusions were sparse and ill defined. Epithelial preponderance over inflammation, together with the nuclear features observed, raised the suspicion of a neoplasm. Histologic assessment showed a PC in a background of severe lymphocytic thyroiditis. The aspirates of the Hurthle cell carcinoma (HCC) showed m any loose, round to oval epithelial cells with moderate to abundant granular cytoplasm and eccentrically placed, enlarged nuclei with coarse chromatin in a colloid deficient background containing lymphocytes (Fig. 4). Some cells were binucleated. No conspicuous macronucleoli were present in the smears stained with H & E. Cyto-nuclear abnormalities were suspicious of a Hurthle cell neoplasm, however a diagnosis of AT was considered due to the presence of lymphoid cells together with Hurthle cells in smears. Serologic tests performed subsequently showed high antibody levels. Histologic examination showed a HCC (Fig. 5) with a surrounding AT.
Toxic features (n = 2) The aspirates showed epithelial cells with anisonucleosis and fire flares in addition to the combination of cytological features suggestive of AT. The background lymphocytic infiltrate was abundant with invasion of follicular cells in one case. Hurthle cells were not seen in either. One patient on presentation was clinically toxic. Subsequently, both patients were found to have biochemically proven hyperthyroidism.
Degree of inflammation (n = 5) Three of four cases in which the aspirates showed insufficient inflammation for a firm diagnosis of AT were available for follow-up. Serology was positive in all three, while one was also confirmed as AT of Hashimoto’s type with fibrosis on histology. The aspirates of the latter case were only moderately cellular and showed a few follicular epithelial cells in small monolayered sheets, a few Hurthle cells exhibiting variation of nuclear size without atypia and
Fig. 4 Sm ears of HCC with discohesive, atypical Hurthle cells with nuclear enlargement, hyperchromasia and mild irregularity in a background of sparse lymphocytes (DQ, original magnification 3 100).
CYTOLO GICAL DIAGNO SIS OF AUTOIM M UNE THYROIDITIS
5
Fig. 5 Histologic sections of HCC showing Hurthle cells with macronucleoli (H& E, original magnification 3 100).
a scattering of a few lymphocytes in a blood stained, colloid deficient background. The appearance was highly suggestive of AT, probably of Hashimoto’s type. One patient presented with a solitary nodule in the area of the right upper pole of the thyroid and the aspirates showed severe inflammation with numerous large and sm all lymphoid cells including activated forms, follicular centre cells and histiocytes including tangible body m acrophages without follicular epithelial cells, Hurthle cells or colloid (Fig. 7). This lesion resembled a reactive lymph node
cytologically but was confirmed histologically to be a nodular form of AT of florid lymphocytic type.
Inadequate cell yield (n = 1) Aspirates of one patient who presented with mild, firm, diffuse enlargem ent of the thyroid, were sparse in spite of several repeat aspirations. This patient was clinically hypothyroid and was subsequently found to have very high antibody levels.
Fig. 6 Sm ears of the M NG showing nuclear atypia in the form of nuclear enlargement and irregularity, and nucleoli (DQ, original magnification 3400).
6
Pathology (1999), 31, February
KUM ARASINGHE AND DE SILVA
Fig. 7 Sm ears of nodular form of AT resembling a reactive lymphnode showing an exclusive population of lymphoid cells with histiocytes (H&E, original magnification 3 100).
In retrospect, when testing was performed in the 12 followup cases where the cytological features were assessed together with biochemical and serological results and the clinical presentation, the diagnosis could be revised in ten (83%) cases, thus m inim izing the indecisiveness (Table 3). In six cases histologic assessment had been done already, in three (50% ) due to high index of suspicion of a neoplasm.
DISCUSSION Although most aspirates of AT show a characteristic combination of cytologic features, those with unusual features may be categorised as atypical or even S, as shown in our results. W hen atypical features are widespread and excessive in spite of strong evidence of AT, histologic assessment should be done as malignancies are well known to occur in a background, of AT. 3,5,6 This was further confirmed as two of the three cases with suspicious cytological features in this series were confirmed to be malignancies occurring in a setting of thyroiditis. There were no unexpected neoplasms among the six which were histologically assessed. Although the aspirates of the PC co-exiting with AT showed highly suspicious features including epithelial preponderance and widespread nuclear grooves, a firm diagnosis was not m ade as it is well known that occasional nuclear grooves can be seen in AT. 9 Absence of convincing intra-nuclear inclusions made the features short of minimal criteria to diagnose a PC. PC is known to occur in a setting of AT, at times. Prom inent Hurthle cell change is a classic feature in Hashimoto’s type of AT, particularly in older patients where, as in younger age groups, the comm on pattern of AT is that of a lymphocytic thyroiditis. Aspirates of the lesion subsequently confirmed to be HCC showed
abundant Hurthle cells. Therefore the differential diagnosis at the initial assessment included Hashimoto’s thyroiditis and a Hurthle cell neoplasm. However suspicion of a neoplasm was high due to the presence of prominent atypia of Hurthle cells in aspirates of a solitary nodule in a young patient. Detailed attention to cyto-nuclear features, such as the size and shape of the cells and nuclei, cell discohesion, location of the nuclei within the cell, nuclear contours and chromatin distribution, and attention to the age and clinical presentation, may avoid a m isdiagnosis of HCC as Hashimoto’s type of AT. Classical m acronucleoli were not seen in smears in contrast to their conspicuous presence in histologic sections in the HCC. Fixation of sm ears with alcohol without a preservative is known to result in shrinkage and dark staining of the nuclei and in non-visualization of nucleoli.3 This may explain the absence of macronuceoli in the smears of the HCC, as they were wet fixed in 95% alcohol. One major problem in the assessment of thyroid cytology is the relative specificity of criteria such as high cellularity, cell discohesion, nuclear inclusions, grooves, nuclear enlargem ent and irregularity, chromatin texture and nucleoli in the diagnosis of neoplastic versus non-neoplastic lesions. W hen such features are widespread, the lesions are regarded as S. A comm on feature favouring neoplasia in both our cases was epithelial cell preponderance. Other features observed in the neoplasms, such as the presence of nuclear crowding, cell discohesion, nuclear grooves and atypical Hurthle cells, varied according to the histologic subtype of the neoplasm. The suspicion was further supported by the clinical presentation, sex and the age of the two patients (Table 2). Degenerate and atrophic changes occurring in the follicular epithelial cells and an associated non-specific lymphoid infiltrate seen in a multinodular goitre (M NG) can be misleading, as observed in our case (Fig. 6). Degenerate
CYTOLO GICAL DIAGNO SIS OF AUTOIM M UNE THYROIDITIS
changes of the follicular epithelial cells in terms of cytonuclear enlargement, nucleoli and cytoplasmic abundance, may be interpreted as atypia or mimic Hurthle cell change. These changes when together with the presence of lymphocytes raised the possibility of AT in the case illustrated (Fig. 6). Thyroid antibody studies perform ed subsequently were negative. It is significant to note that a prominent lymphocyte component seen in a majority of aspirates of AT was not a feature in the aspirates of the lesion subsequently confirmed to be a M NG. The importance of epithelial to lymphocytic ratio 7 and invasion of follicular epithelial cells by lymphocytes8 in the diagnosis of AT have been observed by others but not adequately stressed. Overrunning and invasion of follicular epithelial cells by lymphocytes appears to be an important feature favouring AT against a non-specific lymphoid infiltrate. This feature probably represents invasion of epithelial cells by lymphocytes as seen in histologic sections of some cases of AT. This observation needs to be investigated further as the num bers in this series are not enough to arrive at statistically valid conclusions. However we reiterate that the mere presence of lymphocytes is not diagnostic of AT. Final diagnosis of five of the seven cases with cytologic atypia (not assessed histologically) could be amended as AT on the integrated approach (Table 3). Cytologic atypia in these cases m ay be due to degenerate and atrophic changes occurring in the follicular epithelial cells or Hurthle cells, as was proved in the case histologically confirmed as AT. Abundance or scarcity of lymphoid cells in an aspirate of AT m ay depend on the areas sampled, the stage of the disease and the degree of fibrosis of the gland. M any punctures may partly m inim ize this problem. The case in which the overall cellularity was inadequate in spite of repeated sampling can only be explained by the possibility of the fibrosing phase of autoimm une thyroiditis. This patient had biochemical and serologic evidence of AT and severe hypothyroidism. Distinction between AT co-existing with toxicity (ªHashitoxicosisº) and pure Grave’s disease m ay not be possible cytologically.7,10 It is believed that Grave’s disease and AT are the two major manifestations of AT disease.11 Therefore it is not surprising that features overlap and one may evolve into the other. Follow-up of the two patients in this series may, if the patients develop hypothyroidism, prove that these are actual examples of the toxic phase of AT. In conclusion, FN AB is probably the best alternative to a surgical biopsy to arrive at a morphologic diagnosis of AT, as surgery may not be the treatment of choice. The degree of background inflammation, overall cellularity, cytologic atypia in the form of nuclear enlargement and irregularity,
7
grooves, alteration of the chromatin pattern and nucleoli and co-existing toxicity and malignancies, are potential pitfalls. Epithelial cell preponderance over inflamm ation with various degrees of atypia favour a neoplasm. Lymphocyte preponderance favours AT. The degree and type of cytonuclear atypia vary according to the histologic subtype of the neoplasm and atypia should be assessed bearing this fact in mind. The degree of inflammation may vary according to the type and stage of evolution of AT and a lymphocytic infiltration alone can occur in other thyroid pathologies. Thus, by being aware of the overlapping of cytological features with other thyroid pathologies, and by adhering to an integrated approach with clinicopathologic correlation, some of the potential pitfalls and uncertainties can be minimized and the value of FN AB in the diagnosis of AT can be enhanced. A CKNOWLEDGEMENTS The authors wish to thank staff at Asiri Hospital, Colombo, Sri Lanka, especially Ms Priyanthi Hewage, for assistance in collecting the laboratory results. M s Jayanthi for photographs. Address for correspondence: M. P. K., 37/1 Diyawannawa R oad, Etul Kotte, Sri Lanka. Email ± [email protected]
References 1. Kumarasinghe MP, Sheriffdeen AH. Fine needle sampling without aspiration. Pathology 1995; 27: 330 ± 2. 2. Guidelines of the Papanicolaou Society of Cytopathology for FineNeedle Aspiration procedure and reporting by The Papanicolaou Society of Cytopathology Task Force on Standards of Practice. Diagn Cytopathol 1997; 17(4): 239 ±47. 3. Kini SR. Guides to clinical aspiration biopsy 2nd ed. New York: Igaku-Shoin, 1996; 25, 188, 197, 292, 297. 4. Orell SR, Sterrett GF, Whitaker D, Walters M NI. M anual and atlas of aspiration cytology. 2nd ed. New York: Churchill Livingstone, 1986; 71± 2. 5. Ravinsky E, Safneck JR. Differentiation of Hashimoto’s thyroiditis from thyroid neoplasms in fine needle aspirates. Acta Cytol 1988; 32(6): 854 ±61. 6. Carson HJ, Castelli M J, Gattuso P. Incidence of neoplasia in Hashimoto’s thyroiditis: a fine needle aspiration study. Diagn Cytopathol 1995; 14(1): 38 ± 42. 7. Jayaram G, M arwaha RK, Gupta RK, Sharma SK. Cytomorphologic aspects of thyroiditis. A study of 51 cases with functional, immunological and ultrasonographic data. Acta Cytol 1986; 31(6): 687± 93. 8. Friedman, Shimaoka K, Rao U, Tsukada Y, Gavigan, Tamura K. Diagnosis of chronic lymphocytic thyroiditis (nodular presentation) by needle aspiration. Acta Cytol 1981; 25(5): 513 ± 22. 9. Rupp M , Ehaya H. Nuclear grooves in the aspiration cytology of papillary carcinoma of the thyroid. Acta Cytol 1989; 33(1): 21± 6. 10. Jayaram G, Singh B, Marwaha RK. Grave’s disease. Appearance in cytologic sm ears from fine needle aspirates of the thyroid gland. Acta Cytol 1989; 33(1): 36 ±40. 11. Sm all Rosai J. Ackerman’s surgical pathology, 8th ed. St. Louis: CV Mosby & Co, 1996; 499.
PITFALLS IN CYTOLOGICAL DIAGNOSIS OF AUTOIMMUNE THYROIDITIS M ARIAN P RIYANTHI K UMARASINGHE*
AND
S HARMILA D E S ILVA²
Department of Pathology, Faculty of Medicine, University of Colombo* and National Hospital², Colombo, Sri Lanka
Summary The aims of this study are to document pitfalls in cytologic diagnosis of autoimmune thyroiditis (AT) and highlight possible ways to minimize them. One hundred consecutive thyroid aspirates with features diagnostic or suggestive of AT, performed and reported by the first author, were included in the study. Follow-up was traced and cytologic features responsible for indecisiveness were re-assessed in those reported as suggestive of AT. The features were then correlated with the results of serologic and thyroid function tests and clinical features, and an attempt was made to amend the final diagnosis using an integrated approach. Seventy eight were diagnostic and 22 were suggestive of AT. In the latter 22, features responsible for the indecisiveness were: cytologic atypia, in the form of nuclear enlargement, irregularity and grooves and altered chromatin texture, in 14 (64%); nucleoli with suspicion of a coexisting neoplasm in three (13.6%), two of which showed epithelial preponderance, crowding and discohesion; sparse inflammation in four (18%); a predominant lymphoid population without epithelial cells resembling a reactive lymphnode in one (4.5%); coexisting toxic features in two (9%); and scanty smears in one (4.5%). Eighteen of the 22 suspected of AT had follow-up. Six had been assessed histologically; three with features suspicious of a neoplasm were diagnosed respectively as a papillary carcinoma (PC), Hurthle cell carcinoma (HCC) and a multinodular goitre (MNG) with degenerate changes. The other three were confirmed as AT; one with cytologic atypia, one with sparse inflammation and the third as cytologically resembling a reactive lymphnode. In ten of the remaining 12, the final diagnosis could be revised following an integrated approach with possible reduction of the indecisiveness. Potential pitfalls are: cytologic atypia occurring in AT; abundance or scarcity of background inflammation; low cell yield; and co-existing toxicity and malignancies. Epithelial preponderance over inflammation, nuclear crowding, severe atypia and cell discohesion should raise the possibility of a neoplasm in spite of other features of AT. Awareness of possible pitfalls and adopting an integrated approach, especially in difficult situations, will minimize pitfalls. Key words: Autoimmune thyroiditis, fine needle aspiration biopsy, indecisiveness, integrated approach. Abbreviations: AT, autoimmune thyroiditis; DQ, Diff Quick stain; HCC, Hurthle cell carcinoma; MNG, multinodular goitre; PC, papillary carcinoma; S, suspicious. Accepted 7 September 1998
INTRODUCTION The value of Fine needle aspiration biopsy (FNAB) in the assessment of autoimmune thyroiditis (AT) is widely acknowledged. The term AT embraces the spectrum of a specific inflammatory process occurring as a result of form ation of auto-antibodies indicated by the presence of raised antithyroglobulin and anti-microsomal thyroid antibodies in the serum. Lymphocytic and Hashimoto’s thyroiditis are various stages of the disease within the spectrum of AT. AT m ay show a different cytologic pattern depending on the stage of the disease and age of the patient. A lymphocytic thyroiditis pattern is commonly seen in young patients and the Hashimoto’s pattern occurs in older patients. Cytologic diagnosis of AT may prevent an unnecessary thyroidectomy. There are however many pitfalls in the cytologic assessment leading to diagnostic difficulties, although a firm diagnosis is made on a combination of features rather than on one single feature. The aims of this study are to document potential pitfalls responsible for diagnostic difficulties in the cyto-diagnosis of thyroiditis and to suggest possible ways of minimizing them.
MATERIALS AND METHODS For the year 1995 to 1996, a retrospective analysis was done on 100 consecutive thyroid aspirates with diagnostic or suggestive features of AT. All FNABs were done by the first author using the classical aspiration or the non-aspiration needle jab technique 1 on two or more sites depending on the nature and the size of the thyroid enlargement. Sm ears were wet fixed in alcohol and stained with H&E or air dried and stained with Romanowsky stains. The terminology used for reporting of thyroid aspirations in our setting includes 5 categories: unsatisfactory; benign/no malignant cells; atypical; suspicious including follicular proliferations (S); and malignant. This terminology has recently been recommended by the Papanicolaou Society.2 A firm diagnosis of thyroiditis falls into the broad category of benign lesions, while som e of those suspected of AT may fall into atypical, S or inadequate categories. The features considered mandatory for a firm diagnosis of AT, as described by Kini3 and Orell et al.,4 were the presence of a mixture of sm all and large lymphocytes with or without plasma cells in a colloid deficient background and a variable number of follicular epithelial cells without a significant follicular pattern. Hurthle cells when present along with above criteria, were considered as an important feature, but the presence was not mandatory as they may not be seen in som e types of thyroiditis such as florid lymphocytic thyroiditis. Macrophages and multinucleated cells were additional but inconsistent features. The sm ears were regarded as nondiagnostic but suggestive of AT when one or more features were absent or
0031±3025/99/010001± 07 1999 Royal College of Pathologists of Australasia
2
Pathology (1999), 31, February
KUM ARASINGHE AND DE SILVA
scantily represented or occurred in the presence of unusual features. The sm ears considered as having atypical features were those with epithelial cells showing: irregular nuclear contours; nucleoli; grooves and inclusions of variable quantity and quality; cell discohesion; altered chromatin texture; and nuclear enlargement. W hen one or more of the above atypical features were prominent or wide spread with a suspicion of a neoplasm, the smears were categorised as S. Variation of the nuclear size alone, a well recognised feature of Hurthle cells, was not considered as atypia. The initial cytologic assessm ent of the aspirates was done and a report issued without the knowledge of results of biochemical and serological tests. Clinical presentation was known as all aspirates were performed by the first author. Subsequently, results of serological (anti-thyroglobulin and anti-microsomal antibody) and biochemical (total T3, T4, free thyroxine and ultra sensitive TSH) tests and histological follow-up were traced for those cytologically suspected as thyroiditis (non diagnostic cases) and correlated with the cytologic features. The initial cytologic diagnosis of those lesions which were not assessed histologically was revised, following correlation with the results of the above tests. TABLE 1
RESULTS Seventy eight aspirates were diagnostic and 22 were suggestive of AT. Classical Hurthle cells were seen in 37% (33 diagnostic cases and four suggestive cases). Follow-up was available for 18 cases; six had histologic follow-up with or without results of laboratory tests, and in the remaining 12, results of serologic and/or biochemical tests were available, the final diagnosis being am ended in ten of these 12 cases following the integrated approach. Table 1 shows the features responsible for indecisiveness in the 22 suggestive cases along with the histologic or am ended diagnosis made in a total of 16. Table 2 shows the correlation of cytologic diagnosis with clinical presentation and laboratory results in the six histologically assessed cases. Correlation of cytologic features with clinical presentation and results of serologic and biochemical tests,
Problematic cytologic feature(s) with histologic or amended diagnosis Histologic or amended diagnosis
Problematic cytologic feature(s) Cytologic atypia (n = 14) (1) Nuclear changes resembling PC and with epithelial preponderance and crowding (S) (2) Discohesive Hurthle cells with nuclear enlargement and irregularity and disproportionately low inflammation (S) (3) Marked nuclear enlargement with irregularity and nucleoli (S) (4) Nuclear enlargement with irregularity and nucleoli (5) Nuclear grooves and occasional inclusions
1
PC
1
HCC
1 5 6
MNG AT (3 of 5) AT (3 of 6)
Co-existing toxic features (n = 2)
2
AT + Toxicity
Degree of inflammation (n = 5) (1) Sparse inflammation (2) Exclusive lymphoid cells without epithelial cells
4 1
AT (3 of 4) AT (nodular)
1
AT
22
16
Overall scanty smears (n = 1) Total S = Suspicious of a neoplasm.
TABLE 2 Correlation of cytologic features with clinical presentation and laboratory results in the six histologically assessed cases Clinical features and the age (years)
Problematic cytologic feature(s)
Histologic diagnosis
Serologic results
Functional status
Male, solitary nodule (42) Female, multi-nodular (30) Female, multi-nodular (55) Female, multi-nodular (44) Female, solitary nodule (42) Female, multi-nodular (56)
Atypia, suspicious Atypia, suspicious Atypia, suspicious Atypia No epithelial cells Sparse inflammation
PC HCC MNG AT AT AT
Not done Positive Negative Positive Positive Positive
Not done Euthyroid Not done Euthyroid Euthyroid Hypothyroid
TABLE 3
Correlation of cytologic features with clinical presentation and serologic and functional status of those which were not histologically assessed
Clinical features Female, multi-nodular Female, multi-nodular Female, multi-nodular Female, multi-nodular Male, diffuse, fibrous Male, diffuse Male, nodular Female, multi-nodular
Number of cases 1 1 2 3 1 1 1 2
Problematic cytological feature(s)
Serological results
Atypia Atypia Atypia Atypia Inadequate Toxicity Toxicity Sparse inflammation
Positive Positive Negative Positive Positive Positive Positive Positive
Functional status
Am ended diagnosis
Euthyroid Not done Not done Hypothyroid Hypothyroid Hyperthyroid Hyperthyroid Hypothyroid
AT AT Not done AT AT AT + Toxicity AT + Toxicity AT
CYTOLO GICAL DIAGNO SIS OF AUTOIM M UNE THYROIDITIS
and the amended diagnosis for those which were not histologically assessed (n = 12), is given in Table 3. Smears from five of the ten (50%) cases in which the diagnosis of AT was confirmed histologically or following integrated approach, showed overrunning of epithelial cells by lymphocytes (Fig. 1); a feature which could not be appreciated in four (40%) aspirates with inadequate, overall cellularity or one component cell. Lymphocytic invasion of epithelial cells was also appreciated in one of the two cases with co-existing toxic features. Lymphocytes were intimately associated with malignant epithelial cells in the
3
smears of papillary carcinoma (PC). Problematic cytologic features with the follow-up are discussed in detail below.
Cytological atypia including those suspicious of a neoplasm (n = 14) The smears of the PC showed cellular, branching sheets and three dimensional clusters of crowded cells with sharp borders in a colloid deficient, blood stained background of abundant lymphoid cells (Fig. 2). M any cells were atypical and showed overlapping nuclei with grooves (Fig. 3).
Fig. 1 Classic appearance of AT with numerous lymphocytes overrunning the epithelial cells (H& E, original magnification 3 100).
Fig. 2 Sm ears of PC with surrounding AT, showing epithelial preponderance in a lymphoid background (DQ, original magnification 3 40).
4
Pathology (1999), 31, February
KUM ARASINGHE AND DE SILVA
The third cytologically suspicious case showed clusters of epithelial cells having enlarged, irregular nuclei with nucleoli and moderate to abundant somewhat degenerate cytoplasm (Fig. 6) together with a lymphocytic infiltrate. Lymphocytic invasion of the follicular epithelial cells was absent. This lesion was confirmed to be a multinodular goitre with degenerate changes histologically. The sections showed somewhat irregular enlarged nuclei, nucleoli and abundant foamy cytoplasm of follicular epithelial cells in some areas representing the changes observed in the smears. A non-specific lymphoid infiltrate was noted in relation to degenerate and fibrous areas. One case with atypia was confirmed as AT with atrophic follicular epithelial cells and Hurthle cell change on histology. Fig. 3 Epithelial cells of the PC showing nuclear enlargement, grooves and crowding (DQ, original magnification 3 400).
Intra-nuclear inclusions were sparse and ill defined. Epithelial preponderance over inflammation, together with the nuclear features observed, raised the suspicion of a neoplasm. Histologic assessment showed a PC in a background of severe lymphocytic thyroiditis. The aspirates of the Hurthle cell carcinoma (HCC) showed m any loose, round to oval epithelial cells with moderate to abundant granular cytoplasm and eccentrically placed, enlarged nuclei with coarse chromatin in a colloid deficient background containing lymphocytes (Fig. 4). Some cells were binucleated. No conspicuous macronucleoli were present in the smears stained with H & E. Cyto-nuclear abnormalities were suspicious of a Hurthle cell neoplasm, however a diagnosis of AT was considered due to the presence of lymphoid cells together with Hurthle cells in smears. Serologic tests performed subsequently showed high antibody levels. Histologic examination showed a HCC (Fig. 5) with a surrounding AT.
Toxic features (n = 2) The aspirates showed epithelial cells with anisonucleosis and fire flares in addition to the combination of cytological features suggestive of AT. The background lymphocytic infiltrate was abundant with invasion of follicular cells in one case. Hurthle cells were not seen in either. One patient on presentation was clinically toxic. Subsequently, both patients were found to have biochemically proven hyperthyroidism.
Degree of inflammation (n = 5) Three of four cases in which the aspirates showed insufficient inflammation for a firm diagnosis of AT were available for follow-up. Serology was positive in all three, while one was also confirmed as AT of Hashimoto’s type with fibrosis on histology. The aspirates of the latter case were only moderately cellular and showed a few follicular epithelial cells in small monolayered sheets, a few Hurthle cells exhibiting variation of nuclear size without atypia and
Fig. 4 Sm ears of HCC with discohesive, atypical Hurthle cells with nuclear enlargement, hyperchromasia and mild irregularity in a background of sparse lymphocytes (DQ, original magnification 3 100).
CYTOLO GICAL DIAGNO SIS OF AUTOIM M UNE THYROIDITIS
5
Fig. 5 Histologic sections of HCC showing Hurthle cells with macronucleoli (H& E, original magnification 3 100).
a scattering of a few lymphocytes in a blood stained, colloid deficient background. The appearance was highly suggestive of AT, probably of Hashimoto’s type. One patient presented with a solitary nodule in the area of the right upper pole of the thyroid and the aspirates showed severe inflammation with numerous large and sm all lymphoid cells including activated forms, follicular centre cells and histiocytes including tangible body m acrophages without follicular epithelial cells, Hurthle cells or colloid (Fig. 7). This lesion resembled a reactive lymph node
cytologically but was confirmed histologically to be a nodular form of AT of florid lymphocytic type.
Inadequate cell yield (n = 1) Aspirates of one patient who presented with mild, firm, diffuse enlargem ent of the thyroid, were sparse in spite of several repeat aspirations. This patient was clinically hypothyroid and was subsequently found to have very high antibody levels.
Fig. 6 Sm ears of the M NG showing nuclear atypia in the form of nuclear enlargement and irregularity, and nucleoli (DQ, original magnification 3400).
6
Pathology (1999), 31, February
KUM ARASINGHE AND DE SILVA
Fig. 7 Sm ears of nodular form of AT resembling a reactive lymphnode showing an exclusive population of lymphoid cells with histiocytes (H&E, original magnification 3 100).
In retrospect, when testing was performed in the 12 followup cases where the cytological features were assessed together with biochemical and serological results and the clinical presentation, the diagnosis could be revised in ten (83%) cases, thus m inim izing the indecisiveness (Table 3). In six cases histologic assessment had been done already, in three (50% ) due to high index of suspicion of a neoplasm.
DISCUSSION Although most aspirates of AT show a characteristic combination of cytologic features, those with unusual features may be categorised as atypical or even S, as shown in our results. W hen atypical features are widespread and excessive in spite of strong evidence of AT, histologic assessment should be done as malignancies are well known to occur in a background, of AT. 3,5,6 This was further confirmed as two of the three cases with suspicious cytological features in this series were confirmed to be malignancies occurring in a setting of thyroiditis. There were no unexpected neoplasms among the six which were histologically assessed. Although the aspirates of the PC co-exiting with AT showed highly suspicious features including epithelial preponderance and widespread nuclear grooves, a firm diagnosis was not m ade as it is well known that occasional nuclear grooves can be seen in AT. 9 Absence of convincing intra-nuclear inclusions made the features short of minimal criteria to diagnose a PC. PC is known to occur in a setting of AT, at times. Prom inent Hurthle cell change is a classic feature in Hashimoto’s type of AT, particularly in older patients where, as in younger age groups, the comm on pattern of AT is that of a lymphocytic thyroiditis. Aspirates of the lesion subsequently confirmed to be HCC showed
abundant Hurthle cells. Therefore the differential diagnosis at the initial assessment included Hashimoto’s thyroiditis and a Hurthle cell neoplasm. However suspicion of a neoplasm was high due to the presence of prominent atypia of Hurthle cells in aspirates of a solitary nodule in a young patient. Detailed attention to cyto-nuclear features, such as the size and shape of the cells and nuclei, cell discohesion, location of the nuclei within the cell, nuclear contours and chromatin distribution, and attention to the age and clinical presentation, may avoid a m isdiagnosis of HCC as Hashimoto’s type of AT. Classical m acronucleoli were not seen in smears in contrast to their conspicuous presence in histologic sections in the HCC. Fixation of sm ears with alcohol without a preservative is known to result in shrinkage and dark staining of the nuclei and in non-visualization of nucleoli.3 This may explain the absence of macronuceoli in the smears of the HCC, as they were wet fixed in 95% alcohol. One major problem in the assessment of thyroid cytology is the relative specificity of criteria such as high cellularity, cell discohesion, nuclear inclusions, grooves, nuclear enlargem ent and irregularity, chromatin texture and nucleoli in the diagnosis of neoplastic versus non-neoplastic lesions. W hen such features are widespread, the lesions are regarded as S. A comm on feature favouring neoplasia in both our cases was epithelial cell preponderance. Other features observed in the neoplasms, such as the presence of nuclear crowding, cell discohesion, nuclear grooves and atypical Hurthle cells, varied according to the histologic subtype of the neoplasm. The suspicion was further supported by the clinical presentation, sex and the age of the two patients (Table 2). Degenerate and atrophic changes occurring in the follicular epithelial cells and an associated non-specific lymphoid infiltrate seen in a multinodular goitre (M NG) can be misleading, as observed in our case (Fig. 6). Degenerate
CYTOLO GICAL DIAGNO SIS OF AUTOIM M UNE THYROIDITIS
changes of the follicular epithelial cells in terms of cytonuclear enlargement, nucleoli and cytoplasmic abundance, may be interpreted as atypia or mimic Hurthle cell change. These changes when together with the presence of lymphocytes raised the possibility of AT in the case illustrated (Fig. 6). Thyroid antibody studies perform ed subsequently were negative. It is significant to note that a prominent lymphocyte component seen in a majority of aspirates of AT was not a feature in the aspirates of the lesion subsequently confirmed to be a M NG. The importance of epithelial to lymphocytic ratio 7 and invasion of follicular epithelial cells by lymphocytes8 in the diagnosis of AT have been observed by others but not adequately stressed. Overrunning and invasion of follicular epithelial cells by lymphocytes appears to be an important feature favouring AT against a non-specific lymphoid infiltrate. This feature probably represents invasion of epithelial cells by lymphocytes as seen in histologic sections of some cases of AT. This observation needs to be investigated further as the num bers in this series are not enough to arrive at statistically valid conclusions. However we reiterate that the mere presence of lymphocytes is not diagnostic of AT. Final diagnosis of five of the seven cases with cytologic atypia (not assessed histologically) could be amended as AT on the integrated approach (Table 3). Cytologic atypia in these cases m ay be due to degenerate and atrophic changes occurring in the follicular epithelial cells or Hurthle cells, as was proved in the case histologically confirmed as AT. Abundance or scarcity of lymphoid cells in an aspirate of AT m ay depend on the areas sampled, the stage of the disease and the degree of fibrosis of the gland. M any punctures may partly m inim ize this problem. The case in which the overall cellularity was inadequate in spite of repeated sampling can only be explained by the possibility of the fibrosing phase of autoimm une thyroiditis. This patient had biochemical and serologic evidence of AT and severe hypothyroidism. Distinction between AT co-existing with toxicity (ªHashitoxicosisº) and pure Grave’s disease m ay not be possible cytologically.7,10 It is believed that Grave’s disease and AT are the two major manifestations of AT disease.11 Therefore it is not surprising that features overlap and one may evolve into the other. Follow-up of the two patients in this series may, if the patients develop hypothyroidism, prove that these are actual examples of the toxic phase of AT. In conclusion, FN AB is probably the best alternative to a surgical biopsy to arrive at a morphologic diagnosis of AT, as surgery may not be the treatment of choice. The degree of background inflammation, overall cellularity, cytologic atypia in the form of nuclear enlargement and irregularity,
7
grooves, alteration of the chromatin pattern and nucleoli and co-existing toxicity and malignancies, are potential pitfalls. Epithelial cell preponderance over inflamm ation with various degrees of atypia favour a neoplasm. Lymphocyte preponderance favours AT. The degree and type of cytonuclear atypia vary according to the histologic subtype of the neoplasm and atypia should be assessed bearing this fact in mind. The degree of inflammation may vary according to the type and stage of evolution of AT and a lymphocytic infiltration alone can occur in other thyroid pathologies. Thus, by being aware of the overlapping of cytological features with other thyroid pathologies, and by adhering to an integrated approach with clinicopathologic correlation, some of the potential pitfalls and uncertainties can be minimized and the value of FN AB in the diagnosis of AT can be enhanced. A CKNOWLEDGEMENTS The authors wish to thank staff at Asiri Hospital, Colombo, Sri Lanka, especially Ms Priyanthi Hewage, for assistance in collecting the laboratory results. M s Jayanthi for photographs. Address for correspondence: M. P. K., 37/1 Diyawannawa R oad, Etul Kotte, Sri Lanka. Email ± [email protected]
References 1. Kumarasinghe MP, Sheriffdeen AH. Fine needle sampling without aspiration. Pathology 1995; 27: 330 ± 2. 2. Guidelines of the Papanicolaou Society of Cytopathology for FineNeedle Aspiration procedure and reporting by The Papanicolaou Society of Cytopathology Task Force on Standards of Practice. Diagn Cytopathol 1997; 17(4): 239 ±47. 3. Kini SR. Guides to clinical aspiration biopsy 2nd ed. New York: Igaku-Shoin, 1996; 25, 188, 197, 292, 297. 4. Orell SR, Sterrett GF, Whitaker D, Walters M NI. M anual and atlas of aspiration cytology. 2nd ed. New York: Churchill Livingstone, 1986; 71± 2. 5. Ravinsky E, Safneck JR. Differentiation of Hashimoto’s thyroiditis from thyroid neoplasms in fine needle aspirates. Acta Cytol 1988; 32(6): 854 ±61. 6. Carson HJ, Castelli M J, Gattuso P. Incidence of neoplasia in Hashimoto’s thyroiditis: a fine needle aspiration study. Diagn Cytopathol 1995; 14(1): 38 ± 42. 7. Jayaram G, M arwaha RK, Gupta RK, Sharma SK. Cytomorphologic aspects of thyroiditis. A study of 51 cases with functional, immunological and ultrasonographic data. Acta Cytol 1986; 31(6): 687± 93. 8. Friedman, Shimaoka K, Rao U, Tsukada Y, Gavigan, Tamura K. Diagnosis of chronic lymphocytic thyroiditis (nodular presentation) by needle aspiration. Acta Cytol 1981; 25(5): 513 ± 22. 9. Rupp M , Ehaya H. Nuclear grooves in the aspiration cytology of papillary carcinoma of the thyroid. Acta Cytol 1989; 33(1): 21± 6. 10. Jayaram G, Singh B, Marwaha RK. Grave’s disease. Appearance in cytologic sm ears from fine needle aspirates of the thyroid gland. Acta Cytol 1989; 33(1): 36 ±40. 11. Sm all Rosai J. Ackerman’s surgical pathology, 8th ed. St. Louis: CV Mosby & Co, 1996; 499.