Introduction

Fine-needle aspiration plays a significant role in the evaluation of solitary thyroid nodules. Fine-needle aspiration cytology (FNAC) is a simple, minimally invasive, inexpensive, and easily performed procedure. It is considered a valuable diagnostic method due to its high specificity and sensitivity. The number of thyroidectomies has gradually decreased with the routine use of FNAC in benign diseases (1). Another role of FNAC is to differentiate patients in regarding the options of surgical and conservative treatment (1,2). The Bethesda system is used for standard reporting of thyroid cytology specimens. The defined criteria primarily aim to distinguish papillary carcinoma (PC), the most common thyroid malignancy, from the other benign diseases. However, while the cytomorphological changes are typical, they are not lesion-specific.

In our study, we aimed to determine which cytologic criteria are more valuable in diagnosing PC by investigating the cytologic features defined by FNAC.

Methods

Patients, who had undergone thyroidectomy operation between January 2013 and December 2014 and had preoperative FNAC, were screened. A total of 444 cases were detected. The postoperative histopathological diagnoses were benign, PC, and papillary microcarcinoma, and follicular neoplasm with indeterminate malignancy potential was found in 303 (68.2%), 62 (13,9%), 39 (8.7%), and 40 (9%) patients, respectively. The cytological samples of the cases were examined. The cases with inadequate cytologic preparation, in which the microscope slides were inaccessible due to consultation and the preoperatively biopsied nodule was suspected not to be the nodule that pathology was identified in the postoperative period, were excluded. Among patients who were postoperatively diagnosed with nodular or adenomatous hyperplasia, 62 cases were randomly sampled and were assigned to benign thyroid hyperplasia (BTH) group. PC and papillary microcarcinomas with or above 0.8 cm diameter were selected among the cases postoperatively diagnosed as malignant. A total of 24 cases were included in this group of PC.

The microscope slides belonging to resection and routine cytologic specimens of all cases were brought from the archive. Initially, the cytology preparations were evaluated by the pathologists. Conventional smear samples were stained using the Papanicolaou technique. Light-microscopic examination was performed again by three pathologists together (H.N.U., K.A., and C.B.I.). The Bethesda system for reporting thyroid cytopathology was taken into consideration in the cytological examination of the fine-needle aspiration specimens. The cytologic evaluation criteria were reassessed and graded with the joint decision of pathologists. Cytomorphological criteria, such as cellularity level and presence of colloid were rated as “low” or “high”; honeycomb appearance, lymphocytic ground, hemosiderin-containing histiocytes, three-dimensional cell clusters, micro-follicles, squamous metaplastic cells, Hurthle cells, cellular pleomorphism, nuclear atypia, papillary cell clusters with/without fibrovascular cores, swirling cell layers, large nuclei, nuclear overlapping, long nuclear clefting, intranuclear pseudoinclusion, pale nucleus, perimembranous micronucleolus, psammoma body, and multinuclear giant cells were rated as “present” or “absent”. After completion of the cytological examinations of needle aspiration specimens, resection specimens of all cases were examined with the same methodology and the final diagnoses were confirmed.

Statistical Analysis

Frequency differences between the malignancy and BTH groups of cytology criteria were initially analyzed by using chi-square (pearson) test or Fisher’s exact test. Multivariate analysis was performed with the variables, the frequencies of which were found to be different with the probability ratio of p<0.1 between the two groups in this univariate analysis. Linear regression analysis was used for this purpose. In the linear regression analysis performed with the backward method, the variables associated with the diagnosis of malignancy at the level of p<0.05 were considered significant for predicting the diagnosis of PC.

Results

The demographic data of 86 cases included in the study are shown in Table 1.

All cytomorphological criteria were evaluated in PC and BTH. Resection results were found to be compatible with those of fine-needle biopsy aspiration.

While the most common alterations in PC were found to be long nuclear clefting, nuclear overlapping, large nucleus, nuclear pleomorphism, and insufficient colloid; psammoma bodies, Hurthle cells, and lymphocytic background were not observed at all. All cytologic features were determined in benign thyroid hyperplasia, although with variable rates.

The intranuclear pseudoinclusion was determined with the rates of 58.3%, and 1.6% in the PC and BTH, respectively. Interestingly, the feature was common and was detected very easily in the specimens containing pseudoinclusion; in other cases, it was never seen despite prolonged scanning.

Each criterion was compared by univariate analysis. The results obtained are shown in Table 2.

The variables that remained predictive for the diagnosis of cancer as the result of the backward analysis with the linear regression method following the multivariate analysis performed with the variables with p values <0.1 were cellular pleomorphism, nuclear atypia, papillary structures with fibrovascular cores and intranuclear pseudoinclusion. The cytological changes are shown in Figures 1 to 4. In Table 3, the B coefficients and p values of these variables are given.

Discussion

Differentiation between benign and malignant thyroid nodules is important for avoiding unnecessary surgical procedures and complications as well as determining the treatment protocol. Currently, the initial step in approaching the thyroid nodule is indisputably the FNAC method (2-4).

Based on the defined structural and cytomorphological characteristics, the success rate of detection of PC by FNAC is over 95% (2). Various cytomorphological features have been determined for this purpose. These features might not be possible to co-exist in every sample. Moreover, similar cytomorphological features can also be partially present in BTH. In this study, we investigated which feature (s), and their co-existence might be valuable in the diagnosis of carcinoma by comparing the cytology of PC and BTH. We demonstrated the combination of cellular pleomorphism, atypia, and papillae with cores and intranuclear pseudoinclusion could be useful in diagnosing carcinoma.

In papillary thyroid carcinoma, the sizes and shapes of the cells are variable. The cells may be medium-sized or large; their shape may be cuboidal, columnar, polygonal, fusiform, or even “histiocytoid” (5,6). The last described cellular alteration may lead to a false-positive diagnosis due to the confusion created by the histiocyte clusters showing nuclear atypia in cystic nodular hyperplasia (7). Moreover, radioactive iodine treatment also causes similar changes (8). Cellular pleomorphism and atypia, having high diagnostic value, can be easily identified, however, the evaluation might be rather difficult.

It has been asserted that well-formed papillary structures with fibrovascular cores in BTH characterized by florid hyperplasia can frequently be confused with PC (9,10). On the other hand, in a colloid-rich background, the diagnosis of malignancy can come to the forefront only if the short, nonbranching papillae are co-existent with the typical nuclear alterations (11). The nature of the papillary structures and associated changes might be the factors to consider in the cytological evaluation.

Intranuclear cytoplasmic pseudoinclusions are present in 50-100% of PC aspirations (5). In our study, the rate was determined as 58.3% in PC. It is known that they can also be observed in BTH, although rare. It can be suggested that intranuclear cytoplasmic pseudoinclusions accompany malignancies, although it is not a validated feature if present alone. It has been suggested that a nuclear clefting of 20% or more can be diagnostic for thyroid neoplasm in FNAC in cases that this feature is not available (12). The probability of PC could be eliminated when the clefting rate was below 10%, whereas it was emphasized that the probability of malignant lesions could occur when the clefting rate is 10-19%. The results of our study showed that nuclear clefting was valuable in diagnosing PC, whereas it did not take part in the differential diagnosis.

It has been reported that nuclear crowding and overlapping were frequently at striking levels in PC (5). However, the striking point that we identified in this regard was that the criterion was rather nonspecific due to the 75.8% rate that we found in BTH.

Psammoma body can rarely be found in benign thyroid lesions (4,5,13). LiVolsi (13) reported that the psammoma body was below 1% in benign hyperplasia. For PC, the positive predictive value of psammoma body alone was 50%, whereas it was 100% when taken into account with the other cytologic features (5,14). On the other hand, it is also the fact that psammoma bodies may not be present in every PC.               

Study Limitations

In this study, there were various limitations. Firstly, it had the general limitations of the retrospective method. In addition, the bias can occur as a result of the sampling process.

Conclusion

Although every cytological feature is characteristic for PC, none of them alone is specific; they should be evaluated together with all nuclear and structural features during the diagnostic approach. However, when it is not possible to determine all of them in practice, the togetherness of cellular pleomorphism, nuclear atypia, papillary structures with fibrovascular cores and the intranuclear pseudoinclusion can gain importance.

Ethics

Ethics Committee Approval: It was not taken.

Informed Consent: It was not taken.

Peer-review: Externally and internally peer-reviewed.

Authorship Contributions

Concept: H.N.Ü. Design: H.N.Ü., C.B.I. Data Collection or Processing: H.N.Ü., C.B.I., K.A. Analysis or Interpretation: H.N.Ü., K.A., P.F., C.B.I., M.Z.G. Literature Search: C.B.I., H.N.Ü. Writing: C.B.I., H.N.Ü.

Conflict of Interest: No conflict of interest was declared by the authors.

Financial Disclosure: The authors declared that this study received no financial support.