Clinical presentation.

Papillary carcinomas are included in relatively fast growing nodules. The period lasting from the detection to the first visit in patients who were aware about a palpable tumor was in the range of 3-24 months with a median of 8 months in our practice. This is more fast than in the case of a benign nodule but more slowly compared with medullary carcinoma.

Palpation. As most malignant diseases, papillary carcinoma is presented generally by a firm or hard nodule. Nodules are not as hard as in the event of an anaplastic carcinoma. They are moveable, but in a less extent than benign nodules. In around 20% of cases we found enlarged, metastatic lymph nodes at the time of the diagnosis of papillary carcinoma. Palsy of the

recurrent nerve can be found in aggressive subtypes and in patients who requested evaluation only years after detection of the nodule. 10% of our patients presented with hoarseness.

Functional state. In contrast with other thyroid carcinomas, patients with papillary cancer are not infrequently subclinically hypothyroid. This is the consequence of the well-know association between lymphocytic thyroiditis and papillary carcinoma. There is a long-lasting and yet undetermined debate which is the cause and which is the consequence. Nevertheless, the concomitant thyroiditis is responsible for the usually mild and subclinical hypothyroidism. The presence of dysfunction is an interesting phenomenon but have practically no relevance in the diagnosis.

Ultrasonography.

Features with significantly increased risk of papillary carcinoma are listed in Table. From a practical point-of-view, the blurred borders of the nodule and the presence of microcalcifications are of the greatest diagnostic power.

Most papillary carcinomas occur in hypoechogenic nodules, it means more than 80% of cases. This proportion is less compared with e.g. medullary cancer or secondary thyroid carcinomas.

Irregular borders of a hypoechogenic nodule is the most specific sign of papillary carcinoma. Three types of irregularities have to be discussed. Ill-defined, blurred border is the most important type. It means that we cannot decide where the nodule ends and where the extranodular part of the thyroid begins. This type of irregularity can be observed in the case of de Quervain's thyroiditis, too. Serious differential diagnostic problems may arise in those infrequent cases where the patient does not present typical clinical signs of subacute thyroiditis. Naturally, in this situation FNAC is mandatory.

The lobulated margins of the nodule is another type of irregularity. The cause for this phenomenon is either the infiltrative nature of the tumor or the proliferation of the connective tissue. In the latter a pseudonodular appearance can be observed similarly to that seen in certain cases of a lymphocytic thyroiditis.

The sharp irregular surface of the tumor is the less sensitive and specific type of the irregularity of the borders.

Microcalcification is the second most characteristic sign of papillary carcinoma. It means that small bright punctate granules can be observed within a hypoechogenic nodule. The differential diagnostic of punctate granules includes the so called comet tail artefact and the intranodular fibrosis. The former granule has a small characteristic tail. In the case of intranodular fibrosis not only hyperechogenic granules but hyperchogenic lines can be seen depending on the angle between the fibrotic bundle and the transducer. Not infrequently a clear distinction between various types of hyperechogenic granules cannot be made.

Coarse calcification occur significantly more frequently in the case of papillary carcinoma then in benign nodules. The practical significance is limited. Nevertheless, in certain cases the coarse calcification may be the only suspicious sign. It was of great help in a case where a small hypoechogenic nodule was found in an almost identically hypoechogenic thyroid.

Cystic degeneration is a very common phenomenon in the event of a nodular goiter irrespectively of the nature of the lesion except for medullary and follicular carcinoma, which only rarely contain cystic areas. On the other hand, papillary carcinomas frequently present cystic areas. The presence of microcalcification within the solid part of a cystic nodule increases the likelihood of malignancy.

Although the presence of a halo sign suggests a significantly decreased risk of malignancy overall, the consequences of a detailed analysis indicate that the situation is not so simple. First, if we do not investigate encapsulated nodules, a high proportion of cancers will be missed. Moreover, our results demonstrate that the absence of a halo sign is not an independent risk factor. The absence of a halo decreases the risk of malignancy only because this sign cannot be demonstrated in the most frequently occurring malignant nodules, in hypoechogenic ones. The demonstration of a halo sign is very difficult or even impossible in most hypoechogenic nodules because the US appearance of the capsule in such lesions is identical with the nodule. This is clearly proved by the fact, that for follicular adenomas, which are encapsulated in 100% of the cases, only 8.8% of the hypoechogenic nodules exhibit a halo sign, whereas this feature can be demonstrated by US in 64.1% of all other types of nodules. It is not surprising therefore, that the presence of a halo sign significantly increases the risk of malignancy in cases of moderately hypoechogenic nodules, while in hyperechogenic nodules the presence or absence of a halo sign does not influence the likelihood of malignancy.

The taller-than-wide sign is frequently mentioned in the literature because it also occurs significantly more frequently among malignant lesions than among benign ones, but proved to be of less practical significance in compared with micocalcification or surface irregularities.

The situation is similar as regards type 3 vascular pattern, i.e. the presence of increased intranodular blood flow. This phenomenon occurs significantly more frequently among papillary carcinoma than among benign lesions but the practical value is limited.

Elastography is nowadays a very popular field in thyroidology. Nevertheless, the results gained by different authors are hard to compare because of technical reasons. Moreover, it is questionable whether in the case of a lesion with a benign appearing elastographic pattern we could avoid FNAC.

We must keep in mind, that the task of ultrasonography is not to diagnose a carcinoma, but the detect a nodule with high enough risk of malignancy to perform aspiration cytology.

Cytological diagnosis

The main features of papillary carcinoma are as follows:

• Lack of colloid
• Cellular pattern
• Papillary fragments and monolayered sheets with disturbed pattern
• Atypical cells with mild to moderate pleomorphism
• Nuclear inclusions, grooves and scratchy chromatin structure.
• Distinct cell borders
• Presence of multinucleated cells

The first 2 features are not specific for papillary carcinoma and may be lacking.

Papillary fragments are 3-dimensional figures with sharp, geometric borders. In contrast with hyperplastic papilla, they lack peripheral vacuolization and projections. There is no tendency of cells at the edge of such clusters to dissociate.

Monolayered sheets are frequently found in PC. They are characterized by nuclear crowding and overlapping and loss of polarity of cells.

Nuclear inclusions are the most important clue in the diagnosis of PC. They re

At low power  The aspirates are usually rich in cells and contain little if any colloid. The tumor cells occur in tumor fragments or singly, in small clusters or in cohesive sheets. The multilayered tumor fragments, the papillae  exhibit a sharp edge, with palisading of the cells at the periphery, and a branching structure is often seen at low power in the microscope. The nuclear details in cells from papillary fronds are difficult to analyze in some cases, because of nuclear crowding. In contrast with hyperplastic benign papillary clusters, malignant papillae lack peripheral vacuolization and projections, and also lack the tendency of peripheral cells to dissociate.

Another type of large cell clusters may be observed less often. These monolayered sheets resemble those seen in nodular goiters. Sheets of nodular goiters involve evenly-distributed nuclei without overlapping, while the edges of these sheets are irregular. More important is the presence of naked, pyknotic nuclei close to the sheets. In papillary cancer, the occurrence of dispersed cells is infrequent, and the dispersed cells have abundant cytoplasm. Moreover, cancer cells forming clusters display the characteristic nuclear atypia, and are not evenly distributed. The edges of malignant monolayered sheets may also be irregular.

At high power The cells of papillary cancer vary in size, but are larger than those of nodular goiter cases. The variation in shape is also greater than in benign cases. A great number of the nuclei are oval. This is one of the characteristic features seen in papillary carcinoma. Small nuclei are usually visible within the pale nuclei. The presence of large prominent nucleoli is a rare phenomenon in the case of papillary carcinoma except for the oxyphilic variant.

The chromatin is finely dispersed, giving the nuclei a ground-glass appearance, a feature commonly seen in smears stained by the Papanicolau method, but less frequently by Wright-Giemsa method and only very infrequently in smears stained by the hematoxylin-eosin method.

The nuclei not infrequently exhibit a scratched appearance. This sign is probably caused by the invagination of nuclear membrane as in the case of inclusions and grooves. Another possible explanation is that scratches correspond to condensed chromatin which is a frequent finding in papillary carcinoma.

A peripheral sharp dark ring of condensed chromatin is present in most of the cases stained by the Papanicolau method . Even more important is the presence of nuclear grooves and inclusions within the nuclei. Both of these are artefacts, and reflect intranuclear cytoplasmic invagination. The precise name would be pseudo-inclusions and pseudo-grooves. From a practical point of view, however,  it seems reasonable to reserve the term "pseudo-inclusion" for the inclusions of artefacts seen in benign lesions. (If the smears are not wet-fixed, we can observe "typical" nuclear inclusions. The other type of artefacts may be seen in cases where vacuolization is present, even in euthyroid nodular goiter, for hormonal reasons.)
(In our opinion, this is one of the most important and problematic fields in thyroid cytology and also in pathology. For this reason we deal with it in detail in a distinct chapter.) Briefly, we consider the presence or absence of nuclear inclusions and grooves is to be the most important factor in the diagnosis of papillary cancer. In some cases, the inclusions and/or grooves are rare, but we have seen them in all smears stained by the Papanicolau method, with only one exception. These features may also be observed in all other carcinomas of the thyroid, a fact which is not problematic as concerns its diagnostic power. A matter of greater concern is the observation that a relatively large number of nuclear inclusions and grooves may be observed in oncocytic cells derived from Hurthle-cell adenoma or Hashimoto's thyroiditis.

The occurrence of nuclear inclusions and grooves is greatly influenced by the staining method applied. Their occurrence in hematoxylin-eosin-stained smears is very infrequent. For this reason, this method of staining is not accepted in thyroid cytology. Like other nuclear details, nuclear grooves and inclusions are best seen in Papanicolau-smears. The Wright-Giemsa method is less sensitive than the Papanicolau method in this field. On the other hand, the former has better specificity.

Other features
Psammoma bodies, which are generally regarded as a specific feature for papillary carcinoma, are observed in about 20% of papillary carcinomas. The classical picture of concentrically arranged, multilayered rings, produced by changing the focus of the microscope is rarely seen. On the other hand, benign thyroid lesions, most frequently multinodular goiters, may also contain psammoma bodies. In the absence of cytologic features suggestive of papillary cancer, isolated psammoma bodies are unreliable as a predictor of papillary cancer.
In our practice, a definitive true psammoma body has been observed in only 1 case of papillary cancer. In 12 other cases, we have seen dense deposits resembling psammoma bodies. In all of these cases, other features of the tumor were prominent. Hence, the occurrence of (suspected) psammoma bodies was not of great help.
The presence of multinucleated giant cells is observed in a majority of cases (in our practice in 57% of the cases). Kini et al. reported a similar incidence . In most cases, the presence or absence of multinucleated giant cells is not of great relevance. It could have been of great help in one of our false-negative cases. This patient also had a hyperthyroidism, treated a with thyrostatic drug. The cytologic picture revealed a nuclear enlargement, with a variation of shape. Neither a papillary fragment nor nuclear inclusions or grooves could be observed in smears stained by the hematoxylin-eosin and the Wright-Giemsa methods. One year later, we performed a repeat FNAC and at this time we were able to give the correct cytological diagnosis (based on the occurrence of nuclear grooves in Papanicolau-stained smears). A reviewing of the former smears highlighted two features: a significant number of cells were oval, and some multinucleated giant cells could be observed. Oval cells may be seen only sporadically in hyperthyroidism. Moreover, analysis of the smears from patients with hyperthyroidism demonstrated that the occurrence of multinucleated giant cells is a quite sensitive and very specific feature as concerns the presence of papillary cancer.
Cystic degeneration. The presence of a small number of macrophages is quite often seen in papillary cancer. If the tumor has undergone cystic degeneration, the correct cytodiagnosis may be very difficult or even impossible. In these cases, non-cytological features may be of help. The aspirated fluid is brown in most cases of cystic papillary cancers. In one of our cases, metastatic, cystic papillary cancer was suspected on the basis of the elevated thyroxine content of the cystic fluid. US-guided FNAC has a great role in this field. After evacuation of a cyst, we perform a repeat US at once, with US-guided FNAC on the solid remnant of the nodule. However, in a significant proportion of cystic thyroid nodule cases, the FNAC report remains non-diagnostic. In those cases where the cyst recurs, surgery should be considered as a possibility.
Focal oncocytic metaplasia. This phenomenon is often seen, and is mentioned by most published papers. On the other hand, it is not of major significance, and in no way aids the diagnosis of papillary cancer. It seems that oncocytic metaplasia (including its diffuse form, the oxyphilic cell variant of papillary cancer) occurs somewhat, but not significantly more frequently in papillary cancer than in follicular tumors or nodular goiter. The reason for this may be that degenerative metaplastic changes occur more often in carcinoma cells as compared with non-malignant lesions.
Scattered lymphocytes in smears from papillary carcinoma are a common finding. On the other hand a large number of lymphocytes may cause a serious differential diagnostic problem first of all if follicular cells present oxyphilic metaplasia. The differentiation between a Hashimoto's thyroiditis with or without papillary carcinoma may be impossible in certain cases.
We observed squamous cell metaplasia of papillary carcinoma cells in around 3% of our cases. Differential diagnostic problems arose in one of our cases but the thyroglobulin determination in the wash-out of the needle decided the issue.
The presence of different population of cells on the same smear may cause severe differential diagnostic problems. Fortunately, this is a rare situation. If we gain the cytological sample from the adequate localization than the dominant cell type on the smear comes from the lesion in question and relatively small amount of cells comes from the thyroid ventral to the nodule.

VARIANTS OF PAPILLARY CARCINOMA

Oxyphilic cell variant of papillary cancer
(see chapter on oxyphilic tumors)

Follicular variant of papillary carcinoma

Diagnosis of this type of cancer is more difficult. Papillary structures are not observed, and the dominant follicular clusters resemble those seen in follicular adenoma. The microfollicular structure is preserved in most cases. Colloid is present in a relatively high number of the cases of the follicular variant of papillary cancer. In its typical form, it appears as thick globules. These features involve a risk of a false-negative diagnosis. However, the nuclear details (the presence of nuclear inclusions or grooves) are of great help . Nevertheless, we are of the opinion that, if only the nuclear details are suggestive of papillary cancer, then a definitive cancer diagnosis may well be false-positive (Mesonero 1998). 2 of our 4 false-positive diagnoses were believed to be the follicular variant of papillary cancer, but on histopathological examination both proved to be atypical adenoma without any signs of invasion.
Accordingly, in this field we believe that if only nuclear details are indicative of papillary carcinoma, the correct cytological report should be a suspicion of  (the follicular variant of) papillary cancer. With regard to the relatively poor sensitivity of intraoperative frozen sections in diagnosing this variant of papillary carcinoma, the surgical procedure of choice in our practice when the intraoperative frozen section is not conclusive is lobectomy of the thyroid  (see Chapter Guidelines to surgery).

Columnar and tall cell variant of papillary cancer

These variants are represented by aggressive behaviour and a worse prognosis than for the classical variant of papillary carcinoma. The tall cell variant exhibits cells with a height/width ratio > 2, single nuclei and an abundant basophilic cytoplasm. These features resemble those of Hürthle-cell tumors. The characteristic cytologic features of the columnar cell variant are similar to those of the classical variant except for a higher number of columnar cells and the presence of a complex microfollicular pattern. The stratification of the nuclei covering the tissue fragments and clusters of large cells with vacuolated cytoplasm are also among the important cytologic features mentioned in the literature. Metastatic adenocarcinoma is the most important disease in the differential diagnostics of the columnar cell variant. Immunocytochemistry is not always of help, because the cells from columnar cell carcinoma are less differentiated.

Cytological differential diagnostics

Nodular goiter

This type of lesion displays hyperplastic papillae in a significant number of cases. It is usually possible to make a clear distinction between benign and malignant papillarization  TABLE. It is not the type of the papillary frond that is the important feature, but the presence or absence of the nuclear atypia characteristic of papillary cancer. However, it may be difficult to make a clear distinction when the follicular cells are not well preserved (cystic degeneration) or metaplastic (oxyphilic cell alteration). If well-preserved follicular cells are not seen, or only sporadically, a differentiation between nodular goiter and papillary cancer may be impossible. Additional features may be suggestive of the nature of the disease. The presence of colloid, and the light colour of the aspirated cystic fluid favour a benign disease. On the other hand the dark-brown colour of the cystic fluid is of less relevance. If even repeat FNAC is not diagnostic, then the size of the lesion and the age and particularly the compliance of the patient help us to decide. If the lesion is not larger than 2 cm, and the patient is young, we can recommend follow-up (initially 6 months, and then once a year). If the lesion grows in size, surgery is mandatory. On the other hand, if the uncertainty causes the patient psychological stress, and we are unable to relieve this, then we recommend operation even for lesions as small as 1 cm in maximal diameter. In most cases, providing the patient with detailed information is sufficient for reassurance, and only a few of them then decide to undergo surgery.

Hyperthyroidism

If a patient with Graves-Basedow's disease exhibits a nodule (which is the situation in around 25 % of our cases), FNAC must be performed. The risk of papillary cancer in nodules in Graves-Basedow's disease is similar to that for a euthyroid patient. The interpretation of the cytology is somewhat more difficult, because the atypia caused by hyperthyroidism and/or the thyrostatic agent may contribute to the cytological picture.
It must be borne in mind that either Graves-Basedow's disease or the thyroid nodule itself is a condition which may indicat surgery. From a clinical aspect, if a patient with hyperthyroidism has a nodule, the question of surgery must be assessed very carefully. If there is any doubt about the benign nature of the disease following microscopic analysis, surgery should be proposed.
There is another problem. In autoimmune thyroid disorders, the presence or absence of a nodule is not always clear. Increase in the vascularization of the thyroid, as in hyperthyroidism, causes alterations in the palpation: the thyroid becomes firm. The thyroid is enlarged, and this enhances the extent of small irregularities in the ventral contours of the gland. This is one of the conditions where the palpation may overdiagnose a thyroid nodule. Moreover, the acute phase of hyperthyroidism is the only condition where US may fail to detect all of the lesions.  It is well known that hyperthyroidism demonstrates either diffuse hypoechogenicity or a patchy hypoechogenic pattern. Thyroid malignancies primarily occur in nodules with low echogenicity.) On the other hand, US may reveal discrete echoabnormalities which are not nodules with potentially malignant behaviour, but rather one of the appearance forms of Graves-Basedow's disease. Thus, if there is no doubt that a nodule is present, but doubt arises concerning its benign nature, surgery is mandatory. If the presence of the nodule is questionable and the cytology is not clearly benign, we again consider the indication of surgery. To avoid a failure to detecting papillary cancer (which occurred in one of our cases), we perform a second US 6 months after the beginning of the disease. At this time, the low echogenicity caused by the hyperthyroidism has decreased, and hypoechogenic nodules missed at the first US may be detected.
Now back to the microscope. The atypia caused by hyperthyroidism is reflected by nuclear enlargement, and a great variability in nuclear size. However, the cells are characteristically round. Moreover, nuclear grooves are not present. The characteristic vacuolization caused by hyperthyroidism makes it difficult or impossible to interpret nuclear vacuoles. The cells are arranged in loose cohesive monolayers, which is a clear distinction from sheets of papillary cancer. In 2 of our patients, typical papillary fragments with nuclear details resembling those of papillary cancer were observed. These papillary fronds were compact structures, which we observed in the hyperthyroid phase but neither at the second FNAC, performed in a euthyroid state, nor on histopathological examination. We proposed going on surgery because US revealed discrete lesions in both patients. The histopathology demonstrates benign disease in both cases, and interestingly the final histological report was LT in both cases. The hyperthyroidism was hashitoxicosis.

Hashimoto's thyroiditis

The reported incidence of Hashimoto's thyroiditis in cases of papillary cancer varies greatly in the literature (Kashima 1998). This is connected with the problem of terminology in the various forms of LT (Selzer 1977 , Kashima 1998). Nuclear grooves, peripheral sharp dark rings of condensed chromatin, and of course oxyphilic cell changes, with nuclear atypia (even  pronounced) are frequently encountered in Hashimoto's thyroiditis. There are two conditions where diagnostic difficulties arise. The first is, when there are only a few lymphocytes in the smear, but US reveals a characteristic picture of LT (diffuse hypoechogenic pattern without a discrete nodule). In these cases, US is of great help. Although Hashimoto's thyroiditis is not infrequently associated with small foci of papillary cancer, in our practice no papillary cancer occurred in LT without discrete lesions on preoperative US.
More concerns arise in the second condition, when we see nuclear features resembling papillary cancer in a discrete lesion. Four factors must be kept in mind in the decision-making: the number of lymphocytes in the smear, the size and number of the lesions, the shape of the lesions, and the cell type in which nuclear grooves are observed. TABLE If there are many lymphocytes in the smear, the problem is unfortunately not resolved and some patients with a benign disease may undergo surgery. It may be mentioned here that the characteristic US picture of papillary cancer, i.e. a hypoechogenic nodule with fine granules, may be observed in more than 60 % of patient with the nodular form of Hashimoto's thyroiditis. In the last 2 years we have frequently decided on follow-up in questionable cases. If there are no conclusive signs of papillary cancer, we perform repeat US (and cytology) 3 and 6 months later, and than every year. If the patient accepts our suggestion, we do not give a definitive diagnosis after the first examination. If the patient refuses follow-up, we give a diagnosis of suspicion of papillary cancer. TABLE
This type of problem in the field of thyroid cytology is not emphasized by others. The routine use of US explains why we are faced with this problem relatively frequently.

Follicular adenoma

From a practical point of view follicular adenoma is the only tumor in the thyroid which must be the clearly distinguished from papillary cancer (in all other thyroid tumors surgery can not be avoided). In most cases, the distinction is not difficult. If we apply the Papanicolau method for staining, a clear distinction between follicular adenoma and the follicular variant of papillary cancer is possible on the basis of the occurrence of nuclear inclusions and grooves. These nuclear features are very rare in follicular adenoma. On the other hand, hematoxylin-eosin staining is very insensitive for the detection of  nuclear grooves and inclusions. The Wright-Giemsa staining procedure is better than the hematoxylin-eosin method, but not as good as the Papanicolau method in this field.

Follicular and oxyphilic tumors other than typical follicular adenoma

The suspicion of these tumors is an absolute indication for surgery, so a clear distinction between these tumors and papillary cancer seems to be of little practical relevance. Nevertheless, in the case of a papillary carcinoma total thyroidectomy is the treatment of choice while to perform total thyroidectomy in the case of a benign adenoma is a great failure. Occasionally, these tumors may be difficult to differentiate from papillary cancer. This is particularly true for Hürthle-cell tumors. There may sometimes be numerous nuclear grooves in oncocytic adenoma. The round shape of the cells, and the prominent nucleoli of uniform size may be of great help in avoiding a false-positive cytological report when numerous grooves are observed. Nevertheless, a perfect solution to this differential-diagnostic problem can not be achieved in all Hurthle-cell adenoma cases. A definitive cytological diagnosis of the oxyphilic variant of papillary cancer is acceptable only in cases where characteristic features other than nuclear inclusions and grooves are also be observed.