Why we need a new way of thinking

The main concern about the previous and even the current evaluation process of thyroid nodules is the loss of information.

We need to see a typical evaluation of a patient according to the current protocol of the American and the European Thyroid Association. Firstly, the endocrinologist or internist takes the patient history, palpates the neck of the patient and then the patient is sent for a blood test and an ultrasound examination. The process is followed by an aspiration cytology. The cytopathologist may or may not be aware of the results of other examinations and gives a diagnosis which, except for patients with large goiters and/or hyperthyroidism, decides the fate of the patient. The palpation, the clinical data and even the ultrasound are used only for select patients for cytology. The cytopathologist would be the only physician who is able to integrate all data of the whole evaluation process. However, except for very rare instances, this member of the evaluation team is not able to do this as his profession is the microscopic analysis. The situation may be better in highly specialized and professionally organized thyroid evaluation teams, but the current protocol made by members of such teams does not reflect this assumption.

A similar problem arises in those teams where the ultrasound examination is performed not by the thyroidologist. Not only the cytological diagnosis but even the ultrasound diagnosis may be influenced by the knowledge about clinical and laboratory data. It is very hard to expect to integrate the clinical and laboratory findings from a radiologist who has no special interest in thyroidology.

What are the statistical parameters of thyroid cytology gained from cyto-histological comparisons? FNAC is non-diagnostic in around 10% of cases and in around 30 to 40% of patients the FNAC is equivocal. Depending on the iodine intake, the positive predictive value of FNAC ranges between 33% to 60%. Considering these data far from the ideal, the loss of any information during the evaluation process has professional and also ethical aspects.

We demonstrated many case studies presenting the role of an integrated diagnosis. In this chapter we focus on several issues.

How to solve the issue

The thyroidologist has to perform the ultrasound examination - this is the fundamental way of giving a better ultrasound diagnosis. This way two types of information, i.e. the clinical and the ultrasound data are handled by one specialist. The integration of the cytological data is a more difficult task. A closer cooperation between the clinician and the thyroidologist would be a more simple solution. However, even this assumes an alteration in the responsibility of the single members of the evaluation team with its judiciary and logistical consequences.  
Another possibility would require a change in the postgraduate medical education: we educate thyroidologists who are clinicians with the facility to perform thyroid ultrasound and the whole thyroid cytological examinations including microscopic analysis. This proposal may be a bit ridiculous, I agree. Our current situation is that more than 90% of patients of an endocrinological outpatient institutions are referred because of thyroid disorders. Most of the physicians of a thyroid outpatient unit behave like a manager of logistics - they prescribe examinations, collect the results and give a final proposal, but most of them perform no special examinations. Compare this situation with that of a cardiologist who is able to perform at least cardiological ultrasound examination which is more difficult than a thyroid test. Nowadays, working in a thyroid unit for decades causes a limited challenge, and I assume that it may lead to burnout with greater likelihood compared to other specialists.

 

Follicular proliferation - part 1

Facts and data

1. The risk of carcinoma in a follicular tumor is 3.1% in our institution. This proportion was gained from histopathological analysis: the ratio of follicular adenoma to follicular carcinoma in our surgically treated patient was 1498 to 45 in a 20-year period.
   It is significantly lower than the published data in the literature. The difference is explained by several factors.

Iodine-deficiency. As we previously demonstrated, iodine-deficiency increases the occurrence of follicular adenomas.

Decline of occurrence of follicular carcinoma. Despite the increasing incidence of well-differentiated thyroid carcinomas, follicular thyroid carcinoma is being diagnosed less and less frequently (Sobrinho-Simoes et al.).

Differences in histopathological interpretation of follicular lesions. We cannot exclude the theoretical possibility that we underdiagnose follicular carcinomas compared to other investigators.

2. We have found no clinical proof of histopathological underdiagnosis of follicular carcinoma. 1.498 patients were diagnosed histopathologically with follicular adenoma in the last 20 years in our institution. These patients and more than 500 others with the similar histological diagnosis operated in the previous decades were or are still on regular follow-ups now. More than 25,000 patient-years have elapsed since their operations and we have not found one case during that period which had follicular carcinoma in a recurrent nodule. It means that if underdiagnosis occurred in a few patients, it had minimal if any negative consequences on the results.

3. The   cytological differentiation of a follicular adenoma and a follicular carcinoma is cytologically impossible. What is the consequence of this fact considering the previously mentioned circumstances?
   In order to recognize one follicular carcinoma, we had to operate 33 patients unnecessarily with a benign tumor. Also, around 10% of hyperplastic nodules and Hashimoto's thyroiditis present a cytological pattern indistinguishable from a follicular tumor.

4. We have no tools to distinguish a follicular adenoma from a follicular carcinoma except for histopathology. Although new molecular methods may be promising, but there is no conclusive method in this field yet. The most promising results of molecular biological techniques may give significant impact on the prognosis of thyroid carcinomas or on the diagnosis of papillary carcinomas but it does not impact the differentiation of follicular tumors yet. Also, the use of genetic techniques in differential diagnosis of follicular tumors may be questioned too until the uniform histological interpretation of follicular lesions is not ensured.

5. A malignant tumor must increase in size over time.

6. Basically two histological types of follicular carcinoma exist, the minimally invasive and the widely invasive subtype. While the cytological presentation of the former is identical to a follicular adenoma, the latter is characterized by significant atypia which is easy to identify on cytological smears.

7. We have no reason to suggest that even a few years` delay in the correct therapy of a well-differentiated thyroid carcinoma impairs the prognosis of the patient; therefore a delay of a few years delay can be accepted in the diagnosis and therapy for patients in whom the diagnosis is a low-grade follicular carcinoma. This statement was made by Hamburger in 1996 (Werner &Ingbar's The Thyroid: A Fundamental and Clinical Text), and no one denied it in the past two decades.

Main consequences

The price we pay in order to recognize and operate all follicular carcinomas is high, as we need to operate follicular adenomas and even non-tumorous lesions unnecessarily. Depending on the ratio of follicular carcinomas to adenomas, this cost may be reasonable or unacceptably high.

It is time to reconsider the dogma that the cytological diagnosis of a follicular tumor must lead to surgery in our opinion. The fundamental of this reconsideration is the risk of follicular carcinoma in the event of a cytologically diagnosed or suspected follicular tumor. As the ratio of follicular adenoma and follicular carcinoma differs significantly among institutions, this risk has to be determined individually. E.g. if the risk is 30%, the reconsideration of the dogma is beyond reason. However, we must be aware of the decline of follicular carcinoma cases. Therefore, old histopathological data are not suitable to make decisions.

The following approach is suggested.

• An individual risk management: taking the cytological, ultrasound properties and the clinical presentation into account.

• Suggestion of surgery if the risk is high.

• Consideration of regular follow-ups when patients have the risk of a well-differentiated minimally invasive follicular carcinoma below a cut-off level. This threshold might be in the range of 3% to 10%. The final decision of operation needs to be made by the patient.

• Patients who choose the follow-ups have to undergo yearly ultrasound examinations by the same doctor. If the lesion significantly increases in size, the patient needs to get an operation.

Follicular proliferation - part 2

Facts and data

1. The follicle is the basic unit of the thyroid. Not only follicular tumors, but every thyroid including a healthy one is composed of follicles. Follicular tumors tend to be composed of dominantly microfollicles while non-tumorous lesions are a mixture of micro-, normo- and macrofollicles. A microfollicular proliferation is the main concern. Although we cannot expect a cytopathologist to differentiate a follicular adenoma from a follicular carcinoma, the cytological distinction of a hyperplastic nodule from a follicular tumor is required. A cytopathologist cannot meet this expectation in every case

   •  More than 90% of follicular tumors are composed of mainly microfollicles. As we previously demonstrated, iodine-deficiency increases the proportion of follicular adenomas.

   •  Most hyperplastic nodules contain areas composed of microfollicles.

   •  Around 10-15% of hyperplastic nodules are composed dominantly of microfollicles.

   •  In around 10 to 15% of Hashimoto's thyroiditis presents microfollicular proliferation.

2. According to the current protocol, the cytological diagnosis of a follicular tumor should be followed by surgery. Moreover, the term "tumor" on a medical report causes great concern and fear in patients and most of them wish to be operated no matter what we suggest.

3. A follicular tumor has to be surrounded with capsule which can be demonstrated by ultrasound in more than 90% of cases.
   The demonstration of a halo and/or a perinodular blood flow does not necessarily mean the presence of a follicular tumor because hyperplastic nodules are frequently encapsulated as well. However,   the lack of ultrasound signs of a capsule practically excludes the possibility of a follicular adenoma or follicular carcinoma or oxyphilic variant of both.

Main consequences

In significant number of patients who are diagnosed cytologically, having a follicular tumor as an option is ruled out, but corresponding with the usual approach, we do not take the ultrasound properties into account   after   a cytological diagnosis. Also, if we took the ultrasound into account in the event of a follicular tumor, it would raise serious judiciary concerns. We also need to consider the fact that a cytopathologist will give nothing other than a follicular tumor diagnosis when faced with a pattern fulfilling the criteria of a follicular tumor.

The solution is not very difficult. If the cytopathologist finds a pattern which corresponds to a follicular tumor, the reconsideration of ultrasound is mandatory. In cases where the ultrasound excludes or substantially decreases the possibility of a follicular tumor, a common clinical-ultrasound-cytological diagnosis of a benign follicular proliferation is a reasonable alternative. We unequivocally support this option and this is our everyday practice for more than 15 years.

 

Oxyphilic metaplasia

General considerations:

While oxyphilic metaplasia is a common phenomenon in the thyroid and may occur in each thyroid disorder, extensive oxyphilic metaplasia can be found only in Hashimoto's thyroiditis and in oxyphilic tumors. On the other hand, we have to consider that the cytological material does not fully represent a lesion. E.g. we can aspirate just the area of a hyperplastic nodule which contains predominantly oxyphilic cells. A similar situation may occur in Hashimoto's thyroiditis when we gain material from an area rich in epithelial cells. The opposite happens when an oxyphilic cell tumor does not contain oxyphilic cells exclusively; by convention, thyroid tumors are designated as oncocytic if at least 75% of their constituent cells can be described as oncocytes.
We need to consider that an oxyphilic tumor is more probably malignant than a non-oxyphilic tumor, also, oxyphilic carcinomas present a more aggressive clinical behavior compared to non-oxyphilic carcinomas.

Regarding the cytological presentation, there are two less emphasized differences between oxyphilic and non-metaplastic cells. Firstly, the specificity of intranuclear grooves and in a lesser degree that of intranuclear inclusions is worse in Hürthle-cells. The occurrence of grooves and inclusions is significantly higher in benign oxyphilic cells than in benign non-metaplastic cells. This led to false positive diagnosis in one of our cases (see Case 19 of oxyphilic adenomas). Cells with abundant cytoplasm including Hürthle-cells tend to dissociate, therefore the typical follicular formation is more difficult to reveal in cytological smears. The classic papillary formation is also more difficult to detect in an oxyphilic cell variant of papillary carcinoma compared to the classic variant of this carcinoma.

An oxyphilic tumor may be benign or malignant. A benign tumor is a variant of follicular adenoma. It is characterized by the presence of a complete capsule, the tumorous proliferation of dominantly oxyphilic cells, the lack of nuclear features of a papillary carcinoma and the lack of capsular and vascular invasion. While the definition is clear in the event of an oxyphilic adenoma, there is no general agreement among researchers and authors regarding the categorization of malignant oxyphilic tumors. Basically two categorizations exist. Some authors define oxyphilic carcinoma as a separate entity, while other research groups treat an oxyphilic tumor as either a papillary carcinoma or a follicular carcinoma as oxyphilic variant of one of these entities. The situation is even more confusing because those pathologist, who believe that oxyphilic carcinoma is a distinct entity, use the terms oxyphilic variant of papillary or follicular carcinoma, too, in those cases where the proportion of oxyphilic cells is relatively not so high. Although, this uncertainty in the categorization has minimal if any consequence on the treatment of a patient, hinders the comparison of data of different researchers. We only mention here that even medullary carcinoma cells can produce oxyphilic metaplasia.

 

Consequences

In contrast to non-oxyphilic form of a follicular tumor, the diagnosis of a Hürthle-cell tumor is an absolute indication of surgery because of the difference in clinical behavior of the malignant forms.

The main goal of preoperative diagnosis is to avoid an unnecessary operation in non-tumorous oxyphilic lesions, i.e. in Hashimoto's thyroiditis and in hyperplastic nodules presenting oxyphilic metaplasia. We have a better chance to complete this task when looking at the ultrasound presentation as well. What does it mean in the everyday practice?

A possible common ultrasound-cytological diagnosis plays the most important role in the event of an oxyphilic lesion lacking ultrasound signs of a capsule, i.e. in those nodules which present neither halo sign nor perinodular blood flow. In such cases the risk of an oxyphilic type of a follicular carcinoma or a Hürthle-cell adenoma is very low. Although the specificity of nuclear atypia is worse in oxyphilic cells, the sensitivity is as good in Hürthle-cells as in non-metaplastic cells. It means that the lack of inclusions and grooves makes the possibility of an underlying papillary carcinoma very unlikely.  

To summarize, the cytology is able to exclude the possibility of oxyphilic variant of papillary carcinoma, however it is not able to exclude oxyphilic variant of follicular carcinoma; the ultrasound is able to do that in more than 95% of cases. It means that the risk of malignancy is very low, less than 1.5% according to our unublished data, in those oxyphilic lesions which present neither cytological signs of a papillary carcinoma nor ultrasound signs of a capsule.

It is well-known that Hashimoto's thyroiditis presents discrete hypoechogenic areas in around 90% of cases. In certain cases the only chance to avoid an unnecessary surgery is the common analysis of ultrasound and a cytological pattern.

 

Repeatedly non-diagnostic cytologies

Depending on the iodine intake 5 to 10 % of thyroid FNACs are repeatedly not diagnostic if it was performed by an experienced examiner. Among others, a non-diagnostic report has two negative consequences: the cytology has to be repeated and the report causes concern and fear in the patient. Basically two types of these categories exist.

Cystic lesions

The first group is composed of cystic lesions and this is one of the lessons that focuses on this problem. It is the best to take the type of the cyst into account from which the inadequate material was gained: spongiform-type and central-type cyst are only exceptionally malignant. The risk of malignancy is certainly less than 1% in these cysts. Also, cysts presenting comet-tail artifacts are characteristically benign.  

Parathyroid adenomas are characterized by aspirating watery fluid. The smear of such cysts almost never contains any cell. Instead of a non-diagnostic report, we prefer a combined clinical-ultrasound-cytological report which states that the risk of malignancy is less than 1%.

Solid nodules

Non-diagnostic punctures of a solid nodule belongs to the other group. The ultrasound and clinical findings may be of help in certain cases. Basically, we can determine the risk of malignancy based on ultrasound properties and we can give an estimation of malignancy risk in the final report.