4-1995
The Management of Thyroid Nodules
Jayne A. Franklyn and Michael C. Sheppard, Birmingham, UK
Contents:
Prevalence and pathogenesis
Enlargement of the thyroid is common. The Framingham study in the United
States found that 6.4% of females and 1.6% of males had one or more thyroid
nodules (1), while follow-up for 15 years indicated an annual incidence
of 1.4% over that period (1). The Whickham survey, carried out in the northeast
of England, similarly reported that small goitres (those palpable but not
visible) were present in 8.6% of the population, while obvious goitres
(palpable and visible) were present in 6.9%, goitres being four times more
common in females than in males (2). Follow-up of that population for 20
years has recently revealed a reduction in prevalence of thyroid enlargement
with increasing age; amongst women, 70% had no goitre at either survey,
20% had a goitre at first survey but not at follow-up, 6% had goitres at
both surveys and 4% developed goitres (3). While a decline in goitre prevalence
with age has been reported before (1), such surveys highlight the marked
frequency of thyroid enlargement in the population as a whole. The presence
of nodular thyroid disease may be even more common than such epidemiological
studies suggest in that postmortem studies have shown that up to 50% of
the population have either single or multiple nodules in the thyroid, many
of which are very small (4). Studies employing high resolution ultrasound
scans have similarly revealed discrete nodules in up to 50% of those beyond
the fifth decade of life (5).
Aetiological factors determining the development of either single or
multiple thyroid nodules are complex. TSH has been widely implicated in
the pathogenesis of nodular thyroid disease in view of its role as a potent
stimulator of thyroid growth. Growth stimulating immunoglobulins have also
been implicated (6), suggesting an underlying autoimmune disorder, while
worldwide, iodine deficiency remains a crucial factor in goitre development
(7). Previous exposure to ionizing radiation is also a well documented
risk factor for the development of both benign and malignant thyroid nodules
(8).
In contrast to the marked prevalence of thyroid nodules outlined above,
thyroid cancer is rare, accounting for less than 0.5% of all new malignancies
in England and Wales and less than 0.5% of all cancer deaths (9). Thyroid
cancer is reported to have been found in only between 6 and 14% of single
thyroid nodules selected for surgery on clinical grounds (10), while cancer
risk is even lower in multinodular goitre (11). The picture is, however,
complicated by postmortem studies which describe the presence of occult
cancer in up to 5% of thyroid glands normal to palpation (4). Most occult
cancers are small, being only several millimetres in diameter, and the
substantial prevalence of occult cancer in those who have died from unrelated
disease suggests that this is an entity without clinical significance.
The overwhelming challenge to the clinician is therefore to identify the
small number of patients presenting with significant neoplastic disease
of the thyroid from the majority without, thus sparing up to 90% of patients
unnecessary surgery.
History, examination and laboratory tests
The history of a patient may increase the likelihood of that patient having
a malignant lesion, although for the majority of subjects presenting with
one or more thyroid nodules, the history is not helpful in determining
the nature of the underlying pathology. Prior exposure to radiation is
a risk factor for both benign and malignant nodular thyroid disease (8),
and patients with such a history should be evaluated with care. There are
conflicting reports regarding the effect of age on cancer risk (12) but
thyroid nodules are uncommon in childhood and their presence should be
viewed with suspicion (13). Likewise, there may be an increased risk of
cancer in nodules developing over the age of 65 years, and undifferentiated
thyroid cancer generally occurs in this age group. Sex is itself an important
determinant of neoplasia risk, since thyroid cancer is relatively more
common in thyroid nodules presenting in males rather than females. A family
history of thyroid cancer, especially medullary cancer, may be especially
pertinent and should lead to a suspicion of multiple endocrine neoplasia
syndrome type II or familial medullary thyroid cancer.
A history of symptoms such as dysphagia, dyspnoea or hoarseness may
suggest local invasion from a malignant lesion, although such symptoms
also occur with benign thyroid enlargement, especially in association with
large multinodular goitres. A malignancy rate of 71% has been reported
in a group classified "high risk" on the basis of symptoms suggesting a
compressive lesion, as well as signs such as firmness, fixation to adjacent
structures, vocal cord paralysis and cervical lymphadenopathy (14), although
fewer than 5% of patients presenting with nodular thyroid disease have
such symptoms or signs (15). While it is recognised that thyroid cancer
is less common in multinodular goitre than in clinically solitary thyroid
nodules, review of our own consecutive series of 461 patients presenting
to a specialist thyroid clinic with thyroid enlargement revealed a significant
prevalence of neoplasia in those with multinodular or diffuse thyroid enlargement
as well as in those with solitary nodules, determining that further investigation
of those with multinodular and diffuse goitre, at least amongst a population
referred to hospital, is justified (15).
No laboratory test specifically distinguishes benign from malignant
thyroid enlargement. Circulating concentrations of thyroid hormones and
TSH should, however, be measured, to exclude hyper- or hypothyroidism which
themselves warrant specific treatment, and which if present markedly reduce
the likelihood of any thyroid enlargement reflecting thyroid neoplasia.
Serum concentrations of thyroglobulin are raised in patients with goitre,
but values are similar in those with benign and malignant nodules (16).
Measurement of serum thyroglobulin is therefore of no value in the initial
assessment of the patient presenting with nodular thyroid disease, although
its role in the long-term follow up of patients treated for differentiated
thyroid cancer is clear. Serum calcitonin is the tumour marker for medullary
thyroid cancer and while this diagnosis accounts for less than 10% of thyroid
malignancies, it has recently been suggested that routine measurement of
serum calcitonin has a role in the work up of patients presenting with
nodular thyroid disease (17). While measurement of serum calcitonin in
patients with a family history of medullary thyroid cancer is clearly indicated,
further studies are needed to determine whether the prevalence of the disease
and the relevance of early diagnosis justify routine measurement of circulating
calcitonin (18).
Assessment of thyroid size
In addition to evaluation of the nature of nodular thyroid enlargement
as discussed below, management is also determined in part by the size of
the thyroid lesion and in particular by the presence or absence of tracheal
compression. Plain X-ray examination of the trachea may reveal deviation
or compression of the airway, although comparison of the specificity and
sensitivity of plain radiology with findings from CT scanning of the trachea
has revealed that X-ray findings are often misleading (19). Assessment
of a respiratory flow volume loop is regarded the best method for evaluation
of upper airways obstruction (20) and we have recently performed a prospective
study of this method for assessment of the airway. We found that inspection
of flow volume loops revealed upper airways obstruction in up to one third
of patients presenting with nodular thyroid disease and that symptoms such
as choking or breathlessness were poor predictors of this functional upper
airways obstruction (21). It is as yet unclear whether flow volume loop
evidence of tracheal obstruction in the asymptomatic patient warrants surgical
treatment.
Radionuclide imaging
Technetium-99m (99mTc) pertechnetate is the radionuclide most frequently
employed in thyroid imaging. Pertechnetate ions are trapped by the thyroid
in the same way as iodine but are not organified, in contrast to the less
readily available and more expensive isotope iodine-123 (123I) which is
both trapped and organified. Although the majority of thyroid malignancies
do not concentrate radioisotopes (and thus appear 'cold'), less than 20%
of 'cold' nodules are due to cancerous lesions, the remaining 80% being
due to colloid nodules, haemorrhage, cysts or inflammatory lesions such
as Hashimoto's thyroiditis (22). 'Hot' nodules on iodine scintigrams are
associated with a low incidence of malignancy (23), most reported 'hot'
carcinomas probably representing the existence of a malignant lesion next
to a benign 'hot' lesion (24). The finding of apparently normal isotope
uptake (a 'warm' lesion) may be ambiguous since a hypofunctioning malignant
nodule may be obscured by overlying normal tissue. Radionuclide imaging
may also reveal multifocal abnormalities in patients thought clinically
to have solitary nodules, mixed areas of hyper- and hypofunction being
characteristic of multinodular goitres and less commonly Hashimoto's thyroiditis
(25). While the incidence of cancer is less in multinodular goitres than
in solitary nodules, the presence of 'cold' areas on such scans leads to
concern regarding the underlying pathology since the presence of a multinodular
goitre does not prevent the development of thyroid malignancy as a second
abnormality.
In view of these limitations of radionuclide scanning it is not surprising
that this approach does not specifically distinguish benign from malignant
lesions. This is illustrated by Ashcraft and Van Herle's review of 22 series
(26), in which more than 5000 patients underwent surgery irrespective of
the functional status of the nodule defined by isotope scanning. Of these,
84% were 'cold', 10.5% 'warm' and 5.5% 'hot' lesions. Malignancy was present
in 16% of the 'cold', 9% of the 'warm' and 4% of the 'hot' nodules. Thus
malignant nodules are more likely to be 'cold' than 'hot', but most 'cold'
lesions are benign and the presence of a 'warm' or 'hot' nodule does not
exclude malignancy. There are therefore no specific features which indicate
the benign or malignant nature of a thyroid nodule and this poor specificity
has led many centres to abandon radionuclide scintigraphy in the routine
investigation of those presenting with nodular thyroid disease.
Ultrasound scanning
High resolution ultrasonography can detect lesions in the thyroid as small
as 1 mm in diameter and the technique is therefore better than either physical
examination or radionuclide imaging in detecting nodules (27). A high prevalence
of nodular thyroid disease of up to 50% is revealed by such high resolution
scanning as mentioned above. In addition, in as many as 40% of those patients
thought clinically to have a solitary nodule, multiple nodules in the thyroid
are revealed by ultrasonography (28). Since data on the incidence of malignancy
in nodular thyroid disease are related to clinically palpable goitres,
it is unclear whether such multiple nodules detected ultrasonographically
are associated with a similar or lower risk of malignancy than those nodules
only obvious on clinical examination. Ultrasonography is of value in distinguishing
solid from cystic lesions of the thyroid, ultrasound scanning being reported
to classify nodules as solid, mixed cystic and solid or cystic with greater
than 90% accuracy (29). It has been reported that malignant nodules are
hypoechoic compared with the remainder of the gland, perhaps because of
their disordered architecture, while benign lesions are hyperechoic with
a sonolucent rim ("halo sign") (30). It has become clear, however, that
there are no sonographic criteria which are able to define benign or malignant
disease with good specificity. In Ashcraft and Van Herle's review (31),
more than 1000 patients with thyroid nodules underwent conventional ultrasonography
before proceeding to surgery. Eighty % were classified pre-operatively
as solid or mixed solid cystic lesions and 20% were cystic. At surgery,
17% of all the lesions were malignant. Although solid or mixed cystic lesions
harboured most of the malignancies, 6% of tumours occurred in those lesions
originally thought to be cystic, probably reflecting the observation that
some carcinomas undergo cystic degeneration.
The poor specificity of ultrasonography in terms of indicating the underlying
pathology, together with the difficulties which arise if nodules which
are impalpable are revealed by ultrasound scanning, means that such scanning
is often not considered helpful in patients presenting with nodular thyroid
disease and its routine use has been questioned.
Thyroid hormone suppression
Attempts have been made to distinguish benign from malignant nodules by
seeking a change in size in response to thyroxine, perhaps related to dependency
on TSH for growth. The incidence of malignant disease in patients who demonstrate
an apparent response to thyroid hormone suppression is not known, but it
is well established that some carcinomas may respond to suppression by
reducing in size (32). Conversely a lack of response to a trial of thyroid
hormone suppression is not specific for malignancy, the incidence of cancer
in non-responders being between 12 and 40% (33).
Fine needle aspiration cytology
This technique has emerged as a valuable tool in the diagnosis and management
of nodular thyroid disease. In contrast to the techniques already described,
which provide at best an indirect evaluation of the nature of the disease,
fine needle aspiration provides a means of obtaining thyroid tissue for
direct cytological examination without the need for surgery. The technique
requires practice in order to reduce sampling error and to obtain satisfactory
smears with high frequency and, most importantly, the skills of an expert
cytopathologist.
The procedure can be performed in an outpatient department, is well
tolerated by the patient, and can be repeated. The cytopathologist can
report that an adequate smear contains normal cells only, cells from benign
non-tumorous conditions such as lymphocytic thyroiditis, or suspicious
(indeterminate) or frankly malignant cells. The report that is indeterminate
or suspicious of malignant change should be regarded as an indication to
proceed to surgical exploration - in the same way as a report of unequivocal
malignancy indicating papillary, follicular or medullary carcinoma.
Prolonged evaluation of the technique, especially in Sweden where it
has been applied for more than 25 years, has demonstrated that in the hands
of a skilled cytopathologist excellent diagnostic accuracy (of up to 90%)
can be achieved. A false positive report of malignant change is extremely
rare but the rate of false negatives (i.e. tissue thought to be benign
at cytology but found to show malignant change at surgery) is 5-l0% in
most series (34). This in part is due to difficulty in distinguishing benign
adenomas from malignant follicular neoplasms (34); surgical excision of
all such lesions is therefore required. An analysis from Cusick et al.
(35) suggested, however, that the technique may be less reliable than is
widely accepted. Analysis of their original data revealed an overall accuracy
of 92% (36) but audit of a further prospective evaluation showed an accuracy
of only 69% (35). Conflicting data in reported studies may relate to the
clinical nature of the swellings included (multinodular, isolated or dominant),
frequency of operation, and methods used to calculate accuracy.
Since 1984 we have routinely carried out fine needle aspiration in all
euthyroid patients presenting with thyroid enlargement. We have examined
outcome in 629 consecutive patients followed for a minimum of 12 months
after first aspirate. Patients were classified as having benign or suspicious/malignant
cytological features after their first aspirate. The former group (83%
of the total) was managed conservatively, apart from those requiring surgery
because of thyroid size or patient preference, and the latter group (17%)
proceeded to histological examination. In total, 76% of patients were observed
and 24% underwent partial or subtotal thyroidectomy. Of 529 patients with
benign cytological findings, clinical follow up of all and surgery in a
minority suggested that "true negative" cytology was present in 520. There
were 9 "false negative" reports relating to sampling error - 3 cases of
papillary carcinoma, 2 follicular carcinomas and 4 follicular adenomas.
Seventeen "malignant" cytological reports correctly identified thyroid
malignancy and there was one false positive report; 82 "suspicious" reports
prompting surgery identified 39 "true positive" cases of thyroid neoplasia
(including cancers and follicular adenomas); 43 of this group had benign
colloid goitre. These results indicate a diagnostic accuracy of the cytological
findings of up to 92%.
A diagnosis of thyroid cyst was made in 14% of our series of patients
presenting with solitary nodules on clinical examination. These patients
were investigated only by fine needle aspiration but had they proceeded
to radionuclide scanning, such lesions would have appeared "cold". Approximately
50% of such patients were cured by simple aspiration of cyst fluid, thus
avoiding thyroid surgery. Thus in addition to its diagnostic role, aspiration
of thyroid nodules may be of therapeutic importance.
Non-surgical treatment of nodular thyroid disease
Injection of simple cysts with tetracycline as a sclerosant has been shown
to be of value for recurrent cysts (37). More recently, ethanol sclerotherapy,
with or without ultrasound guidance, has been advocated with reports suggesting
marked reduction in the size of cystic lesions on follow-up with minimal
adverse effects (38,39). Thyroxine therapy in doses which suppress serum
TSH to below the normal range has been extensively studied in terms of
its short-term and long-term effect upon goitre or nodule size. The results
of such studies have been conflicting and most show little or no benefit
from suppressive therapy (reviewed in 40), either in patients with solitary
thyroid nodules or in those with multinodular goitre. Nonetheless, T4 therapy
continues to be widely used in the hope of inducing a reduction in thyroid
size. A trial of T4-suppressive therapy, perhaps for up to one year, has
been judged to be appropriate (40), to determine if a clinically relevant
shrinkage in thyroid size results, although caution needs to be exercised
especially in postmenopausal females in whom it is possible that T4 therapy
is associated with a loss of bone mineral density and hence increased risk
of osteoporotic fracture (41).
A further non-surgical treatment option which has received recent attention
is radioiodine therapy. Radioiodine treatment of multinodular goitre has
been reported to induce reduction in goitre size without any short-term
increase in size leading to compressive symptoms, so that radioiodine treatment
has been judged to be clinically efficacious and safe, although the long-term
risk of inducing hypothyroidism remains unclear (42,43).
Strategy for investigation
The appropriate diagnostic evaluation of the patient presenting with nodular
thyroid disease remains controversial, despite the frequency of the problem.
Several alternative strategies have been proposed ranging from immediate
surgery of all lesions to selective surgery after one or more of a series
of investigations including radionuclide imaging, ultrasound scanning,
a trial of thyroid hormone suppression and fine needle or large needle
biopsy. Molitch et al. (l0) have applied decision analysis to examination
of the value of these different strategies, balancing possible long term
benefits of increased life expectancy after surgery against the risk of
operative complications, and the value of imperfect diagnostic techniques
that alter surgical selection. In view of the good prognosis for most patients
with differentiated thyroid cancer it is not surprising that only small
differences can be demonstrated between the predicted value of each strategy
for investigation and management. At age 20, aspiration cytology provides
a benefit of 0.05% of life expectancy over thyroid hormone suppression,
which in turn outweighs immediate surgery by 0.16%.
Considerations other than life expectancy therefore need to be taken
account of and fine needle aspiration cytology, as a sole investigation,
followed by selective surgery, is emerging as an appropriate scheme for
patient management. This scheme has resulted in a reduction in the number
of radionuclide and ultrasound scans and the number of operations performed
in many clinics. In our own series, surgery was performed for removal of
neoplastic tissue in 40% of cases, compared with a predicted rate of 9.6%
if all solitary nodules have been removed. Al-Sayer et al. (36) and Hamberger
et al. (44) have reported a 25% reduction in frequency of operation for
single thyroid swellings and an increase in the proportion of operations
for neoplastic disease from 31 to 50% after routine use of aspiration cytology.
This approach not only spares many patients with benign disease unnecessary
surgery but leads to substantial financial savings. In centres where large
numbers of patients with solitary nodules are seen and where expert cytological
skills are available, fine needle aspiration cytology is to be recommended
for routine use in the diagnostic evaluation of such patients. In these
circumstances we would recommend that patients with malignant or suspicious
cytological findings, or where clinical suspicion is high, be referred
for selective surgery. Definition of an optimum management strategy does
depend, however, on the long term follow up of patients with negative results
by such centres.
Acknowledgements
We are grateful to Dr Jennifer Young, Senior Lecturer in Pathology for
her advice and expert interpretation of cytological findings. We also acknowledge
the support and assistance of Mr GD Oates, Mr J Watkinson and Mrs J Daykin.