Thyroid and Parathyroid Neoplasms

Abstract

This chapter offers an overview of the epidemiology and aetiology, clinical and histological characteristics and classification, and staging of thyroid and parathyroid neoplasms before providing a more focused and illustrated description of the imaging findings. The indications and the role of ultrasound, ultrasound-guided fine needle aspiration, cross-sectional modalities and ¹⁸F-FDG-PET in the diagnosis and follow-up of thyroid cancer are discussed, with reference to the particular imaging characteristics of different tumour types. The embryology of parathyroid gland development is addressed and the relative advantages of ultrasound, scintigraphy, cross-sectional modalities and percutaneous venous sampling in the management of hyperparathyroidism are described.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Thyroid Neoplasms

1.1 Introduction

The primary role of imaging in thyroid cancer is to distinguish between benign and malignant nodules, with the emphasis on ultrasound (US) and US-guided fine needle aspiration cytology (US-FNAC). Palpable thyroid nodules are present in 3–7% of the population but the incidence of benign nodular thyroid disease detected by high resolution US is probably over 60%, being highest in elderly women. Biologically significant thyroid cancers (those over 10 mm in size or associated with lymphadenopathy) are rare and the prognosis is excellent. By comparison, microcarcinomas are common. The majority of these occult tumours are thought to be clinically insignificant, much in the same way that small prostate cancers appear to be in elderly men.

The investigation and management of thyroid nodules is controversial, but the ideal diagnostic pathway must be a balance between treating all biologically aggressive carcinomata and keeping the number of unnecessary diagnostic hemithyroidectomies and associated morbidity, anxiety and cost to a minimum. Pre-operative staging will depend on the histology of the tumour, but in the majority of cases it is not exhaustive. Imaging also has a role in the follow-up of patients treated for thyroid cancer.

1.2 Epidemiology and Aetiology

Thyroid cancer accounts for ~1% of all malignancies and causes around 0.5% of cancer deaths. The incidence is greater in women, peaking at 25–35 years. Thyroid cancer is rarest in childhood and adolescence but tends to present at a more advanced stage. However, due to a lower incidence of benign thyroid nodules in men and children, the risk of any single nodule being malignant in these groups is increased (Richards et al. 2010).

As most thyroid cancers are indolent in nature, their true incidence is likely to be higher than actually diagnosed; indeed, it is generally accepted that clinically occult papillary carcinomata under 1 cm in size are much more common. In a cadaveric series, microcarcinomas were reported in up to 38% of autopsy cases where death was due to an unrelated cause (Harach et al. 1984). Advances in US and US-FNAC techniques have resulted in greater detection of these occult cancers and may explain why thyroid malignancy has reportedly increased 2.4-fold over the last 30 years while 20-year survival has remained unchanged over the same period (Kent et al. 2007).

There is a geographical variation in the incidence of thyroid cancer, which is 15 times as common in Iceland and Hawaii than Northern Europe, thought to be due to dietary and environmental rather than genetic factors. Thyroid cancer is most common where the diet is iodine-rich but there is an increased proportion of follicular and anaplastic carcinomata in regions where dietary iodine is deplete (Richards et al. 2010).

Familial thyroid cancers occur and there is a genetic predisposition in the multiple endocrine neoplasia (MEN) syndromes, dyshormonogenesis and other inherited conditions such as Cowden disease and Gardner syndrome. An increased incidence of malignancy in nodules coexisting with Graves’ disease has been suggested and Hashimoto’s thyroiditis is associated with primary thyroid lymphoma.

There is a linear dose–response relationship between irradiation and the development of cancer. This is especially significant if received at a young age or if low-dose radiotherapy has been used (such as was once given for conditions ranging from adenoidal hypertrophy to thymic enlargement and tinea capitis) and is associated with a 40-fold increased risk. Higher-dose radiation treatment, including external beam radiotherapy and ¹³¹I treatment for Graves’ disease, is not associated with cancer induction (Richards et al. 2010).

Other aetiological factors include alcohol consumption and pregnancy. It has been postulated that this is due to transformation of indolent microcarcinomas by unknown endogenous or environmental factors. It is uncertain whether follicular adenomata have a tendency for malignant transformation, although it seems unlikely, as these are common lesions whereas follicular carcinoma is rare.

1.3 Classification and Staging

Primary thyroid cancer can arise from several different cell types and histological classification is an important determinant of prognosis. There are four main types of thyroid cancer: tumours arising from thyroid cells, ranging from papillary and follicular to anaplastic; medullary cancer arising from parafollicular C cells; primary thyroid lymphoma (PTL); and rare tumours arising from the stromal cells of the gland, such as sarcoma (Table 1, Hedinger et al. 1988). The International Union Against Cancer (UICC) TNM staging definitions for thyroid cancer are detailed in Table 2 (Sobin et al. 2009). Separate stage groupings are recommended for papillary, follicular, medullary and anaplastic carcinomata. All anaplastic carcinomata are considered stage IV disease.

Table 1 WHO histological classification of thyroid cancer (Hedinger et al. 1988)

Table 2 UICC TNM classification of malignant tumours (Sobin et al. 2009)

1.4 Imaging at Diagnosis

1.4.1 Indications for Imaging

The primary role of imaging in the diagnosis of thyroid cancer revolves around the assessment of nodules with US and FNAC. High frequency linear array transducers give such high spatial resolution that nodules as small as 1–2 mm can be easily detected. Nodules over 4 or 5 mm in size can be characterised with both grey-scale and Doppler and if necessary biopsied for FNAC. Lymph node staging in proven cases of malignancy is also largely by US, although cross-sectional and radionuclide techniques come into play to delineate the extent of local invasion, detect distant metastatic disease, and in the assessment of tumour recurrence. Scintigraphy is largely reserved for the detection of nodal and distant metastases and recurrence.

It is not appropriate to screen for thyroid cancer unless there are clinical factors known to increase the risk of cancer. Clinical assessment (including thyroid function) is all that is required to reassure the majority of patients with palpable nodules, without the need for further investigation. The clinical indications for US are summarised in Table 3.

Table 3 Clinical Indications for US

1.4.2 Sonographic Characterisation of Thyroid Nodules

US can accurately locate, quantify and measure thyroid nodules but also allows for a degree of grey-scale characterisation. Much has been written about the ability of US to predict malignancy (Cappelli et al. 2007; Lin et al. 2005; Papini et al. 2002). The most specific grey-scale characteristic of thyroid cancer is microcalcification (86–96%). This is manifested by multiple small intranodular hyperechoic flecks with little or no posterior acoustic enhancement and is believed to correlate to psammoma bodies (Fig. 1a–c). As with all grey-scale features however, microcalcification has low sensitivity (26–59%) and therefore the overall diagnostic accuracy is reduced to around 85% (Cappelli et al. 2007).

Fig. 1

a–c Punctate calcification on thyroid US (arrow) in three different cases of papillary thyroid carcinoma

As no single US feature has a high enough positive predictive value for cancer to reliably dictate which nodules should undergo FNAC, research has been directed towards the use of multiple US criteria in various combinations (Cappelli et al. 2007; Papini et al. 2002). In general, these studies point to the same few features that are significantly more common in malignant nodules: microcalcification (Fig. 1a–c), irregular margins (Fig. 2), hypoechogenicity compared to background thyroid tissue, tall shape (antero-posterior/transverse diameter ratio >1) and solidity (Cappelli et al. 2005, Fig. 3). The finding of two or more suspicious characteristics significantly increases the risk of cancer, as does the presence of extracapsular extension (Fig. 4) and lymphadenopathy (Kim et al. 2002).

Fig. 2

Irregular, poorly defined nodular margins on thyroid US in a patient with macroscopically invasive follicular thyroid carcinoma

Fig. 3

Transverse thyroid US demonstrating papillary thyroid carcinoma with an antero-posterior dimension (thick arrow) greater than transverse (thin arrow)

Fig. 4

Extracapsular extension into the carotid sheath by an anaplastic carcinoma of the right lobe of the thyroid with thrombosis of the internal jugular vein (arrow). CCA common carotid artery. IJV internal jugular vein

Although the majority of cystic nodules are benign and the majority of cancers are solid, a small proportion (~15%) of papillary thyroid cancer (PTC) can be macrocystic (Hatabu et al. 1991; Kessler et al. 2003, Fig. 5). For this reason it is important to sample the solid component of complex solid/cystic lesions as well as the fluid. Chaotic intranodular vascular pattern (Fig. 6a, b) and defects in or an absence of a hypoechoic halo may also have some predictive value (Fig. 7a, b, Papini et al. 2002).

Fig. 5

Cystic papillary thyroid carcinoma on thyroid US with a mural thickening and irregularity demonstrating punctate calcification (arrow)

Fig. 6

Peripheral vascular distribution in a benign hyperplasic thyroid nodule on colour Doppler (a) and chaotic intranodular vascularity in a small follicular thyroid carcinoma on power Doppler (b)

Fig. 7

a, b Thyroid US of two different follicular lesions demonstrating hypoechoic halos (arrows). Both lesions were follicular adenomata at histology

The risk of any thyroid gland harbouring cancer is no longer thought to depend on the number of nodules present because, although the risk of cancer per nodule is lower in a multinodular goitre, it decreases in proportion to the number of nodules in the gland such that the overall rate of cancer per patient remains similar to that of a patient with a truly solitary nodule (Gharib et al. 2010). Nearly half of all clinically solitary nodules turn out to be multiple at US. As the risk of malignancy in a sonographically proven solitary nodule is highest, these should undergo FNAC (Leenhardt et al. 2002).

It is clearly established that nodule size is a poor indicator of cancer: around 30% of thyroid cancers are non-dominant within a multinodular goitre and the prevalence of cancer appears to be the same in palpable and non-palpable nodules at around 6% (Gharib et al. 2010). Strategies based on FNAC of palpable or dominant nodules are therefore inappropriate. Rather, cytological sampling of a multinodular goitre should be driven by sonographic appearances and often more than one nodule should be targeted.

1.4.3 US-Guided FNAC

US-FNAC of thyroid nodules and regional lymph nodes is simple to perform, has a low complication rate and causes minimal patient discomfort. There are no reported cases of seeding of malignant cells by FNAC of a thyroid cancer. Many differing techniques have been described for both acquiring and preparing FNA samples, but the technique adopted should take into account the reporting cytologist's preference (Baloch et al. 2008). US-FNAC is routinely performed without local anaesthetic at our institution, using either a 23- or 25- gauge venesection needle. Occasionally a spinal needle may be required for deeper lesions. No suction is applied and the material is drawn into the shaft of the needle by capillary action alone. This method produces a cellular sample with minimum haemodilution and a single pass is all that is required in most cases. In cases where suspicion of malignancy is high on US, a second pass is performed for calcitonin immunocytochemistry to aid differentiation between papillary and medullary thyroid cancer. Core biopsy is useful to distinguish between poorly differentiated carcinoma and lymphoma, but is otherwise rarely required.

The reported sensitivity and specificity of thyroid FNAC ranges from 65 to 98% and 72 to 100%, respectively, with a positive predictive value of 50–96% (Cross et al. 2009). Thyroid FNAC samples are designated to six cytological categories (described in Table 4) along with an indication of the implied risk of malignancy (Cross et al. 2009; Ali and Cibas 2010). For a sample to be deemed adequate for diagnosis it must contain at least six groupings of well-preserved thyroid epithelial cells consisting of at least ten cells per group. Inadequate samples are classified as Class 1 and result from poor technique, haemodilution or the aspiration of cystic material. The reported rate of inadequate samples ranges from 2 to 21% with a mean of 17%. Repeat aspiration will result in a diagnostic sample in 50–60% and a delay of 2–3 months is recommended, unless there is clinical concern of malignancy (Gharib et al. 2010). The management of nodules that yield repeatedly inadequate samples should be decided on a case-by-case basis.

Table 4 ACE/AME/ETA thyroid nodule guidelines nomenclature (Cross et al. 2009; Ali and Cibas 2010)

Class 3 samples contain follicular cells and include all follicular patterned lesions. If cells are abundant, with a paucity of colloid, this is suggestive of a follicular adenoma, but the possibility of a hyperplastic nodule, follicular variant papillary thyroid carcinoma or microinvasive follicular thyroid cancer (FTC) can only be excluded histologically, and diagnostic hemithyroidectomy is recommended (Gharib et al. 2010).

The decision to perform FNAC of a thyroid nodule should be a considered one because it commits the patient to a pathway with a significant chance of ending in hemithyroidectomy. This may be as high as 40% if surgery is performed whenever an FNAC sample is anything other than benign (Class 2). The ideal pathway is a balance between the chance of missing carcinomata of potential clinical aggressiveness and reducing the morbidity associated with unnecessary operations while keeping an eye on cost. Inconsistencies in opinion arise from the fact that not enough is yet known about how the diagnosis and treatment of microcarcinomas affect life expectancy and how the benefits of surgery compare to the risks.

The rising prevalence of microcarcinoma and the fact that nodal metastases, extracapsular extension and aggressive histological features are sometimes seen with such small tumours has led to the suggestion that all sonographically visible nodules should undergo FNAC (Richards et al. 2010). Debate over the best strategy for the investigation and management of thyroid nodules has evolved over the last decade with the publication of several consensus statements and guidelines. More recently there has been a shift in attitude with more authorities advocating a less aggressive approach in view of the prevalence of benign nodules, the low incidence of biologically significant thyroid cancer and the fact that papillary carcinomata are so indolent (McDougall and Camargo 2007). Indeed in some institutions, including my own, a size threshold of 8–10 mm is applied to nodules with suspicious sonographic features below which US-FNAC is not performed, although patients are kept under regular sonographic review. The most recent guidelines from the American Association of Clinical Endocrinologists, Associazione Medici Endocrinologi and European Thyroid Association are recommended and provide a clear, comprehensive, evidence-based review which is beyond the scope of this chapter (Gharib et al. 2010).

1.4.4 Lymph Node Staging

US is the technique of choice for the assessment of cervical lymph nodes metastases in thyroid cancer. It has the highest spatial resolution of any imaging modality, allows assessment of internal nodal architecture and vascularity, and can be used to guide FNAC. Furthermore, both primary tumour and nodes can be evaluated at the same sitting. Metastatic nodes from differentiated thyroid cancer tend to be round and hypoechoic, with loss of the normal echogenic hilum and deranged vascularity. Nodal metastases in both papillary and medullary thyroid cancer (MTC) commonly demonstrate punctate calcification (Fig. 8a–c), and cystic components are a feature in 20% of metastatic nodes in PTC (Fig. 9). This may be so dominant as to be misinterpreted for a benign cervical cyst such as a branchial cleft cyst (King 2008).

Fig. 8

a–c Three examples of metastatic cervical lymph nodes from papillary thyroid carcinoma demonstrating punctate calcification (arrows)

Fig. 9

Large metastatic cervical lymph node from papillary thyroid carcinoma which is largely cystic with a small mural nodule demonstrating punctate calcification (arrow)

1.4.5 Magnetic Resonance Imaging and Computed Tomography

Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) characterisation of thyroid nodules and lymph nodes is poor. By comparison, these techniques are invaluable for the assessment of large tumour volume, retrotracheal and retrosternal extension, retropharyngeal and retrotracheal lymphadenopathy and the detection of local soft tissue and vascular invasion (Figs. 10 and 11). They also have a key role in the assessment of distant metastases to the lungs, liver and spine and in planning external radiotherapy (Miyakoshi et al. 2007; King 2008).

Fig. 10

Axial contrast-enhanced CT of the thyroid demonstrating a large thyroid mass involving both lobes, with extracapsular extension and left internal jugular vein invasion (arrow). The trachea is displaced and the oesophagus is inseparable from the mass (star) which was an anaplastic carcinoma at histology

Fig. 11

T1-weighted axial MRI with gadolinium enhancement through the thyroid gland demonstrating an anaplastic carcinoma of the right thyroid with extracapsular extension and oesophageal invasion (arrow)

CT has greater spatial resolution than MRI over large volumes, is less prone to movement and pulsation artefact, and is the modality of choice for the assessment of mediastinal and pulmonary metastases. However, iodine-containing contrast temporarily blocks iodine uptake by follicular cells and will therefore delay subsequent radio-iodine scanning or therapy by at least 6 weeks. For this reason MRI is often the imaging modality of choice. CT remains the most useful adjunct to US for non-iodine avid tumours such as anaplastic carcinoma (AC) and medullary thyroid cancer, or when recurrent disease has become too poorly differentiated to take up iodine (Richards et al. 2010).

Thyroid nodules are common incidental findings on CT and MRI and have an uncertain risk of malignancy. A recent retrospective review of 122 patients with incidental thyroid nodules detected on CT suggested the incidence of malignancy or potential malignancy was around 11% (Shetty et al. 2006), but there is no consensus on how to manage these patients. At our institution, size, age, thyroid function and other clinical risk factors are taken into consideration before committing patients to further work-up. Nodules under 1 cm are observed clinically.

1.4.6 Nuclear Medicine

The most widely used thyroid imaging agents are ¹²³I and ¹³¹I, which are organified in thyroid follicles and therefore highly specific for thyroid tissue. Scintigraphy plays little role in the initial diagnosis of primary thyroid tumours, but radionuclides are useful in the detection of nodal and distant metastases and recurrence. Sensitivity depends on tumour volume and histology, and is reduced by increased total body iodine from diet, the use of intravenous iodinated contrast and some drugs such as amiodarone and carbimazole. ¹²³I produces the best whole body image quality in follow-up studies and is most sensitive for differentiated thyroid cancer metastases (Sarkar et al. 2002, Fig. 12). ¹³¹I plays a major role in treatment.

Fig. 12

¹²³I whole body scintigraphy demonstrating widespread metastatic and nodal metastases from follicular thyroid carcinoma

A different spectrum of radioisotopes is required to assess medullary thyroid cancer. ¹¹¹In-labelled octreotide is the agent of choice (Fig. 13) and is excellent for demonstrating pulmonary metastases, but it is less sensitive for liver and lymph node metastases and false-positive uptake is seen in areas of inflammation and granulomatous disease.

Fig. 13

¹¹¹In-labelled octreotide scintigraphy abnormal uptake in a thyroid bed recurrence in a patient with medullary thyroid carcinoma (arrow).

Higher average 2-[F-18] fluoro-2-deoxy-d-glucose (¹⁸F-FDG) uptake is seen in thyroid malignancies, but there is too great an overlap with benign adenomata for positron emission tomography (PET) to be useful in differentiating malignant from benign nodules (Robbins and Larson 2008). It has been shown to be useful in staging MTC and poorly differentiated thyroid cancers that are iodine non-avid (Fig. 14a–f). The reported sensitivity and specificity for localising MTC metastases when calcitonin levels are raised is 73–88% and 79–80%, respectively (Diehl et al. 2001; Zettinig et al. 2001; Finkelstein et al. 2008). Incidental focal thyroid uptake of both ¹⁸F-FDG and ^99mTc-sestamibi has a high risk of malignancy and these patients should be further assessed with US and FNAC (Eloy et al. 2009; Van den Bruel et al. 2002).

Fig. 14

¹⁸F-FDG-PET in anaplastic carcinoma demonstrating the primary thyroid lesion (arrow) and widespread metastases (a, b), in recurrent papillary thyroid carcinoma (arrow) (c) in a patient with iodine non-avid disease on ¹²³I scintigraphy (arrow) (d) and in a patient with pulmonary metastases (arrow) (e) from iodine non-avid follicular carcinoma on ¹²³I scintigraphy (f)

1.5 Clinical, Imaging and Histological Features, Treatment and Prognosis

1.5.1 Benign Thyroid Pathology

Typical imaging appearances of the various thyroid malignancies are described below but it is also important to mention benign nodular hyperplasia, as the vast majority of thyroid nodules will be hyperplastic. Histologically hyperplastic nodules reflect waves of focal hyperplasia and involution of thyroid tissue and are often filled with microcysts and colloid-laden follicles. Their sonographic appearance is very variable. Most often they are well-defined hypoechoic or isoechoic lesions with a spongiform texture manifested by multiple echogenic striations (Fig. 15). Frank cystic or haemorrhagic degeneration is common (Fig. 16a–c). The presence of colloid indicated by echogenic foci with comet-tail artefact is reassuring and has been reported to be 100% specific for benignity (Fig. 17a, b, Wong and Ahuja 2005 ).

Fig. 15

Longitudinal thyroid US demonstrating multiple well-defined thyroid nodules with a spongiform texture with linear striations (arrow), typical of benign nodular hyperplasia

Fig. 16

Three different examples of haemorrhagic thyroid nodules demonstrating internal echogenic material (arrow) (a), septa (arrow) (b) and fluid/debris levels (arrow) (c)

Fig. 17

Ring-down artefact from colloid (arrows) in a solid hyperplastic nodule (a) and a colloid cyst (b)

Hashimoto’s thyroiditis is associated with PTL and it is important to recognise, as it is often subclinical. Lymphocytic infiltration results in a large hypoechoic and often hypervascular gland in the acute phase. Hashimoto’s thyroiditis is generally a diffuse process and multiple echogenic linear striations are characteristic (Fig. 18). Coexisting focal nodularity in Hashimoto’s thyroiditis should be assessed carefully. If in doubt, the presence of lymphocytes on FNAC is reassuring (Cross et al. 2009).

Fig. 18

Midline transverse US of the thyroid gland demonstrating a diffusely enlarged hypoechoic gland with linear striations (arrow) and a small reactive level VI lymph node (star) typical of Hashimoto’s thyroiditis

Adenomata account for around 10% of thyroid nodules, are typically follicular in cell type and solitary. They are characteristically homogeneous with a hypoechoic halo and are usually isoechoic to background parenchyma (Fig. 7a, b). They can undergo central cystic change (Fig. 19) and may contain foci of coarse calcification. Follicular adenomata can only be differentiated from microinvasive FTC histologically and are therefore treated by a diagnostic hemithyroidectomy (Gharib et al. 2010).

Fig. 19

Well-defined isoechoic lesion with two central cystic foci, which was found to be a follicular adenoma at histology

1.5.2 Papillary Thyroid Carcinoma

PTC is the most common and least aggressive form of thyroid carcinoma. It accounts for around 85% of thyroid malignancy but causes only 0.3% of deaths from thyroid cancer, with a 20-year survival rate of over 90%. PTC has a bi-modal age distribution peaking at 25–30 years and then again at 50–60 years. It affects women three times more than men. Approximately 10–20% of PTC is multifocal. 5% of PTC is familial. It is more common in iodine-rich regions and is the main cancer induced by irradiation in childhood (Richards et al. 2010).

The characteristic feature of PTC is its indolent nature. It typically presents as a neck mass either from the primary tumour or a metastatic lymph node. 25% of cases in young adults and children present with lymph node enlargement. Despite this propensity to metastasise to nodes, regional nodal involvement has no bearing on overall prognosis in patients under the age of 45 years and small nodes may harbour cancerous cells for many years without further spread. Distant metastases occur in 5% of cases, mainly to the lungs and very rarely to bone, liver and brain (Richards et al. 2010).

PTC is typically solid and hypoechoic at US and may show ill-defined margins, tall shape and chaotic intranodular vascularity (Figs. 2, 3 and 6a, b). Although the majority of PTC tumours are solid, a small proportion (~15%) can be macrocystic (Fig. 5). Microcalcification is highly specific but only seen in about 40% (Fig. 1a–c). Metastatic lymph nodes also demonstrate microcalcification and are often cystic (43–70%) (Figs. 8a–c and 9). Occasionally, large, predominantly cystic nodal metastases are seen, often in conjunction with small primary tumours. Extensive punctate calcification may be detectable on CT, and cystic metastases from PTC are characteristically hyperintense on T1-weighted MRI due to a high thyroglobulin content or haemorrhage (King 2008, Fig. 20a, b).

Fig. 20

a, b Coronal T2 and axial T1-weighted MRI of the neck demonstrating large right-sided metastatic cystic lymph nodes (stars) from a papillary thyroid carcinoma (arrow). The nodes are hyperintense on T1-weighted imaging, which is characteristic

Histologically, PTC can be encapsulated or locally infiltrating. Microscopic features include small papillae, ‘ground-glass’ or overlapping ‘shingle-roofed’ nuclei, small, spherical, calcified psammoma bodies and a desmoplastic reaction. Macroscopic calcification and necrosis may occur. Most PTC tumours have some follicular elements which may sometimes almost completely overshadow the papillary formations, but this does not affect prognosis. Several other variants of PTC have been described, including a follicular variant which can be diffuse sclerosing (occurs in children and young adults and has a good prognosis) or diffuse aggressive, an oxyphil cell variant which has biological behaviour closer to follicular carcinoma, and tall cell variant which predominates in older patients and has a more aggressive biological behaviour (Hedinger et al. 1988).

Treatment is by total thyroidectomy and central compartment neck dissection with removal of any sonographically abnormal cervical lymph nodes, followed by radio-iodine ablation. The prognosis in PTC is almost always excellent but is age-related and so, unlike other head and neck cancers, staging includes patient age as well as anatomical extent. In children and the young or early middle-aged, even with relatively advanced primary disease or lymph node metastases, a 100% cure rate can be achieved. In the older age group, the prime determinants of prognosis are tumour size and extent and the presence of distant metastases. Large tumours or extension outside the thyroid gland indicates a poor prognosis in older patients but has no impact on prognosis in younger patients. Similarly, pulmonary metastases are of concern in the elderly but not necessarily so in the young (Richards et al. 2010).

1.5.3 Follicular Thyroid Carcinoma

FTC accounts for 5–15% of thyroid malignancies and is three times more common in women, with a mean age of presentation of 50 years. The relative incidence of FTC is higher in iodine-deficient areas. FTC is associated with dyshormonogenesis, an uncommon cause of congenital hypothyroidism, and with Cowden’s syndrome. Familial cases occur but are rare (Richards et al. 2010).

FTC is also an indolent tumour, but unlike PTC it tends to spread haematogenously to bone, liver or lungs rather than metastasise to regional lymph nodes. It is rarely multifocal. Patients usually complain of a symptomless neck mass, although large tumours may cause symptoms of local compression. 25% of patients with FTC will present with extrathyroidal invasion, 5–10% with cervical lymphadenopathy and 10–20% with distant metastases (Richards et al. 2010).

FTC most closely resembles normal thyroid tissue and when well differentiated can be impossible to distinguish from a follicular adenoma on US or FNAC, which typically yields hypercellular samples devoid of colloid (Fig. 21). The diagnosis of microinvasive FTC requires histological confirmation of vascular invasion by tumour and is therefore often an unexpected finding following hemithyroidectomy for presumed benign disease. Microinvasive FTC is relatively slow-growing and rarely metastasises.

Fig. 21

Longitudinal thyroid US demonstrating a well-defined isoechoic lesion (arrow) which was a microinvasive follicular thyroid carcinoma at histology

Macroscopically invasive FTC is more aggressive and can be diagnosed on cytology. This less well differentiated form typically contains follicular cells, but Askanazy, Hürthle or clear cells may also be found and tumours are subclassified according to the predominant cell type. Macroscopically invasive FTC appears heterogeneous on US, with an irregular margin and absent halo. Frankly invasive FTC is usually poorly differentiated and associated with the worst prognosis. On US it is hypoechoic and irregular and often demonstrates extracapsular spread and vascular invasion, which makes it indistinguishable from AC and PTL (King 2008, Fig. 22).

Fig. 22

Transverse US of the right lobe of the thyroid demonstrating an irregular heterogenous hypoechoic mass which was a macroscopically invasive follicular thyroid carcinoma at histology

The treatment of pre-operatively confirmed FTC is the same as for PTC. While FTC has a good prognosis, it is less favourable than that of PTC. Prognosis depends on the degree of capsular and vascular invasion, the degree of tumour differentiation, tumour size and the presence of distant metastases. Once again, age is an important prognostic factor and is incorporated into the staging groups. Like PTC, the presence of lymph node metastases does not worsen prognosis in patients younger than 45 years but is significant in older patients (Richards et al. 2010).

1.5.4 Medullary Thyroid Cancer

MTC accounts for 3–9% of thyroid cancers and around 13% of thyroid cancer-related deaths. There is a slight female predominance. MTC is a neuroendocrine tumour which arises from parafollicular C-cells and may secrete a number of compounds, usually calcitonin. 75% of MTC is the sporadic form which occurs mainly in the fifth or sixth decade and is usually unilateral. The other 25% of MTC patients have a hereditary form, of which 65% occur in the context of MEN type II. The remainder of cases have a non-MEN familial form of the disease. All types of hereditary tumours are commonly multifocal (Richards et al. 2010).

MEN IIa is the most common form of hereditary MTC. Tumours occur in virtually every case with a peak age of incidence in the second or third decade. These patients are prone to other associated conditions such as phaeochromocytoma (<50%) and parathyroid hyperplasia (<50%). In MEN IIb the disease occurs in early childhood and is very aggressive. In these patients there is an association with phaeochromocytoma, mucosal neuromas, diffuse ganglioneuromas of the gastrointestinal tract, skeletal abnormalities and a ‘Marfanoid habitus’. Non-MEN familial MTC usually arises in the fifth decade of life and is more indolent biologically (Richards et al. 2010).

Patients typically present with hard, painless nodular thyroid enlargement accompanied by cervical and mediastinal lymphadenopathy in 50%. 10–15% will also have distant metastases to lungs, liver and bone at presentation. In advanced tumours, symptoms from local compression and vascular invasion may be present. Excessive or ectopic hormone production may give rise to diarrhoea and flushing. Hereditary cases are usually diagnosed by screening with US and serum calcitonin.

MTC is usually similar in appearances to PTC on US, although punctate calcification is seen in the majority (80–90%), in this case due to amyloid formation (Fig. 23a, b). Microcalcification is also present in 50–60% of metastatic lymph nodes (Fig. 24) and may also be demonstrated on CT (Fig. 25, King 2008). Macroscopically, tumours are often well demarcated but only occasionally encapsulated and have a firm consistency. Calcification, necrosis, haemorrhage, cyst and bone formation are common. Microscopically MTC is a solid cellular tumour without follicles, and a variable amount of stromal amyloid due to calcitonin deposition is seen in upto 82% of cases. Inherited MTC is often smaller than sporadic because it is found by screening, and is accompanied by C-cell hyperplasia (Richards et al. 2010).

Fig. 23

Transverse US of the right lobe of the thyroid demonstrating a small irregular hypoechoic nodule (arrow) in the mid-pole, with echogenic foci representing punctate calcification (a). Histology confirmed medullary thyroid carcinoma. There is also parathyroid enlargement (star) in this patient with MEN IIa syndrome (b)

Fig. 24

Transverse US of the right deep cervical chain demonstrating an enlarged round metastatic lymph node from medullary thyroid carcinoma with punctate calcification (arrow)

Fig. 25

Axial contrast-enhanced CT of the neck demonstrating a metastatic right level II lymph node with a small calcified focus (arrow) from a medullary thyroid carcinoma

Surgery is the only treatment for MTC and patients require careful staging. Total thyroidectomy with routine central and bilateral modified neck dissection is recommended, unless there is evidence of distant metastases. External beam radiotherapy has been used for palliation, but radioactive iodine has no place in treatment.

Despite the propensity to metastasise, MTC is relatively indolent, with a survival of 86% at 5 years and 65% at 10 years and, when confined to the thyroid gland, the prognosis is excellent. Prognostic factors include primary tumour size, the presence of lymph node and distant metastases, advanced age, association with MEN IIB and completeness of the surgical resection. Patients with the best prognosis are those who are diagnosed by provocative screening, prior to the appearance of palpable disease. All patients with MTC, including those with sporadic presentation, should be tested for RET proto-oncogene mutation and, if positive, family members should also be tested. Those with the gene should undergo prophylactic thyroidectomy at an early age (Richards et al. 2010).

1.5.5 Undifferentiated (Anaplastic) Thyroid Carcinoma

AC is rare, representing only 2% of all thyroid malignancies, but it is highly aggressive, with a median survival of only 6 months, and causes the majority of deaths from thyroid cancer. It is slightly more common in females and usually presents after the fifth decade of life. Radiation is a recognised causative factor, reducing the age of presentation (Richards et al. 2010).

AC presents as a rapidly expanding, ill-defined hard neck mass often associated with symptoms and signs of local invasion and compression. Cervical lymphadenopathy is common and half of the patients have distant metastases at presentation. US appearances of AC are of a diffuse, ill-defined heterogeneous mass which usually involves the entire thyroid (Fig. 4). Necrosis and extracapsular spread are common, with invasion of adjacent tissues and vessels often requiring cross-sectional imaging to fully demonstrate (Figs. 10 and 11). ¹⁸F-FDG-PET is useful for staging metastases (Fig. 14a, b).

The important differential diagnosis is of PTL, which sometimes requires histological assessment. AC is likely to arise as a result of transformation of a pre-existing PTC or FTC and areas of more differentiated thyroid carcinoma are often present at histology. Anaplastic transformation of differentiated carcinoma may also occur in metastases.

All patients with AC are considered to have stage IV disease (Sobin et al. 2009). Surgery with adjuvant chemoradiotherapy may be curative for smaller lesions, but treatment is usually palliative. External radiation with dexamethasone cover may slow tumour progression.

1.5.6 Primary Thyroid Lymphoma

PTL is another rare tumour representing 2% of all thyroid malignancies. It occurs most frequently in elderly women and is almost exclusively non-Hodgkin’s lymphoma of a B-cell type. Eighty-five percent of cases occur on a background of Hashimoto’s thyroiditis. It is postulated that chronic antigenic stimulation secondary to the autoimmune thyroiditis may lead to lymphoid proliferation which eventually undergoes transformation to a low-grade MALT lymphoma, with further transformation to large-cell lymphoma, which is the most common type seen clinically (Hedinger et al. 1988; Untch and Olson 2006).

Patients with large-cell thyroid lymphoma usually present with a rapidly growing thyroid nodule, frequently associated with lymphadenopathy, and symptoms of extracapsular spread. Sonographically there is usually a hypoechoic mass with poorly-defined margins and an absent halo, and evidence of Hashimoto’s thyroiditis in the rest of the gland. The nodule is usually solitary but multifocal disease occurs. Diffuse or multifocal PTL is sonographically indistinguishable from AC on both US and cross-sectional imaging (Fig. 26a, b). Patients with the rarer and more indolent MALT lymphoma present with a slow-growing nodule. MALT lymphoma can coexist with large-cell lymphoma and, if too small a sample is obtained, the aggressive component could be missed, resulting in incorrect treatment (Richards et al. 2010).

Fig. 26

Transverse thyroid US of a hypoechoic mass with posterior extracapsular spread extending into the retrotracheal region (a). Axial contrast-enhanced CT of the same patient demonstrating the left-sided thyroid mass with extracapsular spread (arrow) (b). Lymphoma was confirmed at histology

Treatment is based on the lymphoma subtype and the extent of disease and follows protocols for NHL occurring in a nodal site. Primary MALT lymphoma is often treated with radiotherapy alone. Prognostic factors include age, performance status, elevated LDH, the number of extranodal sites involved and the Ann Arbor staging. The 5-year survival rate for patients was 50–86% (Richards et al. 2010).

1.5.7 Other Primary Tumours and Thyroid Metastases

Other primary tumours are rare and include sarcomas, squamous cell carcinoma and primary lymphomas arising from outside the thyroid. At autopsy, 17–24% of patients with known malignancy have thyroid metastases most commonly from renal, melanoma, breast and lung cancers, but these are usually only present when there is other metastatic disease (King 2008). Thyroid metastases are typically large, well-defined, multiple, markedly hypoechoic, found in the lower poles of the gland and associated with cervical lymphadenopathy (Fig. 27). The search for thyroid metastases has increased with the advent of ¹⁸F-FDG-PET, which detects incidental thyroid pathology in 2–3% of cases; the reported risk of malignancy is 14--50%, usually primary thyroid cancers (Untch and Olson 2006).

Fig. 27

Longitudinal thyroid US demonstrating a hypoechoic lesion in the inferior pole (callipers) which was metastasis from squamous cell carcinoma of the lung

1.6 Surveillance

Approximately 10–30% of patients develop recurrent disease, around 80% of which is confined to the neck, with 20% developing distant metastases, most commonly in the lungs. Post-treatment surveillance of thyroid cancer patients revolves around clinical examination, serological assessment of thyroglobulin or calcitonin, and regular chest X-rays. US and radioisotope studies are employed where recurrence is suspected, the latter also yielding information about therapy options. There are several pitfalls, however. Re-growth of normal thyroid tissue may be misinterpreted as recurrence. Stitch granulomata can have very similar appearances to PTC on US, being hypoechoic, with coarse echogenic foci which cast acoustic shadows. Finally, it is not uncommon for small metastatic cervical lymph nodes from PTC to persist for many years after ablation therapy (King 2008).

Both PTC and FTC may de-differentiate with time, losing the ability to accumulate iodine. At this stage there are no contraindications to iodinated contrast and ¹⁸F-FDG-PET becomes a viable alternative to scintigraphy (Fig. 14c–f). ¹⁸F-FDG-PET alone has been reported to be 70–95% sensitive, 25–80% specific and 64–90% accurate in these patients (Lind et al. 2003; Finkelstein et al. 2008).

2 Parathyroid Neoplasms

2.1 Introduction

Primary hyperparathyroidism (PHPT) usually results from a solitary benign parathyroid adenoma, although multiple parathyroid adenomata, parathyroid hyperplasia, rarely parathyroid carcinoma and very rarely ectopic parathormone secretion may be the cause. The need to differentiate between these pathologies, the variability in number and position of parathyroid glands and the morbidity associated with second-look surgery means that pre-operative radiological localisation has an important role. This is even more so with the recent interest in minimally invasive surgical techniques.

2.2 Anatomy and Embryology

Normal parathyroid glands are ellipsoid and around 5 mm in length. They have abundant fatty stroma, making them difficult to detect by most imaging modalities. Around 90% of people have two pairs of glands, although the total number of glands can range from 2 to 6. In parathyroid hyperplasia, activation of embryological rests leads to the development of more supernumerary glands.

Superior parathyroid glands develop with the thyroid from the fourth branchial pouch and migrate caudally during fetal development. Their final position is relatively constant, deep to the upper lobes of the thyroid in around 95%. Inferior parathyroid glands develop from the third branchial pouch and migrate inferiorly with the thymus to lie within a few centimetres of the inferior poles of the thyroid in most people. Their final position is determined by when they disassociate from the thymus and is much more variable with ectopic inferior parathyroid glands occurring in up to 25%. They can lie anywhere between the carotid bifurcation and the mediastinum. Rarer ectopic locations include the retro-oesophageal, prevertebral and pretracheal spaces, within the thyroid or thymus, the aortopulmonary window or even in the pericardium (Johnson et al. 2007).

2.3 Epidemiology and Aetiology

PHPT is the most common endocrinological disorder. Symptomatic PHPT occurs in 1 in 500 women and 1 in 2000 men per year, usually presenting in their fifth to seventh decade, but asymptomatic PHPT detected on biochemical screening is probably even more prevalent. PHPT results from a single parathyroid adenoma in around 90%. The second most common cause is parathyroid hyperplasia, which occurs in around 6% and in which all the parathyroid glands are enlarged. Multiple, usually double, parathyroid adenomata occur in around 4%, with parathyroid carcinoma seen in around 1%, usually in a younger age group (Bilezikian et al. 2009).

The aetiology of parathyroid neoplasia is poorly understood, although previous external beam irradiation of the neck has been implicated. In most cases PHPT is sporadic, but there are both familial and genetic predispositions. Familial isolated hyperparathyroidism and familial hypocalciuric hypercalcaemia are rare diseases affecting multiple glands. Parathyroid hyperplasia is associated with MEN I and IIa. Parathyroid adenomata occur in MEN I but an association with parathyroid carcinoma is unproven. Adenomata and carcinomata are seen in the rare hyperparathyroidism-jaw tumour syndrome, 80% of which present with hypercalcaemia which is due to parathyroid carcinoma in 15% (Richards et al. 2010).

Secondary and tertiary hyperparathyroidism occurs in chronic renal failure when multi-gland hyperplasia or autonomous parathyroid adenomata develop in response to hypocalcaemia.

2.4 Clinical Features

Benign parathyroid enlargement per se is rarely symptomatic, as patients present with the signs or complications of hypercalcaemia such as polyuria, polydypsia and fatigue or renal stones, gastrointestinal ulcers, pancreatitis and bone disease.

The treatment of symptomatic PHPT is surgical and there are clearly-defined criteria for when to operate (AACE and AAES position statement 2005). The management of asymptomatic PHPT is more controversial but depends on calcium levels, bone mineral density, age and patient choice (Bilezikian et al. 2009). In experienced hands, surgery has a 90–95% cure rate in patients with a naive neck, preoperatively diagnosed with a single adenoma. Double parathyroid adenomata are usually bilateral and require both sides of the neck to be explored. Parathyroid hyperplasia also requires bilateral neck dissection and either subtotal parathyroidectomy, removing 3.5 glands, or total parathyroidectomy and autotransplantation of parathyroid tissue into the forearm to allow easy access, should the hyperparathyroidism recur (Johnson et al. 2007).

Parathyroid carcinoma is indolent but can be locally invasive and metastasise. Symptoms such as pain, dysphagia and dysphonia are concerning, especially if associated with very high calcium or parathormone levels. Neither tumour size nor regional lymph node status is prognostically important, but distant metastases are associated with reduced survival. Primary and palliative treatment is surgical, with en-bloc resection including hemithyroidectomy and neck dissection when appropriate. Recurrence can be delayed and is reported in 30–70%. Five-year survival is around 50%, with patients usually dying from the complications of hypercalcaemia (Hedinger et al. 1988).

2.5 Classification and Staging

Parathyroid tumours arise from chief cells, transitional clear cells or can be of mixed cell type. The histological distinction between adenomata and carcinoma is difficult. Standard criteria for malignancy, such as mitoses and capsular invasion may be present in adenomata, as can macroscopic adhesions to surrounding structures. Conversely, carcinomata are often well differentiated. In general, clinical features and macroscopic behaviour are as important to the final pathological diagnosis as histology (AACE and AAES position statement 2005).

There are no staging criteria for parathyroid carcinoma, which is described as either localised, involving the parathyroid gland with or without local invasion, or metastatic. It most frequently metastasises to regional lymph nodes and lung, but may involve liver, bone, pleura, pericardium and pancreas.

2.6 Imaging

The role of imaging in PHPT is one of pre-operative localisation rather than diagnosis, as it is rarely able to differentiate between parathyroid hyperplasia, adenoma and carcinoma. Until recently the standard surgical technique was bilateral neck exploration with visualisation of all four parathyroid glands in order to exclude multi-gland hyperplasia or double adenomata. Pre-operative localisation was therefore reserved for patients undergoing redo operations. With the advent of focussed, minimally invasive techniques, pre-operative imaging is generally recommended in all patients whether or not they have had previous neck surgery (Bilezikian et al. 2009). The combination of scintigraphy and US is highly sensitive for single gland adenomata, but the pre-operative prediction of multi-glandular disease and double adenomata is less successful. Most centres therefore perform bilateral neck exploration whenever there are equivocal, negative or discordant results. Others advocate the use of US-FNA for parathormone (PTH) assay and percutaneous selective venous sampling to reduce the need for bilateral surgery. Cross-sectional imaging has a less central role in imaging parathyroid glands but can be useful in suspected ectopia and post-operative patients.

2.6.1 US, Scintigraphy and ¹⁸F-FDG-PET

US has advantages over scintigraphy in that it can accurately determine the anatomical relationship of an abnormal parathyroid gland to the thyroid, permits evaluation of coexisting thyroid and lymph node pathology and can be used to guide needles for FNA or percutaneous ethanol ablation when required. The reported sensitivity of US for a single parathyroid adenoma is similar to scintigraphy at around 80%, although this is operator-dependent (Haciyanli et al. 2003). As with scintigraphy, sensitivity falls to around 50% for multi-gland hyperplasia and even further for double adenomata to the region of 30% (Sugg et al. 2004).

Abnormal parathyroid glands are usually homogenously hypoechoic with an echogenic capsule and are oval in shape, with a craniocaudal axis, although they can be surprisingly elongated (Fig. 28a–c). Occasionally they can be isoechoic to thyroid and 4% undergo cystic or haemorrhagic degeneration. Typically they are hypervascular and reported to differ from lymph nodes by having a feeding vessel which enters the gland from one of the poles to supply it in a peripheral distribution, rather than the hilar architecture and more central distribution seen in nodes. When detected on power Doppler, the presence of such a polar feeding vessel has been reported to significantly increase the specificity of US for parathyroid enlargement (Rickes et al. 2003); however, in the author's experience, this has been an exceptionally rare finding (Fig. 29a–c).

Fig. 28

Parathyroid enlargement on US. a longitudinal US demonstrating the elongated shape of enlarged parathyroid glands often seen, b longitudinal US demonstrating a parathyroid adenoma with central cystic characteristics and c transverse US demonstrating parathyroid enlargement in tertiary hyperparathyroidism

Fig. 29

Power Doppler studies of three different enlarged benign parathyroid glands. a, b vascular supply from a single pedicle which is not polar and appears to arise from the thyroid gland (arrows) and c a more chaotic pattern of vascularity

Parathyroid adenomata are usually 0.8–3 cm in maximum diameter with a mean of 15 mm. Hyperplastic glands are typically 2–3 times smaller, but it is not uncommon for there to be a dominant gland in multi-gland hyperplasia, which can be misinterpreted as a solitary adenoma. Parathyroid carcinoma can be indistinguishable from adenomata but, when present, evidence of necrosis, haemorrhagic degeneration, deranged vascularity, calcification, local invasion and lymph node metastases are all concerning features (Fig. 30a, b). It is important to remember the association with MEN syndromes and assess the thyroid carefully for malignancy in every case of PHPT (Fig. 23a, b).

Fig. 30

Longitudinal (a) and transverse (b) US of the thyroid demonstrating an irregular heterogenous hypoechoic mass invading into the thyroid gland; this was parathyroid carcinoma at histology

There are a number of pitfalls to be considered when performing US for parathyroid enlargement. Ectopic glands are often beyond the reach of US. Thyroid nodules may cause so much thyroid enlargement that parathyroid glands are displaced or obscured. Thyroid nodules can be pedunculated and mimic parathyroid enlargement, as can the reactive infrathyroid pretracheal lymph nodes seen with autoimmune thyroiditis. Intracapsular and even rarer intrathyroid parathyroid glands can be indistinguishable from thyroid nodules, although the presence of an independent extrathyroidal vascular supply is supportive evidence (Johnson et al. 2007).

US may be used under other circumstances. There is a role for US in pre-operative marking as well as intra-operatively. Occasionally US-FNA may be performed for PTH assay and is highly specific for the presence of parathyroid tissue. Although usually performed to treat tertiary hyperparathyroidism, percutaneous ethanol ablation has a role in patients with PHPT where there are medical co-morbidities or surgery is particularly high-risk. Under US guidance, ethanol is slowly injected into the abnormal parathyroid until blood flow into the gland can no longer be detected on power Doppler.

^99mTc-sestamibi is the radioisotope of choice for imaging parathyroid glands. It is taken up by both thyroid and parathyroid tissue, but the latter is more avid and retained for longer by abnormal parathyroid glands than normal thyroid. These differences are exploited by the use of both initial and delayed ‘wash-out’ acquisitions. 3D single photon emission computed tomography (SPECT) is particularly useful in cases of ectopia (Fig. 31a–c). The concomitant use of ¹²³I or ^99mTc-pertechnetate, which are preferentially taken up by the thyroid, permits subtraction imaging, but this has yet to be proven to be diagnostically advantageous.

Fig. 31

^99mTechnetium-labelled sestamibi scintigraphy: Thyroid retraction study demonstrating bilateral foci of increased uptake representing double parathyroid adenomata (a); axial contrast-enhanced CT of the lower neck (b) and axial SPECT with CT fusion (c) in a patient with a retro-oesophageal ectopic parathyroid adenoma (arrows)

The sensitivity of ^99mTc-sestamibi scintigraphy varies between institutions but is similar to US at around 80%, the main advantage over US being the detection of ectopic glands, particularly in the mediastinum. As with US, sensitivity falls dramatically when a single adenoma is small (50% for glands weighing under 500 mg) or in the presence of multigland hyperplasia (44–73%) or double adenomata (30%) (Johnson et al. 2007). The reason why double adenomata are inherently difficult to detect is unclear and only partly explained by volume. Low mitochondrial activity has been suggested as a factor. False negative results may also occur when there is cystic or haemorrhagic degeneration within an abnormal parathyroid gland.

Interpretation of ^99mTc-sestamibi scintigraphy is complicated by the presence of thyroid disease which can lead to false-positive results. Certain types of thyroid nodule and chronic thyroiditis may take up and retain sestamibi in a similar way to parathyroid tissue. Uptake has also been described in follicular lymph node hyperplasia and chronic lymphadenitis.

Overall sensitivity can be slightly increased by using both US and scintigraphy when compared to either technique (Lumachi et al. 2000), but even when results are concordant, the detection of multigland disease and second adenomata is disappointing, having a reported positive predictive value of only 30% and a negative predictive value for single gland disease of only around 70% (Sugg et al. 2004; Rickes et al. 2003). For this reason, many institutions use intra-operative PTH assay during focussed minimally invasive surgery to reduce the risk of missing a double adenoma or multi-gland disease, even when pre-operative localisation is concordant.

If parathyroid carcinoma is suspected on US or when there are unusually high PTH levels, ¹⁸F-FDG-PET and octreotide scintigraphy are both helpful in staging and may be more sensitive than ^99mTc-sestamibi scintigraphy (Fig. 32a–d).

Fig. 32

Sternal metastasis from parathyroid carcinoma demonstrated on axial CT (a), ¹⁸F-FDG-PET (b) and octreotide scintigraphy (c) (arrows) but not apparent on 99mTc-sestamibi (d)

2.6.2 CT, MRI and Percutaneous Venous Sampling

CT or MRI, which may be co-registered with SPECT, provide cross-sectional anatomical detail which may help surgical planning, but studies are inconsistent as to the benefits in more routine cases. Most centres reserve these techniques for patients with non-concordant US and scintigraphy or where previous surgery has failed because the prevalence of supernumerary and ectopic glands is much higher in this group.

Contrast-enhanced MDCT is preferable to MRI because arterial-phase contrast enhancement can assist in differentiating vascular parathyroid tissue from lymph nodes and because there is less image degradation by pulsation artefact, which is particularly important when assessing the lower neck and mediastinum (Fig. 33a–c). The MRI signal characteristics of abnormal parathyroid glands are variable but typically follow those of lymph nodes, although occasionally there may be cystic or haemorrhagic features or evidence of fibrosis which allows them to be distinguished. Contrast-enhanced MRI is most advantageous when assessing patients with recurrent or persistent disease following parathyroidectomy, because of its ability to demonstrate parathyroid tissue within areas of post-operative fibrosis (Johnson et al. 2007).

Fig. 33

Axial contrast-enhanced CT images demonstrating an enlarged parathyroid gland with a vascular pedicle (arrow) (a) and two examples of enlarged retrotracheal ectopic parathyroid glands (b, c) (arrows)

Pre-operative percutaneous selective venous sampling is considered the gold-standard investigation for patients with recurrent disease following previous surgery and indeterminate results from other imaging modalities, with sensitivities as high as 90% being reported. False negative results are due to disruption of venous anatomy by previous surgery. The complication rate is low (Johnson et al. 2007).