Ultrasonography of the salivary glands: the role of grey-scale and colour/power Doppler M. Carotti1, A. Ciapetti2, S. Jousse-Joulin3, F. Salaffi2

1Departments of , and ABSTRACT aging (MRI) and the RM-scialography 2Rheumatology, Politechnic University of Over the last few years the use of ultra- (3). Over the years, each of these im- the Marche, Ancona, Italy; sonography (US) in the study of sali- aging techniques has acquired a well- 3Service de Rhumatologie, Hopital La vary glands is greatly increased due to defined role in the diagnostic procedure Cavale Blanche - Unité Echo-Doppler, Brest, France. its several advantages and undoubted (4, 5) and, currently, US represents the diagnostic potential in detecting even most frequently used due to its several Marina Carotti, MD Alessandro Ciapetti, MD minimal soft tissue changes in a wide advantages: lack of radiation, low cost, Sandrine Jousse-Joulin, MD range of pathological processes. repeatability, patient friendliness, high Fausto Salaffi, MD, PhD Nowadays, there is general agreement resolution (6). The US allows detailed Please address correspondence to: in considering US as a useful comple- morphological evaluation of major Prof. Fausto Salaffi, ment to other imaging techniques, such salivary glands and became comple- Clinica Reumatologica, as magnetic resonance, in providing mentary to clinical examination for Università Politecnica delle Marche, an accurate assessment of the sali- the study of the inflammatory process Ancona, c/o Ospedale “Carlo Urbani”, vary glands especially in the study of or expanding lesions (Fig. 1-2) (7). US Via dei Colli 52, tumour pathologies. US is also use- 60035 Jesi (AN), Italy. allows the evaluation of the presence, E-mail: [email protected] ful for the evaluation of inflammatory number and topographical position of Received on October 14, 2013; accepted in process affecting salivary glands (e.g. the lesions and, sometimes, the clari- revised form on November 15, 2013. Sjögren’s syndrome) where its accu- fication of their nature (8). The com- Clin Exp Rheumatol 2014; 32 (Suppl. 80): racy and feasibility make it a reliable bined use of high resolution grey-scale S61-S70. method. The useful combination of the and power/colour Doppler (P/CD) tech- © Copyright Clinical and US grey-scale and the colour/power niques and the use of contrast agent add Experimental Rheumatology 2014. Doppler technique provides more valu- useful information and are valuable in able details regarding the presence narrowing the differential diagnosis (9- Key words: and degree of soft tissues blood perfu- 11). US is also widely used as guidance ultrasonography, Sjögren’s syndrome, sion and may be valuable in narrowing for interventional procedures such as colour/power the differential diagnosis. This review aspiration and drainage of abscesses or provides an overview of the main US biopsy for cytological and histological findings observed in a wide range of examinations. In these cases, the use pathological processes that can affect of US as guidance makes the local pro- salivary glands. cedure safer and minimise the risks of complications (e.g. transitory paralysis Introduction of the branches of the or The multiplicity and the relevant diag- skin salivary fistula formation) (12). nostic difficulties of various disorders affecting a specific anatomical struc- Main diseases of the major ture, make it often appropriate to use salivary glands different imaging techniques to obtain Several diseases can affect the major a definite diagnosis. The study of sali- salivary glands and includes traumatic vary glands represents an example due injuries, congenital and inflammatory to the several pathological conditions diseases, benign and malignant tumours may involve them and the overlapping and autoimmune disorders (13). of various clinical features (1, 2). The scialography was the first diagnostic Trauma and salivary fistulas: salivary imaging technique to have been intro- glands, especially the parotid, being su- duced, followed later by direct radio- perficial structures can be damaged af- logical examination, telethermography, ter neck or face trauma. In this case, US , ultrasonography (US), allows the visualisation and the exten- computed (CT) and, more sion of fluid collection in the gland or in Competing interests: none declared. recently, by the magnetic resonance im- the surrounding tissues. US is useful in

S-61 US and colour/power Doppler of the salivary glands / M. Carotti et al.

the sublingual). US shows an increased size of the gland with preserved mar- gins and echogenicity (14, 15).

Pneumatocele: typical of glass blowers or trumpet, is characterised by the pres- ence of gas in the salivary ducts and, sometimes, by the formation of diver- ticula. US examination allows to detect multiple hyperechoic spots or thin lines in the contest of the ducts (14, 15).

Sialolithiasis: is the most common cause of recurrent swelling or the parot- id gland. The affects 1–2% of the population and the salivary gland more frequently involved includes the submandibular (80–94% of cases) and the parotid (6–20%). The stones in the Fig. 1. US scan of in healthy subject. To note the normal echostructure and homogenicity of parenchyma. have predomi- nantly intraductal position, while in the parotid are most frequently placed at the superficial lobe of the gland pa- renchyma and, rarely, at the level of Stenone’s duct. The majority of calculi (90%) are radio-opaque and they are multiple in 25% of patients (16). US has a high degree of accuracy (96%) for their evaluation (17). The stone appears as hyperechoic area with characteristic posterior acoustic shadow. If the stone is located in the main duct an hypoechoic widening of upstream ducts can be ob- served. This finding is best appreciated by stimulating salivation of the subject with lemon juice. US allows also to visualise an hypoechoic soft edges area surrounding the stone which is a sign of an acute inflammation process. The persistence of the obstruction may lead Fig. 2. US scan of submandibular gland in healthy subject. Note the normal echostructure and homo- to an abscess and, subsequently, to a genicity of parenchyma. pseudocyst (14, 15). the follow-up evaluation of haematoma able as anechoic sharply well-defined Inflammatory disease and leakage of saliva after drainage. lesions and are often associated with re- : is divided into acute and inforce of the posterior wall and lateral chronic process according to both the Salivary cysts: primitive genetic sali- acoustic shadows (14, 15). basis of pathogenesis and course of vary cysts are rare, whilst secondary Hypertrophy: usually painless and may the disease. Undoubtedly, the parotid salivary cysts are more common and, be due to constitutional factors (obe- is the most involved gland by sialad- usually, the consequence of ductal diver- sity), race (people of North African, enitis process (80% of cases). In acute ticula (e.g. due to inadequate drainage). Egyptian, etc.), malnutrition, metabolic process the use of US is recommended The sublingual and parotid gland are diseases (diabetes, hypo-hyperthyroid- whilst the sialography is contraindi- the most common involved sites (80% ism) and chronic degenerative diseases cated for the possibility to increase the of cases). Parotid cysts may also results (cirrhosis, spirochaetosis, uremic ca- inflammation. In the acute phase of the from an abscess collection or aseptic in- chexia, sarcoidosis). A secondary hy- inflammatory process the gland appear flammation due to gland stones. On US pertrophy can occur after surgical re- on US increased in its volume with the salivary cysts are easily recognis- moval of the salivary glands (especially contours not well-defined and hetero-

S-62 US and colour/power Doppler of the salivary glands / M. Carotti et al. geneous hypoechoic pattern due to the Neoplastic pathology can be observed (6, 14, 15). The main parenchyma oedema. Sometimes the Neoplastic disease of the salivary limitations of US are represented by the oedema is focal and makes it difficult glands, although sometimes dramatic, difficulty to visualise tumours placed in a differential diagnosis from a nodular is a fairly rare and represents almost the deeper region of the parotid or un- lesion. The evolution of an acute in- 5% of the total tumours affecting the der the part covered by the lower jaw flammation process to an abscess can head and neck. The most involved sites and to evaluate their extension to the frequently occur and is characterised by includes parotid (85%), submandibular para-retropharyngeal tract or the base the presence of liquid or semiliquid col- (12%) and sublingual (3%) glands (14- of the skull. Moreover, small malignant lections with lumpy or starry contours, 16). Usually the frequency of malignant and metastases (less than 20 which tend to burrowing to the surface tumours is greater for the submandibu- mm of diameter) and well-differenti- or into the deeper layers of the gland. lar glands (65–70%) (14, 15). ated malignant neoplasms, sometimes, A significant sign of the inflammation In the study of an expansive pathol- can not be easily distinguished on US is the thickening of the skin overly- ogy of salivary gland US has become a from benign tumours. Therefore, in sus- ing the gland due to an oedema of the priority technique imaging for its well pected neoplastic lesions is necessary to gland wall. In acute inflammatory pro- known properties (17, 18). The rec- resort to other imaging techniques, such cess an increased CD signal can be ob- ognition of a neoplastic process does as MRI which allows the evaluation of served due to blood perfusion. All acute not require clinic difficulties, as it is the margins of the lesion and its exten- inflammatory process may become always superficial and easily palpable. sion towards the surrounding anatomi- chronic with consequent fibrosis and The main problem is the differential cal structures (24). narrowing or stenosis of the ducts and diagnosis between benign and malig- upstream widening ducts. In these cases nant tumour and the evaluation of its Benign tumours: the most common be- CD signal can be visualised due to the extension. Typically, malignant tumour nign tumours of major salivary glands increased blood perfusion. The degree assumes nodular appearance with ir- are the pleomorphic adenomas (mixed of CD signal is related to the severity regularities of the margins, up to a clear tumour), the Warthin’s tumours (cys- of the inflammatory process and can be infiltration of the surrounding tissues. tadenolymphoma) and the oncocy- low when an increasing oedema occurs. The echotexture is inhomogeneous toma. Several typical features of these In the abscesses collections peripheral with hypo-anechoic or hyperchoic as- tumours can be observed on US but, CD signal can be observed. Usually, in pect due to necrotic-haemorrhagic or sometimes, the differential diagnosis chronic inflammation process the most calcification areas, respectively. The between benign and malignant tumours characteristic feature is represented presence of enlargement of regional can be difficult (20, 21). by the changes of ducts calibre in the lymphnodes is extremely important contest of the gland and the spectrum for the evaluation of a suspected ma- (mixed tu- can vary from saccular or rounded ap- lignant process. Further, the presence mour): is the most common tumour of pearance (“apple tree” at sialography) of metastatic-appearing lymph nodes the salivary glands and affect the pa- to wide widening ducts (as “salivary accompanying a tumour in the salivary rotid in 85-90% of cases. In the past it lakes”). At the level of the gland paren- gland strongly suggests a malignancy. was defined as mixed tumour for the chyma fine punctate calcifications and Most of the solid nodular lesions are presence of both epithelial and mesen- lack of homogeneity can be visualised characterised by an increased CD sig- chymal component secreting mucoid, on US. In the granulomatous process nal (80%). Usually, malignant tumours myxoid or chondroid basophilic mate- (tuberculosis, syphilis, actinomicosys) have greater CD signals with respect rial. Microscopically the structure of the granuloma is clearly visible on US to benign tumours (83% of cases). In the tumour is based on nests and cords as a non-homogeneous area with well- malignant tumours venous and arte- appearance, sometimes differentiated defined edges within of an hypo-ane- rial vessels intertwine within the le- in ductal and acinar structures with few choic gland. Often during the exacer- sion given the typical appearance of a liquid component. The pleomorphic bation of inflammation chronic process vascularised nodule with several poles term is usually attributed to an epithe- an increasing volume of the gland and and the absence of a peripheral vascu- lial adenoma characterised by several the presence of hypoechoic areas or larisation. On CD examination malig- pathological features. Usually the pleo- abscess collections can be observed. In nant tumours may show multiple and morphic adenoma affects females, with non-atrophic chronic process CD sig- irregular internal vessels, with hyper- maximum peak incidence between 35 nals are more frequently seen than in vascularity and high resistance arterial and 50 years old and it is described as normal conditions due to the presence flow (9, 22). Studies mainly performed potentially malignant tumour (4–5% of of small vessels not visible in their en- on the parotid gland have shown that a cases) with slow growth and infiltration tirety and surrounding widening ducts. resistive index (RI) of 0.8 and pulsatil- of surrounding tissues. The pleomorphic In the phase of acute exacerbation of ity index (PI) of 1.8 are suggestive for adenoma is often bilateral or multifocal, chronic inflammation blood vessels a benign lesion (23). Rarely (32% of non-adherent to the surrounding tissues shows high CD signal with irregular cases), vessels in common between the and has a typical capsulated nodular calibre and route (9, 14-19). nodule and surrounding parenchyma structure with well-defined margins. In

S-63 US and colour/power Doppler of the salivary glands / M. Carotti et al. the 10% of cases pleomorphic adenoma can shows uneven shapes or sinuous and irregular margins. Usually this tu- mour has a diameter of 2–3 cm, but if untreated can reach wider dimensions. On US the pleomorphic adenoma is eas- ily recognisable because of its peculiar homogeneous hypoechoic echostructure (in 98% of cases) with sharp and regular margins due to the presence of a cap- sule and, sometimes hyperechoic areas due to calcifications. In wider adenoma the margins tend to become irregular and anechoic areas, related to necrotic process, can be observed inside the le- sion (Fig. 3). In these cases US findings are consistent with those of a malignant tumour and, therefore, the presence of lateral cervical metastatic adenopathies is essential in narrowing the differen- Fig. 3. The US longitudinal scan of parotid shows a pleomorphic adenoma within the gland’s paren- tial diagnosis. In the 79% of cases the chyma. Note the lack of homogeneity and the presence of hypo-anechoic necrotic areas. pleomorphic adenoma shows colour pe- ripheral Doppler signal and a trend that follows the outer limiting of the nodule (“basket-like” flow pattern). Inside the lesion is easy to visualise a venous or, sometimes, arterial vascular pole (Fig. 4) (15, 18, 21). US may be used to guide fine-needle aspiration biopsy (25).

Warthin’s tumour (cystadenolymphoma): it represents 5–6% of all epithelial tu- mours. The monomorphic adenoma is the second most common type of be- nign tumour of salivary glands, and, predominantly, affects the parotid (es- pecially the lower pole). Warthin’s tu- mour is frequently seen in male adult- hood and occurs between the fourth to seventh decades (with a peak around the age of forty). It is often bilateral, multifocal and slow-growing and it appears as well rounded or lobulated, encapsulated, cystic-aspect lesion. Fig. 4. US longitudinal scan of parotid gland in a patient with pleomorphic adenoma. Observe an Lesions are multiple or bilateral in hypo-anechoic echostructure of the gland without necrotic process or calcifications inside the lesion 10–15% of patients. On US Warthin’s (a), the vascularisation of the tumour and the vascular pole detected by using contrast agent and PD tumour has an hypo- anechoic “liquid technique (b) and the corresponding contrast-enhanced curve (c). gaps” appearance. The hypo-anechoic appearance corresponds to the histo- in the liquid or semiliquid areas there adenolynphoma and , with logical picture characterised by large is no detectable CD signal (Fig. 5) (15). respect to all the other tumours, are cystic spaces filled with lymphocytes, Oncocytoma: is a rare lesion account- weakly positive at technetium-99m desquamated epithelium and by the ing for 1% of all salivary gland tumours pertechnetate scintigraphy. The onco- cavity and the papillary lesions pro- but similar to the adenolymphoma, usu- cytoma is a benign tumour that rarely truding into the lumen covered by ally is unilateral, painless and without relapse after surgical excision and the pseudostratified cylindrical epithelium. cystic appearance. It represents 1% of pattern of CD signal is similar to that The vascularisation is largely present the tumours of the salivary glands and one found in pleomorphic adenoma at level of the echogenic areas, whilst usually appears as painless lesion. Both (14, 15).

S-64 US and colour/power Doppler of the salivary glands / M. Carotti et al.

Malignant tumours: the most common malignant neoplasms occurring in sali- vary glands are mucoepidermoid car- cinoma and (cylindroma), and un- differentiated carcinoma, epidermoid or . Acinar cell adenocarcinoma and mesenchymal tu- mours () are less common. Less than 30% of focal lesions in the parotid gland are malignant, whereas al- most 50% of focal lesions in the subman- dibular gland are malignant (5, 20, 21).

Mucoepidermoid carcinoma: is the most common malignant tumour of the parotid, accounting for 80–85% of cas- es. Mainly affects females with a peak of incidence between the fourth to fifth decade. It does not have a capsule, has a rather slow growth and can reach a Fig. 5. US longitudinal of parotid gland in a patient with cystadenolymphoma tumour (Warthin’s size between 3 and 5 cm. Usually it is tumour). Note the vascularisation within the tumour detected by CD technique. a painless lesion and difficult to dif- ferentiate from a squamous cell carci- (which faster invades the nearby lymphnodes) and from the adenocarci- noma (which often contains blood or fluid collections within the lesion). It accounts for 5% of all salivary gland tumours. In most cases, the tumour has parotid location but, may be found in other intraoral sites such as the , the alveolar process and the retromolar region. More severe malignant tumours can quickly arise, be locally infiltrative and ulcerated and causing pain and nerve paralysis of the facial nerve, with tendency to metastasise locally or dis- tally and to relapse after surgical resec- tion in 15% of cases (14, 15). Adenoid cystic carcinoma (cylindroma): accounts for 20% and 5% of the sub- mandibular and parotid tumours, re- spectively, with a peak incidence be- tween the sixth to seventh decades. Fig. 6. US longitudinal scan of parotid gland in a patient with adenocarcinoma. Note the irregular Usually it is unilateral and has a slow and lack of continuity of the margins and inhomogeneity of the texture with typical hypo-anechoic growth with tendency to infiltrate the pattern (a). Multiple and irregular vessels are well visualised after administration of contrast agent (b) . facial nerve and to metastasise sur- The corresponding contrast-enhanced curve is visualised (c). rounding lymph nodes. Recurrence is common after . On US the ad- Adenocarcinoma and undifferentiated rapid growth, characterised by skin enoid cystic carcinoma has a typical carcinoma: represent 27% and 35% ulceration and/or facial paralysis and uneven and cribriform echostructure respectively of all salivary glands ma- tend to metastasise early. Both tumours and, sometimes, pseudocystic appear- lignant tumours and originate from the are not encapsulated, their margins are ance (14, 15). Contrast-enhanced MRI malignant transformation of pleomor- irregular and often discontinues and is required to demonstrate perineural phic adenomas which tend to degener- have homogeneous hypoechoic echo- extension which can not be well visu- ate in 4–5% of cases. Adenocarcinoma structure on US (Fig. 6) and haemor- alised on US (21, 24). and undifferentiated carcinoma have rhagic necrosis, blood lakes or cystic

S-65 US and colour/power Doppler of the salivary glands / M. Carotti et al. appearance within the lesion (14, 15, 21, 24).

Epidermoid or squamous cell carci- noma: is a rare tumour characterised by rapid growth, early local invasion and tendency to metastasise. Typical US features includes hypoechoic echo- texture with lakes blood and cystic aspects, not well-defined star-shaped margins and skin retractions (14, 15).

Acinar cell adenocarcinoma: is fairly rare, almost unilateral and rarely ex- ceeds 3 cm in diameter. Its echostruc- ture is similar to that one of adenoma, as it is capsulated and, sometimes, with some tiny cysts inside the lesion. It does not induce marked swelling of the gland or painful symptoms and, clini- Fig. 7. US longitudinal scan of parotid gland in a patient with . The tumour appears as an hypo-anchechoic intraglandular mass with a cystic appearance due to lymphomatous proliferation. cally, it is similar to benign tumours. Survival after radical surgery is 25% of the cases. Histologically it is similar tive disorder and MALT lymphoma is in the salivary glands: to the clear cell tumours of the kidney difficult and the use of histopathology uncommonly salivary glands are sites and the differential diagnosis between must be adviced. The lymphoepithelial of metastases. Primary tumours metas- adenocarcinoma and acinar cell metas- sialadenitis is characterised by a lym- tasising to salivary glands may be locat- tasis from hypernephroma is therefore phocytic infiltrate with follicular hyper- ed in the head and neck region, as well difficult, except in the presence of me- plasia which surrounds and infiltrates as in more distant parts of the body. tastases to regional lymphnodes (20% the salivary ducts, with disorganisation Usually they derived from squamous of cases) (14, 15). and proliferation of epithelial ductal cell carcinoma, malignant melanoma cells and formation of lymphoepithelial of the head and neck, nasopharynx car- Mesenchymal tumours: the lymphoid lesions. Usually, the lobular gland archi- cinoma, carcinoma of the thyroid, oral proliferation of the salivary glands can tecture is preserved. Lymphoepithelial cavity or renal clear cell carcinoma, be reactive or neoplastic. Among the lesions mainly affect the parotid and carcinoma of the , lung and reactive forms of greatest interest are submandibular glands, whilst the minor breast. Metastatic tumours are mostly those represented by the sialadenitis salivary and lacrimal glands may not due to direct invasion by the malig- associated with Sjögren’s syndrome be involved. Lymphomas account for nancy located in immediate proximity. (SS), also known as lymphoepithelial 2–5% of tumours of the salivary glands. Metastases via blood or lymphatic sys- lesion or myoepithelial sialadenitis and Almost 20% of cases are associated tem are rare and can be detected on US the one associated with chronic infec- with SS or lymphoepithelial sialadenitis. as focal defects of the parenchyma with tion with hepatitis C virus (HCV) (5, Lymphomas involve parotid and sub- a round shape and inhomogeneous hy- 20, 21, 24), myastenia gravis, primary mandibular gland in the 70% and 25% poechoic well-defined lesions (14, 15, biliary cirrhosis or lymphoid hyperpla- of cases, respectively, while the minor 26). The US features of submandibular sia cystic (more commonly observed in salivary glands are affected in less than gland metastases are not well described, patients with chronic HIV infection). 10% of cases. On US lymphomas appear although the appearances are likely to Lymphomas of the salivary glands are as an increased gland volume, lack of be similar to those demonstrated in pa- predominantly type B and include the homogeneities, hypoechoic focal or dif- rotid metastatic disease (27). Parotid B-cell lymphoma and marginal zone fuse nodular aspect and not well-defined metastases can vary in appearance but (MALT) lymphoma, follicular lym- margins. The hypoechogenicity is due to tend to be hypoechoic, with heteroge- phoma and diffuse large B-cell lym- the presence of lymphomatous prolifera- neous internal echotexture and poorly phoma. Cystic lymphoid hyperplasia tion that, when of severe degree, confers defined margins (27). can affect both the submandibular and a cystic appearance to the lesion (Fig. 7) parotid gland and unilateral or bilateral (2, 5, 20, 21). However, these features Primary Sjögren’s syndrome swelling in the context of a progres- are not pathognomonic and, on US, lym- Primary Sjögren’s syndrome (pSS) is sive generalised lymphadenopathy can phoma may not be reliably differentiated a chronic autoimmune disorder char- be visualised. Often the distinction be- from other neoplastic or non-neoplastic acterised by lymphocytic infiltrates in tween non-neoplastic lymphoprolifera- salivary gland tumours. the lachrymal and salivary glands (28).

S-66 US and colour/power Doppler of the salivary glands / M. Carotti et al.

Dryness of the eyes (keratoconjunctivi- tis sicca) and () are the main clinical features. Histology shows lymphocytic infiltration and de- struction of the affected glands. pSS concerns patients who have not other connective tissue disease. For this rea- son, it is necessary to know if the sicca syndrome is possibly due to other pa- thologies (e.g. diabetes mellitus, hypo- volemia, sarcoidosis, infection, respira- tory or renal insufficiency) or condi- tions (e.g. smoking, medications). At the moment the minor salivary gland biopsy (MSGB) is considered as the gold standard for diagnosing pSS (50), but further investigations are often necessary and includes: salivary flow measurement, sialochemistry, sequen- tial salivary scintigraphy, and sialogra- phy using liposoluble or hydrosoluble Fig. 8. The sialography in oblique view of a parotid gland in a patient with pSS shows the narrowing contrast material (Fig. 8). In addition, of the main duct, the poor representation of the main branches of the ductal gland and characteristic globular sialectasia pattern of some peripheral branching ductal. to these standard tests for assessment of salivary gland involvement, other methods have been applied such as US, MRI and CT (Fig. 9) (29). These diagnostic tools have widely replaced conventional invasive examinations in scientific research as well as in- clini cal practice. Among all the imaging techniques currently available for the study of salivary glands, US appears to be the most attractive one due to its several advantages and capability to detect wide echostructural abnormali- ties of gland parenchyma (Fig. 10-11). The main echostructural abnormalities detectable on US are represented by the parenchymal inhomogeneity, hypo- anechoic or hypercheoic areas (due to multiple cysts or calcifications, respec- tively), increased or reduced size, irreg- ularity of the margins and the presence of peri-intraglandular lymphnodes. The inhomogeneity of the parenchyma rep- Fig. 9. The CT scan in the parenchyma of parotid glands in patient with pSS shows an increasing size, resents the best parameter for the pSS lack of homogenicity and multiple hypodense cyst-like areas. diagnosis (30, 31). This observation has recently been confirmed in comparison expression of severe parenchymal sub- erwise wide range of previous studies studies with MRI (30, 31). The multi- version with replacement of connective results showing a sensitivity of US be- ple circumscribed or confluent hypo- fibrous tissue (Fig. 12). A study have tween 43% and 90% and a specificity echoic areas and/or multiple cysts have shown higher sensitivity of US with between 83.3 and 100% (30, 33-38). a corresponding histological picture of respect to contrast sialography and sali- In particular, Kawamura et al. (35) and ductal ectasia surrounded by several vary gland scintigraphy (75.3%, 72.7% Salaffi et al. (32, 39) showed that de- linfocytes or widening glandular lob- and 70.1%, respectively), whereas the scriptive and quantitative assessment ules surrounding by linfocytes aggre- specificity was quite similar (83.5%, of the salivary glands by US efficiently gates. The multiple hyperechoic bands 84.9% and 82.3%, respectively) (32). differentiate between pathologic and observed in the later stages can be an These percentages are within the oth- normal glands in patients with pSS.

S-67 US and colour/power Doppler of the salivary glands / M. Carotti et al.

cut-off point of 6 for the proposed US score (total score range from 0 to 16) (Table I) had a good sensitivity/speci- ficity ratio for pSS (75.3%/83.5%) (32). By using this threshold value the posi- tive LR of an abnormal US was 4.58. If a threshold >8.0 was applied the test gained specificity at the cost of a serious loss of sensitivity (sensitivity 54.5%, specificity 97.5%, likelihood ratio 21.5). According to this US scor- ing system (32) the grade 2 corresponds to a clear parenchymal inhomogene- ity, characterised by multiple scattered hypoechogenic areas of variable size (<2mm) and not uniformly distributed. As underlined by several authors (37, Fig. 10. US longitudinal scan of parotid gland in patient with pSS. The parenchyma is heterogeneous 38), the most relevant US sign in pSS is and characterised by multiple scattered hypoechogenic areas and echogenic bands. the parenchymal inhomogeneity which is considered to be the most relevant anatomical structural change in these patients. However, only clear evidence of lack of homogeneity, characterised by multiple scattered hypoechoic are- olaes to multiple cyst-like changes, can be regarded as being of true diagnostic value for the disease. This is especially true considering that mild inhomoge- neity may also be present in patients with xerostomia affected by other pa- thologies. In addition, with regard the presence/absence of parenchymal ho- mogeneity, echogenicity, size of the glands and posterior glandular border the interobserver reliability assessment showed an overall agreement of 89%, 82% and 78%, respectively (k values of Fig. 11. US longitudinal scan of submandibular gland in a patient with pSS. The parenchyma is char- 0.83, 0.79 and 0.71, respectively) and acterised by lack of homogenicity and multiple hypo-anechoic areas. the intraobserver reliability was 91%, 88% and 85%, respectively (k values of They have also demonstrated that US of 92.2% with respect to either symp- 0.85, 0.82% and 0.80%, respectively). abnormalities are strongly correlated tomatic or healthy subjects (33). In a Recently, Cornec et al. (42) confirmed to the histological changes in minor published report by Takagi et al. (40), that morphologic abnormalities of sali- salivary gland tissue (39) and that the US of the salivary glands had a diag- vary glands can be detected early in the proposed US score grading correlates nostic sensitivity of 82% and specificity course of the disease, since US diagnos- well with sialographic gradings (39). of 73%. Therefore, US can be consid- tic properties did not change according To score the changes of the parotid ered an alternative imaging modality to disease duration. Using a weighted gland structure several gradings and to sialography in the diagnosis of pSS. score for each item and a cut-off of 5 classifications have been used (33, 37). Milic et al. (41) found that the US score the proposed US scoring system was So far, the widely used grading score is was positive in 129 (92%) patients with slightly less specific (85.7% vs. 77.9%, the one proposed by De Vita et al. (33). pSS and in 20 (40%) subjects from the respectively) but more sensitive (94.9% This score is ranging from 0 to 6 and as- non-pSS group. By setting the cut-off vs. 98.7%, respectively) than the signs points to the different degrees of point for the US score (range 0–16) at American-European Consensus Group glandular inhomogeneity. Score values 7, the best sensitivity/specificity ratio (AECG) criteria. above 0 showed a sensitivity of 88.8% for pSS was achieved (91.4/94.0%, re- Although the salivary glands have a in pSS and of 53.8% in secondary SS, spectively). In a previous multicentre complex vasculature and intense hyper- as well as a specificity of 84.6% and study we have reported that the optimal aemic responses associated with the se-

S-68 US and colour/power Doppler of the salivary glands / M. Carotti et al.

(p=0.0001). The mean difference be- tween the peak systolic velocity (PSV) values taken from parotids and sub- mandibular glands before and during lemon juice stimulation was statistically significant p ( =0.003 and p=0.01, re- spectively) in the healthy subjects group (46). However, no significant changes in the PSV values were found in the whole group of pSS patients. The variability of RI obtained from the salivary glands be- fore and during lemon juice stimulation was not statistically significant from the one observed in either pSS patients or controls. US abnormalities were detect- ed in the majority of pSS patients and their severity was significantly greater than those recorded in the control group. Among all the C/PDUS parameters only PSV was influenced by the degree of Fig. 12. US longitudinal scan of submandibular gland in a patient with pSS. Note the lack of homo- chronic inflammation, as shown by the genicity of the parenchyma, the not well-defined margins and the presence of hypo-anechoic areas due salivary gland biopsy. These aspects to replacement of fibrous connective tissue. have suggested that PSV may reflect the vascular changes occurring in the Table I. US grading score for the evaluation of salivay glands proposed by Salaffi F et al. salivary glands during the course of an (67). autoimmune disease such as pSS (46). On the basis of these considerations and Grade 0 Normal glands the results of several studies (38, 40, 42, Grade 1 Regular contour, small hypoechoic spots/areas, without echogenic bands, regular or in- 47, 48) there is strong evidence to sup- creased glandular volume (mean values 20 + 3 mm for the parotids and 13 + 2 mm for the port the need to replace both obsolete submandibular glands) and ill-defined posterior glandular border (definite echogenic border with respect to the neighbouring structures) sialography and salivary scintigraphy with US in the AECG criteria (29, 32, Grade 2 Regular contour, evident multiple scattered hypoechogenic areas usually of variable size (<2 40-42, 49). In addition, we also believe mm) and not uniformly distributed, without echogenic bands, regular or increased glandular volume and ill-defined posterior glandular border that US of salivary glands should be included as an additional item in the Grade 3 Irregular contour, multiple large circumscribed or confluent hypoechogenic areas (2–6 mm) recently proposed American College and/or multiple cysts, with echogenic bands, regular or decreased glandular volume and no visible posterior glandular border of Rheumatology (ACR) classification criteria for pSS (50, 51). Grade 4 Irregular contour, multiple large circumscribed or confluent hypoechogenic areas (>6 mm), and/or multiple cysts or multiple calcifications, with echogenic bands, resulting in severe References damage to the glandular architecture, decreased glandular volume and posterior glandular 1. HEBERT G, OUIMET-OLIVA D, NICOLET V, border not visible BOURDON F: Imaging of the salivary glands. Can Assoc Radiol J 1993; 44: 342-9. cretion of saliva (9), poor attention has secretion of saliva as shown by an in- 2. FISCHER T, PASCHEN CF, SLOWINSKI T et al.: Differentiation of parotid gland tumors been paid to C/PDUS in this field. C/ crease of blood flow within the salivary with contrast-enhanced . Rofo PDUS provides useful details in detect- glands more than five times higher than 2010; 182: 155-62. ing soft tissues perfusion and measur- baseline (45). In our previous study (46) 3. KALINOWSKI M, HEVERHAGEN JT, REHBERG ing blood flow velocity in many patho- we have evaluated either the US features E, KLOSE KJ, WAGNER HJ: Comparative study of MR sialography and digital subtrac- logical processes. Recently, C/PDUS of the parotid and submandibular glands tion sialography for benign salivary gland has been used to evaluate the vascular and the blood flow alterations that may disorders. AJNR Am J Neuroradiol 2002; 23: anatomy of the salivary glands and to occur in the salivary glands of both pa- 1485-92. analyse either the physiological changes tients with pSS and subjects suffering 4. BURKE CJ, THOMAS RH, HOWLETT D: Imaging the major salivary glands. Br J Oral in blood flow that occur during salivary from dry mouth not due to pSS (control Maxillofac Surg 2011; 49: 261-9. stimulation in normal subjects (43, 44) group). Abnormal US scores were ob- 5. HALIMI P, GARDNER M, PETIT F: Tumors and the flow alterations in the diseased tained in 86.6% of pSS patients and in of the salivary glands. Cancer Radiother 2005; 9: 251-60. glands of pSS patients (9, 44). Previous 30% of the control group. Moreover, in 6. FALYAS K, LEWIS M, WILLIAMS A et al.: physiologic studies have shown that a pSS patients the US scores were signifi- Diseases of the submandibular gland as dem- marked hyperaemia accompanies the cantly higher than in the control group onstrated using high resolution ultrasound.

S-69 US and colour/power Doppler of the salivary glands / M. Carotti et al.

Br J Radiol 2005; 78: 362-9. Major salivary gland imaging. Radiology GIANNINI FB, PALOMBI C: Salivary gland 7. GRITZMANN N, RETTENBACHER T, HOL- 2000; 216: 19-29. ultrasonography in the evaluation of primary LERWEGER A, MACHEINER P, HUBNER 25. WAN YL, CHAN SC, CHEN YL et al.: Sjögren’s syndrome. Comparison with minor E: Sonography of the salivary glands. Eur Ultrasonography-guided core-needle biopsy salivary gland biopsy. J Rheumatol 2000; 27: Radiol 2003; 13: 964-75. of parotid gland masses. AJNR Am J 1229-36. 8. SCHON R, DUKER J, SCHMELZEISEN R: Neuroradiol 2004; 25: 1608-12. 40. TAKAGI Y, KIMURA Y, NAKAMURA H et al.: Ultrasonographic imaging of head and neck 26. WESTERLAND O, HOWLETT D: Sonoelasto- Salivary gland ultrasonography: can it be an pathology. Atlas Oral Maxillofac Surg Clin graphy techniques in the evaluation and di- alternative to sialography as an imaging mo- North Am 2002; 10: 213-41. agnosis of parotid neoplasms. Eur Radiol dality for Sjögren’s syndrome? Ann Rheum 9. MARTINOLI C, DERCHI LE, SOLBIATI L, 2012; 22: 966-9. Dis 2010; 69: 1321-24. RIZZATTO G, SILVESTRI E, GIANNONI 27. HOWLETT DC: High-resolution ultrasound 41. MILIC V, PETROVIC R, BORICIC I et al.: M: Color Doppler sonography of salivary assessment of the parotid gland. Br J Radiol Ultrasonography of major salivary glands glands. Am J Roentgenol 1994; 163: 933-41. 2003;76: 271-7. could be an alternative tool to sialoscinti- 10. MANSOUR N, STOCK KF, CHAKER A, BAS 28. MOUTSOPOULOS HM: Sjögren’s syndrome: graphy in the American-European classifica- M, KNOPF A: Evaluation of parotid gland le- autoimmune epithelitis. Clin Immunol tion criteria for primary Sjögren’s syndrome. sions with standard ultrasound, color duplex Immunopathol 1994; 72: 162-5. Rheumatology (Oxford) 2012; 51: 1081-5. sonography, sonoelastography, and acous- 29. VITALI C, BOMBARDIERI S, JONSSON R et 42. CORNEC D, JOUSSE-JOULIN S, PERS JO et al.: tic radiation force impulse imaging - a pilot al.; and the European Study Group on Contribution of salivary gland ultrasonogra- study. Ultraschall Med 2012; 33: 283-8. Classification Criteria for Sjögren’s Syn- phy to the diagnosis of Sjögren’s syndrome: 11. ZENGEL P, BERGHAUS A, WEILER C, drome: Classification criteria for Sjögren’s toward new diagnostic criteria? Arthritis REISERM, CLEVERT DA: Intraductally applied syndrome: a revised version of the European Rheum 2013; 65: 216-25. contrast-enhanced ultrasound (IA-CEUS) for criteria proposed by the American-European 43. ARIJI Y, YUASA H, ARIJI E: High-frequency evaluating obstructive disease and secretory Consensus Group. Ann Rheum Dis 2002; 61: color Doppler sonography of the subman- dysfunction of the salivary glands. Eur Radiol 554-8. dibular gland: relationship between salivary 2011; 21: 1339-48. 30. NIEMELÄ RK, TAKALO R, PÄÄKKÖ E et al.: secretion and blood flow. Oral Surg Oral 12. KOVACEVIĆ DO, FABIJANIĆ I: Sonographic Ultrasonography of salivary glands in prima- Med Oral Pathol Oral Radiol Endosc 1998; diagnosis of parotid gland lesions: corre- ry Sjögren’s syndrome. A comparison with 86: 476-81. lation with the results of sonographically magnetic resonance imaging and magnetic 44. CHIKUI T, YONETSU K, IZUMI M, EGUCHI guided fine-needle aspiration biopsy. J Clin resonance sialography of parotid glands. K, NAKAMURA T: Abnormal blood flow to Ultrasound 2010; 38: 294-8. Rheumatology 2004; 43: 875-9. the submandibular glands of patients with 13. BIALEK EJ, JAKUBOWSKI W, ZAJKOWSKI P, 31. NIEMELÄ RK, PÄÄKKÖ E, SURAMO I, Sjögren’s syndrome: Doppler waveform SZOPINSKI KT, OSMOLSKI A: US of the ma- TAKALO R, HAKALA M: Magnetic resonance analysis. J Rheumatol 2000; 27: 1222-8. jor salivary glands: anatomy and spatial rela- imaging and magnetic resonance sialography 45. RINDI G, MANNI E: Fisiologia Umana, 2nd tionships, pathologic conditions, and pitfalls. of parotid glands in primary Sjögren’s syn- ed. Turin, Italy: Unione Tipografico-Editrice Radiographics 2006; 26: 745-63. drome. Arthritis Rheum 2001; 45: 512-8. Torinese, 1982: 1149-54. 14. GALLIPOLI A, SACCO M: Ghiandole salivari. 32. SALAFFI F, CAROTTI M, IAGNOCCO A et 46. CAROTTI M, SALAFFI F, MANGANELLI P, In: ZIVIELLO M, BAZZOCCHI M. Gnocchi al.: Ultrasonography of salivary glands in ARGALIA G: Ultrasonography and colour Editore. Ecografia. Napoli, 1994; 507-17. primary Sjögren’s syndrome: a comparison Doppler sonography of salivary glands in 15. CANDIANI F, RUBALTELLI L: Ghiandole with contrast sialography and scintigraphy. primary Sjögren’s syndrome. Clin Rheumatol salivari. In BAZZOCCHI M. Idelson-Gnocchi Rheumatology (Oxford) 2008; 47: 1244-9. 2001; 20: 213-9. Editore. Eco-Color Doppler in Medicina 33. DE VITA S, LORENZON G, ROSSI G, SABELLA 47. MAKULA E, PORKORNY G, KISS et al.: Interna. Napoli, 1998; 406-16. M, FOSSALUZZA V: Salivary gland echo- The place of magnetic resonance and ultra- 16. SILVERS AR, SOM PM: Salivary glands. graphy in primary and secondary Sjögren’s sonographic examinations of the parotid gland Radiol Clin N Am 1998; 36: 941-66. syndrome. Clin Exp Rheumatol 1992; 10: in the diagnosis and follow-up of primary 17. GRITZMANN N: Sonography of the salivary 351-6. Sjögren’s syndrome. Rheumatology 2000; 39: glands. AJR Am J Roentgenol 1989; 53: 161- 34. ARIJI Y, OHKI M, EGUCHI K et al.: Texture 97-104. 6. analysis of sonographic features of the pa- 48. YONETSU K, TAKAGI Y, SUMI M, NAKAMURA 18. BRUNETON JN, MOUROU MY: Ultrasound rotid gland in Sjögren’s syndrome. AJR Am T, EGUCHI K: Sonography as a replacement in . ORL J Otorhino- J Roentgenol 1996; 166: 935-41. for sialography for the diagnosis of salivary laryngol Relat Spec 1993; 55: 284-9. 35. KAWAMURA H, TANIGUCHI N, ITOH K, glands affected by Sjögren’s syndrome. Ann 19. TRAXLER M, SCHURAWITZKI H, ULM C, KANO S: Salivary gland echography in pa- Rheum Dis 2002; 61: 276-77. SOLAR P, BLAHOUT R: Sonography of non- tients with Sjögren’s syndrome. Artritis 49. BOOTSMA H, SPIJKERVET FKL, KROESE neoplastic disorders of the salivary glands. Rheum 1990; 33: 505-10. FGM, VISSINK A: Toward new classification Int J Oral Maxillofac Surg 1992; 21: 360-3. 36. TAKASHIMA S, MORIMOTO S, TOMIYAMA N, criteria for Sjögren’s syndrome? Arthritis 20. LEE YYP, WONG KT, AHUJA AT: Imaging of TAKEUKI N, IKEZOE J, KOZUKA T: Sjögren’s Rheum 2013; 65: 21-3. salivary gland tumors. Eur J Radiol. 2008; syndrome: comparison of sialography and 50. VITALI C, CAROTTI M, SALAFFI F: Is it the 66: 419-36. ultrasonography. J Clin Ultrasound 1992; time to adopt salivary gland ultrasonogra- 21. THOENY HC: Imaging of salivary gland 20: 99-109. phy as an alternative diagnostic tool for the tumors. Cancer Imaging 2007; 7: 52-62. 37. HOČEVAR A, AMBROŽIČ A, ROZMAN B, classification of patients with Sjögren’s syn- 22. WEI X, LI Y, ZHANG S et al.: Evaluation of KVEDER T, TOMŠIČ M: Ultrasonographic drome? Comment on the article by Cornec et microvascularization in focal salivary gland changes of major salivary glands in primary al. Arthritis Rheum 2013; 65: 1950. lesions by contrast-enhanced ultrasonogra- Sjögren’s syndrome. Diagnostic value of a 51. SHIBOSKI SC, SHIBOSKI CH, CRISWELL LA phy (CEUS) and Color Doppler sonography. novel scoring system. Rheumatology 2005; et al.; for the Sjögren’s International Clin Hemorheol Microcirc. 2013; 54: 259- 44: 768-72. Collaborative Clinical Alliance (SICCA) 71. 38. WERNICHE D, HESS H, GROMNICA-IHLE E, Research Groups American College of 23. BRADLEY MJ, DURHAM LH, LANCER JM: KRAUSE A, SCHMIDT W: Ultrasonography of Rheumatology classification criteria The role of colour flow Doppler in the inves- salivary glands – A highly specific imaging for Sjögren’s syndrome: a data-driven, tigation of the . Clin procedure for diagnosis of Sjögren’s syn- expert consensus approach in the Sjögren’s Radiol 2000; 55: 759-62. drome. J Rheumatol 2008; 35: 285-93. International Collaborative Clinical Alliance 24. YOUSEUM DM, KRANT MA, CHALIAN AA: 39. SALAFFI F, ARGALIA G, CAROTTI M, cohort. Arthritis Care Res 2012; 64: 475-87.

S-70