Abstract
Ultrasound has become an important tool in the evaluation of the neck prior to a neck dissection. Ultrasound is safe, relatively inexpensive, available in the ambulatory setting, and, with the resolution of a few millimeters, is capable of characterizing even small structures in the neck. Ultrasound has proliferated in its use in the neck in the United States because of benign and malignant thyroid disease. This chapter will introduce the basic concepts of ultrasonography, largely from a thyroid perspective, and introduce common methods of evaluating the neck, and discuss normal and abnormal features of organs and structures in the neck that are important to appreciate prior to a neck dissection. There are multiple additional benefits to office-based neck ultrasound. The office utilization allows for direct communication from the physician to the patient in real time. Ultrasound is rapidly moving into outpatient clinical practice in endocrinology and surgery. Proficiency in thyroid ultrasound is now required for endocrine certification, and although not yet mandatory for otolaryngology and general surgery, certification is strongly encouraged.
Keywords: ultrasound, neck dissection, thyroid
14.1 Ultrasonography Basics
Ultrasound (US) is based on the reflection of generated sound waves. There have been tremendous advances in algorithms that transform the captured sound waves into a much more refined image. The depth that the sound waves reach is dependent upon frequency. Lower-frequency sound waves penetrate deeper but typically have poorer resolution. Higher-frequency sound waves do not penetrate as well but have superior resolution. Sound waves are reflected at the boundaries of tissues with different impedance, which is dependent on the density of the tissues and the velocity of the sound within a particular tissue.1 Transducers both generate the sound waves and capture reflected sound. Transducers today commonly have variable frequency capabilities. A common frequency used to evaluate the thyroid lobes and superficial structures is 14 MHz. Slightly lower frequencies up to 12 MHz are often employed to localize parathyroid adenomas. Lower frequencies, from 7 to 10 MHz, are used to look at deep structures in the neck, up to 4 cm from the surface.1
To perform a US of the neck, the patient is typically positioned either supine or reclined up to an angle of 30 to 45 degrees. For some patients in a wheelchair or others who cannot lie supine, US can be performed in the near sitting position with the chin tilted back slightly. One common method to perform a US of the neck is to start in the thyroid gland with a frequency of 12 to 14 MHz. The transducer is placed transversely on the neck and the frequency is optimized for thyroid tissue. The transducer is slowly moved vertically to evaluate the thyroid lobe and nodules in transverse and anteroposterior dimensions. The transducer is then rotated 90 degrees to obtain the axial dimensions of structures of interest. Once the thyroid has been completely evaluated, the transducer is again placed in the transverse orientation over the thyroid and moved caudally, paying particular attention to the area around the hyoid to pick up a possible thyroglossal duct cyst. With the transducer still in the transverse position, it is moved one-half to three-fourths of the width of the transducer laterally and then slowly moved inferior down to the clavicle. The transducer is tilted caudally for the sound beam to characterize lymph nodes or other structures beneath or inferior to the clavicle. This process is repeated moving into the contralateral neck to evaluate the presence of lymph nodes, the size and shape of the submandibular and parotid glands, and for the presence of any abnormal masses.
14.2 Thyroid/Parathyroid Gland
The thyroid gland is often a focus in the planning of a neck dissection. US of a normal thyroid gland is shown in Fig. 14.1. A normal thyroid gland is homogeneous with a density slightly brighter than that of the muscle. Thyroid tissue that arises from the isthmus and extends superiorly is known as a pyramidal lobe and may be present up to 30% of the time.
There are many pathologic states that conceivably could influence the extent of neck dissection. Hashimoto’s thyroiditis is a benign autoimmune disorder that can trigger an extensive inflammatory response potentially causing adhesions and difficulty during a dissection. The most common clinical presentation is a moderately enlarged gland, often one and a half times normal, that is very firm on physical examination. There are several US appearances of a Hashimoto gland.2 The most common has been termed “giraffe skin” because of its mottled appearance with innumerable hypoechoic areas alternating with hyperechoic areas (Fig. 14.2). These nodularities are typically small ranging from 3 to 5 mm. The hyperechoic areas are sometimes confluent, forming benign, well-defined nodules termed “white knights” (Fig. 14.3). Hashimoto’s thyroiditis can also manifest with large areas that are diffusely hypoechoic with some fibrous banding (Fig. 14.4). Often, there is very little normal thyroid tissue present. The fibrous banding often forms pseudonodules, where in the transverse view the hypoechoic area appears exactly as a thyroid nodule; however, when the transducer is rotated to the longitudinal view, the nodule disappears. This form of Hashimoto’s thyroiditis can appear almost identical to thyroid lymphoma (Fig. 14.5) and any biopsy should always include a sample sent for flow cytometry in order to differentiate between the two possible diagnoses. A Hashimoto gland also may present as a hybrid of fibrous banding and hypoe- chogenicity, with hypoechoic areas often encompassing one- fourth or more of the thyroid gland with no normal tissue present and then the remainder of the thyroid lobe appearing relatively normal. Finally, the thyroid gland in early Hashimoto’s thyroiditis may have a near-normal appearance. It is not unusual to see benign lymph nodes around the periphery of a Hashimoto gland. These can be mildly enlarged from 1 to 1.5 mm but are benign appearing with an elongated appearance and a good hilum.
Fig. 14.2 Hashimoto’s giraffe skin (moth-eaten) appearance.
There appears to be a correlation between the appearance of a Hashimoto gland and antithyroid peroxidase (thyroperoxidase [TPOAb]) and thyroglobulin antibody levels. Antibody levels are higher in homogeneously, hypoechoic glands and correlate with the degree of lymphocytic infiltration.3 The prevalence of thyroid cancer in a Hashimoto gland is controversial. A large meta-analysis of over 10,000 papillary thyroid carcinoma (PTC) cases demonstrated that Hashimoto’s gland was far more likely associated with PTC than normal thyroid tissue.4 A second meta-analysis of over 64,000 patients showed a modest increase in PTC.5 However, a recent, well-conducted meta-analysis compared studies categorized into a fine-needle aspiration biopsy (FNAB) group and an archival thyroidectomy group. The prevalence of PTC in the FNAB group was 1.2%, whereas in the archival thyroidectomy group it was 27%, strongly suggesting no real association.6
It is important to ascertain whether or not thyroid cancer is present. The characteristics of malignant or suspicious nodules that would trigger an FNAB will be discussed here, but the details of the workup of indeterminate nodules, etc., are beyond the scope of this publication. Four patterns of thyroid nodules are commonly associated with benign results7: (1) simple cysts with or without colloid clot that do not contain substantial solid element; (2) hyperechoic nodules or white knights (Fig. 14.3) in a Hashimoto gland; (3) giraffe pattern (Fig. 14.2) also in a Hashimoto gland; and (4) nodules resembling a wet sponge or “spongiform” nodules (Fig. 14.6). However, the morphology of these benign lesions should be carefully standardized. As shown in Fig. 14.7, a nodule that closely resembled a spongiform nodule was found to be PTC on biopsy.
There are several US characteristics that suggest malignancy, with microcalcifications less than 1 mm having the highest specificity (Fig. 14.8).8 Intact eggshell calcifications are most often benign, but can be malignant depending upon the continuity of the calcifications. Incomplete eggshell calcifications and those with protruding elements are often malignant.9 However, inspissated colloid can form bright objects that strongly resemble microcalcifications. The two can often be discriminated by shadowing often seen with microcalcifications that is rarely present with inspissated colloid.
Fig.14.3 Hashimoto’s benign hyperechoic nodules (white knights).
Fig. 14.4 Hashimoto’s large hypoechoic areas.
Fig. 14.5 Thyroid lymphoma.
Fig. 14.6 Spongiform nodule.
It should be kept in mind that the presence of coarse calcification does not preclude the diagnosis of thyroid carcinoma, although usually, if malignant, they will be accompanied by microcalcifications (Fig. 14.9). Hypoechogenicity is also a common finding in thyroid carcinoma.10 The degree of hypoechogenicity is highly variable, ranging from an almost completely black appearance similar to the echogenicity of the carotid artery (Fig. 14.8) to nodules that are just barely darker than the surrounding thyroid tissue (Fig. 14.10). The sensitivity of hypoechogenicity is very high, but the specificity is rather low. For example, benign nodules can also be very hypoechoic. The continuity of the capsule is very important. If there is blebbing of the nodule out into surrounding thyroid, or a very poorly defined capsule (Fig. 14.9), this should be immediately concerning for the presence of malig- nancy.11 A nodule that is taller than it is wide has been shown to have a higher propensity for malignancy.11 The physiologic basis for this phenomenon is unclear, but it has been reported to have a specificity as high as 90%.12 Size alone is a relatively poor indicator of malignancy. Generally, there is a positive correlation between the size of a nodule and its risk of malignancy. This is particularly true for nodules that are 2 cm or greater in size.13 However, some studies have shown that nodules greater than 4 cm have a smaller chance of being malignant than nodules that are less than 4 cm.14 In addition, there was no increased risk of malignancy in follicular neoplasms greater than 4 cm in size.15 However, large nodules that are extremely hypoechoic, similar to that of the carotid artery, are highly suspicious and often are lymphoma (Fig. 14.5) or poorly differentiated/anaplastic carcinoma (Fig. 14.11).
Using the above criteria, there are some nodules that appear to be malignant immediately after the transducer is placed on the neck. However, this is not always the case. There are many instances where nodules are found to be malignant on FNAB with only mild hypoechogenicity on US (Fig. 14.7).
Fig. 14.9 Primary papillary thyroid carcinoma with coarse calcifications.
Fig. 14.10 Benign appearing thyroid nodule positive for papillary thyroid carcinoma.
Fig. 14.11 Anaplastic thyroid carcinoma.
Fig. 14.12 Typical rounded, hypoechoic appearance of a parathyroid gland.
Fig.14.13 Flattened, hyperechoic parathyroid gland.
Another structure that is amenable to evaluation by US is the parathyroid gland. There are typically four parathyroid glands, two off the inferior pole of the right and left lobe and two typically found posterior to the midportion of the thyroid lobes commonly referred to as superior parathyroid glands (Fig. 14.12). The appearance of parathyroid glands on US imaging is typically rounded and hypoechoic. However, the gland is often irregularly shaped and may even be slightly hyperechoic (Fig. 14.13). Parathyroid adenomas may show peripheral blood flow unlike central blood flow observed in lymph nodes. There may also be increased blood flow to the parathyroid capsule or adjacent thyroid tissue.16.17 FNAB measuring parathyroid hormone in a saline washout is helpful in identifying an adenoma, particularly if an intrathyroidal parathyroid adenoma is suspected.18 In a large meta-analysis of 20,000 patients, primary hyperparathyroidism (HPT) results from a single adenoma in approximately 88.9% of cases, double adenoma in 4.1%, and four-gland disease/hypertrophy in 5.8%.19
Ultrasound performs relatively poorly when multiple adenomas are present in renal failure or multiple endocrine neoplasia syndromes (MEN), and also in localizing an adenoma prior to reoperation for persistent disease. In a study of 166 patients with HPT as a consequence of renal disease, sestamibi scan and US together failed to detect 61.5% of ectopic glands.20 In 288 patients who required reoperation for persistent HPT, sestamibi provided the best results with 67% true-positive and no falsepositive results. US only had a 48% true-positive and 21% falsepositive results.21 The ability to localize a parathyroid adenoma is reduced in the presence of thyroid nodules.21
14.3 Fine-Needle Aspiration Biopsy
FNAB is the gold standard in the evaluation of a thyroid nodule. FNAB can be performed with the needle either perpendicular or parallel to the transducer. The biopsy can be performed with a 25- or 23-gauge 1- to 1.5-inch needle. The choice of core needle biopsy versus fine-needle aspirate is still debated, but FNAB reveals more nuclear features, has fewer instances of bleeding, and overall is not inferior to core biopsy in terms of diagnostic yield in some22 but not all23 studies. Traditionally, lidocaine anesthesia was avoided because it could obscure nodule landmarks when the biopsy was performed without US guidance. We have found that 1 to 2 mL of lidocaine under the skin and along the probable needle track has markedly decreased patient anxiety and pain involved in the procedure. The only test that should precede an FNAB is a thyroid-stimulating hormone (TSH) determination. A suppressed TSH suggests a toxic adenoma and the diagnosis can be confirmed by thyroid scan and uptake. These lesions are rarely malignant and can be followed with serial US without FNAB. Thyroid scan in the absence of a suppressed TSH is typically not helpful, for benign nodules also are “cold.”
FNAB should be reported according to the criteria established at the Bethesda conference in 2007,24 which have recently been revised.25 As shown in Table 14.1, moving from Bethesda III to Bethesda V dramatically increases the likelihood of malignancy. The most difficult Bethesda criteria to evaluate is Bethesda III and Bethesda IV. These nodules have approximately 30% chance of malignancy. In the past, at least a hemithyroidectomy was performed, with completion if the nodule was positive for malignancy. It is beyond the scope of this chapter to discuss gene classifier systems, but there are now several companies using this technology that are able to determine that a nodule is benign with a high degree of diagnostic accuracy. However, at the present time, these assays are quite poor in predicting malignancy, with a positive predictive value ranging from 15 to 20% to 60 to 70%. The American Thyroid Association (ATA) guidelines for the evaluation of thyroid nodules and well-differentiated thyroid carcinoma recommend a hemithyroidectomy for FNAB positive nodules that are less than 4 cm.26 However, it must be kept in mind that these are just guidelines. Other considerations, such as appearance of the nodule on US or multiple nodules in the contralateral lobe, may result in a total thyroidectomy with a smaller lesion.
It is extremely important that the ipsilateral neck and at least levels II through V be evaluated by US. Metastatic disease as small as 3 mm within the lymph node can be identified and FNAB performed. Lateral neck ultrasonography can often prevent a second surgery. However, thyroid US performed in radiology departments will rarely examine the lateral neck unless this is requested. Reimbursement issues for performing both tests can be troublesome, because the CPT codes for thyroid US and neck US are identical and insurance companies will rarely pay for duplicate charges.
14.4 Lateral Neck Ultrasound
It is very important to perform a thorough and reproducible US of the lateral neck. One approach is to run the transducer in a transverse orientation and adjacent to the trachea from level I through the thyroid lobe down to lower level VI/VII. Then, move the transducer one-half to three-fourths of the width of the transducer laterally while maintaining a perpendicular orientation to the midline. This process should be repeated until the entire lateral neck has been examined to level V posterior to the sternocleidomastoid muscle. For each of these iterations, angle the transducer under the clavicle to look for far inferior lymph nodes. Benign lymph nodes are fairly easy to identify. They are almost always oblong or flattened with a ratio of any orthogonal axis less than 0.5. They also will contain a hilum most commonly represented as a thin hyperechoic band extending into the midportion of the lymph node (Fig. 14.14). Malignant lymph nodes are typically rounded, with all ratios of orthogonal axis to another greater than 0.5 and often approaching 1.0, which represent the characteristics of a sphere (Fig. 14.15). The hilum is effaced by tumor and absent (Fig. 14.15). There is a wide range of echogenicity from completely hypoechoic (Fig. 14.16) to hyperechoic similar to muscle (Fig. 14.17). Calcifications are often present with metastatic PTC (Fig. 14.17). Cystic lymph nodes are almost always malignant, most commonly with squamous cell or thyroid carcinoma (Fig. 14.18). Metastatic squamous cell carcinoma involving lymph nodes in the neck is often more aggressive with obliteration of node architecture and erosion of the capsule (Fig. 14.19).
14.5 Salivary Glands
The salivary glands include the parotid, submandibular, and sublingual gland. It is beyond the scope of this chapter to go into detail on all benign and malignant neoplasms of the salivary glands. In general, the echogenicity is similar between the three and a normal submandibular gland is shown in Fig. 14.20. They are typically hyperechoic and homogeneous, with the degree of hyperechogenicity dependent on the amount of fat present. Fatty glands are more hyperechoic.
The US evaluation of the salivary glands is similar to other structures in the neck. The transducer is initially transverse, fairly parallel to the mandible and canted slightly to visualize beneath. This is especially true when examining the parotid gland, which can extend considerably beneath the mandible. It is important to visualize the contralateral gland to compare size, echogenicity, and presence of lesions.
Table 14.1 The Bethesda System for Reporting Thyroid Cytopathology: risk of malignancy and usual clinical management
|
Diagnostic category |
Risk of Malignancy (%) |
Management |
|
|
I. |
Nondiagnostic or unsatisfactory |
1-4 |
Repeat FNAB (ultrasound guidance) |
|
II. |
Benign |
0-3 |
Clinical follow-up |
|
III. |
AUS or FLUS |
5-15 |
Repeat FNAB (molecular testing) |
|
IV. |
Follicular neoplasm or suspicious for follicular neoplasm |
15-30 |
Lobectomy (molecular testing) |
|
V. |
Suspicious for malignancy |
60-75 |
Lobectomy or total thyroidectomy |
|
VI. |
Malignant |
97-99 |
Lobectomy or total thyroidectomy |
AUS, atypia of undetermined significance; FLUS, follicular lesion of undetermined significance; FNAB, fine-needle aspiration biopsy.
Fig. 14.14 Benign lymph node. Ovoid with defined hilum.
Fig.14.15 Rounded hypoechoic lymph node positive for papillary thyroid carcinoma.
Fig. 14.16 Hypoechoic nodule in the thyroid bed positive for papillary thyroid carcinoma.
Fig.14.17 Rounded hyperechoic lymph node with microcalcifications positive for papillary thyroid carcinoma.
Both benign and malignant neoplasms arising from the salivary gland often have a similar appearance. They are typically hypoechoic and homogeneous with pleomorphic adenoma constituting the most common benign lesion. They are most commonly found in the parotid gland and are predominantly solitary.27 Pleomorphic adenomas can be lobulated and contain calcifications but are poorly vascularized. Warthin’s tumors have a very similar appearance to pleomorphic adenomas, hypoechoic, and oval shaped, but can be bilateral.
Malignant lesions of the salivary glands are predominantly represented by mucoepidermoid carcinoma and adenoid cystic carcinoma. They may present with lobulation and poor borders similar to other malignant neoplasms in the neck28; however, it is not uncommon to present as hypoechoic, well defined, and cystic.29 Probably even more so than thyroid nodules, it is uncommon to be able to distinguish benign and malignant salivary gland tumors by US appearance; therefore, FNAB is mandatory.
Fig. 14.20 Normal submandibular gland.
14.6 Conclusion
US of structures in the neck is safe and relatively inexpensive. Often US precludes more expensive and radiation-intensive imaging. Routine follow-up of thyroid nodules, neck surveillance in thyroid cancer survivors, and evaluation of neck nodes and salivary glands in high-risk individuals have become common in community-based surgical and endocrine practices. However, careful consideration should be given to the decision to pursue routine US-guided FNAB of these lesions. As with any procedure, practitioners need to be impartial in deciding their skill level and whether referral would be prudent.
References
[1] Kremkau F. Diagnostic Ultrasound: Principles and Instruments. St Louis, MO: Elsevier; 2006
[2] Anderson L, Middleton WD, Teefey SA, et al. Hashimoto thyroiditis: part 1, sonographic analysis of the nodular form of Hashimoto thyroiditis. AJR Am J Roentgenol. 2010; 195(1):208-215
[3] LiVolsi VA. Surgical Pathology of the Thyroid. Philadelphia, PA: Saunders; 1990
[4] Lee JH, Kim Y, Choi JW, Kim YS. The association between papillary thyroid carcinoma and histologically proven Hashimoto’s thyroiditis: a meta-analysis. Eur J Endocrinol. 2013; 168(3):343-349
[5] Resende de Paiva C, Gr0nh0j C, Feldt-Rasmussen U, von Buchwald C. Association between Hashimoto’s thyroiditis and thyroid cancer in 64,628 patients. Front Oncol. 2017; 7:53
[6] Jankovic B, Le KT, Hershman JM. Clinical review: Hashimoto’s thyroiditis and papillary thyroid carcinoma: is there a correlation? J Clin Endocrinol Metab. 2013; 98(2):474-482
[7] Anderson L, Middleton WD, Teefey SA, et al. Hashimoto thyroiditis: part 2, sonographic analysis of benign and malignant nodules in patients with diffuse Hashimoto thyroiditis. AJR Am J Roentgenol. 2010; 195(1):216-222
[8] Solbiati L, Osti V, Cova L, Tonolini M. Ultrasound of thyroid, parathyroid glands and neck lymph nodes. Eur Radiol. 2001; 11(12):2411-2424
[9] Kim BM, Kim MJ, Kim EK, et al. Sonographic differentiation of thyroid nodules with eggshell calcifications. J Ultrasound Med. 2008; 27(10):1425-1430
[10] Kim EK, Park CS, Chung WY, et al. New sonographic criteria for recommending fine-needle aspiration biopsy of nonpalpable solid nodules of the thyroid. AJR Am J Roentgenol. 2002; 178(3):687-691
[11] Moon HJ, Sung JM, Kim EK, Yoon JH, Youk JH, Kwak JY. Diagnostic performance of gray-scale US and elastography in solid thyroid nodules. Radiology. 2012; 262(3):1002-1013
[12] Alexander EK, Marqusee E, Orcutt J, et al. Thyroid nodule shape and prediction of malignancy. Thyroid. 2004; 14(11):953-958
[13] Kamran SC, Marqusee E, Kim MI, et al. Thyroid nodule size and prediction of cancer. J Clin Endocrinol Metab. 2013; 98(2):564-570
[14] Shrestha M, Crothers BA, Burch HB. The impact of thyroid nodule size on the risk of malignancy and accuracy of fine-needle aspiration: a 10-year study from a single institution. Thyroid. 2012; 22(12):1251-1256
[15] Ibrahim Y, Mohamed SE, Deniwar A, et al. The impact of thyroid nodule size on the risk of malignancy in follicular neoplasms. Anticancer Res. 2015; 35(3):1635-1639
[16] Mohammadi A, Moloudi F, Ghasemi-Rad M. Preoperative localization of parathyroid lesion: diagnostic usefulness of color doppler ultrasonography. Int J Clin Exp Med. 2012; 5(1):80-86
[17] Mohammadi A, Moloudi F, Ghasemi-rad M. The role of colour Doppler ultrasonography in the preoperative localization of parathyroid adenomas. Endocr J. 2012; 59(5):375-382
[18] Devcic Z, Jeffrey RB, Kamaya A, Desser TS. The elusive parathyroid adenoma: techniques for detection. Ultrasound Q. 2013; 29(3):179-187
[19] Ruda JM, Hollenbeak CS, Stack BC, Jr. A systematic review of the diagnosis and treatment of primary hyperparathyroidism from 1995 to 2003. Otolaryngol Head Neck Surg. 2005; 132(3):359-372
[20] Andrade JS, Mangussi-Gomes JP, Rocha LA, et al. Localization of ectopic and supernumerary parathyroid glands in patients with secondary and tertiary hyperparathyroidism: surgical description and correlation with preoperative ultrasonography and Tc99m-Sestamibi scintigraphy. Rev Bras Otorrinolarin- gol (Engl Ed). 2014; 80(1):29-34
[21] Jaskowiak N, Norton JA, Alexander HR, et al. A prospective trial evaluating a standard approach to reoperation for missed parathyroid adenoma. Ann Surg. 1996; 224(3):308-320, discussion 320-321
[22] Kim SY, Lee HS, Moon J, et al. Fine-needle aspiration versus core needle biopsy for diagnosis of thyroid malignancy and neoplasm: a matched cohort study. Eur Radiol. 2017; 27(2):801-811
[23] Choi YJ, Baek JH, Suh CH, et al. Core-needle biopsy versus repeat fine-needle aspiration for thyroid nodules initially read as atypia/follicular lesion of undetermined significance. Head Neck. 2017; 39(2):361-369
[24] Baloch ZW, LiVolsi VA, Asa SL, et al. Diagnostic terminology and morphologic criteria for cytologic diagnosis of thyroid lesions: a synopsis of the National Cancer Institute Thyroid Fine-Needle Aspiration State of the Science Conference. Diagn Cytopathol. 2008; 36(6):425-437
[25] Cibas ES, Ali SZ. The 2017 Bethesda System for Reporting Thyroid Cytopathol- ogy. Thyroid. 2017; 27(11):1341-1346
[26] Haugen BR, Alexander EK, Bible KC, et al. 2015 American Thyroid Association Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer: The American Thyroid Association Guidelines Task Force on Thyroid Nodules and Differentiated Thyroid Cancer. Thyroid. 2016; 26(1):1-133
[27] Renehan A, Gleave EN, Hancock BD, Smith P, McGurk M. Long-term follow-up of over 1000 patients with salivary gland tumours treated in a single centre. BrJ Surg. 1996; 83(12):1750-1754
[28] Howlett DC, Kesse KW, Hughes DV, Sallomi DF. The role of imaging in the evaluation of parotid disease. Clin Radiol. 2002; 57(8):692-701
[29] Howlett DC. High resolution ultrasound assessment of the parotid gland. Br J Radiol. 2003; 76(904):271-277