The salivary glands are unique in the diversity and complexity of their pathologies. Because fine needle aspiration cytology and frozen section are associated with major diagnostic difficulties, the authors analyzed the use of core needle biopsy (CNB) for the histologic assessment of salivary gland lesions. A systematic observational clinicopathologic quality assessment study was performed over 81 months including 161 CNB procedures in 76 patients with salivary gland pathologies. Adequate samples containing the target tissue were obtained in 73 patients. These samples revealed malignant disease in 45 (62%) patients, benign disease in 26 (36%) patients, and were inconclusive in 2 (3%) patients. Follow-up uncovered no false-positive or false-negative results. On the basis of secondary histologic and clinical follow-up, the statistical parameters were calculated as follows: sensitivity 94%; specificity 100%; accuracy 96%; positive predictive value 100%; negative predictive value 90%. The advantages and potential limitations of CNB in patients with salivary gland masses are discussed. CNB is a reliable biopsy technique for the assessment of salivary gland pathologies, although limitations remain for the subclassification of some neoplastic lesions. The authors recommend CNB as the biopsy technique of choice for a selection of indications.
The presence of a circumscribed mass or diffuse swelling of a salivary gland can represent a major diagnostic and therapeutic challenge. Several features of salivary gland lesions are unique. The mixture of pathologies that may involve the submandibular or parotid gland is characterized by a large and histologically diverse group of benign neoplasms, malignancies and inflammatory lesions. The histopathology and cytology of salivary gland tumours are extremely sophisticated, even for experienced pathologists, demonstrating a highly diverse mix of cell types and growth patterns and overlapping morphologic features. The extent of surgery for salivary gland neoplasms (e.g. save or sacrifice the facial nerve) is not only dependent on the differentiation between benign and malignant disease but also on the exact entity and grading of the malignant tumours. A considerable number of salivary gland pathologies are treated by conservative means and do not require surgical resection (e.g. inflammatory processes, sialadenosis, lymphoma or Warthin’s tumour in the elderly or unfit patient). Despite the small size and superficial location of many salivary gland lesions, surgical resection is virtually impossible under local anaesthesia because the facial nerve cannot be monitored. The surgical approach to some parotid lesions (e.g. tumours involving the deep lobe or the pterygoid fossa) may be highly complex and time-consuming. Revision surgery of the parotid gland (e.g. after superficial parotidectomy) is sophisticated and, due to scarring, associated with a high risk of facial nerve injury.
An accurate preoperative diagnosis of salivary gland lesions is desirable for appropriate patient management and surgical planning. Clinical examination and imaging techniques alone usually cannot reliably distinguish between benign and malignant diseases nor subclassify the pathologic process. Open biopsy is no longer justified because of the high risk of tumour seeding, facial nerve injury, facial scarring and fistula formation. Fine needle aspiration cytology (FNAC) is widely used as a first-line technique for the diagnosis of salivary gland pathologies but many studies have highlighted its limitations including a high rate of false-negative results and poor accuracy for distinguishing between the various types of malignant tumours. Frozen section (FS) has been discussed as an alternative approach to guide treatment intraoperatively but has also been associated with a high rate of diagnostic errors and low sensitivity. In recent years, percutaneous image-guided core needle biopsy (CNB) has gained widespread popularity for tissue sampling particularly of deep-seated masses throughout the body.
Despite the advantages of CNB only a few studies have discussed its application for diagnostic purpose in salivary gland lesions. The authors have analyzed the role of CNB for the evaluation of salivary gland pathologies with emphasis on its reliability in providing a precise subclassification of the disease. The present study is a supplement to the authors’ previously published results on the use of CNB in the head and neck and focuses on the clinical usefulness of CNB in an important potential area of application for this diagnostic instrument.
Material and methods
This is a systematic observational clinicopathologic quality assessment study in an academic tertiary care medical centre. The study was performed in accordance with the guidelines of the Helsinki Declaration of 1975, as revised in 1983 and the study protocol was approved by the local institutional review board.
For purpose of the present study, a selection and update of all patients who underwent CNB at the authors’ institution over a 7-year-period was carried out. Between May 2003 and February 2010, the authors performed CNB in 76 patients with salivary gland pathologies. In 64 patients (84%) a parotid gland mass was biopsied and in 12 (16%) a submandibular gland lesion. The indications for a diagnostic CNB in patients with unclear masses of a salivary gland were as follows: suspected malignant primary salivary gland tumour, lymphoma or metastasis; suspected specific inflammation or sialadenosis; old and/or multimorbid patients who were not candidates for surgery; large tumours with involvement of the deep lobe of the parotid gland; non-compliant patients; and patients who refused surgery. CNB was generally not performed in patients who were fit for surgery and revealed suspected non-specific salivary gland inflammation or suspected benign salivary gland tumours, particularly pleomorphic adenoma, without any signs of malignancy. Patient characteristics of the study population are summarized in Table 1 . Before referral to the authors’ department, 5 patients (7%) had undergone FNAC at another institution; 3 of the 5 FNAC results later turned out to be false-negative, one turned out to be false-positive and one result had correctly indicated malignancy but classification of the neoplasm had been impossible.
|Age (years)||17–100 (mean: 69.0)|
|History of malignancy||22 (28.9%)|
|History of head and neck skin cancers||3 (3.9%)|
|History of local radiotherapy||6 (7.9%)|
|History of surgery at biopsy site||11 (14.5%)|
|Facial nerve dysfunction||13 (17.1%)|
Prior to CNB, each patient underwent a systematic physical checkup and ultrasound examination of the head and neck to evaluate the sonographic features of the salivary glands and potential additional pathologic findings, particularly lymph node enlargement. Colour-coded duplex sonography was applied to evaluate the target lesion’s vascularization and to find a safe pathway for the needle. Sonographic examinations were performed with a LOGIQ 400 ultrasound-scanner (GE Medical Systems, Milwaukee, USA, 7.5–13 MHz linear-array transducer) and a Sonoline G50 ultrasound-scanner (Siemens, Germany, 5–13 MHz linear-array transducer). Ultrasound was the only imaging procedure prior to CNB in 46 patients (61%). 30 patients (40%) underwent additional radiologic examinations including computed tomography (CT) in 21 (27%) and magnetic resonance imaging (MRI) in 9 (12%). Written informed consent for the procedure was obtained. CNB was performed without the attendance of a pathologist under real-time ultrasound-guidance using a freehand technique as described previously. A spring-loaded semi-automatic biopsy gun (Bard Magnum ® , Bard Inc., Covington, USA) was used with side-notch needles (length 100 mm; diameters 12 gauge (2.05 mm), 14 gauge (1.63 mm), 16 gauge (1.29 mm)) and a variable needle throw (forward feed, 15 or 22 mm) depending on the dimension of the target.
The core specimens obtained were immediately assessed by macroscopic inspection and either preserved in formalin for histopathologic examination or in saline for ancillary studies such as immunohistochemistry. Depending on the quality of the specimen, the suspected salivary gland pathology and the dimension of the target lesion multiple needle passes were done in one setting.
Sample quality was assessed by macroscopic inspection of the operator at the time of biopsy and deemed insufficient when CNB did not provide at least one tissue core for histological examination. If sufficient material was obtained, the sample adequacy was determined after microscopic examination from the content of the tissue core and deemed adequate when the target lesion was correctly sampled and inadequate in case of sampling error. If a biopsy core specimen did not contain salivary gland, tumour or lymph node tissue a sampling error was probable and the procedure was deemed inadequate but not a false-negative, because the target lesion had obviously been missed and treatment decisions were not based on these results. A result was deemed inconclusive when the sample was adequate but the pathologist could not differentiate definitely between a benign and a malignant lesion. A result was deemed false-negative (false-positive) when histologic examination of an adequate core specimen indicated benign disease (malignancy) but follow-up revealed malignancy (a benign disease). All insufficient, inadequate, false-positive, false-negative and inconclusive samples were summarized as unsuccessful results. Sensitivity, specificity, accuracy, positive predictive value (PPV) and negative predictive value (NPV) in differentiating malignant from benign disease were calculated with standard methods using Excel software for Windows (Microsoft Corporation, Redmond, WS, USA). The unit of analysis was each patient setting and not each CNB needle pass, as in one setting multiple needle passes may be performed. Statistical parameters were calculated before and after the exclusion of insufficient, inadequate and inconclusive results, because these cases involved re-biopsy but not incorrect treatment. In addition, specific emphasis was placed on the analysis, whether CNB enabled the correct classification of the entity of a salivary gland tumour. No particular distribution of biopsies between parotid and submandibular glands was made.
The 76 study patients underwent 161 biopsy procedures (needle passes). The number of needle passes per patient in one setting ranged from 1 to 5 (mean 2.1; median 2). In 66 patients (87%) the procedure was performed with local anaesthesia and in 10 (13%) under general anaesthesia with a panendoscopy done at the same time. 43 procedures (57%) were performed in inpatients and 33 (43%) in outpatients. Four (5%) patients revealed a diffuse salivary gland swelling and 72 (95%) a circumscribed mass. The maximum diameter of the target lesion in the 72 patients with a circumscribed mass ranged from 1.4 to 9.0 cm (mean 3.7 cm; median 3.0 cm) and was distributed as follows: <2 cm, 8 patients (11%); 2–2.9 cm, 28 patients (39%); 3–3.9 cm, 23 patients (32%); 4–5.9 cm, 8 patients (11%); 6 cm or more, 5 patients (7%). The size of the cutting needle used for biopsy was documented in 120 of the 161 needle passes, of which a 12 G needle was used for 78 (65%), a 14 G needle for 41 (34%) and a 16 G needle for 1 (8%). For the remaining 41 needle passes the needle size could not be determined.
Biopsies performed with the semi-automatic biopsy gun provided sufficient material for histological examination in 75 (99%) patients. One CNB was insufficient, open biopsy of this lesion later revealed non-specific inflammation. In 73 (97%) of the 75 sufficient results the specimens obtained were deemed adequate. In 2 patients (3%) the authors judged that the target tissue was not reached (sampling error) and the samples were deemed inadequate; secondary histologic examination of additional tissue sampling of these two lesions revealed adenocarcinoma (NOS) of the parotid gland in both cases. Histologic examination of the 73 adequate samples revealed malignant disease in 45 (62%) patients (including 33 carcinoma, 10 lymphoma and two other malignant tumours), benign disease in 26 (36%) patients (including 14 non-specific inflammations, seven specific inflammations or sialadenosis and five benign tumours or cysts), and was inconclusive in 2 (3%) patients. In the 73 patients in whom adequate samples were obtained, follow-up and final diagnosis were determined from the result of secondary histologic examination of excised specimens, either for diagnostic purpose or as part of the therapy, in 34 (47%) patients ( Table 2 ) and from the clinical course in 39 (53%) patients ( Table 3 ). Follow-up revealed no false-positive or false-negative results. The two inconclusive results were basal cell tumours that were finally diagnosed as basal cell adenoma and basal cell carcinoma. CNB was deemed successful in 71 (93%) and unsuccessful in 5 (7%) of the 76 study patients. Of the 71 patients in whom the CNB result was deemed successful, 39 were followed-up clinically and the course of the disease was concurrent with the CNB result in all these patients. 32 patients with successful CNB results were followed-up by secondary histologic examinations that exactly matched the CNB result in 25 patients. In four patients, secondary histologic follow-up was concurrent with the CNB result but enabled a more precise classification ( Table 2 ). In three patients, histologic follow-up revealed a different entity of disease than the previous CNB procedure ( Table 2 ). The diagnosis obtained by CNB prevented a surgical intervention in 39 (51%) patients. On the basis of secondary histologic and clinical follow-up the statistical parameters for diagnosing malignancy by CNB in salivary gland lesions were as follows. Including all results ( n = 76), sensitivity was 94%; specificity 100.0%; accuracy 96%; PPV 100%; NPV 90%. After the exclusion of insufficient ( n = 1), inadequate ( n = 2) and inconclusive ( n = 2) results ( n = 71): sensitivity, specificity, accuracy, PPV and NPV were each 100%. During follow-up the authors have recorded no case of tumour recurrence or dissemination that could be attributed to a previous CNB. Two minor complications occurred: one temporary facial weakness after local anaesthesia and one minor bleeding and subsequent haematoma. CNB was well tolerated by all patients who underwent the procedure under local anaesthesia.
|CNB result||n||Final histological diagnosis||n|
|Pleomorphic adenoma||1||→ Pleomorphic adenoma||1|
|Warthin’s tumour||3||→ Warthin’s tumour||3|
|Reactive lymph node||1||→ Reactive lymph node||1|
|→ Actinomycosis *||1|
|→ Bartonellosis (cat scratch disease) *||1|
|→ Warthin’s tumor †||1|
|Squamous cell carcinoma||11||→ Squamous cell carcinoma||10|
|→ Salivary duct adenocarcinoma †||1|
|Adenocarcinoma (NOS)||3||→ Adenocarcinoma (NOS)||2|
|→ Salivary duct adenocarcinoma *||1|
|Adenoid cystic carcinoma||2||→ Adenoid cystic carcinoma||2|
|Mucoepidermoid carcinoma (high-grade)||1||→ Mucoepidermoid carcinoma (high-grade)||1|
|Undifferentiated carcinoma||3||→ Undifferentiated carcinoma||1|
|→ Salivary duct adenocarcinoma *||1|
|→ Carcinosarcoma †||1|
|Basal cell tumour||2||→ Basal cell adenoma||1|
|→ Basal cell adenocarcinoma||1|
|CNB result||n||Clinical course||n|
|Sialadenitis||6||→ Infection, conservative Th||6|
|Tuberculosis||1||→ Tuberculosis, conservative Th||1|
|Cyst||1||→ Benign, no treatment||1|
|Sarcoidosis||3||→ Sarcoidosis, conservative Th||3|
|Actinomycosis||1||→ Infection, conservative Th||1|
|Sialadenosis||1||→ Benign, no treatment||1|
|Auto-immune inflammation||1||→ Sjögren’s syndrome, conservative Th||1|
|Squamous cell carcinoma||5||→ Malignancy, primary RTh/RChTh||4|
|→ Malignancy, palliative ChTh||1|
|Adenocarcinoma (NOS)||7||→ Malignant, treatment refused||1|
|Malignant, primary RChTh||6|
|Poro carcinoma||1||→ Malignant, primary RChTh||1|
|Undifferentiated carcinoma||1||→ Malignant, treatment refused||1|
|Malignant melanoma||1||→ Malignant, treatment refused||1|
|Follicular lymphoma||3||→ LD, RChTh||1|
|→ LD, ChTh||2|
|Diffuse large cell lymphoma||5||→ LD, ChTh||5|
|Acute myeloid leukaemia||1||LD, ChTh||1|
|Plasmazytoma||1||→ LD, ChTh||1|