Lymphoid cell rich fine-needle aspirations of the salivary gland: What is the risk of malignancy?

INTRODUCTION

Fine-needle aspiration (FNA) of the salivary glands is a minimally invasive approach for evaluating mass-forming lesions. The primary aim of FNA is to help guide management by distinguishing between non-neoplastic and neoplastic processes.^[1] The current Milan system establishes six diagnostic categories and serves as a useful tool for standardized reporting of these lesions.^[2-4] Early in embryologic development, lymphocytes colonize the parenchyma of major salivary glands, and subsequently form intraparotid lymphoid aggregates and lymph nodes. A wide array of lesions can arise from these. Around 2–8% of FNAs of the salivary glands have been reported to be from such lymphoid cell rich lesions. The main diagnostic concern in these situations is lymphoproliferative disease. Lymphoma represents 3% of all salivary gland neoplasms and 10% of malignant salivary gland tumors. Higher incidence has been reported in patients with Sjogren syndrome and HIV. They arise primarily in the parotid gland (75–80%), but also in the submandibular gland (15–20%) and infrequently in the sublingual and minor salivary glands.^[5] Pre-operative FNAs yielding a predominant population of lymphoid cells can be challenging, as they could represent a wide variety of disease processes. The goal of this study was to evaluate the incidence of malignancy in lymphoid cell rich FNAs of the major salivary glands, and to characterize surgical outcomes. Our secondary aim was to examine the types of lymphoproliferative disease present.

MATERIAL AND METHODS

This retrospective study was approved by the Institutional Review Board. Our institutional database was queried over a 10-year period for FNAs of mass-forming lesions of the salivary gland with a predominant cytologic finding of lymphoid cells. Cases with scant cellularity, epithelial cells, or diagnostic features of a specific entity (granulomas and chondromyxoid stroma) were excluded from the study. Patients with a clinical history of metastatic malignancy were excluded from the study. Only cases with surgical follow-up were included in the study. Information on clinical history was also collected.

When applicable, cytologic diagnoses were rendered according to the Milan System for Reporting Salivary Gland Cytopathology, where the diagnostic category for these cases was either Non-diagnostic (polymorphous population of lymphoid cells) or Atypia of undetermined significance (lymphoid cells with cytomorphologic atypia).^[6] FNA cytology slides were reviewed and lymphocyte atypia was noted. Lymphocytes were considered atypical if they had one or more of these morphologic findings: Monomorphism, irregular nuclear contours, and abnormal chromatin patterns.^[7] Correlation between cytomorphologic atypia noted in FNAs and subsequent malignancy diagnosis was assessed. For each case, the final surgical diagnosis was classified as one of the following: Non-neoplastic, benign, epithelial – malignant, lymphoma. Diagnosis of lymphoproliferative neoplasms on biopsy and surgical excision specimens was performed by a hematopathologist. The diagnosis of these lymphoproliferative neoplasms was based on the criteria outlined by the 2017 WHO classification of hematopoietic tumors.^[8]

Depending on clinician preference and degree of suspicion for a lymphoproliferative process, flow cytometry was performed concurrently with the FNA in select cases. White blood cells were obtained from fine-needle aspirate material using routine methods, including red blood cell lysis (as applicable). Cases were assessed by a routine lymph node/tissue flow cytometric panel. The routine panel included three tubes. Tube 1 contained CD38-FITC, CD56-PE, CD3-ECD, CD5-PerCP-Cy5.5, CD7-PE-Cy7, CD2-APC, CD19-APC-A700, CD4-APC-A750, CD8-PB, and CD45-KO. Tube 2 contained Kappa-FITC, Lambda-PE, CD19-ECD, CD11c-PerCP-Cy5.5, CD5-PE-Cy7, CD22-APC, CD23-APC-A700, CD10-APC-A750, CD20-PB, and CD45-KO. Tube 3 was a viability tube with 7-AAD and CD45-KO. Events were acquired on the Navios Flow Cytometer (Beckman Coulter). Data were analyzed using FCS Express software (De novo Software) for phenotypic expression and interpreted by a hematopathologist.

RESULTS

A total of 224 lymphoid rich FNA cases were reviewed, and 29 (n = 29) met all the inclusion criteria for the study. Of these, 22 cases were from the parotid gland and seven from the submandibular gland. The indication for FNA was a swelling or nodule in the major salivary glands. None of these patients had a prior history of lymphoma.

The case distribution based on surgical follow-up is summarized in [Figure 1]. Ten cases (n = 10, 35%) were nonneoplastic. This included either intra-parotid or periparotid reactive lymph nodes (n = 4), benign lymphoepithelial cyst (n = 5), and chronic sialadenitis (n = 1). Three cases were benign neoplasms, including pleomorphic adenoma (n = 2) and Warthin’s tumor (n = 1). One case of mucoepidermoid carcinoma (n = 1) was found.

Close

Figure 1:

Surgical follow-up of lymphoid cell rich fine-needle aspirations of the salivary glands.

Export to PPT

Fifteen cases (n = 15; 52%) were lymphomas and subtyping is summarized in [Figure 2]. Low-grade lymphomas represented 54% of cases and included follicular lymphoma Grade 1 or 2 (n = 4; 27%), marginal zone lymphoma (n = 2; 13%), mantle zone lymphoma (n = 1, 7%), and unclassified atypical lymphoid proliferation favor low-grade lymphoma (n = 1; 7%). High-grade lymphomas included diffuse large B cell lymphoma (DLBCL) (n = 6; 40%) and follicular lymphoma Grade 3B (n = 1, 7%). Histologically, nine cases were primary salivary gland lymphomas arising in the parenchyma (n = 9; 60%) and six cases originated in an intraparotid or periparotid lymph node (n = 6; 40%).

Close

Figure 2:

Cases with lymphoma diagnosis (n = 15) on follow-up. *One case was indeterminate and classified as atypical lymphoid proliferation, favor low-grade lymphoma. DLBCL: Diffuse large B cell lymphoma.

Export to PPT

Of the 15 lymphoma cases, lymphoid cells were cytomorphologically atypical on FNA in 12 cases (n = 12/15) [Figures 2 and 3]. Correlation with surgical outcome was assessed [Table 1]. The specificity of atypia on FNA for malignancy was high at 92%. The sensitivity was 69%. One case with atypical lymphoid cells was a benign reactive lymph node on excision. The positive predictive value of cytomorphology for malignancy was 92% and the negative predictive value was 71%. Ancillary studies were available in a few cases and supportive of the diagnosis of lymphoma. This included cell block and immunohistochemistry (n = 7, 47%), flow cytometry (n = 3, 27%), and clonality PCR (n = 1; 7%).

Close

Figure 3:

Example of a case with atypical lymphocytes on fine-needle aspiration (a), with follicular lymphoma on follow-up (b).

Export to PPT

The majority of cases with non-atypical lymphocytes were benign (12/17) with five malignant on follow-up (5/17). The five false negative cases were subsequently diagnosed as: Mucoepidermoid carcinoma (n = 1), mantle zone lymphoma (n = 1), follicular lymphoma Grade 1 or 2 (n = 2), and Grade 3B (n = 1).

DISCUSSION

Abundant lymphoid cells can be seen in salivary gland aspirates in a variety of situations. It could represent sampling of normal lymphoid tissue from an intraparotid or periparotid lymph node or benign neoplasms such as lymphoepithelial cyst and Warthin’s tumor. Malignant epithelial neoplasms with tumor associated lymphoid proliferation include mucoepidermoid carcinoma, acinic cell carcinoma, and lymphoepithelial carcinoma. Importantly, the main concern is sampling of a lymphoma.

Our study is the first to evaluate incidence of malignancy (lymphoma), in FNAs composed exclusively of lymphocytes, from both the parotid and the submandibular glands. Results from selected prior publications^[9-12] are summarized in [Table 2]. A study looking at metastatic and secondary malignancies in the major salivary glands found approximately one-third of their series involved a lymphoproliferative disease (13/36).^[13] Another study specifically examining periparotid and intraparotid lymph node FNAs found that the vast majority (18/20) of the cases were diagnosed as reactive. Two cases were diagnosed as lymphoma on surgical excision, one as marginal zone lymphoma and one as follicular lymphoma. Both studies underscored the role of FNA in the pre-operative diagnosis of salivary gland lymphomas, emphasizing that flow cytometry is a useful adjunct to cytologic preparations.^[14]

In our series of 29 cases with histologic follow-up data, the incidence of lymphoma was 52%. Important caveats of our study are the small sample size, as well as the lack of follow-up surgical data on the cases that were not included in the study. One additional case of epithelial malignancy was identified, mucoepidermoid carcinoma, known to be associated with lymphocytic reaction in the salivary gland. About 54% of the lymphomas were low grade, which are usually difficult to detect in FNAs especially in the absence of prior history of lymphoma (such as in this study). About 46% were high grade lymphomas, including DLBCL and Grade 3B follicular lymphoma. Atypical lymphocyte cytomorphology on salivary gland FNA (monotony, irregular nuclear contours, and abnormal chromatin patterns) had a 92% positive predictive value for malignancy and was highly specific (specificity = 92%). The sensitivity was 69% with five false negative cases that included one epithelial malignancy (epithelial component not sampled on FNA), three low-grade lymphomas, and one follicular lymphoma Grade 3B (high-grade component not sampled on FNA). The benign cytologic diagnosis in these five cases may be secondary to sampling bias in the case with epithelial malignancy and the Grade 3B follicular lymphoma, as well as the subtle mild atypical nature of low-grade lymphomas that can be easily overlooked on cytology.

Lymphomas arising in the salivary gland are essentially B-cell derived, and mostly arise in the parotid. The most common is extra-nodal marginal zone lymphoma or mucosa-associated lymphoid tissue lymphoma (MALT-L), which arises often in patients with Sjogren’s disease. This is followed by follicular lymphoma, which is more commonly associated with a systemic disease. The third most common is DLBCL, which can show concurrent cervical lymph node involvement.^[5] The distribution of lymphoma subtypes in our study was different from what is described in the literature. Namely, MALT-L or extranodal marginal zone lymphoma was only found in 13% of cases, while DLBCL was diagnosed in 40%. None of the 29 patients had a documented history of lymphoma before FNA; therefore, it is more likely that they represent either primary lymphomas of the salivary gland or of peri/ intraparotid lymph nodes. The relative higher incidence of FL and DLBCL could be due to a selection bias, perhaps due to the inclusion of only lymphocyte rich aspirates and importantly the small cohort size. Additional large-scale studies may help clarify the incidence and distribution of primary lymphomas in this setting.

Although only few cases had ancillary studies performed in our series, it is well documented that flow cytometry and immunohistochemistry improve diagnostic yield on FNA. These findings argue for the use of rapid on-site evaluation for salivary gland FNAs, and to triage specimens for ancillary studies, especially in cases where lymphoid cells are not overtly atypical and are abundant.

Four cases in our cohort proved to be epithelial neoplasms on histologic follow-up (two cases of pleomorphic adenoma, one case of Warthin’s tumor, and one case of mucoepidermoid carcinoma). These cases highlight an important pitfall of pre-operative sampling. Several neoplasms of the salivary gland can be accompanied by a polymorphous lymphoid infiltrate (so-called “tumor-associated lymphoid proliferation”),^[15] which may be the only component sampled on the FNA. As such, inclusion of these cases in our study reflects a more realistic calculation of the risk of malignancy for these aspirates.

CONCLUSION

The differential diagnosis of lymphoid cell rich FNAs includes non-neoplastic lesions, benign neoplasms, non-lymphoid malignancies, and lymphoproliferative diseases. The incidence of lymphoma in our cohort of 29 lymphocyte rich FNAs from the parotid and submandibular glands with surgical follow-up is 52%. Cytomorphologic atypia on FNA had a high specificity (92%) and positive predictive value (92%) for the diagnosis of lymphoma on follow-up. Rapid onsite triaging of salivary gland FNA can be helpful in increasing sensitivity of lymphoma detection, particularly when the morphology is non-atypical or without overt atypical features. In such cases where ancillary studies such as flow cytometry are not available to confirm the benign nature of these lymphocyte rich lesions, close clinical follow-up and possibly additional sampling may be helpful to further confirm absence of malignancy.