Ancillary studies on cell blocks from fine needle aspiration specimens of salivary gland lesions: A multi‐institutional study

Abstract Background Ancillary studies are commonly performed on cell blocks prepared from fine‐needle aspiration (FNA) specimens. There are limited studies in application of ancillary studies on cell blocks from salivary gland (SG) FNAs. This multi‐institutional study evaluates the role of ancillary studies performed on cell blocks in the diagnosis of SG lesions, and their impact on clinical management. Method The electronic pathology archives of three large academic institutions were searched for SG FNAs with ancillary studies performed on cell blocks. The patient demographics, FNA site, cytologic diagnosis, ancillary studies, and surgical follow‐up were recorded. If needed, the cytologic diagnoses were reclassified as per the Milan System for Reporting Salivary Gland Cytopathology (MSRSGC). Results 117 SG FNA cases were identified including 3, 10, 11, 6, 23, 4, and 60 cases in MSRSGC categories I, II, III, IVa, IVb, V, VI, respectively with surgical follow‐up available ranging from 27% to 100% within each category. Ancillary studies including histochemistry, immunocytochemistry (IHC), and in situ hybridization (ISH) were beneficial in 60%–100% of cases in each category. Risk of malignancy was 100% in both the suspicious for malignancy (V) and malignant (VI) categories. Ancillary studies improved diagnosis in 60% of non‐neoplastic cases (II, 6/10), 100% of benign neoplasm cases (IVa, 6/6), and 98.3% of malignant cases (VI, 59/60). Conclusion Judicious and case‐based ancillary studies performed on SG FNA cell blocks with sufficient material can improve the diagnostic yield by further characterization of the atypical/neoplastic cells, particularly in MSRSGC categories IVa‐VI.


| INTRODUCTION
Fine-needle aspiration (FNA) is a well-accepted procedure to evaluate salivary gland lesions. [1][2][3][4] It is up to 79% sensitive and 96% specific in detecting malignancy, and up to 96% sensitive and 98% specific in the detecting neoplasia, respectively. 5 Although most commonly occurring salivary gland neoplasms pose little diagnostic challenge on FNA (i.e., pleomorphic adenoma or Warthin tumor), differentiating between non-neoplastic processes, benign lesions, and/or malignancies is not always achievable on routine stains due to cellular heterogeneity and overlapping architectural features. 6,7 In an effort to standardize SG FNA reporting and streamline downstream clinical management, the Milan System for Reporting Salivary Gland Cytopathology (MSRSGC) established six distinct diagnostic categories with associated risk of malignancy (ROM) based on cytomorphologic features. 4,8,9 In the era of precision diagnostics, ancillary studies are often being performed on cytology specimens to provide a specific diagnosis and even prognostic information for optimal patient management. 9 A wide array of ancillary studies such as immunocytochemistry, fluorescence in situ hybridization (FISH), DNA or mRNA in situ hybridization (ISH) can be performed on cell blocks. Salivary gland neoplasia arises from a variety of cell types, which can be delineated utilizing immunocytochemistry. A small panel of immunostains may yield a definitive diagnosis, even with minimal material. For example, p16 a surrogate marker in diagnosing HPV-related squamous-cell carcinoma, allows for a more definitive diagnosis than cytological examination alone. 10 However, cell blocks are not routinely prepared for all SG FNA cases due to utilization of aspirated material for direct microscopic examination and when cell blocks are available, they may be insufficient for ancillary studies.
In this multi-institutional retrospective study, we evaluated the utility of cell blocks with subsequent performance of ancillary studies in the diagnosis of salivary gland lesions classified according to the MSRSGC.

| MATERIALS AND METHODS
The study was conducted after obtaining institutional research Each institution reviewed its own cases individually and classified the cases into the MSRSGC categories. The ancillary studies included in this study were immunohistochemical stains, histochemical stain and stains for detection of mucin, bacterial, fungal and mycobacterial micro-organisms and in situ hybridization. The following data points were recorded for each patient: tumor type, sex, age, biopsy site, FNA diagnosis, cytologic category per MSRSGC, type and results of ancillary studies performed, and surgical follow-up diagnosis when available. The study included the pathology report review only.

| RESULTS
One hundred and seventeen SG FNA specimens met the inclusion criteria. These included 67 male patients and 50 female patients, ranging in age from 2 to 92 years with a mean of 61.1 years and median of 63 years. The parotid gland was the most common site (101 lesions), followed by minor salivary glands (9 lesions), and submandibular gland (7 lesions). The MSRSGC diagnostic category distribution was as follows: 3 (2.6%) cases as non-diagnostic, 10 (8.5%) as non-neoplastic, 11 (9.4%) as atypia of undetermined significance (AUS), 6 (5.1%) as benign neoplasm, 23 (19.7%) as salivary gland neoplasm of uncertain malignant potential (SUMP), 4 (3.4%) as suspicious for malignancy, and 60 (51.3%) as malignant ( Figures 1A-3B Therefore, this study and its finding presents a small number of cases that contained sufficient material for ancillary studies. The amorphous matrix in the background posed diagnostic difficulties, particularly in cystic and hypocellular specimens. Mucicarmine stain and thyroglobulin were used to highlight mucin and colloid in two cystic cases, respectively ( Table 2).
The presence of inflammatory cells, epithelioid histocytes and granulomatous inflammation triggered the initial pathologists to investigate an underlying infectious process. Gram stain, GMS stain, Zeihl Neelsen stain, Warthin-Starry stain, Brown Hopps stain, and spirochete immunostains were utilized in these cases. Although a negative stain cannot exclude an infectious process, a positive stain detecting microorganisms confirms an infectious process. These stains were utilized more often in non-neoplastic cases (Tables 1 and 2 cases. 12 Androgen receptor immunostain was positive in salivary duct carcinoma, while p63 was negative. 13

DATA AVAILABILITY STATEMENT
The data will be available upon request.