Fine-Needle Aspiration Cytology in the Diagnosis of Salivary Gland Tumors Before the Milan System: A Ten-Year Experience From a Tertiary Care Center in Greece

Objective The objective of this study was to determine the diagnostic value of fine-needle aspiration cytology (FNAC) for salivary gland tumors. Methodology A retrospective file analysis of patients with salivary gland pathology, attending the Department of Oral and Maxillofacial Surgery of a tertiary care center in Athens, Greece, over a 10-year-long period, was conducted. Sensitivity, specificity, accuracy, positive prognostic value (PPV), and negative prognostic value (NPV) of FNAC for benign and malignant tumors separately were assessed and compared with histology. Results A total of 82 patients (46 male and 36 female) with salivary gland tumors, submitted to both FNAC and histology, were included. The mean age was 55 years. A total of 73 tumors were histologically diagnosed as benign and nine as malignant. FNAC identified 62 benign and seven malignant tumors but was inconclusive in 13 cases. The most common diagnosis of both histology and FNAC was pleomorphic adenoma. FNAC sensitivity, specificity, accuracy, PPV, and NPV were 98.3% and 100%, 87.5% and 100%, 97.1% and 100%, 98.3% and 100%, and 87.5% and 100% for benign and malignant tumors, respectively. Conclusions FNAC is highly sensitive but moderately specific for the preoperative identification of benign salivary gland tumors. Its use as an initial diagnostic modality is warranted, thanks to its safeness, rapidity, and lack of pain.


Introduction
Salivary gland tumors are clinically diverse, with an estimated overall annual incidence of approximately 2.5-3 cases per 100,000 in the Western world [1]. Salivary gland malignancies account for approximately 3% to 6% of head and neck tumors, with an overall annual worldwide incidence of 0.4-13.6 cases per 100,000, with 80% affecting the parotid gland [2]. The World Health Organization classifies salivary gland epithelial tumors into 10 benign and 23 malignant entities, while nonepithelial represent only about 2% to 5%.
Fine-needle aspiration cytology (FNAC) is a diagnostic tool, used widely thanks to its safeness, costeffectiveness, rapidity, repeatability in case of nondiagnostic results, and lack of pain [3]. Performed by experienced operators and correlated with clinical information, it effectively diagnoses the nature of the lesion. Its varying accuracy has been ascribed to many reasons. Nevertheless, fine-needle aspiration (FNA) has been related to complications, such as neoplastic cell seeding, bleeding, inflammatory reaction, and/or nerve impairment [4][5][6].
Especially in the diagnosis of salivary gland tumors, FNAC is becoming popular, due to the tumors' superficial location and easy accessibility. Given the noncharacteristic clinical and/or radiologic features of these tumors, FNAC is considered superior to combined physical and radiologic examination [5,7,8].
Recently, ultrasound-guided core biopsy has been studied as an alternative for the preoperative evaluation of parotid lesions, but its use as a primary diagnostic tool is restricted by the high incidence of hematoma and facial nerve injury [9].
In 2018, an international group of experts, supported by the American Society of Cytopathology and the International Academy of Cytology, introduced the Milan System for Reporting Salivary Gland Cytopathology (MSRSGC), containing various diagnostic categories (subdivisions of FNAC results), their appurtenant risk of malignancy, and advised management strategy [10].
The objective of the study was to determine the diagnostic value of FNAC for salivary gland tumors among patients, who were attending the Department of Oral and Maxillofacial Surgery of a tertiary care center in Athens, Greece, during a 10-year-long period in the pre-MSRSGC era and compare our findings with those reported in the recent literature.

Materials And Methods
The files of all patients with salivary gland (including parotid, submandibular, and minor oral salivary glands) pathology, diagnosed from 2009 to 2019, were retrieved from the database and analyzed retrospectively. Harvested data included patients' demographics and history, lesion characteristics (location and size), as well as histological and FNAC diagnoses.
FNAs had been performed by specialized clinicians of our department, without ultrasonography assistance, by using a 22-gauge needle, with a minimum of two needle passes in each case, while an experienced cytopathologist was present to assess the slide adequacy immediately after sample collection.
For this study, each tumor was classified as either benign or malignant.

Benign tumors
Benign tumors were classified into four categories: (1) true positive (TP), cases diagnosed as benign tumors both cytologically and histologically; (2) true negative (TN), cases diagnosed as malignancies both cytologically and histologically; (3) false positive (FP), cases initially diagnosed as benign tumors cytologically but not confirmed histologically as benign; and (4) false negative (FN), cases not identified as benign tumors by FNAC, but later diagnosed as benign on histological examination.

Malignant tumors
Malignant tumors were classified into four categories: (1) TP, cases diagnosed as malignant tumors both cytologically and histologically; (2) TN, cases diagnosed as non-malignant tumors both cytologically and histologically; (c) FP, cases histologically diagnosed as benign tumors but misdiagnosed cytologically as malignancies; and (4) FN, cases misdiagnosed cytologically as benign tumors, but later confirmed histologically as malignancies. Subsequently, FNAC diagnoses were compared with the corresponding histology; discrepancies between FNAC and histology on the histological type of the tumors were registered and analyzed. Kappa statistics were used to assess the agreement between FNAC and histology in the diagnosis of both benign and malignant tumors.
Our results were compared with those of the most relevant articles in English, with a sample size of over 40 patients, published over the last 15 years, providing values for sensitivity, specificity, diagnostic accuracy, PPV, and NPV at least for salivary gland malignancies.

Results
A total of 82 salivary gland tumors, submitted to both FNAC and histology, were included. One of these (#31), originally misdiagnosed through FNA as pleomorphic adenoma, was histologically typed as a lymphoepithelial cyst. The age range was 27 to 81 (mean 55) years and the male-to-female ratio was 46:36 (0.78). Patients' data, tumor characteristics, and type of surgery are summarized in Table 1  Of the 82 included tumors, 70 affected the parotid, 10 were the minor oral (six of the hard palate, three of the soft palate, and one of the upper lip), and two were the submandibular salivary glands ( Table 1).
It is worth mentioning that FNAC was nondiagnostic in 11 benign and two malignant tumors; thus, 13 out of the 82 FNAC specimens (17.07%) were inconclusive.

FNAC, fine-needle aspiration cytology
The diagnostic value of FNAC for identifying salivary gland tumors, including sensitivity, specificity, PPV, NPV, and accuracy for both benign and malignant tumors are presented in Table 3.

Discussion
FNAC has been gaining acceptance as an adjunct for preoperative diagnosis of salivary gland pathology, obviating surgery in approximately 33% of the patients, and providing useful information about the extent and time of indicated surgery [10][11][12]. It speeds up the diagnostic process by providing a diagnosis within 24 hours, whereas histopathology may take five to seven days, thus reducing the morbidity and mortality associated with malignancies [13]. FNAC is a safe, cost-effective, and minimally invasive procedure that causes little patient discomfort. However, it carries a low risk of malignant cell dissemination [5,8]. Due to the morphological complexity of salivary gland tumors, the relatively small FNAC specimen may not be representative, with varying proportions (2%-10%) of inadequate smears [14,15]. Ultrasonography-assisted FNAC has been recommended to reduce sampling errors [16,17]. Variables such as the operator's experience, applied technique, promptness of sample evaluation, and the cytopathologist's diagnostic skills influence FNAC accuracy [3,14,16,17].
Especially for the parotid gland, FNAC is included in the diagnostic workup to (a) rule out inflammation, (b) identify systemic disease (reticuloendothelial tumors), (c) exclude direct gland invasion or metastases in patients with oncologic history, (d) evaluate unresectable tumors in poor surgical candidates, and (e) assess patients with low risk of neoplasms, such as children. Preoperative FNAC in parotid malignancies allows staging, determines the surgical plan (resection margins, need for neck dissection, and urgency degree), and assesses the risk of postoperative facial palsy [10,15]. Histology following incisional biopsy is still considered the gold standard to evaluate malignant potential, although potentially leads to fistula or sialocele formation, tumor implantation, or facial nerve damage [5,6].
FNA-induced tissue alterations, such as necrosis, hemorrhage, cellular atypia (squamous metaplasia), cellular proliferation, displacement through the tumor capsule, and/or architecture distortion (stromal hyalinization) that confound subsequent histological diagnosis, have been reported [14]. However, FNAC benefits outweigh the risk of distortion beyond histological recognition, thus not downgrading its diagnostic value.
Our results showed 98.36% and 100% sensitivity, 87.5% and 100% specificity for benign and malignant tumors, respectively; PPV amounted to 98.36% and 100%, NPV to 87.5% and 100%, and diagnostic accuracy to 97.1% and 100%. The relatively high PPV (lack of FPs) is attributed to the relatively small sample size and the low prevalence of malignancies (10.14%); larger studies are more likely to encounter FPs, thus reducing both PPV and specificity.
According to the Cytopathology Resource Committee of the College of American Pathologists, participating laboratories were inadequate to provide a specific diagnosis, although cytologically differentiating benign from malignant tumors (<50% agreement between FNAC and histology) [18]. Tumor type and grade reportedly affect the diagnostic accuracy of FNAC, with specific tumors contributing disproportionately to diagnostic errors; accurate diagnosis may be achieved in high-grade carcinomas, while negative results are mostly encountered in low-grade carcinomas [14,18]. However, precise preoperative typing is not particularly relevant in malignancy, as accurate grading is the most substantial parameter for subsequent surgical planning (radicality of surgery and need for neck dissection) [14,29].
The failure of FNAC to identify specific salivary gland tumors may be ascribed to their morphological complexity (various growth patterns and diverse cell types) and the limited FNA sample that may not be representative of the complete tumor morphologic spectrum [14,15,21,24]. FNAC demonstrates inadequate tumor invasiveness, failing to diagnose carcinomas such as adenoid cystic carcinoma that are highly infiltrative, although appearing bland and nonthreatening cytologically [14]. FNAC also fails to evaluate malignancy signs, such as capsular infiltration, and perineural and perivascular invasion [30].
The pitfalls of FNAC reportedly lie in the diagnosis of pleomorphic and basal-cell adenomas, as well as carcinomas and lymphomas [1,7,13,26]. In pleomorphic adenoma, a few atypical cells may be present, while monomorphic adenoma can be confused with adenoid cystic carcinoma due to the presence of similar stromal components [13]. FNAC accuracy for the diagnosis of largely cystic tumors (Warthin's tumor, mucoepidermoid carcinoma, and adenoid cystic carcinoma) is reportedly small, due to the inadequate architectural information provided through FNAC [6,7,14,23]. Mucoepidermoid carcinoma is particularly challenging to diagnose cytologically, as FN diagnoses may occur due to fluid dilution or bland-looking intermediate cells, which can be misinterpreted as benign.
The limitations of FNAC in differentiating reactive lymphoid hyperplasia from lymphoma are welldocumented. Lymphocyte-predominant lesions significantly contribute to both FN and FP cytological results. Moreover, FNAC fails to distinguish between reactive lymphoid hyperplasia and granulomas originating from either the salivary glands or the intra-/peri-salivary lymph nodes in the parotid and submandibular region [6]. The rate of nondiagnostic (inconclusive) FNAC among our patients was 17.07% (13/82). These were excluded from statistical analysis to minimize the risk of overrating FN results. Although consistent with the reported results in the literature (3%-34%) [6,8,9,14], this rate was partly ascribed to the lack of ultrasound guidance, resulting in low diagnostic yield for deeply located parotid tumors or cystic tumors [14,17], such as Warthin's tumors (8/13 inconclusive cases). However, even in studies with ultrasound-assisted FNA, nondiagnostic results were attributed to complex tissue architecture [3] or inadequate smear cellularity and improper sampling [12]. Moreover, failure to obtain representative specimens may be due to needle positioning outside the tumor or into a central necrosis or hemorrhage area [24]. In our study, all inconclusive FNAC specimens were uninterpretable due to inadequate smears, possibly owing to improper sampling or limited smear cellularity, common in cystic tumors (Warthin's and mucoepidermoid carcinomas); thus, they were excluded from further analysis to avoid bias of the estimated FNAC diagnostic value. However, one should bear in mind that according to the MSRSGC, nondiagnostic FNAC cases present a 25% risk of malignancy [10].
The retrospective nature of this study is a considerable limitation for the following reasons: (1) some patients' files may have inadequate data, and access to a more comprehensive database would improve patient care and facilitate future studies aiming to understand salivary gland pathology and enhance the accuracy of FNAC; and (2) inability to overcome the invalidity of at least some of our inconclusive FNAC results by repeating the FNA procedure.

Conclusions
Based on our findings FNAC is highly sensitive and moderately specific for the diagnosis of benign salivary gland tumors, with high specificity for salivary gland malignancies, establishing the need for surgery but not ensuring histological typing.
FNAC is a relatively safe, rapid, and noninvasive method, when compared with an incisional biopsy but plays a limited role in determining the radicality of surgery (spare or sacrifice the facial nerve). Further investigation is required to elucidate factors, defining the diagnostic accuracy of FNAC, that is optimized when combined with clinical and radiologic evaluation.

Additional Information Disclosures
Human subjects: Consent was obtained or waived by all participants in this study. Animal subjects: All authors have confirmed that this study did not involve animal subjects or tissue. Conflicts of interest: In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.