Diagnostic Utility of the Milan System for Reporting Salivary Gland Cytopathology With Cytohistological Correlation

ABSTRACT

Introduction

Fine needle aspiration cytology (FNAC) is an affordable diagnostic method for assessing salivary gland lesions. However, interpreting cytology results of salivary gland FNACs can be difficult due to the wide variety of lesion morphologies and overlapping features among different tumour subtypes. To address this challenge, the “Milan System for Reporting Salivary Gland Cytopathology” (MSRSGC) was introduced to aid in diagnosis, management, and to determine the risk of malignancy (ROM) across different categories. Our study aims to assess the cytological spectrum of salivary gland lesions on fine needle aspiration cytology and evaluate the diagnostic utility of MSRSGC with histological correlation.

Methods

Cytology slides of FNAC over an 11‐year retrospective period were retrieved, reviewed and categorized based on the MSRSGC. Histological findings were correlated with cytological findings in cases where excision specimens were received for histopathological evaluation.

Results

A total of 111 cytology cases were categorized by MSRSGC. All cases were clinically followed up, and histopathological correlation was available for 34 cases. The overall specificity for diagnosis of neoplastic lesions on cytology was 71.40%, sensitivity was 96%, and accuracy was 90.63%. The specificity for diagnosis of malignant lesions on cytology was 95.83%, sensitivity was 100%, accuracy was 96.29%, positive predictive value was 75%, and negative predictive value was 100%.

Conclusion

FNAC is an excellent, non‐invasive and resource‐efficient tool for diagnosing salivary gland lesions. The MSRSGC six‐tier system aids in managing tumours with complex morphology and overlapping cytological characteristics.

This retrospective study reviews 111 salivary gland fine‐needle aspiration cytology (FNAC) cases collected over an 11‐year period and reclassifies them according to the Milan System for Reporting Salivary Gland Cytopathology. Cytology‐histopathology correlation was performed for excision specimens, confirming the system's accuracy in distinguishing non‐neoplastic, benign, and malignant lesions and its utility for risk stratification and patient management.

1. Introduction

Salivary gland lesions range from non‐neoplastic cysts and inflammatory conditions to tumour‐like and neoplastic lesions. Salivary gland neoplasms (SGNs) are rare, accounting for 3%–10% of head and neck tumours [1], with an incidence of 0.4 to 13.5 per 100,000 population [2]. Malignant SGNs occur in 0.4–2.6 cases per 100,000 population annually [2]. Approximately 80% occur in the parotid gland, 10%–15% in the submandibular gland, and the remainder in the sublingual gland and minor salivary glands of the upper aerodigestive tract [3].

Fine‐needle aspiration cytology (FNAC) is a cost‐effective, minimally invasive, and preferred method for evaluating unexplained salivary gland masses. Aspirations are usually performed with 21–25‐gauge needles, most commonly 23‐gauge, minimizing tissue trauma while yielding adequate cellularity. A meta‐analysis reports FNAC sensitivity of 96% and specificity of 98% for diagnosing neoplasms, though distinguishing benign from malignant tumours shows lower sensitivity (79%) and specificity (96%) [4].

Cytologic findings significantly influence surgical management. High‐grade malignancies receive more aggressive treatment than low‐grade or benign neoplasms. The diversity of lesions and inconsistent reporting made FNAC interpretation challenging for clinicians. In September 2015, at the European Congress of Cytology in Milan, Italy, the American Society of Cytopathology (ASC) and the International Academy of Cytology (IAC) sponsored the development of the “Milan System for Reporting Salivary Gland Cytopathology” (MSRSGC). This six‐tiered system standardized diagnoses to improve communication and patient care. Since its initial publication in 2018, many studies have validated its usefulness in providing consistent prognostic and management guidance [5, 6]. The second edition, published in July 2023, incorporates updates from surgical outcomes, ancillary studies, and the 2022 WHO Classification of Head and Neck Tumours, and removes the prior 60‐cell adequacy threshold, emphasizing clinical context and the cytopathologist's judgement [3].

The present study aims to examine the cytological characteristics of salivary gland lesions in a multispecialty hospital. Cytological findings were used to classify lesions into distinct diagnostic categories according to the MSRSGC. The cytological diagnoses were compared with histological results where available to evaluate the reliability and clinical usefulness of the MSRSGC.

2. Material and Methods

The present study is an 11 years retrospective observational study conducted at the Department of Pathology in a multispecialty hospital in western India. All consecutive FNAC cases referred to the Pathology Department for FNAC of salivary gland lesions between January 2012 and December 2022 were included, irrespective of patient age or gender. Informed consent for FNAC was obtained prior to the procedure in all cases. A total of consecutive 111 FNAC cases of salivary gland lesions with adequately preserved slides were studied; cases with poorly preserved slides were excluded. Papanicolaou‐ and Giemsa‐stained slides were retrieved from archives and independently reviewed and re‐categorized according to the MSRSGC by a consultant and trainee pathologist, with consensus review as needed. Reviewers were blinded to histopathological outcomes. Formalin‐fixed, H&E‐stained tissue sections from follow‐up excisions were examined, and findings correlated with cytology. Clinical details, including presentation and follow‐up, were obtained from patient records.

The spectrum of salivary gland lesions was analysed with respect to clinical and demographic parameters, and cytological diagnoses were correlated with histopathology and follow‐up to assess the MSRSGC's utility in guiding management.

The specificity, sensitivity, and accuracy were calculated for non‐neoplastic lesions (MSRSGC category II) and neoplastic lesions (MSRSGC categories IVA, IVB, V, and VI). Additionally, the specificity, sensitivity, and accuracy were determined for malignant lesions (MSRSGC categories V and VI) and benign lesions (MSRSGC categories II and IVA). The risk of malignancy (ROM) was calculated for all MSRSGC categories.

3. Results

The total number of cases in our study was 111 cases. The FNAC smears of these cases were studied and categorized into the six MSRSGC categories. Cases in our study ranged from 4 to 88 years age. The highest number of cases occurred in the 41–80 years age group (73.87%), and the male‐to‐female ratio in our study was 1.13:1 The most commonly affected salivary gland was the parotid (62.16%), followed by the submandibular gland (17.12%). The majority of cytology cases in this study were classified under category II: Non‐Neoplastic of the MSRSGC (55 cases, 49.55%), followed by category IVA‐Neoplasm: Benign (29 cases, 26.13%) (Table 1). While all cases were followed up clinically, histopathological evaluation was available in 34 cases (31.53%), majority being category IVA neoplasm. Some cases either declined operative intervention or elected for conservative management, and were subsequently maintained on periodic clinical follow up.

TABLE 1.

Distribution of cases based on MSRSGC categories.

Diagnostic category	Number of cases		Risk of malignancy (ROM) (%)
	Only cytology	Both cytology and histopathology available
I. Non‐diagnostic	12 (10.81%)	1 (2.94%)	0
II. Non‐neoplastic	55 (49.55%)	6 (17.65%)	0
III. Atypia of undetermined significance (AUS)	2 (1.80%)	2 (5.88%)	0
IV. Neoplasm
A. Benign	29 (26.13%)	17 (50%)	6
B. Salivary gland neoplasm of uncertain malignant potential (SUMP)	7 (6.30%)	5 (14.70%)	40
V. Suspicious for malignancy (SM)	1 (0.90%)	1 (2.94%)	100
VI. Malignant	5 (4.50%)	2 (5.88%)	100
Total	111	34	—

3.1. Non‐Diagnostic Category (MSRSGC Category I)

25 cases were initially categorized as Non‐Diagnostic. Repeat FNAC were performed, yielding adequate material in 13 cases, which were then reclassified into the appropriate categories. Of the remaining 12 Non‐Diagnostic cases (10.81%), 11 cases resolved with conservative management, and follow‐up FNAC was not requested. However, one case, a 65‐year‐old man, persistently presented as a non‐resolving 4 × 3 cm swelling in the parotid region for 4 months, and the ultrasonography findings were suggestive of a heterogenous, well‐circumscribed lesion. On aspiration, it yielded few adipocytes and very few clusters of benign acinar and ductal cells. This case was categorized as inadequate on multiple FNAC attempts. The patient underwent right superficial parotidectomy, which revealed a well‐circumscribed, encapsulated yellow‐brown solid tumour. Histopathology showed an encapsulated tumour composed of approximately equal proportions of oncocytic cells and mature adipocytes, with oncocytes arranged in acinar, nest, and focal trabecular patterns, lacking atypia, mitoses, necrosis, perineural, lymphovascular, or extracapsular invasion; adjacent salivary gland showed multifocal oncocytic hyperplasia. A final diagnosis of oncocytic lipoadenoma arising in a background of multifocal oncocytosis was rendered (Figure 1).

FIGURE 1.

MSRCGC Category I: Non‐Diagnostic. (a) Aspirate from right parotid gland yielded very few clusters of benign acinar and ductal cells along with occasional fragments of adipose tissue. Papanicolaou stain, ×40 (b) Histopathology of same case was reported as Oncocytic lipoadenoma, against a background of multifocal oncocytosis. Section showed an encapsulated tumour with adjacent salivary gland. H&E, ×40. (c) Higher magnification showed a tumour composed of oncocytic cells with intermixed adipocytes. Oncocytic cells have abundant granular eosinophilic cytoplasm and round nuclei with conspicuous nucleoli. H&E, ×200.

3.2. Non‐Neoplastic (MSRSGC Category II)

Category II presented with lesions that were firm, tender, and palpable swellings, most commonly occurring in the 41 to 60‐year age group. There was a total of 55 cases (49.55%) in this category, of which the most common salivary gland lesions were inflammatory lesions (23 of 38 cases, 60.53%) followed by cystic lesions (12 of 38 cases, 31.58%). The parotid gland was affected most frequently by non‐neoplastic lesions (47.27%). The most common finding in large intraparotid lymph nodes was reactive lymphoid hyperplasia (9 of 17 cases, 52.94%) followed by granulomatous lymphadenitis (6 cases out of 17 cases, 35.29%) and acute suppurative inflammation (2 of 17 cases, 11.76%) (Table 2).

TABLE 2.

MSRCGC category II (non‐neoplastic)—distribution of cases based on cytological diagnoses and site of lesion: (n = 55).

Cytological diagnoses	Parotid gland	Submandibular gland	Intraparotid lymph node	Number of cases
Benign salivary gland tissue	2	1	0	3
Benign Cystic lesion	3	1	0	4
Mucocele	1	3	0	4
Lymphoepithelial cyst	3	0	0	3
Sialadenosis	6	1	0	7
Sialadenitis	2	1	0	3
Acute sialadenitis	1	1	0	2
Chronic sialadenitis	4	4	0	8
Acute Suppurative Inflammation	1	0	2	3
Granulomatous lymphadenitis	1	0	6	7
Reactive lymphoid hyperplasia	0	0	9	9
Cystic lesion (d/d: lymphoepithelial cyst of parotid/Retention cyst)	1	0	0	1
Reactive lymph node in parotid gland (d/d: Lymphoepithelial sialadenitis)	1	0	0	1
Total	26	12	17	55

Histopathological follow‐up was available in six cases of category II, all demonstrating concordance with cytological diagnosis, except for one case. This case appeared as a well‐circumscribed heterogenous lesion on ultrasonography but the aspirate showed only normal salivary gland on the cytology smears. It was diagnosed as sialolipoma on histopathologic evaluation (Figure 2).

FIGURE 2.

MSRCGC Category II: Non‐Neoplastic (a) Aspirate from left parotid gland reported as Non‐Neoplastic; acinar cells, ductal cells, fibrovascular stroma & few adipocytes along. Papanicolaou stain, ×200. (b) Histopathology of same case reported as Sialolipoma, a well circumscribed, partially encapsulated benign lipomatous lesion with component of normal salivary gland tissue. H&E, ×100.

3.3. Atypia of Undetermined Significance (MSRSGC Category III)

The AUS category included two cases that were indefinite for neoplasm. One case showed an abundant mucinous background with scanty cellularity while the other case was indeterminate between a florid reactive hyperplasia and a lymphoproliferative disorder. Histopathological evaluation confirmed these as mucocele and reactive lymphoid hyperplasia, respectively (Table 3).

TABLE 3.

MSRCGC category‐wise distribution of cases based on cytological diagnosis and their histopathological diagnoses.

Sr. no	Cytology diagnosis	Histopathology diagnosis
Category I: Non‐diagnostic
1	Inadequate	Oncocytic lipoadenoma, against a background of multifocal oncocytosis
Category II: Non‐neoplastic
1	Retention cyst	Retention cyst
2	Lymphoepithelial cyst	Lymphoepithelial cyst
3	Normal salivary gland aspirate	Sialolipoma
4	Sialadenitis secondary to sialolithiasis	Chronic sialadenitis secondary to sialolithiasis
5	Necrotizing granulomatous inflammation with abscess	Acute on chronic inflammatory lesion
6	Granulomatous lymphadenitis	Granulomatous lymphadenitis likely Sarcoidosis
Category III: Atypia of undetermined significance (AUS)
1	Abundant extracellular mucin in background with few cells (d/d: low‐grade mucoepidermoid carcinoma).	Mucocele
2	s/o Florid reactive lymphoid hyperplasia (d/d: lymphoproliferative disorder)	Reactive lymphoid hyperplasia
Category IVA: Neoplasm: Benign
1	Pleomorphic Adenoma (10 cases)	Pleomorphic adenoma (10 cases‐concordant)
2	Warthin tumour (4 cases).	Warthin tumour (4 cases‐concordant)
3	s/o Benign salivary gland neoplasm (d/d: Oncocytoma and Pleomorphic Adenoma)	Warthin tumour
4	Basal cell adenoma	Basal cell adenoma
5	s/o Benign salivary gland neoplasm.	Low grade myoepithelial cell rich neoplasm of minor salivary glands (1) Myoepithelial cells neoplasm—Favour myoepithelioma (2) Myoepithelial cell rich pleomorphic adenoma
6	s/o Benign salivary gland neoplasm. (d/d: Pleomorphic adenoma, Myoepithelioma and Basal cell adenoma)	Benign epithelial myoepithelial tumour (invasion not present and low Mib‐1 index)
Category IVB: Neoplasm: SUMP
1	s/o Low grade salivary gland neoplasm (d/d: Acinic cell carcinoma and Oncocytoma.)	Oncocytoma
2	Basaloid neoplasm (d/d: Basal cell adenoma, Basal cell carcinoma and Adenoid cystic carcinoma)	Chronic sialadenitis with ductal cell hyperplasia, fibrosis and sclerosis
3	s/o Salivary gland neoplasm of low‐grade malignancy (d/d: Mucoepidermoid carcinoma low grade type, Acinic cell carcinoma and Carcinoma ex pleomorphic adenoma.)	Acinic cell carcinoma
4	s/o Salivary neoplasm of low‐grade malignancy (d/d: Polymorphous low‐grade Adenocarcinoma and Epithelial—Myoepithelial carcinoma)	Pleomorphic adenoma
5	Adenoid cystic carcinoma. (d/d: basal cell adenoma)	Adenoid cystic carcinoma
Category V: Suspicious for malignancy
1	s/o Low to intermediate grade mucoepidermoid carcinoma	Salivary carcinoma, NOS. Adenocarcinoma, Not Otherwise Specified, intermediate grade
Category VI: Malignant
1	s/o High grade salivary gland neoplasm (d/d: Carcinoma ex Pleomorphic adenoma and Salivary duct Carcinoma)	Salivary duct carcinoma
2	s/o Acinic cell carcinoma	Intraductal papillary mucinous neoplasm, low grade

3.4. Salivary Neoplasm (MSRSGC Category IV to VI)

The benign neoplastic category (MSRSGC category IV A) consisted of 29 cases, with pleomorphic adenoma (PA) being the most common lesion (62.07%), followed by Warthin tumour (13.79%) (Figure 3). The majority of these cases involved the parotid gland (72.41%). Of the 29 cytology cases, histopathological follow‐up was available for 17, and the results were consistent with the cytological diagnoses in this category (Table 3). The remaining 10 patients declined surgery, opted for conservative management, or received surgical treatment at a different institution.

FIGURE 3.

MSRCGC Category IVA: Neoplasm: Benign (a) Aspirate from right parotid gland; Warthin tumour, monolayered sheets as well as crowded clusters of oncocytic cells with lymphoid cells in the background. Papanicolaou stain, ×400 (b) Histopathology of same case; Warthin tumour, tumour showing lymphoid stroma and many cystic spaces lined by large oncocytic cells. H&E, ×100.

There were 7 cases under MSRSGC category IVB‐salivary gland neoplasm of undetermined malignant potential (SUMP) (Figures 4 and 5). All cases required surgical excision for confirmation of diagnosis. Two cases were discordant on histopathology; chronic sialadenitis was mistaken for basal cell adenoma due to ductal cell hyperplasia mimicking basaloid cells, and pleomorphic adenoma of the hard palate was misdiagnosed as a low‐grade malignancy due to resemblance to biphasic tumours like polymorphous adenocarcinoma and epithelial‐myoepithelial carcinoma (Figure 5).

FIGURE 4.

MSRCGC Category IV B‐Neoplasm: Salivary gland neoplasm of uncertain malignant potential (SUMP) (a) Aspirate from left submandibular gland reported as Neoplasm: SUMP, differentials considered were adenoid cystic carcinoma and basal cell adenoma. Highly cellular smears reveal small, uniform, epithelial cells with scanty cytoplasm. Numerous large, hyaline, stromal globules are seen. Papanicolaou stain, ×200. (b) Histopathology of same case reported as Adenoid Cystic Carcinoma, tumour showing a predominant cribriform pattern with few irregular, infiltrating tubules. Intervening thick fibrocollagenous bands are seen within the tumour islands. H&E, ×200.

FIGURE 5.

MSRCGC Category IV B‐Neoplasm: Salivary gland neoplasm of uncertain malignant potential (SUMP) (a) Aspirate from hard palate reported as Neoplasm: SUMP, (differentials considered were Polymorphous adenocarcinoma and Epithelial‐myoepithelial carcinoma); Cells arranged in fronds and pseudopapillary fragments with mild to moderate anisonucleosis. Papanicolaou stain, ×400. (b) Histopathology of same case reported as Pleomorphic adenoma, biphasic tumour, arranged in microtubules with dense eosinophilic secretions. H&E, ×40.

The Suspicious for Malignancy category (MSRSGC category V) had one case which was suspicious for a low to intermediate‐grade mucoepidermoid carcinoma on cytology. Cytological evaluation revealed few cells with nuclear pleomorphism, moderate cytoplasm, and scanty mucin in the background. Histopathological examination later diagnosed the case as Adenocarcinoma, not otherwise specified (NOS), intermediate grade (Figure 6).

FIGURE 6.

MSRCGC Category V: Suspicious for Malignancy (SM). (a) Aspirate from left parotid gland reported as SM, suspicious for low to intermediate grade mucoepidermoid carcinoma. Sheets and clusters of metaplastic epidermoid cells, few mucoid cells, and cystic fluid in the background. Papanicolaou stain, ×100. (b) Histopathology of same case reported as Salivary Adenocarcinoma, Not Otherwise Specified, intermediate grade. Section showed cystic and papillary architecture with stromal invasion and atypia. H&E, ×200.

The Malignant category (MSRSGC category VI) included five cytology cases showing distinct cytomorphological characteristics of malignancy. Among them, one case involved metastatic squamous cell carcinoma in an intraparotid lymph node originating from an oropharyngeal primary; the patient unfortunately succumbed to the disease. Another case involved a 69‐year‐old male with salivary duct carcinoma and lung metastasis who was on palliative treatment and did not undergo surgery. Additionally, a 69‐year‐old male suspected of low‐grade mucoepidermoid carcinoma declined surgical intervention. Consequently, only two cases underwent excision, both showing histopathological concordance. Of these, one case was diagnosed as salivary duct carcinoma. The other case was a 78‐year‐old man who presented with a 2‐month history of a small (1 × 1 cm), non‐tender, firm swelling at the angle of the right mandible. Ultrasonography revealed a solid–cystic lesion measuring 1.4 × 0.8 cm in the superficial lobe of the right parotid gland. FNAC showed cellular smears with sheets, clusters, and papillaroid fragments of polygonal cells having round to oval nuclei and indistinct, finely vacuolated to granular cytoplasm with intracytoplasmic mucin. Mild to moderate anisonucleosis was noted, with no mitosis or necrosis. The cytological impression was acinic cell carcinoma. Histopathological examination of the right superficial parotidectomy specimen was suggestive of low‐grade intraductal papillary mucinous neoplasm, with no evidence of invasion (Figure 7). On immunohistochemistry, p40 highlighted the basal cells at the periphery of the lesion. The columnar cells were positive for CK7 and negative for CK20, CDX2, S100 and NKX3.1. Mib1 labelling was 5%–7% in the highest proliferating area.

FIGURE 7.

Category VI‐ Malignant (a) The Giemsa stained cytology shows sheets, clusters and papillaroid fragments, Giemsa stain, ×200, (b) of polygonal cells with round to oval nucleus, indistinct, finely vacuolated and granular cytoplasm, Giemsa stain, ×400, (c) intracytoplasmic mucin, mild to moderate anisonucleosis, no mitosis or necrosis. Giemsa stain, ×400. (d) Histopathology of same case reported as Intraductal papillary mucinous neoplasm. H&E, ×100.

The overall sensitivity was 96%, specificity was 71.4%, and accuracy was 90.63% in distinguishing between non‐neoplastic and neoplastic salivary gland lesions based on the MSRSGC. Additionally, the study showed 100% sensitivity and 95.83% specificity, with a positive predictive value (PPV) of 75% and negative predictive value of 100% for diagnosing malignant salivary gland lesions on cytology using the MSRSGC.

4. Discussion

The study consisted of diverse cytological cases ranging over a spectrum of non‐neoplastic, most commonly inflammatory lesions to benign and malignant neoplasms.

These cases showed characteristic cytological features in many instances and overlapping cytological features in others and were accordingly categorized into appropriate MSRSGC categories. Most cases in the study belonged to category II‐ MSRSGC Non‐Neoplastic category followed by category IV A‐Neoplasm: Benign which is in agreement with the findings of studies by Kala et al. [7], Sahoo et al. [5] and Bhattacharyya et al. [6]. This predominance underscores the need for accurate cytological evaluation to guide treatment decisions and avoid unnecessary surgery. The parotid gland was most commonly affected, followed by the submandibular gland, consistent with previous studies by Kala et al. [7], Vaithy et al. [8] and Wu et al. [9]. This may reflect the larger size and higher proportion of salivary tissue in the parotid gland, highlighting the importance of careful FNAC evaluation of parotid lesions in clinical practice.

Our hospital caters to a defined population consisting of employees and their dependent family members, enrolled as beneficiaries of the healthcare program. This arrangement supports continuity of care and enables effective follow‐up, as patients typically remain within the system and are more likely to return for scheduled reviews and ongoing management. Histopathological correlation was available in 34 cases. Histological concordance and discordance were evaluated for each category to understand the characteristic features that can aid in diagnosis and to study the possible caveats that one needs to be mindful of.

4.1. Category I

Out of 12 cases, only one case in this category underwent excision and was diagnosed as oncocytic lipoadenoma on histopathology.

Oncocytic lipoadenoma is a rare, well‐circumscribed, benign salivary gland neoplasm lacking predominant organoid salivary gland elements and normal ductulo‐acinic structures. Histologically, it is separated from adjacent parenchyma by a thin fibrous capsule and shows two distinct components: oncocytes arranged in nests and acini resembling oncocytoma, and mature adipose tissue. The proportion of adipose tissue varies widely (5%–80% of tumour volume) [10]. In our case, both components comprised approximately 50% of the lesion (Figure 1).

4.2. Category II

Among the 55 cytology cases in this category, the most common lesions were inflammatory lesions, followed by cystic lesions, with the parotid gland being the most frequently affected site. Bajpai et al. [11] and Dharwadkar et al. [12] also reported inflammatory salivary gland lesions as the commonest aetiology in this category in their respective studies. Non‐Neoplastic category cases were concordant with histopathology findings, except for one case of sialolipoma in which cytological smears showed normal salivary gland tissue (Figure 2). This possibility needs to be kept in mind whenever a long‐standing lesion is identified, especially when sonography reveals a well circumscribed, heterogenous lesion which on aspiration yields mature adipocytes and normal salivary gland tissue, devoid of inflammation [13, 14].

4.3. Category III

There were 2 cases in the AUS category, of which one case had low cellularity and thick mucinous background on cytology, and the other showed florid lymphoid population. Though both cases were suspected to be benign on cytology, a differential of low‐grade mucoepidermoid carcinoma and lymphoproliferative disorder, respectively, could not be ruled out. Milan categorization as AUS category includes cases where an unequivocal decision of non‐neoplastic versus neoplastic cannot be made. The risk of malignancy for this category is 30% [3]. These cases, confirmed histologically as mucocele and reactive lymphoid hyperplasia, highlight the challenges of interpreting borderline cytology and demonstrate the importance of the MSRSGC in guiding appropriate follow‐up and management.

4.4. Category IV A

The most frequent lesion in this category was pleomorphic adenoma (PA) followed by Warthin tumour, predominantly in the parotid gland. Other studies in literature such as those by Bajpai et al. [11], Dharwadkar et al. [12] and Griffith et al. [15] showed similar findings, indicating that the parotid gland is the most commonly affected site for benign salivary gland neoplasms. Four cases were confidently reported as benign salivary gland neoplasm based on smear cellularity and bland cytological features. These cases showed epithelial/myoepithelial cells with or without presence of ground substance in the background and without any cytological or architectural atypia. Differential diagnosis given in these cases includes pleomorphic adenoma, basal cell adenoma, and myoepithelioma. Histopathological evaluation of the lesions showed concordance with the cytological assessment. Assigning a category has helped in clinical decision making for the patients and to plan surgery.

4.5. Category IV B

Our study had 5 concordant and 2 discordant cases under the SUMP category. Chronic sialadenitis can sometimes present as a firm mass, mimicking a neoplasm. A common pitfall in the cytology of chronic sialadenitis is the misinterpretation of basal cell hyperplasia, metaplastic or atrophic ductal cells as a basaloid neoplasm. However, chronic sialadenitis typically lacks the cellularity and larger three‐dimensional epithelial groups seen in aspirates from basaloid neoplasms [16]. Key clinical clues for diagnosing chronic sialadenitis include recurrent pain and swelling with a waxing‐waning pattern, chronicity, and no history of sudden size increase. Multiple FNACs in cases with low cellularity can provide a more accurate representation of the lesion and assist in determining whether the lesion is benign or malignant.

The other discordant case was that of a hard palate swelling, diagnosed as salivary gland neoplasm with differential diagnoses of Polymorphous Adenocarcinoma (PAC) and Epithelial‐Myoepithelial Carcinoma on cytology, but confirmed as Pleomorphic Adenoma (PA) on histopathology (Figure 5). PAC is the second most common malignancy of the minor salivary glands, where it is almost exclusively found. PA can be misdiagnosed as PAC, particularly when fibrillary chondromyxoid stroma is absent [4]. The architectural features observed in cell block preparations can help distinguish PAC and epithelial‐myoepithelial carcinoma from other biphasic salivary gland neoplasms with overlapping cytologic characteristics [4].

4.6. Category V

One case in this category was cytologically classified as SM, suspicious for low to intermediate‐grade mucoepidermoid carcinoma on cytology, but was later diagnosed histologically as intermediate‐grade adenocarcinoma, NOS (Figure 6). This category helps stratify the risk of malignancy (ROM) and alerts the clinician that the specimen cannot be definitively classified as benign or malignant due to limitations such as sparse cellularity or specimen artefacts [16].

4.7. Category VI

There were five malignant cases, all showing distinct cytomorphological features. One case diagnosed as a high‐grade salivary gland neoplasm on cytology was confirmed as salivary duct carcinoma on histopathology. Another, initially identified as acinic cell carcinoma in the parotid, was later confirmed as low‐grade intraductal papillary mucinous neoplasm. (Figure 7). This lesion is considered a low‐grade subtype of salivary gland mucinous adenocarcinoma. Recent studies have linked these tumours to AKT1 E17K mutations and strong, diffuse expression of NKX3.1 [17]. In our case, however, NKX3.1 was negative, while p40 showed positivity in the basal cells. The tumour was well circumscribed, with no evidence of invasion into the duct wall.

4.8. Statistical Overview

Consistent with previous studies, our findings confirm that the MSRSGC effectively distinguishes non‐neoplastic from neoplastic salivary gland lesions and for identifying malignancy (Table 4) [7, 8, 18, 19]. The Risk of Malignancy (ROM) in our study aligned with MSRSGC and other studies (Table 5) [3, 6, 7, 16, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29]. The ROM for categories I, II, and III was zero, likely due to a good correlation between cytology and clinical‐radiological features and resampling and recategorization of category I lesions into appropriate diagnostic groups. A limitation of our study is the small number of cases with histopathological follow‐up, which makes performing accurate and statistically significant analyses across diagnostic categories difficult.

TABLE 4.

Utility of MSRGSC for (I) diagnosing neoplastic and non‐neoplastic lesions and for (II) diagnosing benign and malignant salivary gland lesions.

	Our study	Vaithy et al. [8]	Liang et al. [18]	Tochtermann et al. [19]
I	Utility of MSRGSC for diagnosing neoplastic and non‐neoplastic salivary gland lesions (%)
Specificity	71.40	98.21	100	89.8
Sensitivity	96	84.20	93	97.2
Accuracy	90.63	93.60	—
II	Utility of MSRGSC for diagnosing benign and malignant salivary gland lesions (%)
Specificity	95.83	98.21	95.8	99.3
Sensitivity	100	84.2	82.3	94.6
PPV	75	94.64	90.3	97.9
NPV	100	91.70	92.0	98.2
Accuracy	96.29	93.60	—	—

TABLE 5.

Risk of malignancy based on MSRGSC categories: Comparison with the first and second edition of MSRGSC and other studies.

Authors	Sample size (n)	MSRSGC category‐wise risk of malignancy (%)
		I	II	III	IV A	IV B	V	VI
		Non‐Diagnostic	Non‐Neoplastic	AUS	Neoplasm: Benign	Neoplasm: SUMP	SM	Malignant
ROM Range in first edition of MSRSGC [16]	—	25	10	20	< 5	35	60	90
ROM in second edition of MSRSGC [3]	—	15	11	30	< 3	35	83	> 98
Tochterman G et al. [19]	753	26.7	5.7	34	1.1	21.8	92	99.2
Hafez NH et al. [20]	118	33.3	11.8	37.5	2.1	44.4	60	100
Balmiki Dutta et al. [21]	57	0	0	50	0	50	100	50
Vallonthaiel et al. [22]	190	44	8	0	8	44	81	100
Katta et al. [23]	69	33.33	11.1	100	6.9	50	66.6	87.5
Rohilla et al. [24]	631	0	17.4	100	7.3	50	—	96
Pujani et al. [25]	64	0	10	50	2.5	50	100	100
Karuna et al. [26]	105	0	0	50	2.44	33.33	100	93.33
Savant et al. [27]	199	0	0	33.3	0.8	40.9	100	100
Viswanathan et al. [28]	373	6.7	7.1	38.9	5	34.2	92.9	92.3
Jha et al. [29]	102	42.86	26.67	100	10.17	0	71.42	100
Bhattacharrya et al. [6]	62	0	15.38	50	2.43	—	100	66.6
Kala et al. [7]	172	25	5	20	4.4	33.3	85.7	97.5
Present study	34	0	0	0	6	40	100	100

5. Conclusion

In our experience, MSRSGC categories could enhance communication between cytopathologists and clinicians, guiding management decisions through category‐specific guidelines alongside clinical and radiological correlation. Given the diverse cytology of salivary gland lesions, categorizing ambiguous and challenging cases could also benefit pathologists, as identifying a specific tumour type is not always necessary. The sensitivity, specificity and ROM of our study largely met with those reported using MSRSGC in previous studies. FNAC remains an effective, non‐invasive, and resource‐efficient tool for diagnosing salivary gland lesions. While some cases can present with challenging cytology and diagnostic dilemmas, categorizing them according to MSRGSC can prove valuable for clinicians, ultimately aiding in optimal patient care.

Author Contributions

S.F.K. contributed to data acquisition, data collection, analysis and interpretation, and drafting of the manuscript. R.T.N. contributed to the conceptualization and design of the study, cytological evaluation, and critical revision of the manuscript for important intellectual content. S.C. contributed to literature review, data analysis, and manuscript editing. U.P.C. and P.R.G. contributed to manuscript review and editing. All authors approved the final version of the manuscript and agree to be accountable for all aspects of the work, ensuring integrity and accuracy.

Funding

The authors have nothing to report.

Ethics Statement

All procedures performed in this retrospective data analysis were in accordance with the ethical standards of the Institutional Review Board. Approval was taken from Institutional Ethics Committee Project No. BHMEC/DNB/23/2022 dated 31st January 2023.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.