Necrotizing sialometaplasia: a malignant masquerade but questionable precancerous lesion, report of four cases

Necrotizing sialometaplasia (NSM) is a reactive, self-limiting salivary gland lesion, first described by Abram and colleagues in 1973 [1]. It is known to involve minor salivary gland tissue of the hard palate. However, other mucous glandular tissues, such as the trachea, nasal cavity, and floor of the mouth, have been reported to be involved [2‐5]. In the hard palate, NSM usually appears as an ulcerative mass or swelling accompanied by pain and discomfort [6]. The incidence of NSM has been reported to account for 0.03% of all oral biopsies, but that may be underestimated because of the low recognition on this entity [7]. Histological findings are relatively characteristic, showing extensive coagulative necrosis with preserved lobular architectures and squamous metaplasia of the ductal epithelia [1]. The greatest concern regarding NSM is the risk of misinterpretation as other lesions. The most worrisome histologic mimics are mucoepidermoid carcinoma (MEC) and squamous cell carcinoma. We retrospectively investigated all oral biopsy data in a single institute to determine the real incidence of NSM and its clinicopathologic characteristics. The aim of our study is to raise awareness of this rare disease, which may be overlooked or misdiagnosed as malignancy in a small biopsy.

All oral cavity and hard palate biopsies performed at Seoul National University Boramae Hospital, Seoul, Korea, from 2012 to 2018 were reviewed by two board-certified pathologists (S.A. Shin and J.E. Kim). Immunohistochemistry (IHC) for cytokeratin7, S100, P63, P53 and Ki-67 was performed in some selective cases using an automated immunostainer (Ventana BenchMark XT, Tuscon, AZ) with a standard protocol according to the manufacturer’s recommendation. Clinical findings, radiologic features, surgical procedures and follow-up data were retrieved from electronic medical records.

In this study, we presented 4 cases of NSM after a meticulous search of 726 oral biopsies in a single institution. Although the histologic features were almost identical, except for the presence of dysplasia in one case, the clinicoradiologic findings revealed considerable variations between ours and previously reported cases. According to the literature, NSM can occur in all ages but occurs predominantly after middle age, with a higher incidence in males than in females [6]. However, patients in our series were much younger than middle-aged and showed no sex differences. Mass lesions were present in half of our cases. Conventionally, NSM has been reported to present as a circumscribed ulcer with a diameter of 1 to 3 cm [8]. We suggest that the primary reason for these discrepancies is the difference in the case selection. Many of the previous studies included focal or secondary NSM of the background mucous glands in mass-forming salivary lesions. In this study, we included cases in which NSM was the main pathology, excluding any cases associated with malignancies. As described in the literature, NSM-like morphologic changes can be seen in the periphery of the cancer-involved salivary glands, which raises suspicion that NSM might be a precancerous lesion [9]. We also discovered some cases of salivary gland carcinoma accompanied by squamous metaplasia of the surrounding ductal epithelia (Fig. 4). However, a typical confluent infarct with intact lobular architecture was lacking in those lesions.The etiology of NSM is still uncertain. Naturally, ischemia of the salivary gland lobule resulting from vascular injury and physical or chemical trauma have been suggested to be related to the pathogenesis of NSM [2, 10]. Atherosclerotic changes were proposed as etiologies when the disease was first described [1, 8], and following studies suggested that previous surgeries, dentures, smoking and upper respiratory infection were predisposing factors [6]. An association with local anesthetic injection has also been demonstrated in animal experiments [11]. Recently, some studies have reported cases of NSM associated with eating disorders in young females. Some patients reported a history of consuming ice chips, directly affecting vasoconstriction, while others presented regurgitation of gastric acid due to recurrent vomiting in anorexia nervosa or gastroesophageal reflux disease [12‐16]. In our series, only two patients had evident predisposing factors: wearing a post-orthodontic retainer in case 1 and long-standing inflammation and recent surgery in case 4. The etiology of NSM was unclear in the other two cases because the likelihood of vascular insufficiency is low in young, healthy individuals. In case 4, NSM might have been caused by either recent surgery or longstanding chronic inflammation, as reported in previous studies [6]. However, obvious dysplastic epithelial changes suggest possible premalignant potential and can mimic squamous cell carcinoma.

The main differential diagnosis of NSM include MEC, squamous cell carcinoma, subacute necrotizing sialadenitis (SANS), and mucocele. Abrams and colleagues presented five morphologic criteria of NSM: 1) massive infarction, 2) bland nuclear features, 3) simultaneous metaplasia of the ducts and acini, 4) prominence of inflammatory granulation tissue, and 5) maintenance of the lobular structures [1]. In general, the absence of nuclear pleomorphism or atypism is the most authentic parameter supporting a benign nature, but these features are not applicable in the exclusion of MEC. Low-grade MEC can present relatively bland cytologic features in either epidermoid cells or glandular cells. Some researchers suggested that IHC using a panel of P53, P63 and Ki-67 could be helpful for the differentiation of NSM and MEC [8]. However, ductal epithelial cells in case 4 showed marked atypia corresponding to high grade dysplasia and high Ki-67 and P53. Application of IHC seems to have limited diagnostic value in a differential diagnosis of MEC, because MEC itself lacks any specific or pathognomonic protein biomarkers. Morphologic patterns, such as stromal invasion and disruption of lobular architectures, are more reliable supporting factors in MEC than in NSM. However, these features are not easily recognizable in small biopsy specimens. According to Brannon et al., NSM was misdiagnosed as malignancy in 21% of the preoperative biopsied cases they collected [6]. Although MEC and squamous cell carcinoma are the most common types of misdiagnosed malignancies on the basis of biopsy, NSM has also been misdiagnosed as acinic cell carcinoma, verrucous carcinoma, and ductal carcinoma. The incidence of pure squamous cell carcinoma in the salivary glands is extremely low; this should be kept in mind when squamous cell lesions are encountered in this area [17]. Among the benign entities, SANS and mucocele can be considered in differential diagnoses. SANS is also a benign disease that usually occurs in the minor salivary gland of the hard palate. The principal histopathology of SANS is inflammation without squamous metaplasia. Necrosis is rare in SANS, and if present, it is visible only focally [18].

Although the clinical course of NSM is self-limiting, the premalignant potential of NSM has long been pointed out. Association with other malignancies, such as epithelial-myoepithelial carcinoma or adenoid cystic carcinoma, and even malignant lymphoma have been reported to coexist with NSM [9, 19‐22]. There are still controversial issues regarding the application of strict criteria for NSM in these lesions due to the possibility of overdiagnosing NSM, as shown in some previous reports. Nevertheless, careful evaluation and follow-up is necessary for NSM patients. It should be noted that delayed self-healing may require re-evaluation or resection of the lesion for a correct diagnosis and vice versa.

NSM is a disease that can mimic malignancy, especially in small biopsy specimens. It is important to be aware of this rare disease and understand its clinicopathologic findings to avoid unnecessary intervention. The premalignant potential of NSM should be detemined in future large-cohort studies.