Histopathological comparison of pleomorphic adenomas of the parotid and submandibular gland

Objective: The aim of this study was to investigate and compare the histopathological findings in pleomorphic adenomas (PA) of the parotid and submandibular gland with emphasis on the histological subtype and capsular characteristics. Materials and Methods: The histopathological specimens of all patients with PAs of the parotid and submandibular gland between 2000 and 2020 were re- examined by an experienced head and neck pathologist. Patients without representative slides allowing evaluation of the whole periphery of the PA were excluded from our study sample. Results: Nine hundred and thirty- four patients formed our study sample (327 men, 607 women, male- to- female ratio: 0.53:1). Eight hundred and forty- four cases had a PA in the parotid gland and the remaining 90 in the submandibular gland. Our comparative analysis showed that submandibular PAs are characterized by the consistent presence of an intact anatomical capsule, infrequent occurrence of pseudopodia and satellite nodules, and a low proportion of the high- risk myxoid subtype. Conclusion: Our study highlights significant differences between PAs of the parotid and submandibular glands in their histopathological characteristics. Their differences likely underlie the favorable surgical outcome observed in PAs of the submandibular glands and may explain the propensity of PAs of the parotid glands for local recurrences.

of view. Concerning the pathomechanisms of recurrence, common clinical experience tends to have a confusing effect on explanation attempts. A seemingly uneventful surgical procedure (no exposure of the capsule or macroscopic rupture of the capsule) could result in recurrent nodules in the long term, whereas the intraoperative rupture of a PA does not always lead to a manifest recurrence! A large number of hypotheses point to a multifactorial mechanism based on an interaction of tumor-and surgery-related parameters. However, lack of certainty on this issue complicates the therapeutic management of this disease.
Furthermore, it is known that parotideal PAs show a significant long-term recurrence rate of up to 7% (if followed up for 20 years; Valstar et al., 2020), whereas their submandibular counterparts recur only extremely rarely (Mantsopoulos et al., 2017). In trying to explain this remarkably different biological behavior of the same lesion in different organs, the aim of this study was to compare the cases with PAs of the parotid and the submandibular glands in terms of epidemiological factors (age, gender) as well as the incidence of the various histopathological parameters (histological subtype, intactness of the capsule, surgery-induced capsular defects, pseudopodia (PS), satellite nodules [SN]).

| MATERIAL S AND ME THODS
This study was performed at an academic tertiary referral center specialized in salivary gland diseases (Department of Otorhinolaryngology, Head and Neck Surgery, University of Erlangen-Nuremberg, Erlangen, Germany). For this study, an experienced head and neck pathologist (A.A.) critically reevaluated the histological slides of the pathological specimens of all patients who underwent primary submandibulectomy for PAs of the submandibular gland or parotidectomy for PAs of the parotid gland between 2000 and 2020. Based on tumor size, the number of histological slides per tumor varied from one to more than 10. The 38 cases with PAs in which the whole periphery of the tumor was not histologically ascertainable on at least two slides, based on the tumor size, were excluded from the study sample.
Concerning the histopathological analysis, PAs were divided into 3 histological subtypes on the basis of the stroma cell proportion, according to the classification of Seifert et al., (1976): the classic "mixed" subtype with a stroma content of 30%-50%, the stroma-rich (myxoid) subtype with a stroma content of more than 50%, and the cellular subtype with a stroma content of 30% or less. A pseudopodium represents a "re-encapsulated" secondary nodule separated only by a layer of capsular fibrous tissue from the main tumor mass, but localized within the main tumor capsule. Based on the plan of sectioning, PS may appear either as a mushroom-like capsular protrusion or as a distinct small nodule closely associated with the main tumor and separated from it only by the fibrous capsule, but no normal fatty or salivary tissue is seen between the two nodules. Finally, SN were distinct tumor nodules near the main tumor but outside the main tumor capsule, separated from it by salivary or fatty tissue without any connection to the main tumor.
Statistical analysis was performed using the t test for the epidemiological analysis and the chi-square test for the comparison of histological parameters with 95% confidence intervals (CIs). The software SPSS version 21 for Windows (SPSS, Inc.) was used for the analysis. A p-value of <.05 was considered statistically significant.
The Institutional Review Board (IRB) of the University Hospital of Erlangen approved this study.

| RE SULTS
Nine hundred and thirty-four patients formed our study sample (327 men, 607 women, male-to-female ratio: 0.53:1). Eight hundred and forty-four cases had a PA in the parotid gland, and the remaining 90 had a PA in the submandibular gland. The mean age was 48.6 (12-89 years). No statistically significant differences were detected in gender (p = .906) or age (p = .282) between patients with PAs of the parotid and submandibular glands. Table 1 shows the results of the statistical analysis of the histopathological parameters in our study cases. Comparative analysis showed that PAs of the submandibular gland are characterized by the consistent presence of an anatomically intact capsule, a significantly lower occurrence of PS, a tendency toward a lower incidence of satellite tumors, and a comparatively limited incidence of the fragile myxoid subtype.

| D ISCUSS I ON
A significant number of literature reports try to explain the exact pathogenesis of PA recurrence based on the interaction between several tumor-and treatment-related factors. Although submandibulectomy tends to lead quite frequently to a broad exposure of the capsule of a submandibular PA, it is common experience that submandibular lesions recur extremely rarely (Mantsopoulos et al., 2017). On the contrary, although a vast amount of literature reports are dedicated to the need for preserving a cuff of healthy parotid tissue around a parotid PA , longterm recurrence rates of these lesions are by no means negligible (Valstar et al., 2020). The lack of adequate treatment justifies the characterization of the most common multifocal recurrence of a parotid PA as a so-called "benign tragedy" (Poorten, 2014). This apparent difference in the biological behavior between parotideal and submandibular PA s prompted us to investigate the histological differences ("tumor-related factors") between these lesions in a large single-center cohort in order to obtain information that could help to enlighten the recurrence mechanism.
PA is one of the most common neoplasms of the parotid gland (Mantsopoulos & Iro, 2020). The prefix "pleomorphic" reflects the architectural diversity of these lesions due to highly variable combinations of epithelial and myoepithelial elements, intermingled with a similarly variable mucoid, myxoid, or chondroid mesenchymal component. A review of the relevant literature shows a significant diversity in the incidence of several subtypes (myxoid, mixed, hypercellular; Seifert et al., 1976) in parotideal PAs, probably due to differences in the classification systems (Dulguerov et al., 2017).
Re-examination of our parotid specimens showed that almost half of them belonged to the myxoid subtype (47.7%), with a homogeneous distribution of the remaining lesions in the other categories (hypercellular in 24.4% and "mixed" type in 27.8%). Contrary to this, the majority of submandibular cases manifested the mixed subtype (50%), which showed a statistically significant difference from their parotid counterparts (p = .000). It seems that less stroma content in the submandibular cases warrants more stability and less fragility in comparison with the parotid lesions.
PA is the sole benign lesion in the parotid gland with a frequently instable capsule with herniation of tumor material in the surrounding healthy glandular parenchyma ( Figure 1) Another histopathological feature with clinical relevance refers to the presence of surgically induced defects of the capsule . In the parotid gland cases, close dissection in the vicinity of critical structures (facial nerve) or in the tumor-nerve interface can lead to an accidental rupture of the adenoma capsule. In submandibular cases, common surgical experience shows that adenomas usually lie in the peripheral zone of the gland (Figures 2 and 3). The need for avoiding injury to the marginal mandibular branch of the facial nerve on the cranial aspect and the lingual nerve on the medial aspect of the gland dictates a subcapsular dissection of the gland outside its bed (in the course of submandibulectomy; Mantsopoulos et al., 2017Mantsopoulos et al., , 2018, which exposes a PA lying in the gland periphery with a high risk of rupture. Such microscopic defects could be seen in 4.2% of our parotid and 2.2% of our submandibular cases, without a statistically significant difference between these study groups (p = .255). This points to the fact that the localization of a lesion in the gland is the most important risk factor for a surgically induced defect of its capsule.
In the 1950s, Patey and Thackray already recognized the value of PS in the pathomechanism of the recurrence of Pas (Patey & Thackray, 1958). The incidence of PS in parotideal PAs varies between 28 (Stennert et al., 2001) and 54% (Park et al., 2012)  Provided that the surgical principle "do not see the capsule" is strictly adhered to, it seems that rather the localization and size of the PA determine the extent of capsule exposure: a most frequently eccentrically localized submandibular PA has the same risk of exposure of its capsule as a large PA of the parotid with broad contact to the facial nerve. Leaving surgery-related factors aside, it seems that the rather stable and comparatively beneficial as well as more predictable histological profile of submandibular cases could explain their generally accepted favorable prognosis.

ACK N OWLED G M ENT
Open Access funding enabled and organized by Projekt DEAL.

CO N FLI C T O F I NTE R E S T
None.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/odi.13858.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets used and analyzed during the current study are available from the corresponding author on reasonable request.