A rising number of studies investigating SDC have been published within the last decade. Most of these either focused on clinical or immunohistochemical data. Recognizing different architectural patterns and rare subtypes [18], the diagnostic criteria of SDC have been formulated more widely in the more recent WHO classifications. Hence, from today’s perspective, the incidence of SDC was likely underestimated in the past [9]. Furthermore, improvement of immunohistochemical profiling and clinical studies evaluating targeted therapy options have led to a more precise indication for systemic therapy in the R/M situation [10]. Therefore, the present study is the first describing (a) the proportion of SDC among SGC before and after reclassification of cases initially classified as ANOS in a single-center series; and (b) the frequency of SDC patients with targeted therapy options according to current guidelines (AR expression > 70% in IHC for ADT; HER2 positivity for trastuzumab-based regimens).

A mean age at diagnosis of 66.7 (± 11.7) and a male predominance (84.8%) in the present series are consistent with previous data showing a mean/median age at diagnosis ranging between 60 and 68 years [2,3,4] and 72–83% of SDC patients being male [4, 5, 19]. While the literature shows that most SDC (72–83%) arise from the parotid gland [5, 16, 19, 20], predominance of parotid tumors was particularly high with 93.5% in this series, possibly due to this institution being a center for parotid gland surgery. As SDC is a biologically aggressive entity, advanced pathological T stage (T3/T4: 63.0%) and positive pathological regional lymph node status (79.5%) were to be expected and are consistent with previous data describing advanced T stage in 56–75% [3, 16] and pathological regional lymph node status in 68–82%% of cases [3, 16], respectively.

In accordance with the current ESMO guidelines for SGC [10], most tumor resections included ipsilateral neck dissection (92.9%) and most surgically treated patients received adjuvant radiation therapy (85.7%) in this series. Although cancer registry data, published in 2016, has shown that addition of adjuvant chemotherapy to radiation therapy does not improve OS of patients with high-risk SGC [21], 22.7% of patients in this series received adjuvant chemoradiation therapy. Most of these patients were treated before 2017 and the decision for addition of chemotherapy to adjuvant radiation therapy was made on case-by-case decisions after 2017. Nevertheless, the authors of this study strongly advocate for adhering to current guidelines and therefore avoiding adjuvant chemotherapy in SGC due to the side effects and missing evidence of survival benefits. Furthermore, recently published retrospective data suggest a significant survival benefit for SDC patients with AR expression in > 70% of tumor cells treated with adjuvant ADT (median duration of 12 months) together with radiation therapy compared to radiation therapy alone [11]. Based on this data, 5 recently diagnosed AR-positive patients with advanced tumor stages within the present series received adjuvant ADT together with radiation therapy on case-by-case decisions with the mean follow-up being too short to report sufficient results. Prospective clinical studies investigating the efficacy and safety of adjuvant ADT together with radiation therapy in SDC with a high extent of AR expression seem urgently needed to confirm an improvement of the outcome of these patients.

SDC is associated with an unfavorable survival due to a high rate of recurrence. 5-year OS and 4-/5-year DFS have been shown to range between 40–65% [2,3,4,5,6] and 17–44% [2, 3, 5, 6], respectively. A 5-year OS and a 5-year DFS of 62.8% and 41.0% in the present series are rather favorable compared to the existing literature. Previous studies have reported 5-year LRC and 5-year DC rates of 29–70% and 36–48% [2, 4, 5] being more unfavorable than the 5-year LRC and DC rates of 79.1% and 61.6% found in this study. A potential explanation is that more than one-fourth of the patients in this study were diagnosed between January 1st, 2022 and December 31st, 2023 and therefore have a follow-up of only up to 2 years. As shown in the existing literature, the results of the present study confirm a markedly higher LRC rate compared to the DC rate. This emphasizes the high relevance of systemic disease in SDC and therefore the necessity of effective systemic treatment options. Furthermore, neither locoregional, nor distant recurrence were found after a follow-up of 5 years in this study. Therefore, in contrast to adenoid cystic carcinoma showing a late recurrence after 5 years in 26% of cases [22], it seems that a follow-up of 5 years is sufficient for SDC.

Various independent negative prognostic factors on OS and/or DFS such as higher age, male gender, higher T/N stage, extracapsular spread, perineural invasion, facial nerve palsy, and postoperative radiation therapy have been identified in different studies [3, 16, 19, 23, 24]. This study is the first to reveal the independent negative prognostic influence of positive margins on DFS in SDC (HR = 4.0, p = 0.03), which seems plausible and has been proven for various other solid cancer entities [25,26,27]. Nevertheless, this result has to be emphasized as the surgical therapy of an SDC in the parotid gland poses a particular challenge due to the anatomical proximity to functionally relevant structures such as the facial nerve.

The overall proportion of SDC among primary SGC was 12.1% in the present series. This is markedly higher than the reported SDC proportion rates from European registries, being 5.4% in a Danish cohort of 1,601 patients treated between 1990 and 2015 and 3.5% in a German cohort of 1,680 patients treated between 2009 and 2018 [7, 8]. The main reason for this divergence may be underestimation of the incidence of SDC in cancer registry databases due to an inaccurate histopathological subclassification of SGC. More precisely, 39% of SGC, initially diagnosed as ANOS, were reclassified as SDC by pathologists with special expertise in SGC in a recent study using contemporary immunohistochemical profiling and diagnostic criteria [9]. Transferring this finding to the above-cited cancer registry databases (ANOS: 18.5% of all SGC in German registry and 8.4% in Danish registry) would result in an SDC proportion of around 10% of all SGC in these studies. These results underline the importance of accurately reviewing cases of ANOS by a pathologist specialized in head and neck pathology, especially in the context of established targeted therapies for SDC in the R/M setting and emerging evidence for targeted therapies in the adjuvant situation. Moreover, the present study also shows a rising proportion of SDC after reclassification of cases initially classified as ANOS. While SDC accounted for 7–12% of all primary SGC between 2000 and 2015, 15–18%% of all SGC diagnosed between 2016 and 2023 were SDC. If these data reflect a rise of the SDC incidence also outside of this single-center institution is unclear and needs further research. An accurate histopathological review of cases initially diagnosed as ANOS in large cohorts—ideally in cancer registry databases—of SGC is needed to answer this question. As the median age at diagnosis of SDC compared to other SDC entities is relatively high with up to 68 years [4], an increasing life expectancy in Europe during the last decades [28] may serve as an explanation for the observed rising SDC proportion.

Although various studies have shown that 69–100% and 13–71% of SDC are AR-positive and HER2-positive [12,13,14,15,16], respectively, the present study is the first to evaluate the frequency of SDC patients with established targeted therapy options according to current guidelines. The ESMO guidelines on SGC recommend ADT in case of AR expression in > 70% of tumor cells and trastuzumab-based therapy in case of HER2 positivity in R/M SDC [10]. 56.8% of all SDC in this series showed nuclear AR expression in > 70% of tumor cells. Moreover, 36.4% of all SDC were positive for HER2 (IHC 3 +). This resulted in 70.5% of patients being either eligible for targeted ADT or trastuzumab-based therapy.

Importantly, expression of AR and HER2 was not mutually exclusive as around one-fourth of cases showed AR expression in > 70% of tumor cells and positivity for HER2 (IHC 3 +). This is particularly important as a crosstalk between the AR and HER2 pathways has been shown for prostate carcinoma [29,30,31,32]. In detail, the activation of the HER2 pathway is inhibited by prostate-specific cPAcP in androgen-dependent prostate cancer. Androgen decreases cPAcP resulting in HER2-activated tyrosine phosphorylation of p38-MAPK and ERK1/2 which in turn lead to cell proliferation via the intranuclear androgen-response element. Loss of cPAcP leads to androgen-independent prostate cancer cell proliferation in androgen-independent prostate cancer cells [29, 30]. Crosstalk between the AR and HER2 pathways is also suspected in SDC as patients treated with enzalutamide with AR + /HER2 + tumors experienced a lower clinical benefit rate (22.2%) than patients with AR + /HER2 − tumors (45.8%; p = 0.013) in a phase II trial [33]. Therefore, combination therapies simultaneously targeting the AR and HER2 pathways in SDC may improve patient outcomes in the future.

This study has several limitations. First, clinical data were collected retrospectively. Further, no sufficient FFPE tissue for IHC was available in 2 cases of SDC. Lastly, as FISH results evaluating the HER2 amplification status were only available in 9 cases, HER2 positivity may be underestimated.

Overall, this study displays the proportion of SDC among SGC in a single-center before and after reclassification of cases initially classified as ANOS and suggests a rising absolute and relative incidence of SDC within the last years. Moreover, the results show that the extent of AR and HER2 expression allows for established targeted therapy according to the current guidelines in most cases of SDC and that AR and HER2 expression are not mutually exclusive.