Adenoid Cystic Carcinoma, Clinical Presentation, Current Treatment and Approaches Towards Novel Therapies

Adenoid Cystic Carcinoma in the Context of Head and Neck Cancer - Epidemiology and Risk Factors

Head and neck cancer (HNC) is the 6^th most common cancer worldwide with around 900,000 new cases and >400,000 deaths yearly (5, 6). Important risk factors for HNC are smoking, alcohol, opium, betel quid chewing, oral infections, radiation exposure, personal history of HNC and more recently even viral infections, e.g., human papillomavirus (HPV) and Epstein-Barr Virus (EBV). However, for some HNC e.g., malignant salivary gland tumours (MST), including adenoid cystic carcinoma (AdCC), risk factors are still mainly unknown (1-5, 7). Recently, in many Western countries, due to a decrease in smoking, the spectra of various HNCs have changed. We are seeing a steady decrease in smoking related HNC but instead a parallel rise in HPV associated head and neck cancer has been disclosed (8-10).

HNC namely accounts for various cancers originating from the nasal cavity, the paranasal sinuses, the oral cavity, the pharynx (i.e., the nasopharynx, the oropharynx, and the hypopharynx), the larynx, and the salivary glands and not all these are associated with smoking (5, 11). Nonetheless, HNC are in 90% of the cases squamous cell carcinomas (SCC) present in the mucosa and grow loco-regionally (9, 10). They often spread via the lymph nodes of the neck and/or via the bloodstream, and when the primary tumour cannot be detected, they are called metastatic cancer of unknown primary (CUP) of the head neck region and are also classified as HNC (8, 12).

Clearly HNC varies considerably, as do MST, where adenoid cystic carcinoma (AdCC) is one of the key players, although AdCC is not only limited to MST or to the head and neck region. This review, therefore gives a brief introduction to MST, and evidently, specifically to AdCC, where for both the aetiology is unknown (11, 13-18).

Adenoid Cystic Carcinoma in the Context of Malignant Salivary Gland Tumours

While, HNC is mainly due to smoking, alcohol, or viral infections, this is as mentioned above not the case for MST with a mainly unknown aetiology (1-4, 11, 13-18). There are three paired, major salivary glands in the head and neck region, the parotid gland, the submandibular gland, and the sublingual gland (listed by size), in addition to a plethora of minor salivary glands in the lips, the oral cavity, the throat, and the upper digestive tract.

Salivary gland tumours are mostly benign but around 30% are malignant and the cancer rate increases with a reduced gland size, being approximately 25% in the parotid, 50% in the submandibular, and 80% in the sublingual glands (4, 15). MST thereby represent a heterogenous group with diverse histomorphology, immune profile, and clinical outcome. In 2017, WHO recognized 21 MST types, and mucoepidermoid carcinoma, acinic cell carcinoma, AdCC, and adenocarcinoma as the most common ones (15, 16). Furthermore, due to MST heterogeneity and lack of diagnostic markers, defining a correct diagnosis is challenging as is sometimes distinguishing between benign and malignant tumours, which is especially important in order to offer an optimal treatment (1, 15).

As previously mentioned, the aetiology of MST is mainly unknown, but previous local irradiation, chronic inflammation of the gland, and prior cutaneous neoplasm in the area are potential risk factors (1, 11, 15). The peak age at diagnosis is 60-70 years, (4). Clinical presentation of MST is alike that of benign tumours, initially presenting with asymptomatic masses, but then with the potential spread of disease, and a secondary cancer at diagnosis is unusual unlike in other HNC (4).

Treatment of MST, including AdCC, is similar despite various heterogenicities (1, 4). Radical surgery is the first line of therapy, followed by radiotherapy in cases of high-grade malignancy, high T-stage, uncertain radicality, perineural or perivascular tumour growth. Non-operable cancers are treated with irradiation with/without chemo-therapy (ChT) to obtain local control (18). Approximately 75% of patients are still alive 5-years after diagnosis, but this varies depending on tumour stage at diagnosis (19, 20). Upon local recurrence, re-surgery or re-irradiation can be performed (4). For spread disease, ChT combinations can be used but response rates are generally poor with objective response rates of <10-20%. Evidently, new treatment options are of great importance and this review focuses on AdCC in particular in more detail below (1, 4).

Adenoid Cystic Carcinoma

General background. AdCC is an unusual and rare malignancy, originating from the secretory glands of the body (unspecified) and with yet unknown aetiology (1, 4). It comprises 10% of all major salivary gland neoplasms and 30% of all neoplasms in the minor salivary glands (1). In fact, AdCC is the most common malignant tumour in the minor salivary glands and the second most common in the major salivary glands (2). However, AdCC may also occur at other locations of the head and neck region and more seldom even in the secretory glands outside this area, for example in the oesophagus, the breast, the lung, and the vulva (3). It can affect patients of all ages but is most common in the 5^th or 6^th decades of life and approximately 60-70% of the patients are female (1, 3, 4).

Clinical presentation. Analogous to other MST, AdCC presents with unspecific symptoms, where clinical differentiation from benign tumours is difficult (1, 3, 4). AdCC is characterized by slow growth and perineural invasion, and in many cases it presents a difficult differential diagnosis, due to the lack of good diagnostic markers (15). AdCC, similar to MST, is initially treated aggressively, with radical surgery if possible, often accompanied by postoperative radiotherapy. Nonetheless, in spite of an aggressive therapy regimen, late local relapses (15-85%) and distant metastases (25-55%) are frequent, and notably, these occur late, >5 years or even more 10-15 years (see also below) after completed primary treatment (1). This has been suggested to depend on perineural invasion beyond the surgical margins and microscopic haematogenous dissemination during the disease early stage (21, 22). Due to its slow growth and indolent clinical course AdCC 5-year survival rates are 80-85%, but the 10- and 15-year survival rates decline to 50-60% and 30-35% respectively (2, 23).

Diagnostics

Over the years, there have been many endeavours to identify better markers for diagnosis, prognostication as well as potential targets for targeted therapy, but so far, much is still unknown. As pointed out above the diagnosis of AdCC can be intriguing due to its heterogeneity and its histomorphological likenesses with other MSTs in addition to the presence of polymorphous low-grade adenocarcinoma (PAC) (15, 22). Below some aspects of diagnostics are presented.

Histomorphology. Current diagnostics mainly rely on histomorphology, immuno-histochemistry (IHC) and cytogenetics (15). AdCC presents three separate growth patterns; cribriform, which is the most frequent one, tubular, and solid, where the latter is the far most aggressive type, with the poorest prognosis (1, 3, 15). Examples of these AdCC patterns are presented in Figure 1. Moreover, AdCC shows a biphasic differentiation with both secretory glandular (ductal)- and myoepithelial elements. Therefore, specific staining for both these components, although not exclusive for AdCC, is of importance for its diagnostics (Figure 1) (3, 24).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Different presentations of AdCC revealed using IHC. Magnification, ×20. (A) Haematoxylin and Eosin (HE) staining of a cribriform growth pattern (B) HE-staining of a tubular growth pattern (C) HE-staining of a solid growth pattern (D) Positive p63-staining, a myoepithelial marker of the basaloid cells in AdCC (E) Positive c-KIT (CD117)-staining of the ductal epithelial cells in AdCC (F) MYB-staining, as a surrogate marker for a MYB:NFIB gene fusion, sensitive but non-specific biomarker for AdCC. [The Figure is from the doctoral thesis of Mark Zupancic, with permission from the author (77)].

Molecular traits. In AdCC, it has been shown that a t(6;9) translocation generates a MYB:NFIB gene fusion, leading to an over-expression of the MYB oncoprotein, which can be detected by IHC (22, 25). Notably, the latter has been demonstrated to promote AdCC cell proliferation and half of all AdCC (sometimes up to 80%) have the t(6;9) translocation, but its therapeutic value is yet unknown (3, 26-29). Somewhat less frequently also MYBL1 gene fusions are observed (30). Moreover, some additional biomarkers for example the receptor tyrosine kinase c-KIT (CD117), vascular endothelial growth factor receptor (VEGFR)−3, Ki-67 (proliferation marker), and p53 have also been described to be associated with AdCC (3). In addition, more recently, it was proposed that one could divide AdCC into a group with a more favourable prognosis that has up-regulated TP63 and receptor tyrosine kinases and another with worse prognosis having up-regulation of MYC and MYC-target genes (30).

Presence of viruses. In the past decades, others and we have been interested in investigating whether viruses could be causative agents for MST or AdCC (31-36). In addition, if this indeed was the case it was also of interest to explore whether the presence of a virus potentially has a prognostic value, as it has been shown for HPV-positive tonsillar and base of tongue cancer as compared to HPV-negative tonsillar and base of tongue cancer, for review see e.g. (8). In the above studies however, neither HPV or members of the human polyomavirus (HPyV) family were shown to be frequently found in MST or AdCC and therefore they were not regarded to be causative agents for these diseases (32-36).

Nevertheless, in one of our previous studies where we examined a large cohort of AdCC including 68 patients we did find HPV in three cases (37). Upon further analysis of these three cases none were later confirmed as being an AdCC, since two were later shown to be HPV-positive tonsillar cancers, while the third case was defined as being an HPV-related multiphenotypic sinonasal carcinoma (HMSC) (37, 38). Based on these findings, we suggest that examining for the presence of HPV could be of diagnostic value in suspected AdCC cases where the diagnosis is unclear, since it could be helpful to define a possible differentially correct diagnosis (37).

There could however still be other unknown viruses present in AdCC, but to our knowledge such studies have yet not been initiated to a major extent.

Treatment

Initial therapy of AdCC is surgery as mentioned above, however due to high recurrence rates, 30-75% (23), surgery is currently frequently accompanied by radiotherapy with 54-71 Gy (median 64 Gy) (39). Even so, when the tumour is small and with no nodal disease (T1N0) adding postoperative radiotherapy (PORT) is under debate, notwithstanding it has been proposed that it should still be taken into account when perineural invasion and/or solid histomorphological subtype are present (3).

Unfortunately, randomized controlled trials are infrequent in AdCC, which most likely is due to the rarity of this disease. Nevertheless, retrospective data have shown that progressive-free survival (PFS) is improved when using a multimodal therapy regimen in comparison to single treatment modalities (3, 40). One report presented that patients receiving surgery in combination with radiotherapy had better 5-year local control rates than patients only treated surgery; i.e. 90% vs. 42% (41). Similar data were obtained recently in a large cohort by Chen et al. (40). Another study, which compared radiotherapy alone with surgery accompanied by PORT, also showed improved local control rates upon multimodal treatment (42). Notably, both the 5- and 10-year local control rates were better in the group where surgery and radiotherapy were combined compared to the cohort receiving only radiotherapy; 94% and 91% respectively vs. 56% and 43% respectively (42). However, although PFS is longer using combination treatments, the effects of PORT on OS is still rather uncertain, likely due to that very long follow up times are necessary due to the slow growing nature AdCC (39, 43). Moreover, the additional survival benefit of PORT may still not apply to all AdCC patients and this may be especially true for low-risk patients with limited disease where the role of PORT has to be specified further (40, 44).

With regard to systemic treatment in AdCC, there are no standard therapy recommendations and moreover specific ChT regimens have not shown to be efficient in clinical trials (3). It has also been proposed that the lack of effects of ChT could be due to the indolent biology of AdCC, but of note, also upon progressive and disseminated disease therapy responses to cytotoxic agents are limited (objective response rates <20%) (45). Although, not proven effective, ChT is still being used since we lack other effective treatment options upon disease spread (3, 45).

Recently, we also conducted a study on the effects of various treatments on a large AdCC cohort including 155 patients diagnosed with AdCC between 2000-2022, where 142 had been treated with curative intent (46). Most patients in that study received multimodal treatment i.e., receiving surgery with RT with or without concomitant ChT (CRT), while the remaining patients were divided into two small groups, one with 14 cases receiving only RT or CRT and the other with 10 cases, treated with surgery alone (46). Patients given multimodal therapy had significantly better survival rates as compared to patients receiving single treatment modalities, i.e., the 5-year DFS was 70.7% vs. 37% respectively, and for OS the same trend was observed, 82.8% vs. 60.9%, respectively. This latter report also supports previous data that multimodal therapy options are superior to monotherapy (46).

Nevertheless, several of these studies include small cohorts and the nature of the various tumours are not easily taken into account when comparing the various treatments. Moreover, there is an interest to define whether there is a possibility to further tailor treatment, especially when the tumours are very small. Despite its rarity some studies have also focused on disclosing prognostic markers as well as initiate targeted therapy for AdCC. However, since targeted therapy as such has not been introduced for AdCC in general, it is not presented in this paragraph, but instead in the paragraph below.

Prognostic and Potentially Targetable Markers, Attempts With Targeted Therapy

As emphasized, AdCC is a rare and heterogenic disease and it is difficult to draw general conclusions on prognostic markers, since prospective studies are rare and the cohorts are often small and heterogenous, which is a complicating factor. Moreover, when investigating prognostic markers it is important to take into account not only tumour heterogeneity, but also what type of treatment has been given, especially since there are several reports suggesting that multimodal treatment is superior to monotherapy for AdCC (3, 40-42, 46). This issue has already been discussed above and will not be dealt with separately below. So far, there are no obvious AdCC specific prognostic factors, but one can still identify different types of predictive factors in various categories and some of these are presented below.

Histomorphological markers. As mentioned above AdCC has three separate growth patterns; cribriform, tubular, and solid, where the latter is the far most aggressive type, with adverse clinical outcome (15).

Clinical markers. Clinical characteristics of patients with AdCC are more easily noted and some of these are, e.g., age, sex, performance status, smoking history, tumour stage, perineural invasion. However, these should be followed as mentioned above in association with the type of treatment.

Subsite. In a study by us, focusing on 142/155 cases of AdCC of the head and neck region treated with curative intent, tumour location was a prognostic factor (46). More specifically, prognosis was found to be most favourable in patients with AdCC in the major salivary glands, especially in the parotid gland, and poorest in those with AdCC in the sinonasal area (46). Similar data have been observed in some cohorts, but not in all reports, since in some studies there was no influence of the subsite on disease prognosis (44, 46-50).

Stage of the disease. In a study by us, Stage IV disease was most frequent (35.3%) followed by stage II (28.8%), stage I (22.2%), and stage III (13.7%) (46). Notably, stage IV was most common in AdCC in the nasal cavity and paranasal sinuses (70.8%), as compared to e.g., the submandibular gland with much less stage IV disease (11.9%) (46). Moreover, in this context, not unexpectedly, having stage I and II disease was significantly more favourable than having stage III and IV disease, which was in agreement with several other studies (49, 51-53). However, whether it is easier to detect AdCC at some subsites compared to others and whether this explains why AdCC in some subsites is prognostically more favourable has yet to be determined.

Perineural growth. Perineural growth has also been considered as a prognostic factor. Notably, in a study by us perineural growth was over-represented in cases arising from the major salivary glands (73%), whereas a lower percentage (55.1%) of perineural growth was observed in AdCC arising from other sites in the head and neck region (46). However, in that study, the majority of the patients treated with curative intent had been given multimodal treatment and under that context perineural growth or incomplete surgical margins (more difficult to achieve in cases with perineural invasion) did not confer a survival problem (46). The former was partly in concordance with some but not all other studies by others (54-58). Consequently, although upon receiving multimodal therapy, perineural growth per se could not be considered a prognostic factor by us, this could very well depend on patient treatment or other not yet revealed circumstances (40, 54-56).

Age, sex, and smoking. Age, sex, and smoking have also been studied as potential prognostic markers by others and us (40, 46, 47, 54, 59). In our study, we did not disclose age, sex or smoking as prognostic factors (46). However, there are other reports suggesting that older age is correlated with higher disease stage and worse prognosis (40, 47, 54, 59). Consequently, the influence of age should most likely still be followed further.

Viruses. As mentioned above, so far viruses have not been shown to have a major impact on the aetiology in AdCC except as a means of differential diagnostics and can therefore today not be regarded as potential prognostic factors. However, it is not clear whether any yet unknown viruses contribute to the aetiology of AdCC (8, 32-36).

Molecular traits and attempts with targeted or immunomodulatory therapies. In the era of targeted therapy and precision medicine, the detection of specific molecular alterations is of importance. In this paragraph however the emphasis is more on targetable markers as compared to prognostic markers. This is due to that since AdCC is not only rare but also heterogeneous, it is not straightforward to define whether a specific molecular trait has also a prognostic value. There are a limited number of reviews on the subject and the field is continuously developing, so clearly some of the data applying to other types of cancer may also to some extent apply to AdCC, for reviews see e.g. (3, 60).

The MYB:NFIB gene fusion. As mentioned above, it has been shown that a large proportion of AdCC have a t(6;9) translocation that generates a MYB:NFIB gene fusion, in some cases leading to over-expression of the MYB oncoprotein, which can be detected by IHC (22, 25). This translocation is commonly investigated for during the diagnosis of AdCC and it is known to promote AdCC cell proliferation. Its therapeutic value has still not been resolved, although indirect targeting has been attempted, e.g., by targeting insulin growth factor 1 receptor (IGFR1) (3, 26-29, 60). There are also some experimental models as well as some ongoing clinical trials showing some positive data, but so far there have been no major breakthroughs successfully targeting MYB directly or indirectly (60).

Receptor tyrosine kinase c-KIT (CD117) is another molecular marker of AdCC that has been shown to be produced in high levels in 65-90% of all AdCC and is suggested to be a marker for aggressive disease and poor prognosis and could potentially be possible to target (3, 60, 61). Imatinib is a KIT inhibitor that has been used in four phase II clinical trials; however, only 2/42 AdCC patients had objective tumour responses (3). Moreover adding cisplatin to imatinib did not improve the responses (3, 60). Together the data so far suggest that c-KIT receptors do not play a major role in driving the malignant phenotype of AdCC (3, 60).

Fibroblast growth factor receptor 1 (FGFR1) is also variably over-expressed in AdCC (3, 61). This led to the study of the multikinase inhibitor dovitinib that targets FGFR1-3 in an experimental animal model system showing successful targeting of AdCC (3). This was also followed by the evaluation of dovitinib at the University of Virginia with some success in two AdCC patients (3, 62). However, since dovitinib is a multikinase inhibitor it is not yet clear whether these early positive data are associated to the inhibition of FGFR, other targeted kinases, or a combination of these, moreover the objective success rate was low (3, 60, 62). Other studies have attempted to target components downstream of activated FGFR. for review see (3). One such approach was to use monotherapy with nelfinavir which targets AKT signalling, but this did not give a significant response in patients with advanced AdCC (63).

Epidermal growth factor receptor (EGFR) and/or human epidermal receptor-2 (HER2) are also sometimes over-expressed in AdCC (3, 60, 64). Moreover, targeted therapy, has also been attempted against EGFR, with e.g., gefitinib, a small molecule inhibitor of the EGFR kinase, or e.g. with cetuximab, but no major benefits were observed, suggesting that also these two are only minor contributors to the malignant AdCC phenotype (3, 60, 63-66).

Vascular endothelial growth factor (VEGF) is over-expressed in a considerable number of AdCC (76%), suggesting its potential to be associated with tumour aggressiveness, recurrence and metastasis (3, 60, 67, 68). Moreover over-expression of MYB promotes expression of several genes including VEGF (68). For this reason the VEGF-receptor (VEGFR) has been considered a target and attempts to use VEGFR inhibitors have been made, but these were not so successful (3, 60, 67, 68).

The TP53 gene is not frequently mutated in MST or in AdCC originating from major salivary glands when compared to AdCC at some other sites (69). In fact, in one study it was reported that high p53 expression was observed in 19/21 cases of AdCC of the palate (70). Moreover, experimentally inactivating the MDM2-p53 interaction using a small molecule (MI-773) in a patient derived xenograft (PDX) model caused AdCC tumor regression (71). This molecule has been used in preclinical trials and no recurrences were observed when this molecule was used as adjuvant therapy in mice, while 63% of the mice in the control group showed recurrence (72). As a result a Phase I/II trial is ongoing in patients with MST, where also some AdCC patients are included (NCT03781986).

Immunomodulatory therapy. Lastly, an immunomodulatory therapy has been used and more specifically to target MYB. A TeTMYB vaccine targeting MYB was generated using a full-length MYB complementary DNA (cDNA) bound by two potent CD-4 epitopes derived from the tetanus toxin, which was then cloned into the food and drug administration (FDA)-complient DNA vaccine vector pVAX1 (73). This vaccine has then been successfully used to target MYB for colorectal cancer in animal experimental studies, and subsequentially used in a phase I clinical trial for colorectal cancer and AdCC (NCT03287427) (74-76).