Oral Oncology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Oral Oncology journal homepage: www.elsevier.com/locate/oraloncology

Review

Diagnostic and therapeutic implications of new molecular biomarkers in salivary gland cancers Göran Stenman a,⇑, Fredrik Persson a,b,1, Mattias K. Andersson a,1 a b

Sahlgrenska Cancer Center, Department of Pathology, University of Gothenburg, Gothenburg, Sweden Department of Oncology, Sahlgrenska University Hospital, Gothenburg, Sweden

a r t i c l e

i n f o

Article history: Received 17 March 2014 Received in revised form 22 April 2014 Accepted 26 April 2014 Available online xxxx Keywords: Salivary gland cancer Biomarker Gene fusion Oncogene Chromosome translocation Mutation Therapeutic target

s u m m a r y Salivary gland carcinomas (SGCs) are uncommon tumors, constituting approximately 5% of all cancers of the head and neck. They are a heterogeneous group of diseases that pose significant diagnostic and therapeutic challenges. The treatment of patients with SGCs is mainly restricted to surgery and/or radiation therapy and there is only limited data available on the role of conventional systemic and targeted therapies in the management of patients with advanced disease. There is thus a great need to develop new molecular biomarkers to improve the diagnosis, prognostication, and therapeutic options for these patients. In this review, we will discuss the most recent developments in this field, with focus on pathognomonic gene fusions and other driver mutations of clinical significance. Comprehensive cytogenetic and molecular genetic analyses of SGCs have revealed a translocation-generated network of fusion oncogenes. The molecular targets of these fusions are transcription factors, transcriptional coactivators, and tyrosine kinase receptors. Prominent examples of clinically significant fusions are the MYB–NFIB fusion in adenoid cystic carcinoma and the CRTC1–MAML2 fusion in mucoepidermoid carcinoma. The fusions are key events in the molecular pathogenesis of these tumor types and contribute as new diagnostic, prognostic, and therapeutic biomarkers. Moreover, next-generation sequencing analysis of SGCs have revealed new druggable driver mutations, pinpointing alternative therapeutic options for subsets of patients. Continued molecular characterization of these fusions and their down-stream targets will ultimately lead to the identification of novel driver genes in SGCs and will form the basis for development of new therapeutic strategies for these patients. Ó 2014 Elsevier Ltd. All rights reserved.

Introduction Salivary gland carcinomas (SGCs) are uncommon neoplasms, accounting for up to 5% of all cancers of the head and neck. They include a wide spectrum of histologic entities with variable biological behavior and responsiveness to therapy [1–3]. The diagnosis of SGCs remains challenging, mainly because of the diversity of histologic subtypes and the often overlapping morphological patterns seen in many of these lesions. The primary treatment for patients with localized disease is surgery and/or radiation therapy. A number of recent studies have indicated that radiation plays an important role in local control in both the postoperative setting

⇑ Corresponding author. Address: Sahlgrenska Cancer Center, Department of Pathology, University of Gothenburg, Box 425, SE-405 30 Gothenburg, Sweden. Tel.: +46 31 786 6733. E-mail address: [email protected] (G. Stenman). 1 These authors contributed equally to this work.

and in the definitive setting for unresectable cancers ([3] and references therein). However, the response rates to radiation vary between different histologic subtypes and tumor grades. Chemotherapy is employed almost exclusively with a palliative aim in patients with metastatic and/or recurrent disease. Targeted therapies are currently recommended only for patients in clinical trials. New treatment strategies are therefore needed for the majority of patients with SGCs. Important advances have recently been made in the understanding of the molecular pathogenesis of SGCs. Thus, several recurrent chromosome translocations have been identified and shown to generate a tumor-type specific gene fusion network involving the most common subtypes of SGCs ([4–6] and references therein). The molecular targets of these translocations are tyrosine kinase receptors, transcriptional coactivators, and transcription factors involved in cell cycle regulation and growth factor signaling. The fusions and their downstream targets are new important biomarkers for molecular diagnosis and most importantly also for development of new therapeutic strategies

http://dx.doi.org/10.1016/j.oraloncology.2014.04.008 1368-8375/Ó 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Stenman G et al. Diagnostic and therapeutic implications of new molecular biomarkers in salivary gland cancers. Oral Oncol (2014), http://dx.doi.org/10.1016/j.oraloncology.2014.04.008

2

G. Stenman et al. / Oral Oncology xxx (2014) xxx–xxx

for SGCs [3]. Moreover, recent studies using next generation sequencing and genomic and expression profiling methods have identified several additional biomarkers of potential clinical significance [7,8]. The aim of this review is to discuss the clinical implications of these new biomarkers with focus on their diagnostic and therapeutic significance in SGCs. A summary of these molecular biomarkers and their potential therapeutic targets is presented in Table 1. Mucoepidermoid carcinoma (MEC) MEC is the most common histological subtype of SGC, and includes a wide spectrum of lesions ranging from aggressive high-grade cancers to mainly non-aggressive, low-grade tumors [1]. We have previously shown that MECs, irrespective of histological grade, are characterized by a recurrent chromosome translocation t(11;19)(q21–22;p13) [4,5,9]. The translocations, which are found in a high frequency of MECs originating from diverse anatomical locations, result in fusions involving the MAML2 and CRTC1, or more rarely CRTC3, genes [10–13] (Fig. 1). CRTC1, encodes a CREB (cAMP response element-binding protein) coactivator [14] and MAML2 a Mastermind-like coactivator for Notch receptors [15]. As a consequence of the fusion, the Notch-binding domain of MAML2 is replaced by the CREB-binding domain of CRTC1/3 linked to the transactivation domain of MAML2 [10,11]. The molecular consequences of the fusion are still obscure. However, recent studies have shown that the EGFR-ligand AREG (Amphiregulin) is a downstream target of the fusion and that upregulation of AREG leads to activation of EGFR-signaling in an autocrine manner and increased cell growth and survival of MEC-cells (Fig. 2) [16,17]. In line with this observation, CRTC1–MAML2 positive MEC-cells were shown to be highly sensitive to inhibition of EGFR-signaling in a xenograft model, suggesting that targeting this pathway with small molecule EGFR-inhibitors may offer a new approach to systemic treatment of patients with advanced, unresectable fusionpositive MECs (Fig. 2) [3,17]. Although various grading systems have been introduced to improve the classification of MEC [18–22], their clinical usefulness have been limited because of subjective evaluation criteria and biological heterogeneity among different tumor grades. However, recent molecular genetic studies of large series of MECs have convincingly demonstrated that the fusion is a specific and clinically useful biomarker for this disease [22–26]. These studies have also shown that the highest incidence of the CRTC1–MAML2 fusion is found in low- and intermediate-grade MECs with favorable prognosis. Although less frequently, the fusion may occur in high-grade

MECs with a dismal prognosis [12,26,27]. Interestingly, the latter cases are often associated with deletion of the CDKN2A tumor suppressor gene [26,27]. However, most poorly differentiated MECs are fusion-negative, indicating that they may represent a misclassification of high-grade MEC-like adenocarcinomas not otherwise specified. A recent genome-wide arrayCGH study suggested that MECs may be subclassified in (1) fusion-positive, low- and intermediate-grade MECs with no or few genomic imbalances and favorable prognosis, (2) fusion-positive, high-grade MECs with or without CDKN2A deletions, multiple genomic imbalances, and poor prognosis, and (3) fusion-negative, high-grade MEC-like adenocarcinomas with multiple genomic imbalances and poor prognosis [3,26]. Taken together, available data demonstrate that CRTC1– MAML2 may serve as a specific diagnostic and prognostic biomarker for MEC and that patients with histopathologically confirmed or suspected MECs should be screened by RT-PCR and/or FISH for the CRTC1–MAML2 fusion before being enrolled in clinical trials using targeted therapies. High-grade MEC-like tumors Previous studies have shown that a subset of high-grade, CRTC1–MAML2 negative tumors with a MEC-like phenotype has a t(6;22)(p21;q12) translocation resulting in an EWSR1–POU5F1 gene fusion (Figs. 1 and 2) [28]. The resulting fusion protein is composed of the N-terminal domain of EWSR1 linked to the DNA-binding domain of POU5F1 (a. k. a. OCT4). POU5F1 is a transcription factor with a critical role in embryonic development and maintenance of pluripotency of embryonic stem cells. In agreement with the role of POU5F1 as a master regulator of cellular pluripotency, it should be noted that the EWSR1–POU5F1 positive tumors were more immature compared to the MAML2-positive tumors. Of note, an identical EWSR1–POU5F1 fusion has recently also been found in a subset of deep-seated soft tissue myoepithelial tumors of children and young adults [29]. Adenoid cystic carcinoma (ACC) ACC is the second most common subtype of SGC but may also arise in other secretory glands, such as the breast, and in the sinonasal tract, tracheobronchial tree, skin, and vulva ([1,5] and references in [30,31]). ACC is a slow growing but aggressive cancer with a poor long-term prognosis mainly due to metastatic disease [1]. Eighty to 90% of patients with head and neck ACC die of disease in 10–15 years. The primary treatment of ACC is surgery and/or

Table 1 Molecular biomarkers and potential therapeutic targets in salivary gland cancers. Tumor type

Diagnostic biomarkers

Other biomarkers

Activated oncogenes/ pathways

Potential therapeutic agents

ACC

MYB–NFIB

del(1p), del(6q)

MEC MEC (high-grade) MASC HCCC Ca-ex-Pa

CRTC1–MAML2 EWSR1–POU5F1 ETV6–NTRK3 EWSR1–ATF1 PLAG1-fusions HMGA2-fusions HER2 amplification

del(9p) CDKN2A

MYB–NFIB, EGFR, KIT, BRAF, HRAS, TRKC, FGFR1 CRTC1–MAML2, AREG EWSR1–POU5F1 ETV6–NTRK3, IGF1R EWSR1–ATF1 PLAG1, HMGA2, HER2, MDM2

Cetuximab (EGFR), vemurafenib (BRAF), dacomitinib (panEGFR), imatinib (KIT), AZD7451 (NTRK3/TRKC), dovitinib (FGFR1) Cetuximab (EGFR), erlotinib (EGFR)

Trastuzumab (HER2), nutlin-3 analogs (MDM2-TP53)

SDC

HER2 amplification

HER2, BRAF, androgen receptor, PI3K mTOR

Trastuzumab (HER2), bicalutamide (androgen receptor), vemurafenib (BRAF), temsirolimus (mTOR) Rapamycin, sirolimus (mTOR)

MEK

Trametinib (MEK)

AciCC EMC

t(8q12) t(12q14–15) MDM2/HMGA2 amplification TP53 mutation AR+ Inactivation of PTEN and APC

HRAS

AZD7451 (NTRK3/TRKC), BMS-754807 (IGF1R, IR)

Please cite this article in press as: Stenman G et al. Diagnostic and therapeutic implications of new molecular biomarkers in salivary gland cancers. Oral Oncol (2014), http://dx.doi.org/10.1016/j.oraloncology.2014.04.008

G. Stenman et al. / Oral Oncology xxx (2014) xxx–xxx

3

Fig. 1. Schematic illustration of oncogenic fusion transcripts in mucoepidermoid carcinoma (MEC), adenoid cystic carcinoma (ACC), mammary analog secretory carcinoma (MASC), hyalinizing clear cell carcinoma (HCCC). Functional domains in the encoded proteins are indicated below each transcript. CBD, CREB-binding domain; TAD, transactivation domain; NBD, Notch-binding domain; RRM, RNA-recognition motif; ZF, zinc finger domain; DBD, DNA-binding domain; NRD, negative regulatory domain; EC, extracellular domain; TM, transmembrane domain; PTK, phosphotyrosine kinase domain. Transcripts are not drawn to scale.

radiation therapy. There is currently no effective treatment available for patients with recurrent and/or metastatic disease. We have previously shown that ACC has a signature t(6;9)(q22– 23;p23–24) chromosomal translocation [32,33] resulting in a fusion involving the MYB oncogene and the transcription factor gene NFIB (Figs. 1 and 2) [5,30]. The MYB–NFIB fusion oncoprotein activates transcription of MYB targets of key importance for oncogenic transformation [30,34]. These targets include genes involved in e.g. cell cycle control, DNA repair, and apoptosis (Andersson et al., unpublished data). MYB activation due to gene fusion, or more rarely copy number gain or insertion of the 30 -part of NFIB in the vicinity of the MYB locus, is found in more than 80% of ACCs [6,30,31,35–38]. In contrast, the MYB–NFIB fusion or rearrangements of the MYB locus has not been detected in any other subtype of SGC, indicating that it is a hallmark of ACC [5,6,36,38]. Since MYB-activation is a common event in ACC there is currently no firm evidence to suggest that MYB is a prognostic biomarker. Future studies will show whether different MYB–NFIB fusion

transcript variants generated by alternative splicing and variable breakpoints may have different transforming capacities and thereby also different clinical consequences [39]. A crucial question in this context is whether MYB-negative tumors are true ACCs or whether they may represent SGCs with an ACC-like morphology such as for example basal cell adenocarcinoma or polymorphous low-grade adenocarcinoma. Comprehensive molecular and histopathological analyses of such cases may shed further light on this issue. In particular, RNA-sequence analysis will reveal whether MYB-negative tumors have variant gene fusions that might characterize a subset of ACCs. In addition to being a validated, diagnostic biomarker for ACC, MYB–NFIB and MYB and the signaling pathways they activate are novel potential therapeutic targets in ACC [3,5,6]. So far, most studies using conventional chemotherapy and molecularly targeted therapies of patients with progressive ACC has been disappointing. Consequently, these patients are left with no reliable treatment options. Therefore, studies are now in progress in several

Please cite this article in press as: Stenman G et al. Diagnostic and therapeutic implications of new molecular biomarkers in salivary gland cancers. Oral Oncol (2014), http://dx.doi.org/10.1016/j.oraloncology.2014.04.008

4

G. Stenman et al. / Oral Oncology xxx (2014) xxx–xxx

Fig. 2. Fusion oncoproteins and aberrantly activated signaling pathways in salivary gland cancers. Potential therapeutic agents and their points of action are illustrated.

laboratories to develop inhibitors of MYB, MYB–NFIB, and their molecular targets [3,5,40–42]. However, targeting transcription factors has proven notoriously difficult and hence there are currently only limited opportunities to directly target MYB and MYB–NFIB. Alternative ways to target these oncogenic drivers, include inhibition of the signaling pathways they are acting in or inhibition of their down-stream targets. The latter alternative seems less attractive because MYB has over 10,000 known target genes in MCF-7 breast cancer cells [43] and it will be very difficult to identify the key oncogenic targets operating in ACC. This is further exemplified by the tyrosine kinase (TK) receptor KIT which is a well known MYB target that is highly overexpressed but not mutated in MYB-positive ACCs [44]. However, clinical studies have shown that the TK-inhibitor imatinib has no major effects on patients with progressive head and neck ACC [45,46], indicating that KIT is not a driver in this disease. Therefore, approaches to inhibit the expression of the pathways regulating MYB and MYB– NFIB are likely to be more promising. Such efforts will hopefully lead to the emergence of potent therapies that will significantly improve the survival of ACC patients with progressive disease. Previous studies have shown that FGF2 (fıbroblast growth factor 2) can cooperate with v-Myb and maintain hematopoietic progenitor cells in a proliferative, undifferentiated state [47], suggesting that the FGFR signaling pathway (Fig. 2) may be a possible therapeutic target also in ACC. Thus, a recent ongoing phase II study of the multitargeted FGFR1/3 inhibitor dovitinib has shown some promise with objective partial responses and stable disease with acceptable toxicity in patients with progressive metastatic ACC [48]. The final result from this study and from another ongoing Canadian study (http://www.clinicaltrials.gov/ ct2/show/NCT01678105) will show whether treatment with dovitinib is an option for a subset of ACC patients. Recent studies have also shown that the TrkC/NTRK3 signaling pathway is activated in ACC and that treatment of patient-derived ACC xenografts with the TrkC kinase inhibitor AZD7451 results in growth inhibition, suggesting that TrkC kinase inhibition may be a

potential therapeutic option in ACC [49]. Moreover, subsets of breast and head and neck ACCs were recently shown to have mutations in RAS pathway genes (Fig. 2), including BRAF and HRAS, indicating that the BRAF inhibitor vemurafenib may be effective in ACC patients with activating BRAF kinase mutations [44]. Two recent exome and whole genome sequencing studies [7,8] have identified new potential molecular targets in ACC. Both studies confirmed previous genome-wide arrayCGH studies [31] showing that ACC has a quiet genome with few genomic alterations, consistent with the view that the MYB–NFIB gene fusion is the major oncogenic event in ACC [5,6,30,31]. Notably, these studies revealed a wide mutational diversity and a low exonic somatic mutation rate in ACC with mutations in genes encoding chromatin-state regulators (e.g. SMARCA2, CREBBP, KDM6A), DNA damage response (e.g. TP53, UHRF1, ATM), protein kinase A signaling (e.g. RYR3, RYR2, PTPRG), and FGF–IGF–PI3K signaling (e.g. PIK3CA, FOXO3, INSRR). These results were partly confirmed in a smaller sequencing study of 196 cancer-related genes in a clinically defined subset of ACCs [50]. Future studies will establish the role of these pathways as molecular targets for therapy in ACC patients. Mammary analog secretory carcinoma (MASC) MASC of the salivary glands is a recently described subtype of SGC, with strong histologic and immunohistochemical resemblance to secretory breast carcinoma (SBC) [51–53]. MASC and SBC also share a t(12;15)(p13;q25) translocation, which results in an ETV6–NTRK3 gene fusion (Figs. 1 and 2) [51–53]. The fusion encodes a chimeric tyrosine kinase that activates two major effector pathways (Fig. 2), the Ras–MAP kinase pathway and the phosphatidyl inositol-3-kinase (PI3K)–AKT pathway [54–56]. ETV6–NTRK3 is expressed in more than 90% of all MASCs and is an important molecular biomarker that defınes this entity and distinguishes it from other SGCs, including acinic cell carcinoma, low-grade cribriform cystadenocarcinoma, cystadenocarcinoma NOS, and low-grade MEC. In fact, recent molecular cytogenetic

Please cite this article in press as: Stenman G et al. Diagnostic and therapeutic implications of new molecular biomarkers in salivary gland cancers. Oral Oncol (2014), http://dx.doi.org/10.1016/j.oraloncology.2014.04.008

G. Stenman et al. / Oral Oncology xxx (2014) xxx–xxx

studies have indicated that most non-parotid acinic cell carcinomas probably represent MASC [57]. MASC is regarded as a lowgrade tumor with an overall favorable outcome, but there are also cases reported with high-grade transformation and an accelerated clinical course and dismal prognosis [58]. Of note, the ETV6–NTRK3 fusions have been found also in other neoplasms, including acute myeloid leukemia, congenital mesoblastic nephroma and fibrosarcoma [56], suggesting that the ETV6–NTRK3 fusion protein has transforming activity in cells of different lineages. Interestingly, Sorensen and co-workers have shown that an intact IGF1R/insulin receptor (INSR) signaling axis is necessary for oncogenic transformation induced by ETV6–NTRK3 [59] and that transformation of mammary epithelial cells by ETV6–NTRK3 can be blocked by targeting the IGF1R/INSR signaling pathway [60]. Treatment with the small molecule IGF1R/INSR kinase inhibitors BMS-754807 and BMS-536924 can block ETV6–NTRK3 induced oncogenic transformation in vitro and significantly reduces tumor growth in vivo (Fig. 2). These findings indicate that the IGF1R/INSR signaling pathway is an interesting new therapeutic target in ETV6–NTRK3 positive cancers and that fusion-positive patients with MASC may benefit from treatment with IGF1R/INSR inhibitors. Hyalinizing clear cell carcinoma (HCCC) HCCC is a unique low-grade SGC with typical clear-cell morphology and pattern of hyalinization often with focal mucinous differentiation [61]. The differential diagnosis is wide and includes for example MEC, epithelial-myoepithelial carcinoma, and clear cell myoepithelial carcinoma. HCCC has usually an excellent prognosis and only occasional cases with metastatic spread are seen. Antonescu and coworkers recently described a recurrent t(12;22) (q13;q12) translocation in HCCC [61,62]. This translocation, which was originally identified in soft tissue, clear cell sarcoma, results in a fusion between the EWSR1 and ATF1 genes (Figs. 1 and 2) [63]. Rearrangements of EWSR1 (detected by FISH) have been found in up to 87% of HCCC [61,64] and recently also in a high frequency of clear cell odontogenic carcinomas (CCOC) [65]. The latter tumors are morphologically very similar or identical to HCCC and the finding of EWSR1–ARF1 fusions also in these tumors raise the question of whether CCOCs in fact are central HCCCs [65]. In contrast, EWSR1–ATF1 has not been detected in any other clear cell mimics, demonstrating that it is robust and diagnostically usefulness molecular marker for HCCC. Carcinoma ex pleomorphic adenoma (Ca-ex-PA) Ca-ex-PA is defined as a carcinoma arising from a primary or recurrent benign pleomorphic adenoma (PA). They constitute approximately 12% of all SGCs [1]. The malignant component is often an adenocarcinoma not otherwise specified (NOS) or an undifferentiated carcinoma, but may in principle be any other histological subtype of SGC, including salivary duct carcinoma, MEC, or ACC. Ca-ex-PA may be subclassified into non-invasive, minimally invasive, and invasive tumors. The two former typically have an excellent prognosis whereas the latter usually are high-grade, aggressive tumors with an unfavorable prognosis. We have previously shown that Ca-ex-PA, as expected, display the prototypic PA-specifıc gene fusions involving the transcription factor genes PLAG1 and HMGA2 (Fig. 2) [4–6,66–68]. Subsets of these tumors also show amplifıcation of MDM2 and HMGA2 in 12q13–15, mutations of TP53, and/or amplification of HER2 (ERBB2) as molecular markers of malignant transformation [66,67,69]. The findings of TP53 mutations or MDM2 amplifıcation suggest that targeting the p53-pathway with small-molecule inhibitors of MDM2

5

(e.g., MI-219 or Nutlin-3 analogs), which can reconstitute TP53function in tumors with MDM2-amplifıcation/overexpression, may be an attractive way of treating tumors with these molecular abnormalities (Fig. 2) [70]. Most Ca-ex-PAs with HER2 amplifıcation are salivary duct carcinomas developing within PAs. Of note, there are anecdotal reports showing that these patients may significantly benefit from anti-HER2 treatment with trastuzumab [71]. Salivary duct carcinoma (SDC) SDCs account for approximately 10% of all SGCs and are often diagnosed in males above 50 years of age. It is one of the most aggressive SGCs and may occur de novo or as SDC-ex-PA (see above) [1]. SDC shows strong histologic resemblance to high-grade ductal carcinoma of the breast [72]. The primary treatment of SDC is surgical excision with lymph node dissection and subsequent radiotherapy with or without chemotherapy. Local recurrences as well as regional lymph node and distant metastases are common. Most patients succumb to the disease within 3–4 years after the initial diagnosis [73,74]. Similar to invasive ductal carcinoma of the breast, SDC show overexpression and amplification of the HER2 gene [73–76]. Overexpression is found in approximately one third of the cases and the majority of these have amplification of HER2 [77–79]. SDC patients with HER2 overexpression and amplification may be targeted with trastuzumab (Fig. 2). Indeed, there are several recent reports indicating positive responses following treatment with anti-HER2 agents, such as trastuzumab, with improved disease-free and overall survival in patients with progressive disease [73,74,79,80]. In contrast to ductal carcinoma of the breast, SDC do not express estrogen or progesterone receptors. However, since the majority of SDCs do express high levels of the androgen receptor (AR), androgen deprivation therapy has emerged as a possible targeted approach for these patients (Fig. 2). Thus, small series of AR + SDC patients have shown clinical benefit following androgen deprivation therapy [72,79]. Interestingly, recent studies of the mutation spectrum of SDC have revealed that subsets of HER2-negative SDCs have mutations in PIK3CA (20–33%) or hemi- or homozygous deletions of PTEN (50–59%) [75,79,81,82], suggesting that the PI3K/Akt/mTOR pathway may be an attractive therapeutic target in these patients. Indeed, studies of a few patients have confirmed that they indeed may benefit from treatment with temsirolimus, an inhibitor of this pathway [75]. Another subset of SDCs were recently shown to have activating BRAF V600E kinase mutations (7%), indicating that BRAF inhibitors such as vemurafenib may be effective in these patients (Fig. 2) [83]. In all, these observations further underscore the diversity of driver mutations observed in SDCs and calls for personalized genetic profiling of these patients to determine the optimal treatment regimen. Acinic cell carcinoma (AciCC) AciCC is a low-grade, slow-growing tumor generally associated with a good prognosis [1]. However, a subset of cases may occasionally develop recurrent and/or metastatic disease or undergo high-grade transformation. Next to nothing is known about the genetic profile of AciCC. Only 11 tumors have been reported with abnormal karyotypic profile and the only common change observed was trisomy 8 in three cases [84]. No gene fusions or recurrent mutations have been identified. A first step in the understanding of the molecular pathogenesis of AciCC was recently reported when it was shown that mice with constitutive activation of the Wnt and mTOR signaling pathways by conditional inactivation of the Apc and Pten tumor suppressor genes in salivary glands develop tumors similar to human AciCCs

Please cite this article in press as: Stenman G et al. Diagnostic and therapeutic implications of new molecular biomarkers in salivary gland cancers. Oral Oncol (2014), http://dx.doi.org/10.1016/j.oraloncology.2014.04.008

6

G. Stenman et al. / Oral Oncology xxx (2014) xxx–xxx

[85]. Treatment of AciCC-bearing mice with the mTOR inhibitor rapamycin resulted in complete regression of the tumors (Fig. 2). Immunohistochemical analysis of human AciCC samples revealed that mTOR signaling is also activated in human AciCCs [82,85], indicating that mTOR inhibitors such as rapamycin or temsirolimus might be useful for treating patients with AciCC. Epithelial-myoepithelial carcinoma (EMC) EMC is a rare biphasic SGC with some morphologic resemblance to benign pleomorphic adenoma [1]. It is a low-grade tumor of presumed intercalated duct origin that usually has an excellent prognosis. Recent studies have demonstrated that these tumors have hotspot mutations in the HRAS gene in up to 25% of the cases [86,87]. Notably, such mutations were not detected in high-grade EMC or in EMC-ex-PA. These data suggest that future clinical trials, testing for example MEK-specific inhibitors, e.g. trametinib (Fig. 2), in patients with EMC will require careful patient selection and biomarker screening. Conclusions and future perspectives SGCs represent a heterogeneous group of cancers with different histologies, molecular background, and biologic behavior. The diagnoses are challenging, because of the diversity of histologic subtypes and the overlapping morphological patterns between many of these lesions. Patients with localized disease are generally treated with surgery and/or radiation therapy and there is only limited data on the role of systemic therapies in patients with progressive metastatic disease. To improve the classification of SGCs and to develop new therapeutic options for patients with highgrade tumors and/or advanced disease we clearly need new molecularly based diagnostic, prognostic, and therapeutic biomarkers. Indeed, recent molecular genetic studies have revealed that at least six different subtypes of SGCs are characterized by translocationgenerated gene fusions. These fusions are tumor-type specific and clinically important diagnostic, and for certain entities also prognostic, biomarkers but most importantly they are novel targets for development of new molecularly targeted therapies. The discovery of this gene fusion network is the most important breakthrough in our understanding of the molecular pathogenesis of SGCs that has occurred during the last decades. Moreover, nextgeneration sequencing studies of SGCs have revealed new druggable driver mutations, thus pinpointing new therapeutic options for subsets of patients. Based on our current knowledge, we suggest that patients with the aforementioned SGCs should be screened for known gene fusions and driver mutations before being enrolled in clinical trials with targeted agents. Previous studies of SGCs have been seriously hampered by the lack of relevant animal and in vitro models. Efforts are now underway in several laboratories to develop such in vivo and in vitro models. These will no doubt greatly facilitate future studies of the pathogenesis of SGCs and the development of new therapeutic avenues for these patients. Conflict of interest statement None declared. Funding This work was supported by Grants from the Swedish Cancer Society and BioCARE – a National Strategic Research Program at University of Gothenburg.

References [1] Barnes L, Eveson JW, Reichart P, Sidransky D. World health organization classification of tumours. Pathology and genetics of head and neck tumours. Lyon: IARC Press; 2005. p. 209–74. [2] Laurie SA, Licitra L. Systemic therapy in the palliative management of advanced salivary gland cancers. J Clin Oncol 2006;24:2673–8. [3] Carlson J, Licitra L, Locati L, Raben D, Persson F, Stenman G. Salivary gland cancer: an update on present and emerging therapies. Am Soc Clin Oncol Educ Book 2013:257–63. [4] Stenman G. Fusion oncogenes and tumor type specificity – insights from salivary gland tumors. Semin Cancer Biol 2005;15:224–35. [5] Stenman G, Andersson MK, Andren Y. New tricks from an old oncogene: gene fusion and copy number alterations of MYB in human cancer. Cell Cycle 2010;9:2986–95. [6] Stenman G. Fusion oncogenes in salivary gland tumors: molecular and clinical consequences. Head Neck Pathol 2013;7(Suppl. 1). S12–9. [7] Ho AS, Kannan K, Roy DM, Morris LG, Ganly I, Katabi N, et al. The mutational landscape of adenoid cystic carcinoma. Nat Genet 2013;45:791–8. [8] Stephens PJ, Davies HR, Mitani Y, Van Loo P, Shlien A, Tarpey PS, et al. Whole exome sequencing of adenoid cystic carcinoma. J Clin Invest 2013;123:2965–8. [9] Nordkvist A, Gustafsson H, Juberg-Ode M, Stenman G. Recurrent rearrangements of 11q14–22 in mucoepidermoid carcinoma. Cancer Genet Cytogenet 1994;74:77–83. [10] Tonon G, Modi S, Wu L, Kubo A, Coxon AB, Komiya T, et al. t(11;19)(q21;p13) Translocation in mucoepidermoid carcinoma creates a novel fusion product that disrupts a Notch signaling pathway. Nat Genet 2003;33:208–13. [11] Enlund F, Behboudi A, Andren Y, Oberg C, Lendahl U, Mark J, et al. Altered Notch signaling resulting from expression of a WAMTP1–MAML2 gene fusion in mucoepidermoid carcinomas and benign Warthin’s tumors. Exp Cell Res 2004;292:21–8. [12] Tirado Y, Williams MD, Hanna EY, Kaye FJ, Batsakis JG, El-Naggar AK. CRTC1/ MAML2 fusion transcript in high grade mucoepidermoid carcinomas of salivary and thyroid glands and Warthin’s tumors: implications for histogenesis and biologic behavior. Genes Chromosomes Cancer 2007;46: 708–15. [13] Von Holstein SL, Fehr A, Heegaard S, Therkildsen MH, Stenman G. CRTC1– MAML2 gene fusion in mucoepidermoid carcinoma of the lacrimal gland. Oncol Rep 2012;27:1413–6. [14] Conkright MD, Canettieri G, Screaton R, Guzman E, Miraglia L, Hogenesch JB, et al. TORCs: transducers of regulated CREB activity. Mol Cell 2003;12:413–23. [15] Wu L, Sun T, Kobayashi K, Gao P, Griffin JD. Identification of a family of mastermind-like transcriptional coactivators for mammalian notch receptors. Mol Cell Biol 2002;22:7688–700. [16] Coxon A, Rozenblum E, Park YS, Joshi N, Tsurutani J, Dennis PA, et al. MECT1– MAML2 fusion oncogene linked to the aberrant activation of cyclic AMP/CREB regulated genes. Cancer Res 2005;65:7137–44. [17] Chen Z, Chen J, Gu Y, Hu C, Li JL, Lin S, et al. Aberrantly activated AREG–EGFR signaling is required for the growth and survival of CRTC1–MAML2 fusionpositive mucoepidermoid carcinoma cells. Oncogene 2013. [18] Auclair PL, Goode RK, Ellis GL. Mucoepidermoid carcinoma of intraoral salivary glands. Evaluation and application of grading criteria in 143 cases. Cancer 1992;69:2021–30. [19] Goode RK, El-Naggar AK. Mucoepidermoid carcinoma. In: Barnes L, Eveson JW, Reichart P, Sidransky D, editors. World health organization classification of tumours pathology and genetics of head and neck tumours. Lyon: IARC Press; 2005. p. 209–74. [20] Aro K, Leivo I, Makitie AA. Management and outcome of patients with mucoepidermoid carcinoma of major salivary gland origin: a single institution’s 30-year experience. Laryngoscope 2008;118:258–62. [21] Ellis GL, Auclair PL. Tumors of the salivary glands. Washington DC: American Registry of Pathology; 2008. [22] Seethala RR, Dacic S, Cieply K, Kelly LM, Nikiforova MN. A reappraisal of the MECT1/MAML2 translocation in salivary mucoepidermoid carcinomas. Am J Surg Pathol 2010;34:1106–21. [23] Behboudi A, Enlund F, Winnes M, Andren Y, Nordkvist A, Leivo I, et al. Molecular classification of mucoepidermoid carcinomas–prognostic significance of the MECT1–MAML2 fusion oncogene. Genes Chromosomes Cancer 2006;45:470–81. [24] Okabe M, Miyabe S, Nagatsuka H, Terada A, Hanai N, Yokoi M, et al. MECT1– MAML2 fusion transcript defines a favorable subset of mucoepidermoid carcinoma. Clin Cancer Res 2006;12:3902–7. [25] Clauditz TS, Gontarewicz A, Wang CJ, Munscher A, Laban S, Tsourlakis MC, et al. 11q21 Rearrangement is a frequent and highly specific genetic alteration in mucoepidermoid carcinoma. Diagn Mol Pathol 2012;21:134–7. [26] Jee KJ, Persson M, Heikinheimo K, Passador-Santos F, Aro K, Knuutila S, et al. Genomic profiles and CRTC1–MAML2 fusion distinguish different subtypes of mucoepidermoid carcinoma. Mod Pathol 2013;26:213–22. [27] Anzick SL, Chen WD, Park Y, Meltzer P, Bell D, El-Naggar AK, et al. Unfavorable prognosis of CRTC1–MAML2 positive mucoepidermoid tumors with CDKN2A deletions. Genes Chromosomes Cancer 2010;49:59–69. [28] Moller E, Stenman G, Mandahl N, Hamberg H, Molne L, van den Oord JJ, et al. POU5F1, encoding a key regulator of stem cell pluripotency, is fused to EWSR1 in hidradenoma of the skin and mucoepidermoid carcinoma of the salivary glands. J Pathol 2008;215:78–86.

Please cite this article in press as: Stenman G et al. Diagnostic and therapeutic implications of new molecular biomarkers in salivary gland cancers. Oral Oncol (2014), http://dx.doi.org/10.1016/j.oraloncology.2014.04.008

G. Stenman et al. / Oral Oncology xxx (2014) xxx–xxx [29] Antonescu CR, Zhang L, Chang NE, Pawel BR, Travis W, Katabi N, et al. EWSR1– POU5F1 fusion in soft tissue myoepithelial tumors. A molecular analysis of sixty-six cases, including soft tissue, bone, and visceral lesions, showing common involvement of the EWSR1 gene. Genes Chromosomes Cancer 2010;49:1114–24. [30] Persson M, Andren Y, Mark J, Horlings HM, Persson F, Stenman G. Recurrent fusion of MYB and NFIB transcription factor genes in carcinomas of the breast and head and neck. Proc Natl Acad Sci USA 2009;106:18740–4. [31] Persson M, Andren Y, Moskaluk CA, Frierson Jr HF, Cooke SL, Futreal PA, et al. Clinically significant copy number alterations and complex rearrangements of MYB and NFIB in head and neck adenoid cystic carcinoma. Genes Chromosomes Cancer 2012;51:805–17. [32] Stenman G, Sandros J, Dahlenfors R, Juberg-Ode M, Mark J. 6q- and loss of the Y chromosome – two common deviations in malignant human salivary gland tumors. Cancer Genet Cytogenet 1986;22:283–93. [33] Nordkvist A, Mark J, Gustafsson H, Bang G, Stenman G. Non-random chromosome rearrangements in adenoid cystic carcinoma of the salivary glands. Genes Chromosomes Cancer 1994;10:115–21. [34] Ramsay RG, Gonda TJ. MYB function in normal and cancer cells. Nat Rev Cancer 2008;8:523–34. [35] von Holstein SL, Fehr A, Persson M, Therkildsen MH, Prause JU, Heegaard S, et al. Adenoid cystic carcinoma of the lacrimal gland: MYB gene activation, genomic imbalances, and clinical characteristics. Ophthalmology 2013;120:2130–8. [36] Brill 2nd LB, Kanner WA, Fehr A, Andren Y, Moskaluk CA, Loning T, et al. Analysis of MYB expression and MYB–NFIB gene fusions in adenoid cystic carcinoma and other salivary neoplasms. Mod Pathol 2011;24:1169–76. [37] Mitani Y, Rao PH, Futreal PA, Roberts DB, Stephens PJ, Zhao YJ, et al. Novel chromosomal rearrangements and break points at the t(6;9) in salivary adenoid cystic carcinoma: association with MYB–NFIB chimeric fusion, MYB expression, and clinical outcome. Clin Cancer Res 2011;17:7003–14. [38] West RB, Kong C, Clarke N, Gilks T, Lipsick JS, Cao H, et al. MYB expression and translocation in adenoid cystic carcinomas and other salivary gland tumors with clinicopathologic correlation. Am J Surg Pathol 2011;35:92–9. [39] O’Rourke JP, Ness SA. Alternative RNA splicing produces multiple forms of cMyb with unique transcriptional activities. Mol Cell Biol 2008;28:2091–101. [40] Ramsay RG, Barton AL, Gonda TJ. Targeting c-Myb expression in human disease. Expert Opin Ther Targets 2003;7:235–48. [41] Gonda TJ, Leo P, Ramsay RG. Estrogen and MYB in breast cancer: potential for new therapies. Expert Opin Biol Ther 2008;8:713–7. [42] Pattabiraman DR, Gonda TJ. Role and potential for therapeutic targeting of MYB in leukemia. Leukemia 2013;27:269–77. [43] Quintana AM, Liu F, O’Rourke JP, Ness SA. Identification and regulation of cMyb target genes in MCF-7 cells. BMC Cancer 2011;11:30. [44] Wetterskog D, Wilkerson PM, Rodrigues DN, Lambros MB, Fritchie K, Andersson MK, et al. Mutation profiling of adenoid cystic carcinomas from multiple anatomical sites identifies mutations in the RAS pathway, but no KIT mutations. Histopathology 2013;62:543–50. [45] Hotte SJ, Winquist EW, Lamont E, MacKenzie M, Vokes E, Chen EX, et al. Imatinib mesylate in patients with adenoid cystic cancers of the salivary glands expressing c-kit: a Princess Margaret Hospital phase II consortium study. J Clin Oncol 2005;23:585–90. [46] Pfeffer MR, Talmi Y, Catane R, Symon Z, Yosepovitch A, Levitt M. A phase II study of imatinib for advanced adenoid cystic carcinoma of head and neck salivary glands. Oral Oncol 2007;43:33–6. [47] Bartunek P, Pajer P, Karafiat V, Blendinger G, Dvorak M, Zenke M. BFGF signaling and v-Myb cooperate in sustained growth of primitive erythroid progenitors. Oncogene 2002;21:400–10. [48] Dillon P, Moskaluk C, Fracasso P, Petroni G, Thomas C. Phase II study of dovitinib (TKI258) in patients with progressive metastatic adenoid cystic carcinoma. J Clin Oncol 2013:31. [49] Ivanov SV, Panaccione A, Brown B, Guo Y, Moskaluk CA, Wick MJ, et al. TrkC signaling is activated in adenoid cystic carcinoma and requires NT-3 to stimulate invasive behavior. Oncogene 2013;32:3698–710. [50] Ross JS, Wang K, Rand JV, Sheehan CE, Jennings TA, Al-Rohil RN, et al. Comprehensive genomic profiling of relapsed and metastatic adenoid cystic carcinomas by next-generation sequencing reveals potential new routes to targeted therapies. Am J Surg Pathol 2014;38:235–8. [51] Skalova A, Vanecek T, Sima R, Laco J, Weinreb I, Perez-Ordonez B, et al. Mammary analogue secretory carcinoma of salivary glands, containing the ETV6–NTRK3 fusion gene: a hitherto undescribed salivary gland tumor entity. Am J Surg Pathol 2010;34:599–608. [52] Fehr A, Loning T, Stenman G. Mammary analogue secretory carcinoma of the salivary glands with ETV6–NTRK3 gene fusion. Am J Surg Pathol 2011;35: 1600–2. [53] Skalova A. Mammary analogue secretory carcinoma of salivary gland origin: an update and expanded morphologic and immunohistochemical spectrum of recently described entity. Head Neck Pathol 2013;7(Suppl. 1). S30–6. [54] Wai DH, Knezevich SR, Lucas T, Jansen B, Kay RJ, Sorensen PH. The ETV6– NTRK3 gene fusion encodes a chimeric protein tyrosine kinase that transforms NIH3T3 cells. Oncogene 2000;19:906–15. [55] Tognon C, Knezevich SR, Huntsman D, Roskelley CD, Melnyk N, Mathers JA, et al. Expression of the ETV6–NTRK3 gene fusion as a primary event in human secretory breast carcinoma. Cancer Cell 2002;2:367–76. [56] Lannon CL, Sorensen PH. ETV6–NTRK3: a chimeric protein tyrosine kinase with transformation activity in multiple cell lineages. Semin Cancer Biol 2005;15: 215–23.

7

[57] Bishop JA, Yonescu R, Batista D, Eisele DW, Westra WH. Most nonparotid ‘‘acinic cell carcinomas’’ represent mammary analog secretory carcinomas. Am J Surg Pathol 2013;37:1053–7. [58] Skalova A, Vanecek T, Majewska H, Laco J, Grossmann P, Simpson RH, et al. Mammary analogue secretory carcinoma of salivary glands with high-grade transformation: report of 3 cases with the ETV6–NTRK3 gene fusion and analysis of TP53, beta-catenin, EGFR, and CCND1 genes. Am J Surg Pathol 2014;38:23–33. [59] Morrison KB, Tognon CE, Garnett MJ, Deal C, Sorensen PH. ETV6–NTRK3 transformation requires insulin-like growth factor 1 receptor signaling and is associated with constitutive IRS-1 tyrosine phosphorylation. Oncogene 2002;21:5684–95. [60] Tognon CE, Somasiri AM, Evdokimova VE, Trigo G, Uy EE, Melnyk N, et al. ETV6–NTRK3-mediated breast epithelial cell transformation is blocked by targeting the IGF1R signaling pathway. Cancer Res 2011;71:1060–70. [61] Weinreb I. Hyalinizing clear cell carcinoma of salivary gland: a review and update. Head Neck Pathol 2013;7(Suppl. 1). S20–9. [62] Antonescu CR, Katabi N, Zhang L, Sung YS, Seethala RR, Jordan RC, et al. EWSR1–ATF1 fusion is a novel and consistent finding in hyalinizing clear-cell carcinoma of salivary gland. Genes Chromosomes Cancer 2011;50:559–70. [63] Zucman J, Delattre O, Desmaze C, Epstein AL, Stenman G, Speleman F, et al. EWS and ATF-1 gene fusion induced by t(12;22) translocation in malignant melanoma of soft parts. Nat Genet 1993;4:341–5. [64] Shah AA, LeGallo RD, van Zante A, Frierson Jr HF, Mills SE, Berean KW, et al. EWSR1 genetic rearrangements in salivary gland tumors: a specific and very common feature of hyalinizing clear cell carcinoma. Am J Surg Pathol 2013;37:571–8. [65] Bilodeau EA, Weinreb I, Antonescu CR, Zhang L, Dacic S, Muller S, et al. Clear cell odontogenic carcinomas show EWSR1 rearrangements: a novel finding and a biological link to salivary clear cell carcinomas. Am J Surg Pathol 2013;37:1001–5. [66] Roijer E, Nordkvist A, Strom AK, Ryd W, Behrendt M, Bullerdiek J, et al. Translocation, deletion/amplification, and expression of HMGIC and MDM2 in a carcinoma ex pleomorphic adenoma. Am J Pathol 2002;160:433–40. [67] Persson F, Andren Y, Winnes M, Wedell B, Nordkvist A, Gudnadottir G, et al. High-resolution genomic profiling of adenomas and carcinomas of the salivary glands reveals amplification, rearrangement, and fusion of HMGA2. Genes Chromosomes Cancer 2009;48:69–82. [68] von Holstein SL, Fehr A, Persson M, Nickelsen M, Therkildsen MH, Prause JU, et al. Lacrimal gland pleomorphic adenoma and carcinoma ex pleomorphic adenoma: genomic profiles, gene fusions, and clinical characteristics. Ophthalmology 2014. [69] Nordkvist A, Roijer E, Bang G, Gustafsson H, Behrendt M, Ryd W, et al. Expression and mutation patterns of p53 in benign and malignant salivary gland tumors. Int J Oncol 2000;16:477–83. [70] Shangary S, Wang S. Small-molecule inhibitors of the MDM2-p53 protein– protein interaction to reactivate p53 function: a novel approach for cancer therapy. Annu Rev Pharmacol Toxicol 2009;49:223–41. [71] Sharon E, Kelly RJ, Szabo E. Sustained response of carcinoma ex pleomorphic adenoma treated with trastuzumab and capecitabine. Head Neck Oncol 2010;2:12. [72] Jaspers HC, Verbist BM, Schoffelen R, Mattijssen V, Slootweg PJ, van der Graaf WT, et al. Androgen receptor-positive salivary duct carcinoma: a disease entity with promising new treatment options. J Clin Oncol 2011;29. e473–6. [73] Nashed M, Casasola RJ. Biological therapy of salivary duct carcinoma. J Laryngol Otol 2009;123:250–2. [74] Prat A, Parera M, Reyes V, Peralta S, Cedres S, Andreu J, et al. Successful treatment of pulmonary metastatic salivary ductal carcinoma with trastuzumab-based therapy. Head Neck 2008;30:680–3. [75] Piha-Paul SA, Cohen PR, Kurzrock R. Salivary duct carcinoma: targeting the phosphatidylinositol 3-kinase pathway by blocking mammalian target of rapamycin with temsirolimus. J Clin Oncol 2011;29. e727–30. [76] Locati LD, Bossi P, Licitra L. How many therapeutic options are there for recurrent or metastatic salivary duct carcinoma. J Clin Oncol 2012;30:672 [author reply-3]. [77] Williams MD, Roberts DB, Kies MS, Mao L, Weber RS, El-Naggar AK. Genetic and expression analysis of HER-2 and EGFR genes in salivary duct carcinoma: empirical and therapeutic significance. Clin Cancer Res 2010;16:2266–74. [78] Suzuki S, Dobashi Y, Minato H, Tajiri R, Yoshizaki T, Ooi A. EGFR and HER2– Akt–mTOR signaling pathways are activated in subgroups of salivary gland carcinomas. Virchows Arch 2012;461:271–82. [79] Nardi V, Sadow PM, Juric D, Zhao D, Cosper AK, Bergethon K, et al. Detection of novel actionable genetic changes in salivary duct carcinoma helps direct patient treatment. Clin Cancer Res 2013;19:480–90. [80] Nabili V, Tan JW, Bhuta S, Sercarz JA, Head CS. Salivary duct carcinoma: a clinical and histologic review with implications for trastuzumab therapy. Head Neck 2007;29:907–12. [81] Griffith CC, Seethala RR, Luvison A, Miller M, Chiosea SI. PIK3CA mutations and PTEN loss in salivary duct carcinomas. Am J Surg Pathol 2013;37:1201–7. [82] Ettl T, Schwarz-Furlan S, Haubner F, Muller S, Zenk J, Gosau M, et al. The PI3K/ AKT/mTOR signalling pathway is active in salivary gland cancer and implies different functions and prognoses depending on cell localisation. Oral Oncol 2012;48:822–30. [83] Bollag G, Hirth P, Tsai J, Zhang J, Ibrahim PN, Cho H, et al. Clinical efficacy of a RAF inhibitor needs broad target blockade in BRAF-mutant melanoma. Nature 2010;467:596–9.

Please cite this article in press as: Stenman G et al. Diagnostic and therapeutic implications of new molecular biomarkers in salivary gland cancers. Oral Oncol (2014), http://dx.doi.org/10.1016/j.oraloncology.2014.04.008

8

G. Stenman et al. / Oral Oncology xxx (2014) xxx–xxx

[84] Mitelman F, Johansson B, Mertens F. Mitelman Database of Chromosome Aberrations and Gene Fusions in Cancer; 2014. . [85] Diegel CR, Cho KR, El-Naggar AK, Williams BO, Lindvall C. Mammalian target of rapamycin-dependent acinar cell neoplasia after inactivation of Apc and Pten in the mouse salivary gland: implications for human acinic cell carcinoma. Cancer Res 2010;70:9143–52.

[86] Chiosea SI, Miller M, Seethala RR. HRAS mutations in epithelial–myoepithelial carcinoma. Head Neck Pathol 2013. [87] Cros J, Sbidian E, Hans S, Roussel H, Scotte F, Tartour E, et al. Expression and mutational status of treatment-relevant targets and key oncogenes in 123 malignant salivary gland tumours. Ann Oncol 2013;24:2624–9.

Please cite this article in press as: Stenman G et al. Diagnostic and therapeutic implications of new molecular biomarkers in salivary gland cancers. Oral Oncol (2014), http://dx.doi.org/10.1016/j.oraloncology.2014.04.008