Sinonasal cancers (SNCs) include different tumors of the nasal cavities, maxillary, sphenoidal, ethmoidal, and frontal sinuses. Epithelial SNCs include different histological subtypes: the most common is squamous cell carcinoma (SCC), either keratinizing or non-keratinizing, followed by adenocarcinoma (intestinal-type or non-intestinal type), sinonasal undifferentiated carcinoma (SNUC), sinonasal neuroendocrine carcinoma (SNEC), NUT carcinoma, lymphoepithelial carcinoma, teratocarcinosarcoma, and minor salivary gland tumors.
1. Squamous Cell Carcinoma
Sinonasal SCC (SNSCC) is the most common histological subtype (60–75%) of the skull base, with an incidence of 35–58%
[1] and a 5-year mortality rate of ~40%
[2].
The genetic characterization of SNSCC is showing promising results. In 2015, Udager et al.
[3] analyzed the presence of pathogenic somatic mutations in SNSCC and showed a high prevalence of EGFR alterations (88%) in the inverted papillomas (IPs) and IP-associated SNSCC cases (77%). In contrast, no EGFR alterations were observed in the non-IP-associated SNSCC and in other papillomas. The most common EGFR alteration identified was exon 20-insertion (ins), involving residues located between A767 and V774. Other less common EGFR alterations were deletion-insertion in exon 19 and single nucleotide substitution in exon 19. In addition, in de novo SNSCC EGFR gene amplifications have been documented in about 30% of cases
[3][4].
Since several therapies are approved for treating EGFR-mutant non small-cell lung cancer (NSCLC), unique treatment opportunities may open. The potential utility of first-generation EGFR-inhibitors (gefitinib and erlotinib) and second-generation EGFR-inhibitors (neratinib, afatinib, dacomitinib) in the context of SNSCC has been investigated and has shown limited results
[3]. This could be explained by the high prevalence of EGFR exon 20-ins, which are resistant to these drugs
[3], but are more susceptible to new target therapies, including amivantamab
[5] and mobocertinib
[6], recently studied and approved in NSCLC. Trials with these molecules in SNSCC are desirable. Even poziotinib (HM781–36B), an irreversible EGFR inhibitor, has been studied in different clinical trials, showing efficacy in NSCLC
[7] and in recurrent and/or metastatic head and neck SCC (R/M-HNSCC)
[8].
ERBB2 copy number gain is another genetic alteration found in SNSCC with an incidence of 21% and elevated protein expression levels of 7%. The ERBB2 amplification and overexpression correlated with higher tumor stage (T4), intracranial dissemination, and worse outcomes
[9]. Several agents have been to treatHER2-overexpressing breast cancer and metastatic gastric or gastroesophageal junction adenocarcinoma. The efficacy of anti-HER2 agents might also be tested in SNSCC.
Further non-actionable genomic alterations were found in SNSCC. A higher frequency of p53 expression in SNSCC was reported by several studies, ranging from 33.3% to 100%
[10]. Brown et al.
[11] identified CDKN2A inactivation in 72.4% of the SNSCC, through mutation and subsequent loss of heterozygosity or focal ‘deep deletion’ of the gene locus. At the same time, it was not detected in sinonasal papillomas. Overexpression of TrkB
[12] was identified in 70.4% of SNSCC analyzed and was associated with poor prognosis. SOX2 amplifications were identified in 35% of SNSCC
[13]. Other SNSCC minor molecular alterations are TERT copy number gains (27.6%) without TERT promoter mutations, NFE2L2 mutation, CCND1 and MYC copy number gain
[11] and CARD11 mutation
[13]. Finally, 3.2% of sinonasal tumors showed a deficiency of mismatch repair proteins and/or high microsatellite instability (dMMR/MSI-H), which may confer clinical benefit to immune checkpoint inhibitors (ICIs) treatment
[14].
To conclude, the DEK::AFF2 fusion-associated carcinoma was recently detected as a distinct variant of SNSCC
[15][16]. In a patient with DEK::AFF2 fusion-associated carcinoma, an exceptional response to ICIs was identified
[15][17].
2. Intestinal-Type Adenocarcinoma (ITAC)
Intestinal-type adenocarcinoma (ITAC) is the most frequent adenocarcinoma of the skull base and occurs predominantly in the ethmoid sinuses (40–85%)
[18][19]. Franchi et al.
[20] suggested that ITAC could arise from premalignant intestinal metaplasia of respiratory and/or glandular epithelium.
ITACs are named for their histologic resemblance to adenocarcinoma of the intestinal tract. Since the late 1990s and early 2000s, researchers have considered ITAC and colorectal adenocarcinoma molecular pathways to overlap in different studies
[21][22].
TP53 is the most frequently mutated gene (40–50%, up to 86%) and no target drugs are available. However, there are ongoing trials on the potential role of WEE1 inhibitors (such as adavosertib
[23]) in p53-mutated or deficient cancer cells. p53 status may be used to predict response to chemotherapy
[24][25].
KRAS and HRAS mutations have been found in one of 12 (8%) and in five of 31 (16%) ITACs, respectively
[26][27]. The frequency of KRAS mutations in sinonasal carcinomas is lower than the 30–45% reported in colorectal cancer
[28]. The KRAS mutations primarily consist of base pair changes in three hotspots, corresponding to codons 12 and 13 in exon 1 and codon 61 in exon 2
[29]. No specific target agents are available for these types of KRAS mutations. Pérez et al.
[30] analyzed 31 ITACs for the presence of HRAS mutations: G12V alteration appears to be the most frequent in the HRAS gene (16%). HRAS mutations were related to a worse prognosis. In another study, no HRAS mutations were found
[31]. Tipifarnib, a farnesyltransferase inhibitor that disrupts HRAS function, has been investigated in metastatic HNSCC with high mHRAS variant allele frequency, showing promising efficacy
[32]. NRAS mutations have been infrequently described in ITAC
[33].
EGFR amplifications and/or overexpression are present in a substantial subset of ITACs with a colonic differentiation pattern
[34]. EGFR gene copy number gains occur in 38–55% of the cases, mostly in the context of a whole chromosome 7 gain. High-level amplification is reported to be rather infrequent, between 2% and 16%. The frequency of EGFR alterations observed in ITAC is lower than in colorectal cancer, lung cancer, or HNSCC
[35] and SNSCC. EGFR overexpressed ITAC could be potentially treated with EGFR inhibitors.
Most ITACs carry genetic alterations in four different pathways: Wnt/b catenin, DNA damage response (ATM, BRCA 1 and 2), MAPK and PI3K pathways. This means many ITACs might be treated with specific inhibitors of these pathways. Promising specific therapies targeting the Wnt pathway are currently under investigation in phase I clinical trials
[36][37]. Treatment with PARP inhibitors may be considered for ATM, BRCA1 or BRCA2-mutant ITACs. PIK3CA mutations may be susceptible to PIK3CA inhibitors (alpelisib), mTOR inhibitors or new molecules such as AKT inhibitors (capivartesib and ipatasertib).
Although emphasis is placed on these four signaling pathways, other potentially actionable mutations have been found. BRAF mutations have been rarely seen in a subset of ITAC. MET inhibitors represent another interesting treatment option since MET-activating mutation can be found in up to 64% of ITACs 49; other possible opportunities could be trametinib or cobimetinib in NF1-mutated, anti-HER2 (such as trastuzumab, trastuzumab-deruxtecan and trastuzumab-emtansine) in ERBB2-mutated
[38], anti-IDH1 in IDH1- mutated ITAC
[33]. However, at the moment, no efficacy data are present in the literature about targeted agents agnostically used in ITAC treatment.
3. Non-Intestinal Type Adenocarcinomas (N-ITAC)
Sinonasal non-intestinal type adenocarcinoma (N-ITAC) is an extremely rare adenocarcinoma, which morphologically presents neither intestinal-type nor salivary-type adenocarcinoma aspects
[39]. According to immunohistochemistry, this type of tumor shows respiratory-type features.
Different variants of N-ITAC are commonly divided into two categories: low grade (with a particular subset of seromucinous adenocarcinoma) and high grade (blastomatous, oncocytic/mucinous, apocrine, poorly differentiated and undifferentiated types)
[40]. There is also a very rare distinct form of N-ITAC, the renal cell-like adenocarcinoma
[41]. Differences between these histological subtypes are related to the expression of different biomarkers detected using IHC
[42].
Few studies have analyzed the mutational landscape of N-ITAC. Yom et al.
[29] noted a small subset of N-ITAC cases showing p53 overexpression, whereas other cases did not show any genetic abnormalities in KRAS, APC, CTNNB1, DNA mismatch repair genes, or TP53. Another study by Franchi et al.
[43] reported that two cases contained a BRAF V600E mutation detected by direct sequencing. BRAF inhibitors may be a therapeutic option for a small quote of N-ITAC with EGFR overexpression and BRAF mutations.
4. Sinonasal Neuroendocrine Carcinoma (SNEC)
SNECs are rare poorly differentiated carcinoma with neuroendocrine differentiation, characterized by poor prognosis and a high tendency to relapse. According to the new WHO classification, the diagnostic term of neuroendocrine carcinoma can be applied only to poorly differentiated epithelial neuroendocrine neoplasms
[44]. Actually, the SNEC standard of care management is represented by the combination of surgical resection, systemic chemotherapy and radiation therapy. However, the treatment efficacy remains sub-optimal; therefore, the molecular landscape should be explored to increase survival rates by discovering new potential therapeutic targets
[45].
SMARCB1-deficient carcinomas have also been described among SNEC. They represent an aggressive and poor-prognosis subgroup of sinonasal tumors, characterized by INI1 loss mostly due to homozygous SMARCB1 deletion
[46]. SMARCB1/INI-1 (also known as BAF47) is a core subunit of the SWI/SNF complex, and acts as a tumor suppressor by regulating gene transcription and cell proliferation. SWI/SNF tumor suppressor proteins act as antagonists of the polycomb enhancer gene of zeste homolog 2 (EZH2), whereby the EZH2 oncogene is constitutively activated in INI-1-deficient tumors and regulates histone methylation resulting in tumor-suppressor gene silencing, oncogenic transformation, metastasis development, and drug resistance
[47][48][49]. Recently, in a phase II basket trial
[50], a selective inhibitor of EZH2, tazemetostat, showed clinical activity in patients with advanced epithelioid sarcoma with loss of INI-1/SMARCB1.
5. Sinonasal Undifferentiated Carcinomas (SNUC)
Sinonasal Undifferentiated Carcinomas (SNUC) are highly aggressive epithelial tumors with uncertain histogenesis, lacking squamous or glandular differentiation; diagnosis is often challenging and is usually made by exclusion
[51]. Because of their aggressive clinical behaviour, they are usually diagnosed as locally advanced, mainly from dural and/or orbital invasion
[52][53]. Owing to their chemosensitivity the standard approach is based on neoadjuvant chemotherapy followed by either chemoradiation or surgery followed by postoperative radiotherapy
[54]. However, the prognosis remains poor, with a median OS of 22 months
[55].
IDH2 mutations are the most frequent genetic alterations in SNUC. The positivity of IDH2 11C8B1 on IHC in sinonasal carcinomas would be highly predictive of the presence of IDH2 R172S/T mutations in around 70% of cases
[56]. In a study
[57], 88% (14 of 16) of SNUCs had IDH2 R172X mutations, a global methylation phenotype. and an increase in repressive trimethylation of H3K27. These epigenetic alterations severely reduce gene expression, thus preventing cellular differentiation
[58]. In another study
[59], scholars performed an NGS on 11 cases of SNUCs, identifying IDH2 R172X mutations in 55% of cases, R172S, R172T, and R172M. Several concomitant oncogenic alterations, such as PIK3CA, mTOR, SOX2, and SOX9 were also identified. Using both IHC and NGS, other scholars
[45] demonstrated the presence of mutations in IDH2 in SNUCs with 11/36 (31%) cases affected, with R172S and R172G as sequence variants. The most important copy number alterations in the IDH2-mutated tumors were gains on chromosome arm 1q and combined loss of 17p and gain of 17q and loss of 22q. To note, these IDH2 mutations act both as positive prognostic and potentially predictive biomarkers. IDH2 is an interesting potential target for IDH inhibitors
[60].
6. NUT Carcinoma (NC)
NC is a rare and aggressive subtype of poorly differentiated squamous carcinoma, genetically defined by the rearrangement of the NUT (recently renamed NUTM1) gene. In approximately 70% of cases, NUTM1 is involved in a balanced translocation with the BET family gene BRD4 on chromosome 19p13.1 [t (15; 19) (q14; p13.1)], forming the BRD4-NUT fusion oncogene. In the remaining 30% of cases, the NUTM1 gene is fused with BRD3 (25%) on chromosome 9 [t (9; 15) (q34.2; q14)], the histone methyltransferase NSD3 on chromosome 8 [t (8; 15) (p11.23; q14)] or ZNF532 on chromosome 18 [t (15; 18) (q14; q23)]
[61]. The outcome of the patients with NC is often dismal, with a median survival of only 6.7 months
[62]. Unfortunately, all the chemotherapeutic agents tested, including doxorubicin-based regimens, have not shown improved outcomes
[63]. Based on these data, there is a clear need to find new therapeutic strategies for this aggressive cancer. Recently, several studies evaluated the efficacy of the BET inhibitors (BETi), drugs with acetyl-histone mimetics compounds that target BRD4-NUT by competitively inhibiting its binding to chromatin. The first proof of the clinical activity of a BETi in NCwas presented by OTX015/MK-8628
[61][64][65]. Other phase I trials are currently evaluating BETi in NC
[61], like birabresib
[66] and molibresib
[67]. Despite these promising results, not all patients with NC respond to the BETi.
The histone deacetylase inhibitors (HDACi) represent another therapeutic approach for NC. Schwartz et al.
[68] identified that the expression of BRD4-NUT is associated with globally decreased histone acetylation and transcriptional repression, which could be restored by treating the NC with histone deacetylase inhibitors (HDACi). A child was treated with the histone deacetylase inhibitors vorinostat, showing an objective response after 5 weeks of therapy
[69].
7. Teratocarcinosarcoma (TCS)
TCS are aggressive tumors arising primarily in the sinonasal area and anterior cranial base. They are extremely rare, with less than 100 cases ever reported in the literature. They have different features of malignant teratoma, epithelial cells, neural cells, and mesenchymal elements [70].
Rooper et al. found a loss of SMARCA4 expression in 18 cases of 22 sinonasal TCS (82%) and variable positivity for Claudin-4 [71]. Complete loss of SMARCA4 expression in 68% of TCS by IHC, with NGS confirmation of biallelic SMARCA4 inactivation in three cases. These results provide important information about the emerging role of SMARCA4 in SNCs. They particularly suggest that TCS is on a spectrum with SMARCA4-deficient sinonasal carcinomas which show overlapping morphology and molecular characteristics, further readjusting the classification of high-grade sinonasal tumors [70].
8. Sinonasal Lymphoepithelial Carcinoma (SLEC)
Lymphoepithelial carcinoma (LEC) was described for the first time in literature by Schminke
[72] and Regaud
[73] in 1921. Sinonasal lymphoepithelial carcinoma (SLEC) is an extremely rare neoplasm with approximately 40 cases recognized in the literature. It can be considered an SCC morphologically similar to nonkeratinizing nasopharyngeal carcinoma, an undifferentiated subtype, with a reactive intermixed lymphoplasmacytic infiltrate
[74].
There are no data in the literature on altered molecular pathways in this very rare type of sinonasal tumor and there is no evidence of potential molecular targets. However, the neoplastic microenvironment is characterized by an important nonneoplastic lymphoplasmacytic infiltrate cells (including CD8
+T lymphocytes) between and around tumor nests and high expression of PD-1/PD-L1. Even though data from studies on LEC of other head and neck sites show that MSI and loss of expression of the DNA mismatch repair proteins are not common, there is a potential role for immunotherapy in SLEC
[75][76].
9. Immune-Check Point Inhibitors
9.1. Immuno-Markers in Sinonasal Cancers
9.1.1. Deficient Mismatch Repair Proteins (d-MMR) and Microsatellite Instability (MSI)
Only a few studies have addressed the MSI/MMR status in sinonasal carcinomas, with a resulting frequency of MSI for ITACs of 2% [77] and between 2–21% in d-MMR/MSI for SNSCCs [78][79][80].
In a recent study
[79] just three of 131 (2.3%) SNSCC showed d-MMR expression, whereas the other 128 (97.7%) cases showed intact expression of all four MMR proteins. All three d-MMR cases showed concurrent loss of MLH1 and PMS2 expression. The scholars also tried to analyze themutual relationship with other cancer and/or subject characteristics. In particular, these three tumors did not have a synchronous or metachronous inverted sinonasal papilloma component, nor did they display HPV positivity, EGFR mutation, and EGFR copy number gain.
9.1.2. PD-L1 Expression
Riobello et al. [81] analyzed the expression of PD-L1 in 53 SNSCC and 126 ITAC samples. Membranous PD-L1 staining in at least 5% of tumor cells was observed in 34% (18/53) of SNSCC and 17% (22/126) of ITAC. Expression in >50% of tumor cells was frequent in SNSCC (14/53; 26%) in contrast to ITAC (4/126; 3%). Surprisingly, the nuclear expression of PD-L1 was exclusively observed in papillary/colonic-type ITAC; both SNSCC and ITAC with >5% PD-L1 expression had significantly worse disease-free survival, when treated with standard therapeutic options.
9.1.3. Tumor Microenvironment: Cytokines and Tumor Infiltrating Leucocytes (TILs)
In a series of SNCs [82], the scholars analyzed different high-grade tumors. Among them, 16 were SNUCs, four SMARCB1-deficient sinonasal carcinomas, one SMARCA4-deficient carcinoma, five high-grade neuroendocrine carcinomas, one NC, one TCS, and two sinonasal N-ITAC. They focused on the expression of major histocompatibility complex molecules, the leukocyte infiltrates, and chemokines expression, finding that chemokines CXCL8 and CXCL5 were upregulated in high-grade sinonasal carcinomas, influencing leukocyte activation and trafficking, angiogenesis, metastasis, and cancer cells proliferation. On the other hand, several chemokines such as CCL28 and CCL14 were downregulated in SNUCs and high-grade neuroendocrine carcinomas compared with normal tissue. Targeting migration-related chemokines and their receptors in sinonasal tumors might be beneficial for immunotherapy.
9.2. The Efficacy of ICIs in Sinonasal Cancers
Most of the data on the potential efficacy of ICIs in various histological subtypes of sinonasal tumors come from case reports. Interestingly, the responses observed are not strictly related to PD-L1 expression, d-MMR phenotype, MSI or the presence of TILs.
A case report
[83] presented two immunotherapy applications in SNCs and their relationship with other therapeutic strategies. The first patient was a 23-year-old man, treated with pembrolizumab in the second line for relapsing NC. After four cycles the patient underwent a partial response, but then a local progression of the disease was registered. He was offered hypofractionated stereotactic radiotherapy, and pembrolizumab was continued until a local complete response. The other patient was a 29-year-old man with a late local relapse of an SNSCC. Treatment with nivolumab and reirradiation was able to obtain a response, thus supporting the activity of this combination.
A phase II study with pembrolizumab and cetuximab is ongoing to treat R/M HNSCC, including SNSCC (NCT03082534). Another large phase II trial with nivolumab and ipilimumab is ongoing in patients with rare tumors, including SCC and adenocarcinoma of nasal and sinonasal sites (NCT02834013); similarly, in rare cancers, a phase II trial with nivolumab (AcSé trial) is ongoing, including SNCs (NCT03012581).