Systemic Treatments for Recurrent or Metastatic Sinonasal Carcinomas: A Retrospective Multicenter Study

ABSTRACT

Introduction

Sinonasal cancers are rare and heterogeneous and pose a therapeutic challenge at an advanced stage due to the lack of data on appropriate systemic management.

Methods

This retrospective multicenter study analyzed 83 patients with recurrent or metastatic sinonasal carcinomas ineligible for curative treatment, treated in France between 2012 and 2021 (excluding squamous cell carcinomas and adenoid cystic carcinomas). Data were extracted from the REFCOR (French Network of Expertise on Rare Head and Neck Cancers) and Onco‐Occitanie databases, enabling standardized case collection.

Results

The main histological subtypes were intestinal‐type adenocarcinomas (42.2%), undifferentiated sinonasal carcinomas (19.3%), and olfactory neuroblastomas (14.5%). The predominantly male (85.5%) cohort, with a median age of 60 years at diagnosis, mainly received first‐line platinum‐based doublet chemotherapy (platinum‐etoposide, platinum‐5FU). Regardless of histologies and chemotherapy protocols, median overall survival was 13.6 months [10.8; 20.3], with median progression‐free survival of 5.3 months [4.4; 6.1]. Complete and partial response rates were 8.3% and 30.6%, respectively, with overall disease control in 73.6% of cases. For adenocarcinomas, platinum‐5FU and platinum‐taxane regimens were preferred. For sinonasal undifferentiated carcinomas, olfactory neuroblastomas, and neuroendocrine carcinomas, platinum‐etoposide was most frequently used. Targeted therapies and immunotherapy were rarely used.

Conclusion

Improving prognosis requires new strategies, including genomic sequencing and enrollment in clinical trials.

1. Introduction

Cancers of the nasal cavity and sinuses account for 3%–5% of head and neck (HN) cancers, with an annual incidence of 0.56–1 case per 100 000 inhabitants [1, 2, 3]. Patients with this condition average between 50 and 60 years of age, with the disease affecting men more frequently (58%–63% of cases) [2, 4]. The most common histologies are squamous cell carcinomas (SCC, 50%–80%) and adenocarcinomas (AC, 10%–20%), although other subtypes such as olfactory neuroblastomas (ONB, 7%), melanomas (7%), adenoid cystic carcinomas (ACC) (7%), and sinonasal undifferentiated carcinomas (SNUC, 3%) are also observed [1, 4].

The prognosis for sinonasal cancers remains generally poor, with 5‐year survival rates ranging from 35% to 70%, depending on the histological subtype [1, 3, 4, 5]. Standard curative treatment is complete surgical resection followed by adjuvant radiotherapy. Systemic treatments (administered as neoadjuvant or adjuvant therapy) are used to improve oncological outcomes, particularly in terms of recurrence‐free survival and overall survival.

Where tumors are large and resectable or not immediately resectable, induction chemotherapy is often offered, with protocols such as TPF (docetaxel‐platinum‐5FU) for SCC, or platinum‐etoposide for high‐grade ONB and neuroendocrine carcinomas (NEC) [6].

In advanced stages, when local treatment is no longer possible or in cases of metastatic disease, palliative systemic treatments are administered, mainly to improve or maintain patients' quality of life [7]. Targeted therapies, such as neoangiogenesis inhibitors (antivascular endothelial growth factor—anti‐VEGF) and therapies directed against epidermal growth factor receptors (antiepidermal growth factor receptor—anti‐EGFR; antihuman epidermal growth factor receptor 2—anti‐HER2), are used less frequently, as are c‐KIT inhibitors, anti‐BRAF, EZH2 inhibitors, and BET (Bromodomain and Extra‐Terminal domain) inhibitors [8]. Immunotherapies, particularly immune checkpoint inhibitors (such as antiprogrammed cell death protein 1—anti‐PD1), are also considered, especially in cases of SCC [9, 10, 11, 12].

While most research focuses on initial and curative management, treatment options in advanced stages remain poorly defined, with studies often being small, monocentric, and limited to a single histological subtype. The lack of prospective studies has led to recommendations based on extrapolation from neoadjuvant and adjuvant treatments.

This study, conducted by REFCOR (French Network of Expertise on Rare Head and Neck Cancers), aims to describe the different therapeutic strategies and their results in recurrent and metastatic sinonasal cancers that are not amenable to locoregional treatment, considering histology and treatment line. The analysis focused on the different histological subtypes, excluding SCC due to the abundance of data already available, and ACC due to their distinct histopathological, biological, clinical, and prognostic characteristics.

2. Materials and Methods

2.1. Study Design and Participants

This is a multicenter retrospective and observational study conducted in French cancer centers that are members of the REFCOR network. Data come from two main databases: REFCOR, a French national multicenter prospective database containing clinical, histopathological, and evolutionary information on rare HN cancers; the Onco‐Occitanie regional database dedicated to collecting clinical and therapeutic data on cancer patients in the Occitanie region (France). Eligible participants were adults aged 18 years or older diagnosed with primary sinonasal cancer among the following histological subtypes: NUT (Nuclear protein in testis) carcinoma, INI1 (Integrase Interactor 1)‐deficient carcinoma, SMARCA4‐deficient carcinoma, intestinal‐type AC (ITAC), non‐ITAC, ONB, SNUC, NEC, HPV (Human Papillomavirus)‐related multiphenotypic carcinoma, biphenotypic sinonasal sarcoma, all at an initial or secondary metastatic stage, or locally advanced and not amenable to curative locoregional treatment. SCC and ACC subtypes, sarcomas (other than biphenotypic sinonasal) and melanomas were excluded. Patients enrolled were diagnosed between January 2012 and December 2021.

The study was conducted in accordance with the principles of the Declaration of Helsinki, and patient data were anonymized, secured, and registered with the French national Data Protection Authority (CNIL, no. F20221003103154). Patients signed an informed consent form.

2.2. Objectives and Assessments

The main objective of this study was to describe the different systemic treatments used in different centers for recurrent and metastatic sinonasal cancers, in the absence of multimodal curative options. The secondary objectives were to describe overall survival (OS), progression‐free survival (PFS), and tumor response rate according to histological subtype and treatment lines, as well as in the population with meningeal infiltration (MI) on imaging.

The study also aimed to describe the demographic and oncological characteristics of patients, including histological subtype, stage, presence of primary or secondary metastases, and previous treatments received.

OS was defined as the time between starting the treatment line and death; patients who were alive were censored at last follow‐up. PFS was defined as the time between starting the treatment line and progression or death; patients who were alive and progression‐free were censored at the start of the next treatment line or at last follow‐up in case of absence of progression or new treatment line.

Complete response (CR) corresponds to the disappearance of all lesions on imaging. Partial response (PR) is defined as a significant reduction in tumor lesions (according to RECIST criteria). Progressive disease (PD) is characterized by an increase in the size of known lesions or the appearance of new lesions. Stabilization corresponds to a situation where there is neither complete nor partial response, nor progression.

2.3. Procedures

An electronic case report form was created specifically for this study. It was completed by consulting each participant's electronic medical record and was filled in either by investigating physicians or by clinical research associates. Standardized data collection procedures were applied using predefined variables. When information was missing or unclear, the medical records were reweds again by the investigators to clarify discrepancies and ensure the completeness and accuracy of the dataset.

Data collected included demographic information, tumor characteristics, treatments received, and clinical and radiological results.

2.4. Statistical Analysis

Population characteristics were described using descriptive statistics. Quantitative variables were summarized by median, minimum and maximum values, along with the number of missing data; qualitative variables were presented as frequencies and percentages.

Survival rates were estimated using the Kaplan–Meier method, with 95% confidence intervals. Analyses were performed using Stata version 16 (StataCorp LLC, College Station, TX).

3. Results

3.1. Patient Characteristics

The study included 83 patients from 13 French centers (Figure 1).

FIGURE 1.

Patient selection strategy for the study and reasons for exclusion.

Patient characteristics are detailed in Table 1. Median age at primary diagnosis was 60 years old, predominated by males (85.5%). Among risk factors, 48.2% were smokers (current and/or former) and 15.1% had excessive alcohol consumption (defined as alcohol intake exceeding 30 g/day for men and 20 g/day for women). The primary cancer site was mainly ethmoidal (64.6%), followed by nasal (19.5%) and maxillary (11%). The most common histological subtype was ITAC (42.2%), followed by SNUC (19.3%) and ONB (14.5%). The majority of patients (78.3%) had T4 disease and 59% had distant metastases.

TABLE 1.

Patient characteristics.

	Total (N = 83)
Age at primary diagnosis (Y) (n = 83)
Median (interval)	60.0 (19.0; 81.0)
Age at diagnosis (n = 83)
< 65 years	57 (68.7%)
≥ 65 years	26 (31.3%)
Sex (n = 83)
Male	71 (85.5%)
Female	12 (14.5%)
Performance status (n = 79)
0–1	72 (91.1%)
2–3	7 (8.9%)
Smoker (n = 56)
No	29 (51.8%)
Yes	27 (48.2%)
Alcohol consumption (n = 53)
No	45 (84.9%)
Yes	8 (15.1%)
Localization of primary tumor (n = 82)
Ethmoid	53 (64.6%)
Nasal	16 (19.5%)
Maxillary	9 (11.0%)
Sphenoid	2 (2.4%)
Frontal	2 (2.4%)
Histological type (n = 83)
ITAC	35 (42.2%)
SNUC	16 (19.3%)
ONB	12 (14.5%)
NEC	8 (9.6%)
Non‐ITAC	6 (7.2%)
INI1‐deficient carcinoma	3 (3.6%)
NUT carcinoma	1 (1.2%)
HPV‐related multiphenotypic carcinoma	1 (1.2%)
Biphenotypic sinonasal sarcoma	1 (1.2%)
T at diagnosis of incurable disease (n = 83)
Tx	6 (7.2%)
T0–T3	12 (14.5%)
T4a	9 (10.8%)
T4b	56 (67.5%)
N at diagnosis of incurable disease (n = 83)
Nx	16 (19.3%)
N0–N1	54 (65.1%)
N2–N3	13 (15.6%)
M a at diagnosis of incurable disease (n = 83)
Mx	5 (6.0%)
M0	29 (34.9%)
M1	49 (59.0%)

Abbreviations: HPV, human papillomavirus; INI1, integrase interactor 1; ITAC, intestinal‐type adenocarcinoma; NUT, nuclear protein in testis; SNUC, sinonasal undifferentiated carcinoma.

AJCC, American Joint Committee on Cancer (8th edition); M, metastasis; N, node; T, tumor.

Seventy patients (84.3%) had previously received curative treatment, including surgery (n = 55, 78.6%) and/or radiotherapy (n = 67, 97.1%) and/or chemotherapy (n = 32, 46.4%). Platinum‐based protocols were the most commonly used: cisplatin (n = 12, 37.5%), cisplatin‐etoposide (n = 7, 21.9%), TPF (n = 6, 18.7%), carboplatin‐etoposide (n = 4, 12.5%), cisplatin‐5FU (n = 2, 6.25%).

Median time between diagnosis and incurable relapse was 21.5 months [3.9; 206.0]. At diagnosis of incurable disease, 41% (n = 34) of patients had locally advanced disease, 43.4% (n = 36) had both locally advanced and metastatic disease, and 15.7% (n = 13) had metastatic without locoregional disease. The most common metastatic sites were bones (n = 23, 46.9%), lung (n = 20, 40.8%), followed by distant lymph nodes (n = 14, 28.6%), brain (n = 10, 20.4%), and liver (n = 10, 20.4%) (Table A1).

3.2. Treatments

In the incurable setting, median number of treatment lines received was 2 [1.0; 6.0]. All patients received at least one line of chemotherapy, 16.9% (n = 14) received targeted therapy, 13.3% (n = 11) received immunotherapy, and 7.2% (n = 6) participated in a clinical trial.

Details of response rates according to the nonchemotherapy agents used are available in Table A2.

For the first line, the most frequently used protocols were platinum‐based doublets, in particular platinum‐etoposide (n = 26, 31.3%), platinum‐5FU (n = 25, 30.1%), and platinum‐taxane (n = 13, 15.7%). Six patients (7.2%) received targeted therapy concomitant with chemotherapy: anti‐EGFR therapy (n = 4) or anti‐VEGF therapy (n = 2). No patients received immunotherapy; none were enrolled in a clinical trial.

3.3. Survival Data

Median follow‐up was 64.5 months [42.6; not reached]. Median OS for first‐line treatment was 13.6 months [10.8; 20.3] (Figure 2), with median PFS of 5.3 months [4.4; 6.1] (Figure 3). One‐year, 2‐year, and 3‐year OS rates were, respectively, 53.4% [42.0; 63.5], 30.4% [20.4; 40.9], and 15.9% [8.6; 25.2]. Objective response rate was 38.9%, and for 73.6% of patients the disease was controlled (response or stability). However, after a period of time, the majority of patients (n = 72, 92.3%) experienced PD after first‐line treatment.

FIGURE 2.

Overall survival for first‐line treatment all histologies combined.

FIGURE 3.

Progression‐free survival for first‐line treatment all histologies combined.

In second‐ and third‐line treatments, chemotherapy remained the most commonly used protocol (n = 48, 84.2% for second‐line treatment; n = 28, 87.5% for third‐line treatment). Other systemic treatments were less common: immunotherapy (n = 6, 10.5% as second line; none in third line) and targeted therapy (n = 4, 7% as second line; n = 7, 21.9% as third line).

In second‐line treatment, median OS was 10.3 months [5.5; 13.0]. Median PFS was 3.1 months [2.1; 4.2], objective response rate was 19.5%, with a disease control rate of 58.6%. In third line treatment, median OS was 5.3 months [2.8; 10.1], with median PFS of 2.5 months [1.8; 3.1].

Table 2 presents median survival and objective response rates according to lines of treatment.

TABLE 2.

Median survival and objective response rate.

	L1 (n = 83)	L2 (n = 57)	L3 (n = 32)
OS median [95% CI] (months)	13.6 [10.8; 20.3]	10.3 [5.5; 13.0]	5.3 [2.8; 10.1]
PFS median [95% CI] (months)	5.3 [4.4; 6.1]	3.1 [2.1; 4.2]	2.5 [1.8; 3.1]
CR rate	8.3% (n = 6)	6.5% (n = 3)	0.0% (n = 0)
Response rate: CR + PR	38.9% (n = 28)	19.5% (n = 9)	6.7% (n = 2)
Tumor control rate: CR + PR + SD	73.6% (n = 53)	58.6% (n = 27)	30.0% (n = 9)
Primary resistance rate: PD	26.4% (n = 19)	41.3% (n = 19)	66.7% (n = 20)

Abbreviations: CR, complete response; L1, Line 1 treatment; L2, Line 2 treatment; L3, Line 3 treatment; OS, overall survival; PD, progressive disease; PFS, progression‐free survival; PR, partial response; SD, stable disease.

3.4. Analysis According to Histological Subtypes

3.4.1. AC

The AC group (n = 41) included ITAC (n = 35) and non‐ITAC (n = 6). Chemotherapy was the main first‐line treatment (n = 41), combined with targeted therapy in 9.8% of cases. Platinum‐based doublets were the most commonly used treatments: platinum‐5FU (n = 19, 46.3%) and platinum‐taxane (n = 9, 22.0%), with oxaliplatin predominantly used in the platinum‐5FU regimen (n = 12, 63.2%) (see Table A3).

In first‐line treatment, median OS for the AC group was 14.4 months [8.6–19.0], with PFS 4.9 months [3.0–6.1].

For ITAC, patients treated with platinum‐5FU doublets (n = 19, 54.3%) and platinum‐taxane (n = 8, 22.9%) had median OS, respectively, of 11.0 months [5.9–14.4] and 19.0 months [2.1–34.5] for first‐line treatment, and median PFS, respectively, of 5.2 months [2.6–6.1] and 3.0 months [1.5–11.6] (see Table A4).

For non‐ITAC, various protocols (platinum monotherapy, TPF, platinum‐taxane, platinum‐gemcitabine, taxane monotherapy, 5FU) were used with equal frequency (see Table A1).

Details of response rates according to the chemotherapy regimen used are available in Table A5.

3.4.2. SNUC

SNUCs (n = 16) were mainly treated first line with platinum‐based doublets, with protocols such as platinum‐etoposide (n = 7, 43.8%) and platinum‐5FU (n = 5, 31.3%). In first‐line treatment, median OS was 12.1 months [8.7–28.1], with median PFS of 5.9 months [3.9–13.8] (see Table A6).

3.4.3. ONB and Other Subtypes

ONBs (n = 12) were mainly treated first line with platinum‐etoposide (n = 11, 91.7%). NECs (n = 8) received a similar regimen, while rare subtypes (INI1‐deficient carcinomas, NUT carcinomas, HPV‐related multiphenotypic carcinoma, and biphenotypic sarcomas) were treated with platinum doublets (taxane, gemcitabine, or 5FU). No ONBs patient was treated with Lutathera. OS, PFS, and response rates for first‐line treatment are detailed in Table A7.

3.5. Brain and Meningeal Infiltration

Contiguous brain infiltration was present in 42.7% (n = 32) of patients, and imaging evidence of MI was present in 47.2% (n = 34). Among MI patients, 16.1% (n = 5) had symptomatic carcinomatous meningitis. Intrathecal chemotherapy administered in three patients (methotrexate or dacarbazine) did not control the disease, with progression occurring within 4 months in all cases. The OS, PFS, and response rates according to MI status are detailed in Table A8.

4. Discussion

In this study, we reported a multicenter description detailing systemic treatment modalities for locally advanced or metastatic sinonasal carcinomas (excluding SCC and ACC). To our knowledge, this is the largest cohort reported to date.

The majority of patients were men with a median age of 60 years, with tumors mainly located in the ethmoid, nasal, and maxillary sinuses. These characteristics are partially consistent with the literature, although our cohort excludes SCC [1, 2, 4]. Overall prognosis was poor, with median OS of 13.6 months and median PFS of 5.3 months with first‐line chemotherapy, reflecting the advanced stage of the disease. ONBs stood out with longer survival (median OS: 34.4 months), whereas AC, SNUC, and other subtypes had shorter survival, consistent with data available in the literature [1, 3, 4].

Platinum‐based doublet regimens were the most commonly used for first‐line treatment. For AC, platinum‐5FU and platinum‐taxane regimens achieved tumor control rates, respectively, of 86.7% and 44.4%. The platinum‐etoposide regimen was preferred for SNUC, ONB, and NEC, with response rates of 80% for SNUC and 50% for NEC. The choice of these chemotherapy protocols was primarily extrapolated from those used in induction or adjuvant treatment [13, 14, 15, 16, 17]. These results support REFCOR recommendations [6] for the using of these protocols in the context of advanced disease.

Beyond the first line treatment, management was heterogeneous, reflecting a lack of standardization. Targeted therapies and immunotherapy were rarely used, being reserved for specific contexts.

To date, no targeted therapy or immunotherapy has been approved for sinonasal carcinomas, and the data available remain limited. Extrapolating data from studies carried out in HNSCC [9, 18], treatments such as cetuximab, pembrolizumab, and nivolumab [12, 19] are used for SCC of the sinuses.

For other histological subtypes, targeted therapies have shown potential interest, mainly by extrapolation from known indications in digestive neoplasms (based on molecular similarities, particularly in the signaling pathways involved, such as EGFR or VEGF) or by identification of specific biomarkers.

Immunotherapy might also prove interesting, particularly for SNUC and ONB with a favorable tumor microenvironment [20], as explored in the phase II BARON trial (Bintrafusp Alfa in Recurrent/Metastatic Olfactory Neuroblastoma, NCT05012098).

Systematic molecular sequencing is emerging as an essential tool for identifying actionable molecular targets [21]. Different studies [8, 22] have investigated the various potentially actionable molecular pathways and examined data on clinical application. In the case of ITAC, several recurrent genetic alterations have been identified. The main pathways involved include p53 (tumor protein p53), EGFR, PI3K/AKT (phosphoinositide 3‐kinase/protein kinase B), Wnt (wingless‐related integration site), and MAPK (mitogen‐activated protein kinase). In particular, overexpression and/or amplification of the EGFR gene is observed in 38%–55% of cases, particularly in colonic differentiation forms. This molecular profile suggests a possible sensitivity to EGFR inhibitors, already used in other types of cancer such as HNSCC, colorectal, and lung cancers.

As regards SNUCs, among the most notable genetic alterations, mutations of the IDH2 (isocitrate dehydrogenase 2) gene have been identified in 11%–83% of cases depending on the cohort. Treatment with enasidenib, a specific IDH2 inhibitor, is already approved by the Food and Drug Administration (FDA) in treating refractory or relapsed acute myeloid leukemia. In addition, a clinical trial evaluating enasidenib in patients with IDH‐2‐mutated sinonasal tumors is currently ongoing in the United States (NCT06176989). In undifferentiated carcinomas without IDH2 mutations (IDH2 wild‐type), other genetic abnormalities have been identified, such as ARID1A (AT‐rich Interaction Domain 1A) mutations or SMARCB1 (also known as INI1) loss. Both of these elements are involved in the SWI/SNF (switch/sucrose nonfermentable) complex that plays a key role in chromatin remodeling. These abnormalities could make tumor cells sensitive to EZH2 (Enhancer of Zeste Homolog 2) inhibitors, such as tazemetostat.

This study has several limitations. Its retrospective design is associated with missing data and precludes comparative analyses. A selection bias may also be present, as several centers, including some with large patient volumes, were unable to participate and patient inclusion in the REFCOR registry was not systematic. In addition, only patients treated in university hospitals (CHU) and comprehensive cancer centers (CLCC) were included. However, given the rarity of sinonasal cancers and the complexity of their management, care is usually initiated in these expert centers. Finally, the relatively small cohort (83 patients) limited the statistical power of the analyses.

However, these data highlight the importance of care in expert centers and the need to standardize therapeutic approaches.

5. Conclusion

This study highlights differences in prognosis and treatment efficacy according to histological subtypes and underscores the need for prospective studies to fully integrate personalized medicine and immunotherapy strategies into the management of advanced sinonasal carcinomas.

Author Contributions

Study concept: Marie Degrange, Victor Sarradin, Mathilde Morisseau, François Régis Ferrand, and Sarah Bétrian. Study design: Marie Degrange, Victor Sarradin, Mathilde Morisseau, and Sarah Bétrian. Data acquisition: Marie Degrange, Justin Michel, Thibaut Reverdy, Sylvain Morinière, Amaury Daste, Maxime Humbert, Caroline Viala, Pauline Corbaux, Tiffany Darbas, Valentin Favier, Camille Evrard, Diane Evrard, and Delphine Farneti. Quality control of data and algorithms: Marie Degrange, Victor Sarradin, Mathilde Morisseau, and Sarah Bétrian. Data analysis and interpretation: Marie Degrange, Victor Sarradin, and Sarah Bétrian. Statistical analysis: Mathilde Morisseau. Manuscript preparation: Marie Degrange. Manuscript editing: Marie Degrange, Victor Sarradin, Mathilde Morisseau, Justin Michel, Thibaut Reverdy, Sylvain Morinière, Amaury Daste, Maxime Humbert, Caroline Viala, Pauline Corbaux, Tiffany Darbas, François Régis Ferrand, Valentin Favier, Camille Evrard, Diane Evrard, Delphine Farneti, and Sarah Bétrian. Manuscript review: Marie Degrange, Victor Sarradin, Mathilde Morisseau, Justin Michel, Thibaut Reverdy, Sylvain Morinière, Amaury Daste, Maxime Humbert, Caroline Viala, Pauline Corbaux, Tiffany Darbas, François Régis Ferrand, Valentin Favier, Camille Evrard, Diane Evrard, Delphine Farneti, Sarah Bétrian.

Funding

The authors have nothing to report.

Ethics Statement

Approved by the French National Data Protection Authority (CNIL, no. F20221003103154).

Conflicts of Interest

The authors declare no conflicts of interest.

Appendix A.

See Tables A1, A2, A3, A4, A5, A6, A7, A8.

TABLE A1.

Metastatic sites among patients with M1 disease.

If M1, sites of métastases (n = 49)	n (%)
Bone	23 (46.9%)
Lung	20 (40.8%)
Distant lymph nodes	14 (28.6%)
Brain	10 (20.4%)
Liver	10 (20.4%)
Others	6 (12.5%)

TABLE A2.

Nonchemotherapy systemic agents across all treatment lines and histologies.

	Total (N = 83)
Targeted therapy (n = 83)	14 (16.9%)
Cetuximab	8
Bevacizumab	2
Aflibercept	2
Trastuzumab	1
BET inhibitor	1
Immunotherapy (n = 83)	11 (13.3%)
Nivolumab	10
Atezolizumab‐Selicrelumab	1
Clinical trial (n = 83)	6 (7.2%)

TABLE A3.

Description of first‐line treatments administered to the AC group.

	AC (N = 41)	ITAC (N = 35)	Non‐ITAC (N = 6)
Chemotherapy (n = 41)	41 (100.0%)	35 (100.0%)	6 (100.0%)
Type of chemotherapy (n = 41)
Platinum + 5FU	19 (46.3%)	19 (54.3%)	0 (0.0%)
Platinum + Taxane	9 (22.0%)	8 (22.9%)	1 (16.7%)
Platinum	4 (9.8%)	3 (8.6%)	1 (16.7%)
TPF	2 (4.9%)	1 (2.9%)	1 (16.7%)
Taxane	2 (4.9%)	1 (2.9%)	1 (16.7%)
Platinum + Etoposide	1 (2.4%)	1 (2.9%)	0 (0.0%)
5FU or Capecitabine	1 (2.4%)	0 (0.0%)	1 (16.7%)
Gemcitabine	1 (2.4%)	1 (2.9%)	0 (0.0%)
FOLFIRI	1 (2.4%)	1 (2.9%)	0 (0.0%)
Platinum + Gemcitabine	1 (2.4%)	0 (0.0%)	1 (16.7%)
Targeted therapy (n = 41)	4 (9.8%)	4 (11.4%)	0 (0.0%)
Type of targeted therapy (n = 4)
Anti‐EGFR	2 (50.0%)	2 (50.0%)	0 (0.0%)
Other	2 (50.0%)	2 (50.0%)	0 (0.0%)

Abbreviations: AC, adenocarcinoma; EGFR, epidermal growth factor receptor; ITAC, intestinal‐type adenocarcinoma; TPF, docetaxel‐platinum‐5FU.

TABLE A4.

Summary of overall survival, progression‐free survival and response rate for first‐line treatment for the two most frequently used chemotherapies for ITAC.

	ITAC all chemos (n = 35)	Platinum‐5FU (n = 19)	Platinum‐taxane (n = 8)
Median OS [95% CI] (months)	12.2 [6.1; 17.5]	11.0 [5.9; 14.4]	19.0 [2.1; 34.5]
Median PFS [95% CI] (months)	4.9 [3.0; 6.0]	5.2 [2.6; 6.1]	3.0 [1.5; 11.6]
CR rate	0% (n = 0)	0% (n = 0)	0% (n = 0)
Response rate: CR + PR	33.3% (n = 10)	46.7% (n = 7)	0% (n = 0)
Tumor control rate: CR + PR + SD	76.7% (n = 23)	86.7% (n = 13)	50.0% (n = 4)
Primary resistance rate: PD	23.3% (n = 7)	13.3% (n = 2)	50.0% (n = 4)

Abbreviations: CR, complete response; OS, overall survival; PD, progression; PFS, progression‐free survival; PR, partial response; SD, stability.

TABLE A5.

Response rate for first‐line treatments administered to the AC group according to type of chemotherapy.

	AC (N = 36)	ITAC (N = 30)	Non‐ITAC (N = 6)
All types of chemo (n = 36)
PR	12 (33.3%)	10 (33.3%)	2 (33.3%)
SD	15 (41.7%)	13 (43.3%)	2 (33.3%)
PD	9 (25.0%)	7 (23.3%)	2 (33.3%)
Platinum + 5FU (n = 15)
PR	7 (46.7%)	7 (46.7%)	0 (0%)
SD	6 (40.0%)	6 (40.0%)	0 (0%)
PD	2 (13.3%)	2 (13.3%)	0 (0%)
Platinum + Taxane (n = 9)
SD	4 (44.4%)	4 (50.0%)	0 (0%)
PD	5 (55.6%)	4 (50.0%)	1 (100.0%)
Platinum (n = 3)
SD	2 (66.7%)	2 (100.0%)	0 (0%)
PD	1 (33.3%)	0 (0%)	1 (100.0%)
TPF (n = 2)
PR	1 (50.0%)	0 (0%)	1 (100.0%)
PD	1 (50.0%)	1 (100.0%)	0 (0%)
Platinum + Etoposide (n = 1)
PR	1 (100.0%)	1 (100.0%)	0 (0%)
Platinum + Gemcitabine (n = 1)
PR	1 (100.0%)	0 (0%)	1 (100.0%)
Taxane (n = 2)
SD	2 (100.0%)	1 (100.0%)	1 (100.0%)
5FU or Capecitabine (n = 1)
SD	1 (100.0%)	0 (0.0%)	1 (100.0%)
Gemcitabine (n = 1)
PR	1 (100.0%)	1 (100.0%)	0 (0.0%)
FOLFIRI (n = 1)
PR	1 (100.0%)	1 (100.0%)	0 (0.0%)

Abbreviations: AC, adenocarcinoma; ITAC, intestinal‐type adenocarcinoma; PD, progression; PR, partial response; SD, stability; TPF, docetaxel‐platinum‐5FU.

TABLE A6.

Summary of overall survival, progression‐free survival and response rate for first‐line treatment for the two most frequently used chemotherapies for SNUC.

	SNUC all chemos (n = 16)	Platinum‐VP16 (n = 7)	Platinum‐5FU (n = 5)
Median OS [95% CI] (months)	12.1 [8.7; 28.1]	ND	ND
Median PFS [95% CI] (months)	5.9 [3.9; 13.8]	6.5 [2.3; 14.4]	ND
CR rate	23.1% (n = 3)	40.0% (n = 2)	25.0% (n = 1)
Response rate: CR + PR	61.6% (n = 8)	80.0% (n = 4)	50.0% (n = 2)
Tumor control rate: CR + PR + SD	77.0% (n = 10)	100.0% (n = 5)	75.0% (n = 3)
Primary resistance rate: PD	23.1% (n = 3)	0% (n = 0)	25.0% (n = 1)

Abbreviations: CR, complete response; OS, overall survival; PD, progression; PFS, progression‐free survival; PR, partial response; SD, stability; SNUC, sinonasal undifferentiated carcinoma.

TABLE A7.

Summary of overall survival, progression‐free survival and response rate for first‐line treatment for ONB and NEC.

	ONB (n = 12)	NEC (n = 8)
Median OS [95% CI] (months)	34.4 [8.9; 38.4]	9.9 [1.5; 12.0]
Median PFS [95% CI] (months)	8.1 [1.6; 12.5]	4.5 [1.4; 5.7]
CR rate	8.3% (n = 1)	25% (n = 2)
Response rate: CR + PR	25% (n = 3)	50% (n = 4)
Tumor control rate: CR + PR + SD	75% (n = 9)	62.5% (n = 5)
Primary resistance rate: PD	16.7% (n = 2)	25% (n = 2)

Abbreviations: CR, complete response; NEC, neuroendocrine carcinoma; ONB, olfactory neuroblastoma; OS, overall survival; PD, progression; PFS, progression‐free survival; PR, partial response; SD, stability.

TABLE A8.

Summary of overall survival, progression‐free survival and response rate for first‐line according to MI status on imaging.

	No MI (n = 38)	MI (n = 34)
Median OS [95% CI] (months)	14.4 [11.1; 28.3]	10.5 [6.1; 16.6]
Median PFS [95% CI] (months)	5.4 [3.3; 9.0]	5.2 [3.5; 6.5]
CR rate	11.1% (n = 4)	3.8% (n = 1)
Response rate: CR + PR	36.1% (n = 13)	50% (n = 13)
Tumor control rate: CR + PR + SD	69.4% (n = 25)	80.8% (n = 21)
Primary resistance rate: PD	30.6% (n = 11)	19.2% (n = 5)

Abbreviations: CR, complete response; MI, meningeal infiltration; OS, overall survival; PD, progression; PFS, progression‐free survival; PR, partial response; SD, stability.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.