Nasoethmoidal adenocarcinomas are rare tumors with an incidence of 0.19/100 000 in Western Europe [1]. Risk factors for development of these tumors are well defined [2]. For the past 30 years, the standard treatment reported in the literature has been a combination of surgery and post-operative radiotherapy [3-24]. Five-year survival rate is currently around 60% and has not improved since the nineties [3-24].

Sinonasal Intestinal Type Adeno Carcinoma (ITAC) is histologically similar to colorectal adenocarcinoma [25]. The genetic characterization of colorectal adenocarcinoma has enabled new targeted treatments. Since 2004, numerous studies have demonstrated the efficacy of monoclonal antibodies (mAbs) targeting the Epidermal Growth Factor Receptor (EGFR) in patients with metastatic colorectal adenocarcinoma [26-29]. The efficacy of these recent targeted therapies in colorectal adenocarcinoma is restricted to patients with no somatic mutation of KRAS [30].

The aim of this study was to investigate molecular similarities between ITAC and colorectal adenocarcinoma in order to offer new targeted therapies for ITAC.

Forty-one cases of patients with ethmoidal carcinoma occurring between 1983 and 2007 were retrieved from the database at Rouen University Hospital. Of the 41 patients included, there were 37 men and four women. Mean age at presentation was 63.6 years (range: 40.7-86.4 years). Occupational exposure was documented for 32 patients, with 31 cases of wood exposure and one of leather exposure. Distribution was 1 T1, 18 T2, 6 T3 and 16 T4 according to the 2002 TNM classification of the American Joint Committee on Cancer. There were no patients with metastatic neck lymph node (N0) or distant visceral metastasis.

To ensure uniform tumor classification, all patient biopsies and/or surgical specimens were reassigned according to the new 2005 World Health Organization classification for ethmoidal adenocarcinoma.

Thirty-eight out of the 41 tumor specimens had sufficient tissue for immunohistochemistry and mutation analysis.

Immunohistochemical study was performed on 5 micron deparaffinized tissue sections from formalin-fixed and paraffinembedded (FFPE) tissue with antibodies directed against CDX2 (rabbit monoclonal antibody (MoAb), EPR2764Y, Cell Marque®), CK20 (mouse MoAbKs20.8, Dako®), CK7 (mouse MoAb OV-TL 12/30, Dako®) and EGFR (mouse MoAb, CONFIRM anti-EGFR 3C6, Ventana®). Immunolabeling was revealed with Ultraview DAB revelation kit (Ventana Medical Systems®) using automated BenchMark® XT PLC (Ventana Medical Systems®).

KRAS c.34 G, c.35G, c.37G, c.38G, and EGFR c.2155G, c.2156G, c.2369C, c.2573T, and c.2582T mutation hot spots were assessed with a single multiplex primer extension assay with SNaPshot® technology. Briefly, tumor genomic DNA was isolated using the RecoverAll™ Total Nucleic Acid Isolation Kit for FFPE Tissues (Applied Biosystems®) as recommended by the manufacturer. KRAS exon 2 and EGFR exons 18, 20 and 21 were then simultaneously PCR amplified using specific primers, and after purification of PCR products, hot spot mutation specific primers (listed in Table 1) were extended with fluorescently labeled dideoxynucleotides using ABI PRISM SNaPshot® Multiplex Kit. Extended primers were finally separated in an automated sequencer (ABI PRISM 3130xl Genetic analyzer) and data analysis was performed using GeneMapper® software version 4.0 (Applied Biosystems®).

Table 1: List of SNaPshot® primers.

Deletions, insertions, and duplications in EGFR exons 19 and 20 were detected using a fragment analysis assay. In short, EGFR exons 19 and 20 were PCR amplified in multiplex using 5’ fluorescently labeled specific primers to allow PCR product separation in an automated sequencer (ABI PRISM 3130xl Genetic analyzer). Data analysis was then performed using GeneMapper® software version 4.0 (Applied Biosystems®).

Three out of the 38 tumors analyzed were non-intestinal ethmoidal adenocarcinomas and the remaining 35 tumors were classified as ITAC (33 men and two women) and wood exposure was found in 29 (85%) of these patients.

Non-intestinal adenocarcinomas

CDX 2 immunohistochemistry was negative for all three tumors. EGFR positivity was low (1+) for all three tumors. All three cases were CK20 negative and two of the three cases disclosed CK7 immunostaining. No EGFR or KRAS mutation was observed.

Intestinal type adenocarcinomas

Immunohistochemistry showed CDX2 positivity in 31 of the 35 ITAC tumors (89%) (Figure 1). There was no difference in CDX2 expression since all positive ITAC disclosed the same high level of positivity. Of the four ITAC samples with no CDX2 expression, one was decalcified introducing artefact in the immunohitochemistry technique.

Figure 1: strong and diffuse CDX 2 immunostaining in ethmoidal intestinal type adenocarcinoma (Obj X20).

EGFR was positive in 29 of the 35 ITAC tumors (83%). Five cases disclosed no EGFR expression and one case was non-evaluable. EGFR expression was 1+, 2+, and 3+ respectively in 19 (56%), 7 (21%), and 3 (9%) cases (Figure 2). Thirty-three out of the 35 ITAC cases expressed CK20 (Figure 3). CK7 immunostaining was positive in 15 out of 35 ITAC cases.

Figure 2: Strong (3+) and focal EGFR immunostaining in ethmoidal intestinal type adenocarcinoma (Obj X20).

Figure 3: CK 20 expression.

EGFR genotyping was successful in all patients and no mutation was found. The same investigation for KRAS revealed mutations in five patients (14%): 1 C34G, 3 C35G and 1 C38G (Figure 4). All five patients with mutations had undergone wood exposure.

Figure 4: 35G>A; p.G12D KRAS mutation in ethmoidal intestinal type adenocarcinoma shown by SNaPshot® Technique.

Ethmoidal adenocarcinomas are rare lesions with low incidence [2] estimated at 0.19/100 000 in Western Europe [31]. Our study population with ethmoidal carcinomas was similar to that investigated in the most important series reported in the literature [2]. In most cases, our patients were 63-year-old men who had been exposed to toxic wood. At first diagnosis, the majority of lesions had large extensions (40% T4).

The new 2005 WHO classification separates ethmoidal adenocarcinomas into non-intestinal type and intestinal type adenocarcinomas (ITAC) due to the latter’s histologic similarity with colorectal adenocarcinomas [32]. ITAC are more common in woodworkers [1], and this characteristic was present in our series.

CDX 2

As recently reported [33], CDX 2 is a highly specific and sensitive marker of ITAC. In our study, this marker was present in 89% of ITAC and was missing in all non-intestinal ethmoidal adenocarcinomas. Therefore, CDX 2 could be a useful tool for the diagnosis of ITAC.

Cytokeratin expression: Cytokeratin 20 is a strong marker of colorectal origin and is routinely used to discriminate metastatic tumors [34]. In contrast, CK7 is expressed in gynecological, lung, breast and genitourinary tract tissues, but can also be expressed in rectal adenocarcinoma [35]. In our series, 94% of ITAC disclosed CK20 positivity and 42% were CK7 positive like some colorectal carcinomas. The possibility of the double CK7-CK20 positivity of ITAC favors use of CDX2, which appears to be more specific.

EGFR

According to the literature, EGFR is expressed in 60 to 80% of colorectal adenocarcinomas [26,30,36-41]. In our study, EGFR expression, evaluated by immunohistochemistry, was present in 83% of ITAC. This value is similar to that observed in colorectal adenocarcinomas. For Franchi et al. [42], EGFR expression by immunohistochemistry was found in only one third of ITAC. Nevertheless, patients with intestinal type compared to non-intestinal type adenocarcinomas are more exposed to toxic wood. Exposure to wood was present in 75% of our study population, whereas only 25% of the Franchi et al. population were woodworkers. However, in the latter study, EGFR expression increased to 43% in those patients with wood exposure. The differences reported regarding intestinal type compared to non-intestinal type adenocarcinomas could be explained by exposure to toxic wood. Thus, based on immunohistochemistry EGFR expression, we suggest that similarities might exist between patients previously exposed to toxic wood products, resulting in the development of intestinal type adenocarcinomas.

In all our cases, tumor genotyping disclosed wild type EGFR gene status. To the best of our knowledge, we report this data on ITAC for the first time. In colorectal adenocarcinoma, EGFR mutations are rare and incidence is below 5 % [33].

If KRAS mutations (in codon 12 or 13) in colorectal adenocarcinoma are well known occurring in 40% of cases [43], there is a paucity of literature (Table 2) for KRAS mutations in ITAC [44-48]. The findings in the literature are diverse, probably due to the differences in the populations studied. In most recent articles KRAS mutation rate is evaluated at around 15%. This incidence of mutation is lower than in colorectal adenocarcinoma, which is around 40%.

Table 2: distribution of KRAS mutations in ethmoidal intestinal type adenocarcinoma according to the literature.

The histopathologic similarity between ITAC and colorectal adenocarcinoma is well established. We report here, phenotype (CDX2) and genotype (EGFR and KRAS) similarities between these two types of adenocarcinoma.

Given the efficacy of anti-EGFR antibodies in the treatment of patients with wild type KRAS metastatic colorectal adenocarcinoma, we could expect similar efficiency in ITAC. Therefore, anti-EGFR targeted therapy could be proposed in order to enhance both overall survival and progression free survival in patients with wild type KRAS metastatic ITAC. New therapeutic trials are warranted to validate the strength of anti-EGFR targeted therapy in ITAC.

The authors are grateful to Nikki-Sabourin-Gibbs, Rouen University Hospital, for editing the manuscript.