Association of Oncogenic Mutations in Patients With Advanced Cutaneous Squamous Cell Carcinomas Treated With Cetuximab

Key Points

Question  What is the incidence of HRAS, KRAS, NRAS, BRAF, and EGFR mutations in patients with advanced cutaneous squamous cell carcinomas and the correlation with response to cetuximab?

Findings  In this retrospective study, samples from 31 patients with histologically confirmed advanced cutaneous squamous cell carcinoma treated with cetuximab were analyzed. Only 2 RAS mutated samples were identified. There was no correlation with response to cetuximab treatment.

Meaning  The incidence of RAS, BRAF, and EGFR mutations is very low in cutaneous squamous cell carcinomas. Thus, the search for mutations appears unnecessary before initiating a cetuximab treatment for advanced cutaneous squamous cell carcinomas.

Importance  Cetuximab was recently proposed for advanced cutaneous squamous cell carcinomas (cSCC); however, its efficacy is inconsistent and identification of predictive biomarkers for response is necessary.

Objective  To search for somatic mutations of the HRAS, KRAS, NRAS, BRAF, and EGFR genes in patients with advanced cSCC treated with cetuximab; and to investigate the efficacy and tolerance of cetuximab according to these mutations.

Design, Setting, and Participants  A multicentric and retrospective study of 31 patients (22 men, 9 women) with histologically confirmed advanced cSCC carried out in 1 department of dermatology and 2 departments of medical oncology in France between January 2008 and December 2014. The median age of participants was 86 years (range, 48-96 years).

Interventions  Mutational status was determined by pyrosequencing method, allelic discrimination, or Sanger sequencing. Patients were treated by single-agent cetuximab.

Main Outcomes and Measures  The primary end point was the incidence of somatic mutations of the RAS, BRAF, and EGFR genes and association of cetuximab efficacy with these mutations was investigated by using Fisher test. Secondary end points were the disease control rate (DCR) at week 6, the progression free-survival (PFS), overall survival (OS), and safety profile of cetuximab.

Results  Thirty-one samples of cSCC from 31 patients were analyzed. Only 2 RAS mutated samples (6.5%) were identified. The first harbored a NRAS point mutation (c.35G>A) in codon 12, resulting in a p.G12D substitution. The second sample presented a HRAS point mutation (c.38G>T) in codon 13, resulting in a p.G13V substitution. No mutation of KRAS, BRAF, and EGFR genes at the investigated loci was found. Two patients with NRAS and HRAS mutations showed a partial and complete response to cetuximab, respectively. The mean duration of follow-up was 19 months. At week 6, the disease control rate was 67.8%. The median OS was 13 months and the median PFS was 9 months. All patients could continue cetuximab treatment without dose reduction.

Conclusions and Relevance  Even in elderly patients with advanced cSCC, cetuximab was efficacious and well-tolerated. This suggests that cetuximab is certainly warranted in the treatment of advanced cSCC. However, it is also important to identify tumor specific mutations that may determine response to treatment and prognosis for the disease. We have identified here that the incidence of RAS, BRAF, and EGFR mutations is low in cSCC.

Cutaneous squamous cell carcinoma (cSCC) represents approximately 20% to 30% of nonmelanoma skin cancer. In France, the incidence is estimated to be at least 30 per 100 000 persons per year and has increased by 10% in recent years. The median age at diagnosis of cSCC is 70 years.¹ Cutaneous squamous cell carcinoma has a good prognosis when it is diagnosed at an early stage. Conversely, the 10-year survival rates are below 20% for patients with local lymph node metastases of cSCC and below 10% for those with distant metastases.¹ The treatment of unresectable or metastatic cSCC remains highly challenging. Investigation of systemic therapies for advanced cSCC has been limited to a few prospective trials. Cisplatin-based, combination chemotherapy is the most commonly used treatment, with an overall response rate of up to 80%.^2,3 Other systemic therapies such as 13-cis-retinoic acid, bleomycin, doxorubicin, or immunotherapy (interferon-alpha 2a) have been used.^4-6

Although these agents have shown efficacy in unresectable or metastatic cSCC, the adverse effects of the combination approach the limit in elderly patients.

There is a rationale, from preclinical data, to blockade the programmed cell death protein 1 (PD-1) and/or the PD-1 ligand (PD-L1) pathway in inoperable cSCC. Case reports and case series describe persons of locally advanced or metastatic cSCC to PD-1 inhibitors.^7-9 Before using checkpoints inhibitors, such as PD-1or PD-L1, data from ongoing trials (such as NCT02760498) are necessary and expected.

The epidermal growth factor receptor (EGFR) is a membrane-bound tyrosine kinase receptor that is highly expressed in normal epidermal keratinocytes and in many epithelial tumors, including cSCC.^10,11 Cetuximab, a human-mouse chimeric monoclonal antibody that competitively inhibits the EGFR, is used for the treatment of wild-type KRAS metastatic colorectal cancer (mCRC)¹² and for advanced or metastatic head and neck SCC combined with radiotherapy and with cisplatin/fluorouracil.^13-15 The response rate to cetuximab in unselected mCRC is lower than 30%, underlining the need to identify a predictive biomarker.¹⁶ A lack of response to anti-EGFR therapy has been demonstrated in mCRC harboring KRAS or NRAS mutations. KRAS and NRAS mutations are independently associated with lower progression-free survival (PFS) and overall survival (OS) in mCRC.^17-19 Cetuximab was recently proposed in advanced cSCC when surgery or radiotherapy could not be used.^20,21 However, there is presently no available data concerning oncogenic mutations in cSCC related to response to cetuximab therapy.

Somatic mutations of the HRAS, KRAS, NRAS, BRAF, and EGFR genes in patients with unresectable or metastatic cSCC treated with cetuximab were examined to correlate response to treatment to mutational status of each tumor.

We conducted a retrospective multicentric study in 3 French centers (Department of Dermatology, University Hospital of Nice; Department of Oncology, Centre-Antoine-Lacassagne, Nice; Department of Oncology, Cannes Hospital, Cannes).

The main objective was to evaluate the incidence of HRAS, KRAS, NRAS, BRAF, and EGFR mutations in unresectable or metastatic cSCC and to correlate the cetuximab response to this mutational status. The secondary objectives were to assess (1) the disease control rate (DCR), defined as the percentage of patients who have achieved complete response (CR), partial response (PR), or stable disease at week 6; (2) the objective response rate (ORR) at week 6 of treatment with cetuximab; (3) the PFS, defined as the number of days between the first dose of cetuximab and the earliest day of progression, start of another anticancer treatment, or death; and (4) the OS, defined as the time between the first infusion of chemotherapy and last known patient update.

The eligibility criteria were (1) cSCC confirmed by pathological evaluation; (2) cSCC locally advanced and surgically unresectable or metastatic with documented progression; (3) cSCC treated with single-agent cetuximab between January 2008 and December 2014; and (4) a sample of tumor accessible before cetuximab treatment.

All patients signed written informed consent for the research of mutational status and they were not compensated for their participation. The unresectable criteria was determined by a multidisciplinary committee composed of dermatologists, surgeons and radiation therapists, who evaluated the inability to achieve complete resection and/or surgical impairment of critical cosmetic or functional outcomes. The institutional review board, University Hospital of Nice, determined that ethics committee was not required owing to the retrospective and noninterventional nature of the study. The patients with a recurrent primary cSCC who had prior surgery, radiotherapy, and chemotherapy were eligible.

Exclusion criteria included prior therapy with an agent that targets EGFR, prior radiotherapy within the last 4 weeks, and no available biopsy before cetuximab treatment.

Cetuximab was administered as an intravenous infusion at an initial dose of 400 mg/m², followed by weekly 1-hour infusions of 250 mg/m². Patients received pretreatment with an oral antihistamine. In cases of infusion reactions or dermatologic toxic effects, dose modifications were planned. Cetuximab could be continued as long as the response or the stabilization persisted. After 3 months of weekly treatment, cetuximab was administered every 15 days.

Adverse events and grades were recorded according to National Cancer Institute Criteria, version 4.0.

Tumor response was assessed at week 6, clinically or radiologically according to RECIST criteria.²²

Formalin-fixed and paraffin-embedded tumor (FFPE) samples from surgical resection of primary tumor or recurrent tumor were obtained.

Genomic DNA was extracted from FFPE tissue containing more than 50% neoplastic cells as determined by a pathologist. The DNA extraction was performed using the automated Maxwell 16 Instrument with the Maxwell 16 FFPE Plus LEV DNA Purification Kit (Promega), following the manufacturer’s instructions.

Genotyping of codons 12 and 13 of the exon 2 of KRAS was performed using allelic discrimination. The design of sequences of the TaqMan probes (Applied Biosystems) was provided by Pierre Laurent-Puig, MD, PhD.¹⁹ The 7 hot-spot mutations (c.34G>C [p.G12R]; c.35G>C [p.G12C]; c.34G>A [p.G12S]; c.35G>A [p.G12D]; c.35G>C [p.G12A]; c.35G>T [p.G12V]; c.38G>A [p.G13D]) located in codons 12 and 13 of exon 2 of the KRAS gene were screened. Specific probes for each mutated and nonmutated allele were labeled using the fluorescence reporter dyes FAM and VIC at their 5′-end, respectively.

The genotyping of the exons 3 and 4 of KRAS and the exons 2 and 3 of NRAS gene was performed using the pyrosequencing method. Primers chosen for PCR amplification were designed using Pyromark Assay Design software (version 2.0, Qiagen). The primers used were:

• KRAS 3F 5′- AATTGATGGAGAAACCTGTCTCTT-3′,

• KRAS 3R 5′biotin- TCCTCATGTACTGGTCCCTCATT-3′,

• KRAS 4F 5′- AGGCTCAGGACTTAGCAAGAAGTT-3′

• KRAS 4R 5′biotin-AGTTATGATTTTGCAGAAAACAGA-3′,

• NRAS 2F 5′-CTTGCTGGTGTGAAATGACTG-3′,

• NRAS 2R 5′biotin-TCTGGATTAGCTGGATTGTCAGT-3′,

• NRAS 3F 5′-TTTGTTGGACATACTGGATACAGC-3′, and

• NRAS 3R 5′biotin-CGCAAATGACTTGCTATTATTGA-3′.

The pyrosequencing reaction was performed according to the manufacturer's recommendations using a Q24 Pyromark (Qiagen). The pyrosequencing primers were 5′- CTTGGATATTCTCGACAC-3′ for KRAS exon 3, 5′-AATTCCTTTTATTGAAACAT-3′ for KRAS exon 4, 5′-GTGGTGGTTGGAGCA-3′ for NRAS exon 2 and 5′-GGACATACTGGATACAGCT-3′ for NRAS exon 3.

Genotyping of EGFR exons 18, 19, 20, and 21 was performed by pyrosequencing method with PyroMark Q24 and CE-IVD-marked Therascreen EGFR Pyro Kit (Qiagen) following manufacturer’s instructions.

The genotyping of exons 2 and 3 of HRAS gene was performed by Sanger sequencing using the following primers:

• HRAS 2F 5′-CAGGAGACCCTGTAGGAGGA-3′,

• HRAS 2R 5′- CCTATCCTGGCTGTGTCCTG-3′,

• HRAS 3F 5′- AGAGGCTGGCTGTGTGAACT-3′, and

• HRAS 3R 5′- TGGTGTTGTTGATGGCAAAC-3′.

Categorical values at baseline were expressed as counts and percentages whereas continuous values were expressed as medians. The OS and PFS were estimated by the Kaplan-Meier method. Associations between efficacy outcomes and HRAS, KRAS, NRAS, BRAF, and EGFR mutations were tested by using Fisher exact test.

All statistical analyses were performed with α = 5% in bilateral hypothesis using R statistical software for windows (version 3.2.2, R Foundation).

Among the 31 patients, the median age was 86 years (range, 48-96 years) (Table 1). Seventy one percent of tumors were located on the head and neck. Twelve (39%) patients had a local disease, 13 (42%) lymph node disease, and 6 (19%) distant metastases. In our cohort, most of the patients were pretreated; 9 (29%) patients had previous surgery alone, 1 (3%) patient had previous radiotherapy alone and 14 (45%) patients had surgery plus radiotherapy or chemotherapy. Cetuximab was the first line of treatment for only 7 (23%) patients. Of the 31 patients, 15 (48%) were immunosuppressed (history of steroid use for more than 6 months [n = 1], organ transplant [n = 3], cancer [n = 3], diabetes [n = 8]).

Two mutated samples (6.5%) were identified. One harbored a NRAS point mutation (c.35G>A) in codon 12, resulting in a p.G12D substitution. The second sample presented a HRAS point mutation (c.38G>T) in codon 13, resulting in a p.G13V substitution (Figure 1). We did not find any mutation of KRAS, BRAF, and EGFR genes at the investigated loci. The patient with NRAS mutation showed a PR to cetuximab and had a medical history of breast and cervical cancer. The patient with HRAS mutation had a CR to cetuximab and had a medical history of head and neck SCC. Both patients had poorly differentiated cSCC (Table 2).

The mean duration of follow up was 19 months (range, 1-36 months). At week 6, the DCR was 67.8%. Of the 31 patients 2 (6.5%) achieved a CR, 13 (41.9%) patients a PR, 6 (19.4%) had stable disease, and 10 (32.2%) progressive disease (PD) (Table 2). The objective response rate was 48.5%. For 12 patients with local disease, the DCR was 75.0% (2 CR, 6 PR, and 1 stable disease). For 19 patients with metastatic or lymph node disease, the DCR was 63.1% with no complete response, 7 PR, and 5 stable disease. For 9 patients treated with surgery alone, the ORR and DCR were 44.4% and 55.5%, respectively, with 4 PR, 1 stable disease, and 4 PD. For 14 patients previously treated with surgery plus radiotherapy or chemotherapy, ORR and DCR were 42.9% and 64.3%, respectively, with 6 PR, 3 stable disease, and 5 PD. As expected, for 7 previously untreated patients, the cetuximab response was excellent. Indeed, the ORR and DCR were 71.4% and 100%, respectively with 1 CR, 4 PR, and 2 stable disease.

The median OS was 13 months (range, 1-36 months), and the median PFS was 9 months (range, 0-36 months). At 18 months and 36 months 13 (43%) and 9 (29%) patients were still alive, respectively (Figure 2).

Sixteen (52%) patients developed cetuximab-induced grades 1 to 2 folliculitis. One patient, who received 4 cetuximab infusions followed by CR, had a grade 3 folliculitis. All patients continued cetuximab treatment without dose reduction. No deaths related to treatment occurred.

Our study confirms the very low frequency of somatic mutations of HRAS, KRAS, NRAS, BRAF, and EGFR in cSCC. No EGFR mutations were found in our samples. This is consistent with the April 2016 COSMIC database,²³ which listed alterations of EGFR in only 1% of cSCC cases, and with a low (2.5% to 5%) overall incidence of EGFR mutations in cSCC.^24-26 Notably, we did not screen for overexpression of EGFR in our cohort because immunohistochemical data from the literature have shown a high level of EGFR expression in cSCC without any correlation with cetuximab efficacy.^20,27,28 The variation in the reported incidence of RAS mutations in cSCC ranges from 3% up to 30%.^20,29

Our results are in accordance with another study²⁰ of 36 patients showing a low rate of mutations. Indeed, we identified only 1 case of HRAS mutation (1 of 31, 3%), and 1 case of NRAS mutation (1 of 31, 3%). The HRAS gene is described to be mutated in 8% to 9% of patients with cSCC.^26,30 A phase 2 study,²⁰ assessing cetuximab as first-line single-drug therapy in 36 patients with unresectable cSCC, found only 1 HRAS mutation and thus has the same rate of HRAS mutation as our series. Notably, the mutation in HRAS p.G13V was reported in cSCC treated by BRAF inhibitors.³¹ Our patient was not treated by targeted therapy or chemotherapy. Using next generation sequencing Ion AmpliSeq Cancer Hotspot Panel (v2) in this tumor, we confirmed the HRAS mutation and did not find other mutations. The mutation in NRAS p.G12D identified in 1 of our patient has never been described in cSCC but was reported in melanoma and in colorectal cancer.³² As reported in other studies, no mutation in KRAS or BRAF genes was found in our population.^20,30,33,34

In mCRC, KRAS mutations are strongly associated with lack of cetuximab benefit whereas no correlation between the mutation status and response to treatment could be observed in non–small-cell lung cancer.³⁵ Notably, the only 2 patients in our study harboring NRAS or HRAS mutations showed PR and CR to cetuximab. Overall dependence on EGFR signaling for growth and survival, and subsequent EGFR inhibition effect, may not be the same in cSCC and mCRC. Comparative genomic hybridization analysis of our HRAS mutated tumor showed a gain of chromosome 7 and chromosome 11, chromosomes that contain the EGFR and HRAS genes, respectively. The log ratio of fluorescence intensity of tumor DNA vs control DNA was of 0.24. This value can be estimated to correspond to a duplication of chromosome 7 (trisomy 7) that might explain the efficacy of cetuximab in this patient. Another explanation might rely on the hypothesis that HRAS and NRAS mutations may have no oncogenic potential in cSCC tumors.

These results should be interpreted with caution because they come from only 2 patients and may be purely coincidental (owing to genomic instability in these tumors). Therefore, it is possible that the tumor development may be entirely owing to other driver mutations. The patients identified with NRAS and HRAS mutations had a history of other cancers in CR (breast and/or cervical cancer and HNSCC), which may be confounding and perhaps even suggest that these patients have germline susceptibility to cancer.

In our study of elderly patients with mostly pretreated unresectable and metastatic cSCC, single-agent cetuximab showed a 67.8% DCR. These data are consistent with those from the literature, especially in comparison to 2 prospective phase 2 studies. Indeed, Maubec et al,²⁰ in a clinical trial including 36 treatment-naïve patients with unresectable cSCC treated with cetuximab, reported comparable results, with a DCR of 69% at 6 weeks. Moreover, Foote et al³⁶ published data using panitumumab, another monoclonal anti-EGFR antibody and reported a DCR at 6 weeks of 69%. The safety profile was also good (only 1 grade 3 folliculitis). In the work published by Sadek et al,² using cisplatin, fluorouracil, and bleomycin, the objective response rate was 84% but with a high rate of toxic effects. Therefore, compared with cisplatin chemotherapy, cetuximab is generally well tolerated and may be considered as a therapeutic option, especially in elderly patients. The efficacy observed in this study is consistent with previous data and encourages the use of EGFR inhibitors for unresectable or metastatic cSCC.

A limitation of our research is the sole focus on RAS, BRAF, and EGFR mutations and the lack of full genome sequencing, which is an expensive technique and not suitable for daily practice. Others have examined molecular determinants of response to cetuximab that may be relevant in cSCC. Maubec et al²⁰ reported that combined FcγRIIa-131H/H and/or FcγRIIIa-158V/V polymorphisms were not associated with the response to cetuximab. Several recent studies evaluated recurrent genomic alterations in metastatic cSCC. Li et al³⁷ performed targeting sequencing of 504 genes in 29 patients with metastatic cSCC. They found previously identified recurrently altered genes, such as TP53, CDKN2A, NOTCH1/2, and also a wide spectrum of oncogenic mutations affecting RAS/RTK/PI3K pathways. This suggests potential additional therapeutic targets in cSCC. Moreover, another study³³ of 39 aggressive cSCC cases identified 3 novel candidates’ tumor suppressor genes (NOTCH2, PARD3, and RASA1) with putative links to cancer or differentiation which were identified as possible drivers in cSCC. The mutational spectrum of cSCC seems dominated by tumor suppressor genes, as in head and neck squamous cell carcinomas. These data combined with ours, highlight that whole exome sequencing is urgently needed to address the question of what the recurrent mutations in cSCC are.

This study confirms the efficacy and acceptable tolerance of cetuximab as a single-agent in advanced cSCC in the elderly population. The incidence of HRAS, KRAS, NRAS, BRAF, and EGFR mutations is very low in cSCC and does not predict the response to cetuximab treatment. Therefore, looking for mutations in these genes is not useful before initiating cetuximab therapy for cSCC, but ultimately mutational data are needed to better define the genetic landscape of this disease.

Corresponding Author: Henri Montaudié, MD, Department of Dermatology, University Hospital of Nice, 151 route de Saint Antoine de Ginestiere, Hôpital Archet 2, 06200 Nice, France (montaudie.h@chu-nice.fr).

Published Online: March 4, 2017. doi:10.1001/jamadermatol.2017.0270

Author Contributions: Drs Montaudié and Picard had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Saudes, Poissonnet, Lacour, Passeron, Montaudié.

Acquisition, analysis, or interpretation of data: Picard, Pedeutour, Peyrade, Duranton-Tanneur, Chamorey, Cardot-Leccia, Sudaka, Ettaiche, Benchetrit, Weinbreck, Dadone, Passeron, Montaudié.

Drafting of the manuscript: Picard, Saudes, Duranton-Tanneur, Cardot-Leccia, Ettaiche, Benchetrit.

Critical revision of the manuscript for important intellectual content: Pedeutour, Peyrade, Chamorey, Sudaka, Poissonnet, Weinbreck, Dadone, Lacour, Passeron, Montaudié.

Statistical analysis: Chamorey, Ettaiche.

Obtained funding: Peyrade, Saudes, Lacour.

Administrative, technical, or material support: Pedeutour, Duranton-Tanneur, Cardot-Leccia, Benchetrit, Weinbreck, Montaudié.

Supervision: Peyrade, Lacour, Montaudié.

Conflict of Interest Disclosures: Dr Frederic Peyrade is a Merck board member. No other disclosures are reported.