Gut Online First, published on July 10, 2014 as 10.1136/gutjnl-2014-307234

GI cancer

ORIGINAL ARTICLE

FCGR polymorphisms and cetuximab efficacy in chemorefractory metastatic colorectal cancer: an international consortium study Ravit Geva,1 Loredana Vecchione,2 Konstantinos T Kalogeras,3 Benny Vittrup Jensen,4 Heinz-Josef Lenz,5 Takayuki Yoshino,6 David Paez,7 Clara Montagut,8 John Souglakos,9 Federico Cappuzzo,10 Andrés Cervantes,11 Milo Frattini,12 George Fountzilas,3 Julia S Johansen,4 Estrid Vilma Høgdall,13 Wu Zhang,5 Dongyun Yang,5 Kentaro Yamazaki,14 Tomohiro Nishina,15 Demetris Papamichael,16 Bruno Vincenzi,17 Teresa Macarulla,18 Fotios Loupakis,19 Jef De Schutter,2 Karen Lise Garm Spindler,20 Per Pfeiffer,21 Fortunato Ciardiello,22 Hubert Piessevaux,23 Sabine Tejpar2 ▸ Additional material is published online only. To view please visit the journal online (http://dx.doi.org/10.1136/ gutjnl-2014-307234). For numbered affiliations see end of article. Correspondence to Professor Sabine Tejpar, Laboratory of Molecular Digestive Oncology, University Hospital Gasthuisberg, Herestraat 49, Leuven B—3000, Belgium; [email protected] RG and LV contributed equally. Received 14 March 2014 Revised 21 June 2014 Accepted 24 June 2014

To cite: Geva R, Vecchione L, Kalogeras KT, et al. Gut Published Online First: [ please include Day Month Year] doi:10.1136/ gutjnl-2014-307234

ABSTRACT Objective We aimed to better clarify the role of germline variants of the FCG2 receptor, FCGR2A-H131R and FCGR3A-V158F, on the therapeutic efficacy of cetuximab in metastatic colorectal cancer (mCRC). A large cohort with sufficient statistical power was assembled. Design To show a HR advantage of 0.6 in progressionfree survival (PFS) for FCGR2A-HH versus the rest and FCGR3A-VV versus the rest, with an 80% power, 80 Kirsten Rat Sarcoma Viral Oncogene Homolog (KRAS) wild-type (KRAS-WT) and 52 KRAS-WT patients are required, respectively. This leads to a total sample size of 952 and 619 patients, respectively. Samples were collected from 1123 mCRC patients from 15 European centres treated with cetuximab alone or in combination with chemotherapy. Fc gamma receptor (FCGR) status was centrally genotyped. Two additional externally genotyped series were included. Results Incidences of FCGR2A-HH and FCGR3A-VV in KRAS-WT patients were 220/660 (33%) and 109/676 (16.1%) respectively. There was no difference in median PFS (mPFS) for KRAS-WT patients with FCGR2A-HH (22.0 weeks; 95% CI18.8 to 25.2) versus non-HH (22.0 weeks; 95% CI 19.4 to 24.6) or for FCGR3A-VV (16.4 weeks; 95% CI 13.0 to 19.8) versus non-VV (23 weeks; 95% CI 21.1 to 24.9) ( p=0.06). Median overall survival, response rate and disease control rate assessments showed no benefit for either HH or VV. Conclusions No differences in mPFS were found between the FCGR polymorphisms HH and the others and VV versus the others in KRAS-WT mCRC patients refractory to irinotecan, oxaliplatin and 5-fluorouracil treated with cetuximab. We cannot confirm the effects of other IgG1 antibodies, which may be weaker than previously suggested. Other markers may be needed to study the actual host antibody response to cetuximab.

INTRODUCTION Cetuximab, a chimeric immunoglobulin 1 (IgG1) monoclonal antibody (moAb) that targets the

Significance of this study What is already known on this subject?

▸ Germline variants in the FCG receptor (FCGR), FCGR2A-H131R and FCGR3A-V158F, have been suggested to influence the efficacy of rituximab and trastuzumab in lymphoma and breast cancer. ▸ The role of FCGR polymorphisms in chemorefractory metastatic colorectal cancer patients treated with cetuximab is still unclear. Several studies have been published but with conflicting results. ▸ The main reason for the conflicting and confounding results were the small sample sizes, the non-cross-validation between techniques and the lack of correction for multiple testing.

What are the new findings?

▸ No differences in median progression-free survival (mPFS) for chemorefractory KRAS wild-type (KRAS-WT) metastatic colorectal cancer (mCRC) patients treated with cetuximab were observed for FCGR2A-HH versus non-HH or for FCGR3A-VV versus non-VV. ▸ No benefit in terms of response rate (RR), disease control rate (DCR), and median overall survival (mOS) was observed for FCGR2A-HH versus non-HH or for FCGR3A-VV versus non-VV. ▸ An advantage in DCR and mOS for the FCGR3A-FF subgroup compared to the others was observed in KRAS-mutant patients, but there was no advantage in RR or in mPFS. The inconsistencies in the comparisons of response, progression-free survival and OS for this genotype in KRAS-mutant patients indicate that these findings are exploratory at best, without a current clinical implication.

Geva R, et al.Article Gut 2014;0:1–8. doi:10.1136/gutjnl-2014-307234 1 Copyright author (or their employer) 2014. Produced by BMJ Publishing Group Ltd (& BSG) under licence.

GI cancer

Significance of this study How might it impact on clinical practice in the foreseeable future?

▸ FCGR polymorphisms do not predict response to cetuximab in chemorefractory KRAS-WT mCRC patients. ▸ The antibody-dependent cell-mediated cytotoxicity (ADCC) does not seem to play a role in cetuximab response. ▸ Other markers may be needed to better study the effect of the immune system to cetuximab activity and response in chemorefractory KRAS-WT mCRC.

extracellular domain of the epidermal growth factor receptor (EGFR), has shown efficacy in metastatic colorectal cancer (mCRC) patients in combination with chemotherapy as first and second lines,1–4 and in monotherapy as third line.5 Although initially developed for all patients with mCRC, it became clear from the analysis of several studies in 2008 that the benefit of this drug is restricted to the KRAS wild-type (WT) population.6 Although the mechanisms of action of cetuximab are not completely known, it is likely that the competition in binding the EGFR with some of its natural ligands can be responsible for the inhibition of cell proliferation, angiogenesis and metastasis.7 Antibody-dependent cellular cytotoxicity (ADCC) has been identified as an important mechanism of tumour cell death. It is mediated by the bifunctional structure of IgG molecules. As a chimeric IgG1 monoclonal antibody (moAb), cetuximab has an antigen-binding fragment (Fab) that engages the tumour cell antigen and a crystalline fragment (Fc) that binds a Fc gamma receptor (FCGR) on an effector cell, such as natural killer (NK) cells, monocytes, or macrophages. The ADCC is initiated when the Fab and the Fc portions of the moAb bind the tumour cell antigen and the FCGR. FCGRs are composed of three distinct classes: FCGR1, FCGR2 and FCGR3. The second two can be further divided into FCGR2A/FCGR2B and FCGR3A/ FCGR3B. FCGR1 has high affinity for IgG and can bind to monomeric IgG, FCGR2 and FCGR3, which have weaker affinity for the monomeric form, need multimeric immuno complexes to effectively interact.8 Single nucleotide polymorphisms (SNP) that affect affinity to human IgG have been demonstrated for FCGR2A and FCGR3A, with a histidine(H)/arginine(R) polymorphism at position 131 for FCGR2A and a valine(V)/phenylalanine(F) polymorphism for FCGR3A at position 158.9 The strongest evidence for a role of FCGR polymorphisms in the efficacy of therapeutic IgG moAbs comes from studies on follicular lymphoma patients treated with rituximab in combination with the cyclophosphamide, hydroxydaunorubicin, oncovin and prednisone (CHOP) regimen as first-line treatment. Several papers suggested an advantage for FCGR3A VV over the other SNPs.10–13 Other publications suggested a similar effect from FCGR3A VV or FCGR2A HH with trastuzumab in breast cancer.14 15 However, these results are less consistent, and recent data dispute the validity of this effect.16 17 Following these conflicting results, FCGR has not been adopted as a predictive biomarker by the medical community for the treatment of breast and lymphoma. Several studies attempted to explore the role of FCGR SNPs in the treatment of cetuximab in mCRC. The studies performed are diverse in treatment lines and selection of explored genotypes, thus resulting in conflicting 2

and confusing associations. Most were small with an insufficient statistical power, did not correct for multiple testing and did not consider the KRAS status. Moreover, the use of different genotyping techniques without cross-validation contributed to these confounding results. We aimed to follow the associations found in lymphoma and breast cancer for FCGR3A-VV and FCGR2A-HH and test their relevance in mCRC patients treated with cetuximab. A large database with sufficient statistical power was generated using the data from a European consortium18 to which additional cohorts from Japan and the USA were added. Our large sample size together with the cross-validation of techniques and the multiple testing contributed to the obtaining of conclusive results.

MATERIALS AND METHODS Patients The current study investigated chemorefractory mCRC patients treated with cetuximab alone or in combination with chemotherapy. The primary endpoint was median progression-free survival (mPFS) in the FCGR3A-VV KRAS WT subgroup versus the non-VV subgroup, and the FCGR2A-HH subgroup versus the non-HH subgroup. Sample size calculation assumptions included the mPFS in KRAS-WT patients to be 24 weeks and the incidence of KRAS mutation to be 40%. Based on Zhang et al,19 the incidence of a favourable FCGR2A-HH genotype was expected to be 14%, and that of FCGR3A-VV genotype to be 14% as well. To obtain an improvement in mPFS in the favourable subgroups versus the rest, with a HR of 0.6, a total of 80 KRAS-WT FCGR2A-HH and 52 KRAS-WT FCGR3A-VV patients were needed. This yielded a calculated total of 952 and 619 patients, respectively. Fifteen European centres (a European consortium) and two additional cohorts from Japan and the USA contributed to the final analysis. Clinical data on treatment were collected from a total of 1123 patients (899 belonging to the consortium, 91 from Japan and 133 from the USA), including 645 men (57.4%) and 476 women (42.4%) with an age range of 22–85 years. The patients’ demographics are described in table 1. The European consortium was comprised of 15 European centres that were contacted to provide predefined clinical parameters, normal tissue samples, or DNA extracted from blood, and at least two slides of formalin-fixed paraffin embedded (FFPE) tumour tissue, and one consecutive haematoxylin-eosinstained slide. Nine of these centres had already participated in a previous European consortium and six new members were added.20 The specimens of two additional cohorts (91 patients from Japan and 133 patients from the USA) were analysed separately, but their clinical data were combined. Tumour response was evaluated by the investigators according to the Response Evaluation Criteria in Solid Tumours21 or WHO criteria, and classified as complete response (CR), partial response (PR), stable disease (SD) or progressive disease. The response rate (RR) was classified as CR plus PR. Disease control rate (DCR) was classified as RR plus SD. Progression-free survival (PFS) was reported by the centres in weeks, and measured as the time from the first day of cetuximab administration until the observation of radiological or clinical progression, or death from any cause. This study was conducted after obtaining approval by the local ethics review boards.

DNA extraction Tumour and normal DNA were extracted from FFPE slides after pathological review and macrodissection using the QIAamp DNA FFPE Tissue Kit. Germline DNA was extracted from Geva R, et al. Gut 2014;0:1–8. doi:10.1136/gutjnl-2014-307234

GI cancer Table 1

Table 1

Patients characteristics

Characteristics Age, years Median Range Sex Female Male Primary tumour localisation Right colon Left colon Rectum Missing Number of previous lines of chemotherapy 0 1 2 3 4 5 6 7 8 Previous chemotherapy Irinotecan Missing Oxaliplatin Missing Fluoropyrimidine Missing KRAS status WT Mutant Missing Treatment characteristics Irinotecan-based cetuximab regimen Missing Oxaliplatin-based cetuximab regimen Missing Cetuximab monotherapy or with 5-FU Missing Response characteristics and outcome Whole population Best response CR PR SD PD Missing mPFS mOS KRAS WT Best response CR PR SD PD

Number of patients

Per cent

Characteristics Missing mPFS

60 22–85 476 645

42.4 57.4

127 283 214 499

11.3 25.2 19.1 44.4

3 154 652 177 87 30 13 2 5

0.3 13.7 58.1 15.8 7.7 2.7 1.2 0.2 0.4

1106 7 894 54 1068 51

98.5 0.6 79.6 4.8 95.1 4.5

740 355 28

68.9 31.6 2.5

904 4 50 4 174 4

80.5 0.4 4.5 0.4 15.5 0.4

8 221 457 389 48 17.4 weeks

0.7 19.7 40.7 34.6 4.3 (95% CI 16.6 to 18.3 weeks) (95% CI 40.8 to 47.1 weeks)

44 weeks

8 203 293 208

Continued

28 21.8 weeks

mOS

53 weeks

KRAS MUT Best response PR SD PD Missing mPFS

13 154 173 15 12 weeks

mOS

31.2 weeks

Per cent 3.8 (95% CI 19.9 to 23.7 weeks) (95% CI 48.3 to 57.6 weeks)

3.7 43.4 48.7 4.2 (95% CI 10.9 to 13 weeks) (95% CI 27.9 to 34.4 weeks)

CR, complete response; mOS, median overall survival; mPFS, median progression free survival; PD, progressive disease; PR, partial response; SD, stable disease.

blood samples of 349 patients by the DNeasy Blood and Tissue Kit (Qiagen, Belgium).

KRAS mutational analysis The Sequenom platform was used to assess KRAS mutations. KRAS mutations in exons 12 (G12D, G12A, G12V, G12S, G12R and G12C), 13 (G13D), 59 (A59T), and 61 (Q61H, Q61P, Q61R and Q61L) were analysed as previously described.18

FCGR2A-H131R and FCGR3A-V158F genotyping and technical validation

1.1 27.4 39.6 28.1 Continued

Geva R, et al. Gut 2014;0:1–8. doi:10.1136/gutjnl-2014-307234

Number of patients

The mutations, histidine/arginine-131 of CD32A (FCGR2A-131 H/R; rs1801274) and valine/phenylalanine-158 of CD16A (FCGR3A-158V/F; rs396991), are also known as FCGR2A 535A>G and FCGR3A 818A>C polymorphisms. To avoid confounding nomenclature, online supplementary figures S1 and S2 include illustrations from the Mutation Discovery database that show the sequence and flanking regions of the two SNPs that we analysed. An allele-specific qPCR (Taqman) was performed by means of specific predesigned probes for the FCGR2A-131H/R and FCGR3A-158V/F polymorphisms (part No. C__9077561_20 and C__25815666_10, respectively, Applied Biosystems, Foster City, California, USA) and run on 7500 real-time PCR system (Applied Biosystems). The 25 mL reaction solution was incubated in a 96-well plate at 95°C for 10 min, followed by 50 cycles of 92°C for 15 s and 60°C for 1 min. An internal validation of the assays was performed by directly sequencing a set of 12 samples using primers (available upon request) that were able to detect the SNPs of interest and distinguish them from pseudogenes (see online supplementary figures S3 and S4). All the Taqman results were confirmed without exception. External validation consisted of sending 18 samples to an independent centre (Leids Universitair Medisch Centrum (LUMC), The Netherlands)22 to confirm our results. Genotyping was performed on a Taqman 7500 (Applied Biosystems) with the same predesigned assays used in our analysis. A 100% compatibility was found between our centre and the LUMC, thus providing an independent technical replicate. The samples from Japan and those from the USC Norris

3

GI cancer Table 2

Distribution of FCGR polymorphisms Group

Distribution of FCGR polymorphisms within the groups FCGR 2A status HR Count % within FCGR 2A % of Total HH Count % within FCGR 2A % of Total RR Count % within FCGR 2A % of Total Total Count % within FCGR 2A % of Total FCGR 3A status VF Count % within FCGR 3A % of Total VV Count % within FCGR 3A % of Total FF Count % within FCGR 3A % of Total Total Count % within FCGR 3A % within group % of Total

Europe

Japan

USA

Total

status

397 87.4 39.6

28 6.2 2.8

29 6.4 2.9

454 100.0 45.3

status

226 70.2 22.6

61 18.9 6.1

35 10.9 3.5

322 100.0 32.1

status

188 83.2 18.8

2 0.9 0.2

36 15.9 3.6

226 100.0 22.6

status

811 80.9 80.9

91 9.1 9.1

100 10.0 10.0

1002 100.0 100.0

status

372 79.8 36.3

36 7.7 3.5

58 12.4 5.7

466 100.0 45.5

status

126 75.4 12.3

4 2.4 0.4

37 22.2 3.6

167 100.0 16.3

status

308 78.8 30.1

51 13.0 5.0

32 8.2 3.1

391 100.0 38.2

806 78.7 100.0 78.7

91 8.9 100.0 8.9

127 12.4 100.0 12.4

1024 100.0 100.0 100.0

status

Distribution of FCGR polymorphisms according to the KRAS status Count FCGR 2A HR HH RR VF FCGR 3A VV FF

KRAS WT

KRAS M

n

Per cent

n

Per cent

291 220 149 296

44.1 33.3 22.6 43.8

158 94 75 163

48.3 28.7 22.9 48.9

109 271

16.1 40.1

57 113

17.1 33.9

FCGR, Fc gamma receptor; RR, response rate; WT, wild-type.

Comprehensive Cancer Center were locally genotyped and were not available as DNA to us. The Japanese group used a method that combines the multiplex PCR and sequence-specific oligonucleotide probes with microsphere-based suspension array technology (Luminex, Austin, Texas, USA).23–25 The samples from the USA were analysed for FCGR2A H131R polymorphism by PCR followed by allele-specific restriction enzyme digestion,24 and for the FCGR3A V158F polymorphism by an allele-specific PCR reaction.25 4

The incidence of the genotypes in the European consortium as well as those in the Japanese and US series is reported in table 2.

Statistical analysis The median PFS was our primary end point. RR, DCR and median overall survival (mOS) were our secondary outcome variables. Additional genotype grouping analyses were performed but regarded as exploratory. Geva R, et al. Gut 2014;0:1–8. doi:10.1136/gutjnl-2014-307234

GI cancer Table 3 Progression-free survival (PFS) and OS within Fc gamma receptor (FCGR) polymorphisms Total

FCGR 2A HH Non-HH

KRAS WT

mPFS (95% CI)

mOS (95% CI)

mPFS (95% CI)

mOS (95% CI)

18 (15.8 to 20.1) 17.4 (16.2 to 18.6) NS

46.1 (40.9 to 51.3) 43.3 (38.7 to 47.8) NS

22 (18.7 to 25.2) 22 (19.4 to 24.5) NS

55.4 (47.2 to 63.6) 53.4 (47.4 to 59.3) NS

39.6 (32.1 to 47) 44.1 (40 to 48.2) NS

16.4 (13 to 19.7) 23 (21.1 to 24.8) NS

50 (40.3 to 59.8) 54 (49 to 58.9) NS

FCGR 3A VV Non-VV

14.1 (11.6 to 16.5) 18 (16.8 to 19.1) p=0.039 KRAS (codon12,13,59,61,146)=WT

Survival values reported in weeks for the different FCGR polymorphisms. mOS, median overall survival; mPFS, median progression-free survival; NS, non-significant; WT, wild-type.

A p value below 0.05 was considered to indicate statistical significance. The PFS and OS in the various subgroups were summarised with the use of Kaplan–Meier curves, and Kaplan– Meier estimates of OS and PFS probabilities are shown where applicable. The difference between these groups was compared with the use of the log rank test, with the HR and its 95% CI from a Cox regression model with a single covariate. The association between biological and clinical features was investigated using Pearson’s χ2 test or Fisher’s exact tests.

RESULTS Study population Data on 1123 patients were collected, exceeding the required target population. Patients’ characteristics and oncological outcomes are presented in table 1. The general FCGR status within the groups and according to KRAS status is provided in table 2. FCGR2A distribution did not concur with the expected distribution presented by Zhang et al19 (14% for FCGR2A HH). Polymorphism spread in the combined analysis of the three groups was 45.3% for HR, 32.1% for HH and 22.6% for RR. FCGR3A polymorphism for VF, VV and FF was 45.5%, 16.3% and 38.2%, respectively. Moreover, as reported in table 2, the genotype incidence between the Japanese series and the European consortium is significantly different for all genotypes, while the comparison between our consortium and the US series reveals a significantly more frequent incidence of the VV genotype compared to the non-VV genotype in the US series. As reported below, however, these genotype frequency differences are not associated with differences in biological effects (see online supplementary table S1). No clear association was seen between KRAS status and FCGR distribution. The only significant association was found for the FCGR3A-FF versus the non-FF (Fisher’s Exact test 1-sided p=0.034).

Outcome of FCGR genotypes in KRAS WT population No difference in mPFS was observed for FCGR2A HH versus non-HH, with HH reporting 22 weeks (95% CI 18.7 to 25.2 weeks) and non-HH reporting 22 weeks (95% CI 19.4 to 24.5 weeks). A difference was observed for FCGR3A-VV versus non-VV. FCGR3A-VV resulted in a mPFS of 16.4 weeks (95% CI 13.0 to 19.7 weeks) versus 23 weeks for non-VV (95% CI 21.1 to 24.8 weeks). Nevertheless, no statistical significance was reached. ( p=0.06). Comparisons of the median PFS in the European consortium series with the Japanese and the US ones (see online supplementary table S1) revealed the same lack of Geva R, et al. Gut 2014;0:1–8. doi:10.1136/gutjnl-2014-307234

effect in all of them, despite differences in genotype incidences, as expected. As reported in tables 3 and 4, our secondary endpoints analysis for mOS, RR and DCR failed to show any benefit for FCGR2A-HH versus non-HH or for FCGR3A-VV versus non-VV.

Exploratory analysis: genotype combinations or different groupings The outcomes of FCGR polymorphism subgroups in KRAS mutant patients are described in table 5. Exploratory analysis showed an advantage in DCR for the FCGR3A-FF subgroup compared with the others (Pearson χ2 test, p=0.042). A similar advantage in mOS was observed for the FF subgroup compared with the others (Pearson χ2 test, p=0.03) but not for RR and mPFS. This finding is exploratory and not part of the main design. We also highlight the inconsistency between the response, PFS and OS comparisons for this genotype. Exploratory combination analyses following Zhang et al19 and Pander et al22 analysis are presented in table 6.

DISCUSSION The current study represents the largest retrospective analysis for evaluating the association between FCGR2A and FCGR3A genotypes and clinical outcome in mCRC patients treated with cetuximab. Validated central lab analysis was used in order to generate robust results. Our study confirms the findings from our previous European Consortium publication18 by showing that KRAS mutations are associated with a lack of response to cetuximab in patients with mCRC refractory to 5-FU, irinotecan and oxaliplatin. The study results failed to support the primary study hypothesis, and no correlation was seen between mPFS and FCGR2A-HH compared with the others, and FCGR3A-VV compared with the others. No other clinically significant correlation was observed between FCGR2A and FCGR3A SNPs and cetuximab efficacy, thus leading to the conclusion that these polymorphisms cannot be considered as predictive markers of response to cetuximab in chemorefractory mCRC KRAS-WT patients. There was a numerical difference in the mPFS in the FCGR3A-VV group compared to the non-VV group among the KRAS-WT patients suggesting a possible detrimental effect. This, however, reached only a borderline ( p=0.06) significance and cannot be considered as clinically relevant. Cohort studies on breast cancer and lymphoma patients supported the hypothesis that FCGR polymorphisms affect the efficacy of trastuzumab and rituximab, with the most consistent 5

GI cancer Table 4 Response rates and disease control rates within Fc gamma receptor (FCGR) polymorphisms

Table 5

Best response FCGR2A HH vs non-HH FCGR 2A status (HH vs not HH) Non-HH Count % within FCGR 2A status HH Count % within FCGR 2A status Total Count % within FCGR2A status

CR

PR

SD

PD

Total

5 1.2

130 30.5

168 39.4

123 28.9

426 100.0

3 1.4

55 25.8

93 43.7

62 29.1

213 100.0

8 1.3

185 29.0

261 40.8

185 29.0

639 100.0

Disease control FCGR2A HH vs non-HH FCGR2A status (HH vs non-HH) Non-HH Count % within FCGR2A status (HH vs non-HH) % of total HH Count % within FCGR2A status (HH vs non-HH) % of total Total Count % within FCGR3A status (HH vs non-HH % of total

PD

CR+PR+SD

123 28.8 19.2

303 71.1 47.4

62 29.1 9.7

151 70.8 23.6

185 29.0 29.0

454 71.0 71.0

Best response FCGR3A VV vs not VV FCGR 3A status (VV vs not VV) Non-VV Count % within FCGR 3A status VV Count % within FCGR 3A status Total Count % within FCGR3A status

CR

PR

SD

PD

Total

6 1.1

157 28.7

231 42.2

153 28

547 100

2 1.9

30 28.6

35 33.3

38 36.2

105 100

8 1.2

187 28.7

266 40.8

191 29.3

652 100

Disease control FCGR3A VV vs non-VV FCGR3A status (VV vs non-VV) Non-VV Count % within FCGR3A status (VV vs non-VV) % of total VV Count % within FCGR3A status (VV vs non VV) % of total Total Count % within FCGR3A status (VV vs non-VV) % of total PD, progressive disease; PR, partial response.

6

Exploratory analysis for KRAS-mutant patients KRAS M

PD

CR+PR+SD

167 30.2 25.3

386 69.8 58.4

44 40.7 6.7

64 59.3 9.7

211 31.9 31.9

450 68.1 68.1

FCGR 2A HR HH RR FCGR 3A VF VV FF Total

RR (%)

DC (%)

mPFS (95% CI)

mOS (95% CI)

4.6 4.4 2.8

50.3 48.4 51.4

12.0 (11.5 to 12.5) 13.0 (10.3 to 15.7) 12.0 (8.3 to 15.7)

30.4 (26.6 to 34.3) 37.7 (31.1 to 44.3) 32.0 (24.7 to 39.3)

5.1 5.4 1.8 3.8

48.1 35.7* 56.8‡ 48.4

12.0 12.0 13.0 12.0

28.0 30.0 38.1 31.2

(11.0 to 13.0) (9.0 to 15.0) (10.7 to 15.3) (11.0 to 13.0)

(24.2 to (24.4 to (31.8 to (27.9 to

31.8) 35.6)† 44.5)§ 34.5)

Efficacy variables in the KRAS-M population. Survival is reported in weeks. *Pearson χ2 p=0.035 for VV worse, FF better. †Log rank (Mantel–Cox) p=0.048 for VV worse, FF better. ‡Pearson χ2 p=0.042 (FF vs non-FF, FF better). §Log Rank (Mantel-Cox) p=0.030 (FF vs non-FF, FF better). CR, complete response; DC, disease control; FCGR, Fc gamma receptor; mOS, median overall survival; mPFS, median progression-free survival; RR, response rate; SD, stable disease.

results showing an advantage for FCGR3A-VV compared to VF or FF.11 13–15 The mCRC studies that examined the correlation between cetuximab and FCGR status mostly included small patient populations. Zhang et al,19 with a total of 39 mCRC patients analysed, they suggested advantage in RR and PFS for HH and FF versus RR or VV ( p=0.003 and p=0.004, respectively). Increasing the cohort to 130 patients who participated in the phase II open-label multicentre clinical trial (IMC 0144), however, failed to replicate those results.26 Graziano et al20 analysed 110 chemorefractory patients treated with irinotecan and cetuximab. Those authors investigated several possible associations between genetic variants and clinical outcomes. Their multivariate analysis revealed associations between OS and EGFR for the intron-1 S/S and EGF 61 G/G genotypes, but no association was detected with FCGR status. Paez et al27analysed a diverse group of 104 mCRC patients of whom 92 were treated with cetuximab and chemotherapy, the majority with irinotecan-based chemotherapy, and six with oxaliplatin regimens. Only 56% of those patients were chemorefractory and had failed at least two previous lines. That study failed to show a significant association between the different FCGR polymorphisms and response, PFS or KRAS mutation status. By contrast, Bibeau et al28 analysed 69 patients and described an advantage in PFS for VV FCGR3A patients compared to F carriers (6.9 vs 3.2 months, p=0.047). Those authors reported the association with FCGR status to be maintained in KRAS-mutant (KRAS–M) patients for whom PFS was 5.5 months for VV patients versus 2.8 months for F carriers, p=0.039, although only 27 KRAS-M patients were analysed. FCGR and KRAS were independent predictors of efficacy in their multivariate analysis. Rodríguez et al29 recently analysed 106 mCRC patients treated with cetuximab for FCGR status, as well as for the EGFR downstream mutations KRAS, tested in codons 12 and 13, Neuroblastoma RAS Viral (V-Ras) Oncogene Homolog (NRAS), tested in codons 12, 13 and 61, PIK3CA tested in exon 20 and V-Raf Murine Sarcoma Viral Oncogene Homolog B (BRAF) tested in exon 15. As expected, DCR was correlated with KRAS-WT tumours (61% vs 39%, p=0.049). In patients with EGFR downstream mutations, FCGR2A-HH subtype demonstrated a higher DCR than the others (67% vs 33%, p=0.017). This was maintained in the multivariate analysis. Geva R, et al. Gut 2014;0:1–8. doi:10.1136/gutjnl-2014-307234

GI cancer Table 6 Genotype combinations in KRAS wild-type. Study

Effects assessed

Zhang et al

19

Pander et al22

HH and FF DCR mPFS mOS RR and VV DCR mPFS mOS VV and VF mPFS FF mPFS

Effects observed

Effects observed in our consortium

75.1% 24 weeks (95% CI 21.5 to 26.4 weeks) 56.8 weeks (95% CI 52.7 to 60.9 weeks)

70.6% 23 weeks (95% CI 21 to 24.9 weeks) 54.7 weeks (95% CI 50.3 to 59.1 weeks)

68.8% 20 weeks (95% CI 16.1 to 23.8 weeks) 54.4 weeks (95% CI 43.9 to 64.9 weeks)

63% 18 weeks (95% CI 15.4 to 20.5 weeks) 46.5 weeks (95% CI 39 to 54 weeks)

8.2 months (95% CI 6.7 to 10.3 months)

20.7 weeks (95% CI 18 to 23.4 weeks)

12.8 months (95% CI 10.3 to 14.7 months)

22 weeks (95% CI 19.5 to 24.4 weeks)

DCR, disease control rate; mOS, median overall survival; mPFS, median progression-free survival; RR, response rate.

In our current study, an exploratory analysis in patients with KRAS-M demonstrated an advantage for FCGR3A-FF compared to VF and VV in DCR and OS. This finding was inconsistent with the results of RR and with PFS, that showed no advantage for the FF subgroup. As already mentioned, cetuximab has no efficacy in tumours with mutations in KRAS18 since no inhibition of the EGFR pathway is observed because of KRAS constitutive activation. Nevertheless, in those tumours, cetuximab might trigger the activation of the immune system. Ongoing trials that asses activity of cetuximab in KRAS-mutant patients, based on FCGR genotype might help addressing this questions.30 Finally, a first-line setting analysis was performed by Pander et al22 as part of the phase III CAIRO2 study with chemotherapy and bevacizumab alone, or with the addition of cetuximab in mCRC patients. That substudy assessed the association of the EGFR, EGF, CCND1, FCGR2A and FCGR3A polymorphisms with PFS in KRAS-WT mCRC patients treated with cetuximab. The authors analysed 246 patient subsets of KRAS-WT of which 127 patients were treated with cetuximab and 119 were not. The analysis reported an unexpected decreased mPFS in FCGR3A-VV or VF compared with the FF genotype (8.2 vs 12.8 months, respectively; HR-1.57 (95% CI 1.06 to 2.34), p=0.025). We were able to replicate these findings when we performed the same comparison in our chemorefractory cetuximab-treated cohort. The size of our 740-patient KRAS WT cohort is the largest to be reported thus far, and helps in obtaining a more accurate assessment of the incidences of FCGR SNPs in mCRC patients. Those incidences were different from previously reported values,19 22 but our study still had sufficient numbers of patients to maintain the power required for the primary analysis. We conclude that differences in FGCR affinity to cetuximab secondary to naturally occurring genotypes were not related to different outcomes in chemorefractory mCRC patients who were treated with cetuximab. We are unable to explain why these polymorphisms cannot predict cetuximab response, but at least two main hypotheses can be considered. First, the patients included in our study had received more than one line of chemotherapy. Conventional chemotherapy suppresses the activation and the number of NKs and macrophages, thus inhibiting moAb-mediated cell death. Moreover, late stages of disease may also be associated with anergy, that is, the lack of response of the immune system to cancer antigens, independently from the effect of chemotherapy. Second, in metastatic solid tumours, the biodistribution of moAb might be limited Geva R, et al. Gut 2014;0:1–8. doi:10.1136/gutjnl-2014-307234

due to the tumour burden, the low vascularity and the high interstitial pressure. In this context, the immune system environment may be ineffective due to the anergy and the lack of distribution of the moAb itself. These factors may negatively affect the immune response and underpower the role of polymorphisms. Nevertheless, cetuximab still remains a very effective drug, even in late lines of treatments, thus indicating that mechanisms other than the immune system mediate its effectiveness. Overall, we can conclude that our negative findings suggest that ADCC does not play a role in cetuximab efficacy in mCRC. However, we cannot exclude the possibility that the immune response may be one of the mechanisms behind cetuximab activity in this setting. Further studies for investigating the impact of innate and adaptive immune response to cetuximab treatment are therefore recommended. Finally, our study supports the growing recognition that the introduction of new biomarkers into clinical practice can only be supported by large, well-powered analyses. Author affiliations 1 Department of Oncology, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel 2 Laboratory of Molecular Digestive Oncology, University Hospital Gasthuisberg, Leuven, Belgium 3 Hellenic Cooperative Oncology Group (HeCOG), Athens, Greece 4 Herlev Hospital, Copenhagen, Denmark 5 USC Norris Comprehensive Cancer Center and Hospital, California, USA 6 Department of Gastroenterology and Gastrointestinal Oncology, National Cancer Center Hospital East, Kashiwa, Japan 7 Hospital Santa Creu i Sant Pau, Barcelona, Spain 8 Department of Oncology, University Hospital del Mar -IMIM, Barcelona, Spain 9 Department of Medical Oncology, University General Hospital of Heraklion and Lab of Tumor Cell Biology, School of Medicine, University of Crete, Heraklion, Greece 10 Ospedale Civile di Livorno, Livorno, Italy 11 Department of Hematology and Medical Oncology, INCLIVA, University of Valencia, Valencia, Spain 12 Laboratory of Molecular Pathology, Institute of Pathology, Locarno, Switzerland 13 Department of Pathology, Herlev University Hospital, Copenhagen, Denmark 14 Division of Gastrointestinal Oncology, Shizuoka Cancer Center, Shizuoka, Japan 15 Department of Gastrointestinal Medical Oncology, National Hospital Organization Shikoku Cancer Center, Matsuyama, Japan 16 B. O. Cyprus Oncology Centre, Nicosia, Cyprus 17 Università Campus Bio-Medico, Rome, Italy 18 Hospital Vall d’Hebron, Barcelona, Spain 19 U.O. Oncologia Medica 2, Azienda Ospedaliero-Universitaria Pisana, Istituto Toscano Tumori, Pisa, Italy 20 Department of Oncology, Vejle Hospital, Vejle, Denmark 21 Department of Oncology, Odense University Hospital, Odense, Denmark 22 Division of Medical Oncology, Department of Experimental and Clinical Medicine, Second University of Naples, Naples, Italy 23 Cliniques universitaires Saint-Luc, Brussels, Belgium 7

GI cancer Acknowledgements We thank the patients, their families and all the investigators who participated in the study, for their contributions to this report. We thank S Yuki from Hokkaido University Hospital, K Yamaguchi from Saitama Cancer Center, K Muro at Aichi Cancer Center Hospital, Eiji Shinozaki at Cancer Institute Hospital of Japanese Foundation for Cancer Research, and Shinya Kajiura at University of Toyama for their contribution. Esther Eshkol is thanked for editorial assistance. Contributors RG, LV, HP and ST designed the work, provided and profiled the samples, acquired and analysed the data, interpreted the data, wrote the manuscript. KTK, BVJ, H-JL, TY, DP, CM, JS, FC, AC, MF, GF, JSJ, EVH, WZ, DY, KY, TN, DP, BV, TM, FL, KLGS, PP, FC provided patients’ samples, read and approved the manuscript. JDS profiled the samples. Funding LV is supported by an ESMO fellowship for Translational Research. TY is supported by the Grant-in-Aid for Cancer Research (21 S4-5) from the Ministry of Health, Labour and Welfare of Japan. FC is sponsored by a grant from Associazione italiana per la ricerca sul cancro (AIRC) and from Ministero dell’Istruzione, dell’Università della Ricerca (MIUR) and progetti di ricerca di interesse nazionale (PRIN). DP is supported by the Instituto de Salud Carlos III (FIS/1101711). CMV is supported by PI12/00989 and AECC 11964. ST is a Senior Clinical Investigator of the Fund for Scientific Research Flanders (FWO-Vlaanderen) and has received research grants from the Belgian Federation against Cancer (Stichting tegen Kanker) and from the Belgian National Cancer Plan (Nationaal Kankerplan).

10

11

12

13

14

15

16

Competing interests All the conflicts of interest reported are outside of this work. AC has been giving lectures and attending advisory boards for Merck Serono and Roche. FC has been attending advisory boards for Merck Serono, Bayer, Roche and has been receiving research grants from Astra Zeneca, Merck Serono and Bayer. KY has been giving seminar presentations for Merck Serono and Bristol Meyer Squibb. H-JL has been attending advisory boards for Merck Serono and Bristol Meyer Squibb and has been receiving honoraria from Merck Serono. FL has been attending advisory boards for Amgen, Sanofi Aventis, has been giving educational presentations for Roche and Amgen and lectures for Sanofi Aventis, Roche, Bayer and has been receiving research grants from Roche and Merck Serono. TN has been receiving honoraria from Merck Serono. TY has been receiving honoraria from Chugai, Takedo, Merck Serono and research fundings from Daiichi, Sankyo, Taiko, Bayer, Eli Lilly, Pfizer, Chugai, Yokult. ST has been giving lectures and attending advisory boards for Merck Serono.

17

Ethics approval local ethics review boards.

21

Provenance and peer review Not commissioned; externally peer reviewed.

18

19

20

22

REFERENCES 1 2

3

4

5 6

7

8

9

8

Van Cutsem E, Köhne C-H, Hitre E, et al. Cetuximab and chemotherapy as initial treatment for metastatic colorectal cancer. N Engl J Med 2009;360:1408–17. Bokemeyer C, Bondarenko I, Makhson A, et al. Fluorouracil, leucovorin, and oxaliplatin with and without cetuximab in the first-line treatment of metastatic colorectal cancer. J Clin Oncol 2009;27:663–71. Cunningham D, Humblet Y, Siena S, et al. Cetuximab monotherapy and cetuximab plus irinotecan in irinotecan-refractory metastatic colorectal cancer. N Engl J Med 2004;351:337–45. Saltz LB, Meropol NJ, Loehrer PJ, et al. Phase II trial of cetuximab in patients with refractory colorectal cancer that expresses the epidermal growth factor receptor. J Clin Oncol 2004;22:1201–8. Karapetis CS, Khambata-Ford S, Jonker DJ, et al. K-ras mutations and benefit from cetuximab in advanced colorectal cancer. N Engl J Med 2008;359:1757–65. Allegra CJ, Jessup JM, Somerfield MR, et al. American Society of Clinical Oncology provisional clinical opinion: testing for KRAS gene mutations in patients with metastatic colorectal carcinoma to predict response to anti-epidermal growth factor receptor monoclonal antibody therapy. J Clin Oncol 2009;27:2091–6. Mendelsohn J. The epidermal growth factor receptor as a target for cancer therapy. Endocr Relat Cancer 2001;8:3–9. http://www.ncbi.nlm.nih.gov/pubmed/11350723 (accessed 9 Jun 2014). Robertson MJ, Ritz J. Biology and clinical relevance of human natural killer cells. Blood 1990;76:2421–38. http://www.ncbi.nlm.nih.gov/pubmed/2265240 (accessed 9 Jun 2014). Van Sorge NM, van der Pol W-L, van de Winkel JGJ. FcgammaR polymorphisms: Implications for function, disease susceptibility and immunotherapy. Tissue Antigens 2003;61:189–202. http://www.ncbi.nlm.nih.gov/pubmed/12694568 (accessed 9 Jun 2014).

23

24

25

26

27

28

29

30

Kim DH, Jung HD, Kim JG, et al. FCGR3A gene polymorphisms may correlate with response to frontline R-CHOP therapy for diffuse large B-cell lymphoma. Blood 2006;108:2720–5. Weng W-K, Levy R. Two immunoglobulin G fragment C receptor polymorphisms independently predict response to rituximab in patients with follicular lymphoma. J Clin Oncol 2003;21:3940–7. Persky DO, Dornan D, Goldman BH, et al. Fc gamma receptor 3a genotype predicts overall survival in follicular lymphoma patients treated on SWOG trials with combined monoclonal antibody plus chemotherapy but not chemotherapy alone. Haematologica 2012;97:937–42. Ahlgrimm M, Pfreundschuh M, Kreuz M, et al. The impact of Fc-γ receptor polymorphisms in elderly patients with diffuse large B-cell lymphoma treated with CHOP with or without rituximab. Blood 2011;118:4657–62. Musolino A, Naldi N, Bortesi B, et al. Immunoglobulin G fragment C receptor polymorphisms and clinical efficacy of trastuzumab-based therapy in patients with HER-2/neu-positive metastatic breast cancer. J Clin Oncol 2008;26:1789–96. Tamura K, Shimizu C, Hojo T, et al. FcγR2A and 3A polymorphisms predict clinical outcome of trastuzumab in both neoadjuvant and metastatic settings in patients with HER2-positive breast cancer. Ann Oncol 2011;22:1302–7. Hurvitz SA, Betting DJ, Stern HM, et al. Analysis of Fcγ receptor IIIa and IIa polymorphisms: lack of correlation with outcome in trastuzumab-treated breast cancer patients. Clin Cancer Res 2012;18:3478–86. Ahn J, Jung M, Koo JMB. Prediction of response in advanced breast cancer (ABC) patients who received trastuzumab-paclitaxel (TP): HER2 FISH ratio, expression of class III beta tubulin (bTubIII), and immunoglobulin G fragment C receptor (FCGR) polymorphisms. J Clin Oncol 2010;28(15 Suppl):1101. De Roock W, Claes B, Bernasconi D, et al. Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on the efficacy of cetuximab plus chemotherapy in chemotherapy-refractory metastatic colorectal cancer: a retrospective consortium analysis. Lancet Oncol 2010;11:753–62. Zhang W, Gordon M, Schultheis AM, et al. FCGR2A and FCGR3A polymorphisms associated with clinical outcome of epidermal growth factor receptor expressing metastatic colorectal cancer patients treated with single-agent cetuximab. J Clin Oncol 2007;25:3712–18. Graziano F, Ruzzo A, Loupakis F, et al. Pharmacogenetic profiling for cetuximab plus irinotecan therapy in patients with refractory advanced colorectal cancer. J Clin Oncol 2008;26:1427–34. Tsuchida Y, Therasse P. Response evaluation criteria in solid tumors (RECIST): new guidelines. Med Pediatr Oncol 2001;37:1–3. Pander J, Gelderblom H, Antonini NF, et al. Correlation of FCGR3A and EGFR germline polymorphisms with the efficacy of cetuximab in KRAS wild-type metastatic colorectal cancer. Eur J Cancer 2010;46:1829–34. Itoh Y, Mizuki N, Shimada T, et al. High-throughput DNA typing of HLA-A, -B, -C, and -DRB1 loci by a PCR-SSOP-Luminex method in the Japanese population. Immunogenetics 2005;57:717–29. Jiang XM, Arepally G, Poncz M, et al. Rapid detection of the Fc gamma RIIA-H/R 131 ligand-binding polymorphism using an allele-specific restriction enzyme digestion (ASRED). J Immunol Methods 1996;199:55–9. http://www.ncbi.nlm.nih. gov/pubmed/8960098 (accessed 9 Jun 2014). Leppers-van de Straat FG, van der Pol WL, Jansen MD, et al. A novel PCR-based method for direct Fc gamma receptor IIIa (CD16) allotyping. J Immunol Methods 2000;242:127–32. http://www.ncbi.nlm.nih.gov/pubmed/10986395 (accessed 9 Jun 2014). Lurje G, Nagashima F, Zhang W, et al. Polymorphisms in cyclooxygenase-2 and epidermal growth factor receptor are associated with progression-free survival independent of K-ras in metastatic colorectal cancer patients treated with single-agent cetuximab. Clin Cancer Res 2008;14:7884–95. Paez D, Paré L, Espinosa I, et al. Immunoglobulin G fragment C receptor polymorphisms and KRAS mutations: are they useful biomarkers of clinical outcome in advanced colorectal cancer treated with anti-EGFR-based therapy? Cancer Sci 2010;101:2048–53. Bibeau F, Lopez-Crapez E, Di Fiore F, et al. Impact of Fc{gamma}RIIa-Fc{gamma} RIIIa polymorphisms and KRAS mutations on the clinical outcome of patients with metastatic colorectal cancer treated with cetuximab plus irinotecan. J Clin Oncol 2009;27:1122–9. Rodríguez J, Zarate R, Bandres E, et al. Fc gamma receptor polymorphisms as predictive markers of Cetuximab efficacy in epidermal growth factor receptor downstream-mutated metastatic colorectal cancer. Eur J Cancer 2012;48:1774–80. http://www.clinicaltrials.gov/ct2/show/NCT01450319

Geva R, et al. Gut 2014;0:1–8. doi:10.1136/gutjnl-2014-307234