Purpose:

Amphiregulin (AREG) and epiregulin (EREG) are ligands of EGFR. Predictive information for anti-EGFR treatment in metastatic colorectal cancer (mCRC) was observed, but data for other agents is limited.

Experimental Design:

Ligand mRNA expression; RAS, BRAF, PIK3CA mutations; and EGFR expression were assessed by qRT-PCR, pyrosequencing, and IHC, respectively, in mCRC tumor tissue of patients participating in the randomized controlled trials FIRE-1, CIOX, and FIRE-3. Normalized mRNA expression was dichotomized using median and third quartile. Overall (OS) and progression-free survival (PFS) were estimated by Kaplan–Meier method including univariate and multivariate Cox regression analyses. Penalized spline regression analysis tested interaction of mRNA expression and outcome.

Results:

Of 688 patients with available material, high AREG expression was detected in 343 (>median) and 172 (>3rd quartile) patients. High AREG expression was associated with significantly higher OS [26.2 vs. 21.5 months, HR = 0.80; 95% confidence interval (CI), 0.68–0.94; P = 0.007], PFS (10.0 vs. 8.1 months, HR = 0.74; 95% CI, 0.63–0.86; P = 0.001), and objective response rate (63.1% vs. 51.6%, P = 0.004) compared to low expression at both threshold values. This effect remained significant in multivariate Cox regression analysis (OS: P = 0.01, PFS: P = 0.002). High AREG mRNA expression interacted significantly with the efficacy of cetuximab compared with bevacizumab (OS: P = 0.02, PFS: P = 0.04) in RAS WT mCRC.

Conclusions:

High AREG mRNA expression is a favorable prognostic biomarker for mCRC which interacted significantly with efficacy of anti-EGFR treatment.

Translational Relevance

Amphiregulin (AREG) and epiregulin (EREG) are ligands of EGFR and were considered as predictive biomarkers for treatment targeting EGFR in metastatic colorectal cancer (mCRC). Interaction of EGFR ligand expression with anti-EGFR treatment was never shown, as control arms without anti-EGFR agents were often missing in these investigations. For the first time, we demonstrated significant beneficial treatment interaction of AREG mRNA expression with cetuximab-compared with bevacizumab-containing regimen and cytotoxic treatment without biologicals using similarly ascertained data from three randomized controlled trials (FIRE-1, CIOX, and FIRE-3). Moreover, a subgroup of patients with primary tumors of the right colon and high AREG expression benefitted from anti-EGFR treatment. Therefore, predictive information of AREG mRNA expression was confirmed and validated, and AREG mRNA expression might be considered as a predictive biomarker for cetuximab treatment independently from primary tumor sidedness.

Antibodies targeting EGFR significantly improve outcome of patients with RAS wild-type (WT) metastatic colorectal cancer (mCRC; 1, 2). The beneficial effect of anti-EGFR antibodies on survival appeared to be limited to left-sided tumors. Present research is focusing on the investigation of additional biomarkers to provide a mechanistic basis for the prediction of anti-EGFR treatment efficacy.

Amphiregulin (AREG) and epiregulin (EREG) are ligands of EGFR, which promote activation of the MAPK signaling pathway including cell proliferation and invasion (3). It was shown that EGFR ligands were regulated by autocrine loops with activation by integrin α6β4 and RAS-independent methylation of intragenic regions (3–6). Moreover, lower ligand expression was observed in right-sided, BRAF V600E–mutated (MUT), and microsatellite instable tumors (7, 8). High ligand expression was associated with better prognosis regardless of treatment and higher susceptibility to anti-EGFR antibodies in patients with RAS WT mCRC, which implicated a potential role of predictive biomarkers (7, 9–13).

Despite their benefit, major problems interpreting these biomarkers were of methodologic nature (e.g., normalization to reference genes and threshold values to dichotomize high and low expression). Moreover, the effect of ligand expression according to primary tumor sidedness was not investigated yet owing to limited sample sizes or missing clinical data. We aimed to address these issues and performed a combined biomarker analysis for AREG and EREG mRNA expression in the randomized controlled trials FIRE-1, CIOX (FIRE-2), and FIRE-3.

FIRE-1 was a phase III trial that compared weekly application of 5-fluorouracil, leucovorin, and irinotecan (FOLFIRI) to a modified protocol of irinotecan plus oxaliplatin (IROX) in first-line treatment of mCRC (14). The CIOX phase II trial tested the combination of capecitabine and oxaliplatin (CAPOX) or irinotecan (CAPIRI) plus cetuximab as first-line treatment of mCRC (15). The FIRE-3 phase III trial showed superiority of biweekly FOLFIRI plus cetuximab versus bevacizumab in first-line treatment of KRAS WT (and later RAS WT) mCRC (16).

Prognostic and predictive effects of AREG and EREG mRNA expression were assessed in a large collective of patients with mCRC treated with cytotoxic agents (FIRE-1), cetuximab (CIOX, FIRE-3), or bevacizumab (FIRE-3). Two different threshold values were applied for exclusion of threshold specific effects. Subgroup analyses were performed to assess the influence of primary tumor location and IHC EGFR expression.

Trial designs and patients

Trial designs, treatment protocols, efficacy, and safety of all trials were published elsewhere (14–16). All trials were conducted before RAS mutations were identified as relevant biomarker in mCRC. This retrospective analysis included only patients with available qRT-PCR data for AREG and EREG mRNA expression. All patients provided written informed consent for trial participation and translational research. This analysis was conducted in accordance with the Declaration of Helsinki (1996). Analyses of mRNA expression data in clinical trials were approved by the local ethics committee of the University of Munich (Munich, Germany; registry-nos. 090–04, 545–11, 186–15, respectively).

End points

Primary and secondary endpoints of the original studies were published previously (14–16). This analysis investigated prognostic and predictive effects of high EGFR ligand mRNA expression to overall survival (OS), progression-free survival (PFS), and objective response rates (ORR). ORR was defined as percentage of complete and partial remissions (FIRE-1: UICC criteria; CIOX, FIRE-3: RECIST 1.0). PFS and OS were defined as described in their respective protocols and expressed as median values. Patients alive at the end of their study were censored at the last time point of patient contact.

Mutational analysis, extraction of mRNA, and qRT-PCR analysis

DNA mutational analysis was performed by pyrosequencing as described previously (13). mRNA was isolated following the instructions of the RNeasy FFPE Tissue Kit (Qiagen). Transcripted into cDNA were 150 ng (FIRE-1), 20 ng (CIOX), and 150–500 ng (FIRE-3) of mRNA, respectively. mRNA expression was determined in duplicates using the LightCycler 480 and Universal Probe Library system (Roche). β-Actin and GAPDH (not FIRE-1) were used as reference. A total of 5 × 104 molecules cDNA were used as positive and RNase-free water as negative control. ΔΔCp values were calculated as described previously (13).

Normalization of ΔΔCp gene expression between trials

To correct for methodologic bias, logarithmic ΔΔCp expression values were considered as real numbers. The difference of each ΔΔCpo of a trial and trial-specific minimum min(ΔΔCp) was divided by the trial-specific expression range of maximum and minimum ΔΔCp.

formula

In addition, results were recalculated with unnormalized expression values to gather variance deriving from normalization procedure.

Membranous EGFR expression by IHC

EGFR expression was assessed by IHC in FIRE-1 and CIOX using a prediluted monoclonal mouse antibody clone 3C6 (Ventana Medical Systems) on a Ventana BenchMark XT autostainer. Two independent reviewers (FIRE-1: A. Stahler and J. Neumann; CIOX: C. Kapaun and J. Neumann) assessed the staining intensities from 0 to 3 and the percentage of stained cells. A maximum score of 300 could be achieved by multiplication of these parameters. As published before, a score analogous to Neumann and colleagues dichotomized EGFR-positive (score > 22.5) and -negative (score < 22.5) tumors (13, 17).

Statistical analysis

Mean expressions were compared by the t test for AREG mRNA expression (parametric distribution) in clinical and molecular patient subgroups. Two threshold values (median and third quartile of normalized mRNA expression) dichotomized high and low ligand expression. Kaplan–Meier method estimated OS and PFS; comparisons were made using the log-rank test. Univariate and multivariate Cox regression analyses were performed in subgroups for OS and PFS. Group comparisons including ORR were performed using the χ2 test. Test on interaction was performed with penalized spline regression analysis of AREG expression as a continuous variable. P values < 0.05 (two-sided) were considered statistically significant. SPSS PASW 23.0 (SPSS) and R v3.6.1 software were used for statistical analyses.

Trial population

The total trial population of FIRE-1, CIOX, and FIRE-3 consisted of 1,408 patients. Of these, 688 patients (48.9%) provided material for mRNA extraction and gene expression analysis (Fig. 1). Data for mutations in RAS (no KRAS codon 59 and NRAS exon 2–4 data in CIOX), BRAF, and PIK3CA genes by pyrosequencing were available from all trials. IHC expression of EGFR was assessed in FIRE-1 and CIOX. Baseline characteristics of trial subsets were comparable, but differences were observed for primary tumor side (P = 0.02), RAS [P < 0.001, prospective (FIRE-3) versus retrospective (FIRE-1, CIOX) assessment of KRAS mutations], BRAF (P = 0.04), and EGFR (P < 0.001) status (Table 1; Supplementary Fig. S1).

Figure 1.

CONSORT diagram of the investigated population. Available mRNA expression data for AREG and EREG of the randomized controlled trials FIRE-1, CIOX, and FIRE-3 were pooled and analyzed using two threshold values, which dichotomized the population into high and low expression for AREG and EREG, respectively. EREG data were excluded due to high variance. Prognostic analyses were performed in the complete dataset. Treatment interaction and predictive effects of high AREG expression were investigated in the RAS WT subpopulation. Impact of AREG expression according to sidedness and EGFR expression was investigated in patients treated with cetuximab after exclusion of BRAF- and PIK3CA-mutated tumors.

Figure 1.

CONSORT diagram of the investigated population. Available mRNA expression data for AREG and EREG of the randomized controlled trials FIRE-1, CIOX, and FIRE-3 were pooled and analyzed using two threshold values, which dichotomized the population into high and low expression for AREG and EREG, respectively. EREG data were excluded due to high variance. Prognostic analyses were performed in the complete dataset. Treatment interaction and predictive effects of high AREG expression were investigated in the RAS WT subpopulation. Impact of AREG expression according to sidedness and EGFR expression was investigated in patients treated with cetuximab after exclusion of BRAF- and PIK3CA-mutated tumors.

Close modal
Table 1.

Baseline characteristics of mRNA expression subsets of FIRE-1, CIOX, and FIRE-3.

FIRE-1CIOXFIRE-3
n = 192n = 113an = 383b
FOLFIRImIROXCAPIRI + CetuximabCAPOX + CetuximabFOLFIRI + CetuximabFOLFIRI + Bevacizumab
n = 101n = 91n = 56n = 54n = 189n = 192P
Age, median (range) 63 (42–75) 63 (25–76) 63 (32–75) 61 (43–77) 64 (38–76) 66 (31–76) 0.18 
 <65 years, n (%) 59 (58.4) 55 (60.4) 32 (57.1) 31 (57.4) 95 (50.3) 87 (45.3)  
 ≥65 years, n (%) 42 (41.6) 36 (39.6) 24 (42.9) 23 (42.6) 85 (45.0) 102 (53.1)  
 Missing, n (%) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 9 (4.8) 3 (1.6)  
Sex 
 Female, n (%) 37 (36.6) 26 (28.6) 16 (28.6) 15 (27.8) 57 (30.2) 66 (34.4) 0.71 
 Male, n (%) 64 (63.4) 65 (71.4) 40 (71.4) 39 (72.2) 123 (65.1) 123 (64.1)  
 Missing, n (%) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 9 (4.8) 3 (1.6)  
ECOG status, n (%) 
 0–1 98 (97.0) 82 (91.2) 53 (96.4) 50 (92.6) 176 (97.9) 185 (97.9) 0.05 
 2 3 (3.0) 8 (8.8) 2 (3.6) 4 (7.4) 4 (2.1) 4 (2.1)  
 Missing, n (%) 0 (0.0) 1 (1.1) 1 (1.8) 0 (0.0) 9 (4.8) 3 (1.6)  
Primary tumor location, n (%) 
 Right 14 (13.9) 14 (15.4) 16 (28.6) 21 (38.9) 47 (24.9) 54 (28.1) 0.02 
 Left 76 (75.2) 68 (74.7) 40 (71.4) 33 (61.1) 142 (75.1) 136 (70.8)  
 Missing, n (%) 11 (10.9) 9 (9.9) 0 (0.0) 0 (0.0) 0 (0.0) 2 (1.0)  
Sites of metastases, n (%) 
 1–2 84 (83.2) 73 (80.2) 42 (75.0) 41 (75.9) 143 (75.7) 147 (76.6) 0.67 
 ≥3 16 (15.9) 16 (17.6) 14 (25.0) 13 (24.1) 33 (17.4) 41 (21.3)  
 Missing, n (%) 1 (0.9) 2 (2.2) 0 (0.0) 0 (0.0) 13 (6.9) 4 (2.1)  
RAS status, n (%) 
 Wild-type 43 (42.6) 51 (56.0) 29 (51.8) 37 (68.5) 121 (64.0) 134 (69.8) <0.001 
 Mutated 58 (57.4) 40 (44.0) 27 (48.2) 17 (31.5) 57 (30.2) 51 (26.6)  
 Missing, n (%) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 11 (5.8) 7 (3.6)  
BRAF status, n (%) 
 Wild-type 100 (99.0) 84 (92.3) 49 (87.5) 49 (90.7) 127 (67.2) 139 (72.4) 0.04 
 Mutated 1 (1.0) 7 (7.7) 7 (12.5) 5 (9.3) 18 (9.5) 13 (6.8)  
 Missing, n (%) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 44 (23.3) 40 (20.8)  
PIK3CA status, n (%) 
 Wild-type 95 (94.1) 85 (93.4) 50 (89.3) 44 (81.5) 176 (93.1) 170 (88.5) 0.07 
 Mutated 6 (5.9) 6 (6.6) 6 (10.7) 10 (18.5) 13 (6.9) 22 (11.5)  
 Missing, n (%) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)  
EGFR IHC score, n (%) 
 <22.5 63 (62.4) 61 (67.0) 17 (30.4) 15 (27.8) 0 (0.0) 0 (0.0) <0.001 
 ≥22.5 38 (37.6) 30 (33.0) 35 (62.5) 37 (68.5) 0 (0.0) 0 (0.0)  
 Missing, n (%) 0 (0.0) 0 (0.0) 4 (7.1) 2 (3.7) 189 (100.0) 192 (100.0)  
Normalized AREG expression, median (range) 0.50 (0.0–1.0) 0.44 (0.1–0.9) 0.48 (0.1–1.0) 0.49 (0.0–0.7) 0.50 (0.0–1.0) 0.51 (0.0–0.8)  
 <Median, n (%) 46 (45.5) 57 (62.6) 31 (55.4) 27 (50.0) 89 (47.1) 90 (46.9) 0.13 
 ≥Median, n (%) 55 (54.5) 34 (37.4) 25 (44.6) 27 (50.0) 100 (52.9) 102 (53.1)  
 <3rd quartile, n (%) 74 (73.3) 77 (84.6) 42 (75.0) 41 (75.9) 134 (70.9) 143 (74.5) 0.27 
 ≥3rd quartile, n (%) 27 (26.7) 14 (15.4) 14 (25.0) 13 (24.1) 55 (29.1) 49 (25.5)  
Normalized EREG expression, median (range) 0.49 (0.0–1.0) 0.42 (0.1–0.8) 0.38 (0.1–1.0) 0.37 (0.0–0.7) 0.27 (0.0–0.8) 0.29 (0.0–1.0)  
 <Median, n (%) 27 (26.7) 50 (54.9) 17 (30.4) 18 (33.3) 123 (65.1) 130 (67.7) <0.001 
 ≥Median, n (%) 74 (73.3) 41 (45.1) 39 (69.6) 36 (66.7) 66 (34.9) 62 (32.3)  
 <3rd quartile, n (%) 46 (45.5) 55 (60.4) 36 (64.3) 34 (63.0) 170 (89.9) 176 (91.7) <0.001 
 ≥3rd quartile, n (%) 55 (54.5) 36 (39.6) 20 (35.7) 20 (37.0) 19 (10.1) 16 (8.3)  
FIRE-1CIOXFIRE-3
n = 192n = 113an = 383b
FOLFIRImIROXCAPIRI + CetuximabCAPOX + CetuximabFOLFIRI + CetuximabFOLFIRI + Bevacizumab
n = 101n = 91n = 56n = 54n = 189n = 192P
Age, median (range) 63 (42–75) 63 (25–76) 63 (32–75) 61 (43–77) 64 (38–76) 66 (31–76) 0.18 
 <65 years, n (%) 59 (58.4) 55 (60.4) 32 (57.1) 31 (57.4) 95 (50.3) 87 (45.3)  
 ≥65 years, n (%) 42 (41.6) 36 (39.6) 24 (42.9) 23 (42.6) 85 (45.0) 102 (53.1)  
 Missing, n (%) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 9 (4.8) 3 (1.6)  
Sex 
 Female, n (%) 37 (36.6) 26 (28.6) 16 (28.6) 15 (27.8) 57 (30.2) 66 (34.4) 0.71 
 Male, n (%) 64 (63.4) 65 (71.4) 40 (71.4) 39 (72.2) 123 (65.1) 123 (64.1)  
 Missing, n (%) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 9 (4.8) 3 (1.6)  
ECOG status, n (%) 
 0–1 98 (97.0) 82 (91.2) 53 (96.4) 50 (92.6) 176 (97.9) 185 (97.9) 0.05 
 2 3 (3.0) 8 (8.8) 2 (3.6) 4 (7.4) 4 (2.1) 4 (2.1)  
 Missing, n (%) 0 (0.0) 1 (1.1) 1 (1.8) 0 (0.0) 9 (4.8) 3 (1.6)  
Primary tumor location, n (%) 
 Right 14 (13.9) 14 (15.4) 16 (28.6) 21 (38.9) 47 (24.9) 54 (28.1) 0.02 
 Left 76 (75.2) 68 (74.7) 40 (71.4) 33 (61.1) 142 (75.1) 136 (70.8)  
 Missing, n (%) 11 (10.9) 9 (9.9) 0 (0.0) 0 (0.0) 0 (0.0) 2 (1.0)  
Sites of metastases, n (%) 
 1–2 84 (83.2) 73 (80.2) 42 (75.0) 41 (75.9) 143 (75.7) 147 (76.6) 0.67 
 ≥3 16 (15.9) 16 (17.6) 14 (25.0) 13 (24.1) 33 (17.4) 41 (21.3)  
 Missing, n (%) 1 (0.9) 2 (2.2) 0 (0.0) 0 (0.0) 13 (6.9) 4 (2.1)  
RAS status, n (%) 
 Wild-type 43 (42.6) 51 (56.0) 29 (51.8) 37 (68.5) 121 (64.0) 134 (69.8) <0.001 
 Mutated 58 (57.4) 40 (44.0) 27 (48.2) 17 (31.5) 57 (30.2) 51 (26.6)  
 Missing, n (%) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 11 (5.8) 7 (3.6)  
BRAF status, n (%) 
 Wild-type 100 (99.0) 84 (92.3) 49 (87.5) 49 (90.7) 127 (67.2) 139 (72.4) 0.04 
 Mutated 1 (1.0) 7 (7.7) 7 (12.5) 5 (9.3) 18 (9.5) 13 (6.8)  
 Missing, n (%) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 44 (23.3) 40 (20.8)  
PIK3CA status, n (%) 
 Wild-type 95 (94.1) 85 (93.4) 50 (89.3) 44 (81.5) 176 (93.1) 170 (88.5) 0.07 
 Mutated 6 (5.9) 6 (6.6) 6 (10.7) 10 (18.5) 13 (6.9) 22 (11.5)  
 Missing, n (%) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)  
EGFR IHC score, n (%) 
 <22.5 63 (62.4) 61 (67.0) 17 (30.4) 15 (27.8) 0 (0.0) 0 (0.0) <0.001 
 ≥22.5 38 (37.6) 30 (33.0) 35 (62.5) 37 (68.5) 0 (0.0) 0 (0.0)  
 Missing, n (%) 0 (0.0) 0 (0.0) 4 (7.1) 2 (3.7) 189 (100.0) 192 (100.0)  
Normalized AREG expression, median (range) 0.50 (0.0–1.0) 0.44 (0.1–0.9) 0.48 (0.1–1.0) 0.49 (0.0–0.7) 0.50 (0.0–1.0) 0.51 (0.0–0.8)  
 <Median, n (%) 46 (45.5) 57 (62.6) 31 (55.4) 27 (50.0) 89 (47.1) 90 (46.9) 0.13 
 ≥Median, n (%) 55 (54.5) 34 (37.4) 25 (44.6) 27 (50.0) 100 (52.9) 102 (53.1)  
 <3rd quartile, n (%) 74 (73.3) 77 (84.6) 42 (75.0) 41 (75.9) 134 (70.9) 143 (74.5) 0.27 
 ≥3rd quartile, n (%) 27 (26.7) 14 (15.4) 14 (25.0) 13 (24.1) 55 (29.1) 49 (25.5)  
Normalized EREG expression, median (range) 0.49 (0.0–1.0) 0.42 (0.1–0.8) 0.38 (0.1–1.0) 0.37 (0.0–0.7) 0.27 (0.0–0.8) 0.29 (0.0–1.0)  
 <Median, n (%) 27 (26.7) 50 (54.9) 17 (30.4) 18 (33.3) 123 (65.1) 130 (67.7) <0.001 
 ≥Median, n (%) 74 (73.3) 41 (45.1) 39 (69.6) 36 (66.7) 66 (34.9) 62 (32.3)  
 <3rd quartile, n (%) 46 (45.5) 55 (60.4) 36 (64.3) 34 (63.0) 170 (89.9) 176 (91.7) <0.001 
 ≥3rd quartile, n (%) 55 (54.5) 36 (39.6) 20 (35.7) 20 (37.0) 19 (10.1) 16 (8.3)  

Note: P values < 0.05 were considered significant and are displayed bold.

Abbreviation: ECOG, Eastern Cooperative Oncology Group.

aMissing treatment information for 3 patients.

bMissing treatment information for 2 patients.

AREG mRNA expression according to baseline characteristics

The distributions of normalized expression were parametric for AREG and nonparametric for EREG (Supplementary Fig. S2). Normalization failed to harmonize the collective in regard to EREG mRNA expression owing to high variance. Thus, EREG expression was excluded from further analyses (Table 1). Mean AREG expression was comparable for all baseline characteristics including sidedness, EGFR expression, and MSI-H status. However, median AREG and EGFR expression were lower in detailed primary tumor locations of the right colon (Supplementary Fig. S3). A trend for lower AREG expression was observed in RAS WT/BRAF MUT tumors (Supplementary Table S1).

Normalized AREG mRNA expression is a strong prognostic biomarker

In individual trial data analyses, OS was more pronounced and PFS was significantly longer in the presence of high AREG expression (Supplementary Fig. S4).

In pooled data analysis of all trials, high AREG expression was significantly associated with longer OS and PFS regardless of threshold value (Fig. 2). The favorable prognostic effect of normalized AREG expression was confirmed in multivariate Cox regression analyses including baseline and molecular characteristics for OS [HR = 0.46 (95% CI, 0.25–0.85); P = 0.01] and PFS [HR = 0.40 (95% CI, 0.22–0.71); P = 0.002; Supplementary Table S2).

Figure 2.

Kaplan–Meier plots of high versus low AREG expression. A, OS, when AREG expression dichotomized by median, all trials. B, PFS, when AREG expression dichotomized by median, all trials. C, OS, when AREG expression dichotomized by third quartile, all trials. D, PFS, when AREG expression dichotomized by third quartile, all trials.

Figure 2.

Kaplan–Meier plots of high versus low AREG expression. A, OS, when AREG expression dichotomized by median, all trials. B, PFS, when AREG expression dichotomized by median, all trials. C, OS, when AREG expression dichotomized by third quartile, all trials. D, PFS, when AREG expression dichotomized by third quartile, all trials.

Close modal

ORR was significantly higher in the presence of high AREG expression [median: 63.1% vs. 51.6%, OR = 1.61 (95% CI, 1.16–2.21), P = 0.004; third quartile: 71.5% vs. 52.6%, OR = 2.26 (95% CI, 1.53–3.34); P < 0.0001]. Comparable results were observed for unnormalized expression values (Supplementary Table S3).

Impact of normalized AREG mRNA expression in relation to molecular biomarkers

Next, the prognostic impact of normalized AREG expression was related to biomarkers activating the MAPK signaling pathway directly (IHC EGFR expression, RAS, BRAF MUT) or indirectly (PIK3CA MUT).

Among 688 patients, RAS, BRAF, or PIK3CA MUT and EGFR positivity were detected in 253 (36.7%), 51 (7.4%), 62 (9.0%), and 140 (20.3%) patients, respectively. The favorable prognostic effect of high AREG expression was observed in RAS, BRAF, PIK3CA WT, and EGFR-positive tumors for OS and PFS (Supplementary Table S4).

Normalized AREG mRNA expression is a predictor of efficacy for cetuximab in RAS WT mCRC

Across all trials, 415 patients had RAS WT tumors and of these, 187 (45.1%) were treated with a cetuximab-containing regimen. In this subgroup, OS (36.6 vs. 23.5 months), PFS (10.6 vs. 7.8 months), and ORR (83.3 vs. 63.9%) were significantly higher in the presence of high AREG expression (dichotomized by median). Positive effects of high AREG expression were not observed in RAS WT patients treated with bevacizumab or without biologicals (Table 2).

Table 2.

Treatment-dependent outcome of amphiregulin mRNA expression in RAS WT patients of FIRE-1, CIOX, and FIRE-3.

Cetuximab + ChemotherapyBevacizumab + ChemotherapyChemotherapy
AREGAREGAREGAREGAREGAREGAREGAREGAREGAREGAREGAREG
≥median<median≥3rd quartile<3rd quartile≥median<median≥3rd quartile<3rd quartile≥median<median≥3rd quartile<3rd quartile
n = 102n = 85n = 64n = 123n = 68n = 66n = 35n = 99n = 40n = 54n = 22n = 72
Overall survival, months 36.6 23.5 37.1 24.5 27.5 23.8 28.6 23.8 23.2 21.9 23.2 21.8 
HR (95% CI) 0.60 (0.43–0.83) 0.61 (0.43–0.87) 0.93 (0.65–1.33) 0.96 (0.65–1.43) 1.50 (0.94–2.40) 0.78 (0.46–1.32) 
P, log-rank 0.002 0.006 0.71 0.85 0.09 0.35 
Progression-free survival, months 10.6 7.8 11.2 8.6 11.3 10.3 11.5 10.7 7.5 7.7 7.5 77 
HR (95% CI) 0.66 (0.49–0.88) 0.77 (0.56–1.05) 0.92 (0.65–1.30) 1.03 (0.70–1.53) 0.82 (0.53–1.27) 0.86 (0.51–1.43) 
P, log-rank 0.006 0.10 0.63 0.87 0.37 0.55 
Objective response rate, % 83.3 63.9 83.6 69.3 68.2 55.9 76.5 57.1 47.5 46.3 59.1 43.1 
P, χ2 0.006 0.06 0.20 0.06 1.00 0.23 
Cetuximab + ChemotherapyBevacizumab + ChemotherapyChemotherapy
AREGAREGAREGAREGAREGAREGAREGAREGAREGAREGAREGAREG
≥median<median≥3rd quartile<3rd quartile≥median<median≥3rd quartile<3rd quartile≥median<median≥3rd quartile<3rd quartile
n = 102n = 85n = 64n = 123n = 68n = 66n = 35n = 99n = 40n = 54n = 22n = 72
Overall survival, months 36.6 23.5 37.1 24.5 27.5 23.8 28.6 23.8 23.2 21.9 23.2 21.8 
HR (95% CI) 0.60 (0.43–0.83) 0.61 (0.43–0.87) 0.93 (0.65–1.33) 0.96 (0.65–1.43) 1.50 (0.94–2.40) 0.78 (0.46–1.32) 
P, log-rank 0.002 0.006 0.71 0.85 0.09 0.35 
Progression-free survival, months 10.6 7.8 11.2 8.6 11.3 10.3 11.5 10.7 7.5 7.7 7.5 77 
HR (95% CI) 0.66 (0.49–0.88) 0.77 (0.56–1.05) 0.92 (0.65–1.30) 1.03 (0.70–1.53) 0.82 (0.53–1.27) 0.86 (0.51–1.43) 
P, log-rank 0.006 0.10 0.63 0.87 0.37 0.55 
Objective response rate, % 83.3 63.9 83.6 69.3 68.2 55.9 76.5 57.1 47.5 46.3 59.1 43.1 
P, χ2 0.006 0.06 0.20 0.06 1.00 0.23 

Note: Survival times are displayed as medians. P values < 0.05 are considered significant and displayed bold.

AREG expression interacted significantly with cetuximab, not bevacizumab, in patients with RAS WT tumors regarding OS (P = 0.02) and PFS (P = 0.04). However, interaction of AREG expression with the respective trial (CIOX or FIRE-3) indicated trial-dependent effect sizes (OS: P = 0.03; PFS: P = 0.05).

AREG expression in context of cetuximab treatment and EGFR expression

To investigate the impact of high AREG expression in the context of isolated EGFR expression and cetuximab treatment, we selected 38 patients of CIOX with RAS, BRAF, and PIK3CA WT tumors and available IHC EGFR expression data.

In this specific subgroup, high AREG expression was significantly associated with higher ORR and remarkably prolonged OS and PFS in EGFR-positive tumors, but also numerically prolonged OS and PFS in EGFR-negative tumors (Table 3).

Table 3.

Efficacy of cetuximab in RAS/BRAF/PIK3CA WT patients with high versus low AREG expression according to EGFR expression and sidedness.

EGFR statusSidedness
EGFR positiveEGFR negativeRight-sided tumorsLeft-sided tumors
n = 25n = 13n = 26n = 96
AREGAREGAREGAREGAREGAREGAREGAREG
≥median<median≥median<median≥median<median≥median<median
Endpointsn = 13n = 12n = 7n = 6n = 10n = 16n = 53n = 43
CR/PR, n (%) 8 (61.5) 3 (23.1) 5 (71.4) 4 (66.7) 7 (70.0) 7 (43.8) 40 (75.5) 26 (60.5) 
SD/PD, n (%) 0 (0.0) 7 (58.3) 2 (28.6) 2 (33.3) 1 (10.0) 7 (43.8) 5 (9.4) 10 (23.3) 
Missing ORR, n (%) 5 (38.5) 2 (16.7) 0 (0.0) 0 (0.0) 2 (20.0) 2 (12.5) 8 (15.1) 7 (16.3) 
P (χ20.004 1.00 0.17 0.08 
OS, months (95% CI) 39.9 (11.4–68.3) 11.3 (0.0–25.9) 40.4 (22.3–58.6) 18.0 (13.8–22.2) 24.5 (n/a) 16.1 (0.9–31.2) 39.9 (35.0–44.7) 26.9 (16.5–37.3) 
HR (95% CI) 0.43 (0.16–1.16) 0.35 (0.09–1.26) 0.32 (0.11–0.87) 0.70 (0.44–1.11) 
P (log-rank) 0.10 0.11 0.03 0.13 
PFS, months (95% CI) 11.2 (6.9–15.5) 4.3 (3.3–5.3) 11.8 (3.6–20.0) 5.9 (3.6–8.1) 8.5 (2.3–14.7) 4.3 (3.7–5.0) 10.7 (9.2–12.1) 8.0 (6.5–9.5) 
HR (95% CI) 0.23 (0.08–0.64) 0.23 (0.05–0.95) 0.47 (0.21–1.09) 0.61 (0.40–0.93) 
P (log-rank) 0.005 0.04 0.08 0.02 
EGFR statusSidedness
EGFR positiveEGFR negativeRight-sided tumorsLeft-sided tumors
n = 25n = 13n = 26n = 96
AREGAREGAREGAREGAREGAREGAREGAREG
≥median<median≥median<median≥median<median≥median<median
Endpointsn = 13n = 12n = 7n = 6n = 10n = 16n = 53n = 43
CR/PR, n (%) 8 (61.5) 3 (23.1) 5 (71.4) 4 (66.7) 7 (70.0) 7 (43.8) 40 (75.5) 26 (60.5) 
SD/PD, n (%) 0 (0.0) 7 (58.3) 2 (28.6) 2 (33.3) 1 (10.0) 7 (43.8) 5 (9.4) 10 (23.3) 
Missing ORR, n (%) 5 (38.5) 2 (16.7) 0 (0.0) 0 (0.0) 2 (20.0) 2 (12.5) 8 (15.1) 7 (16.3) 
P (χ20.004 1.00 0.17 0.08 
OS, months (95% CI) 39.9 (11.4–68.3) 11.3 (0.0–25.9) 40.4 (22.3–58.6) 18.0 (13.8–22.2) 24.5 (n/a) 16.1 (0.9–31.2) 39.9 (35.0–44.7) 26.9 (16.5–37.3) 
HR (95% CI) 0.43 (0.16–1.16) 0.35 (0.09–1.26) 0.32 (0.11–0.87) 0.70 (0.44–1.11) 
P (log-rank) 0.10 0.11 0.03 0.13 
PFS, months (95% CI) 11.2 (6.9–15.5) 4.3 (3.3–5.3) 11.8 (3.6–20.0) 5.9 (3.6–8.1) 8.5 (2.3–14.7) 4.3 (3.7–5.0) 10.7 (9.2–12.1) 8.0 (6.5–9.5) 
HR (95% CI) 0.23 (0.08–0.64) 0.23 (0.05–0.95) 0.47 (0.21–1.09) 0.61 (0.40–0.93) 
P (log-rank) 0.005 0.04 0.08 0.02 

Note: Survival times are displayed as medians. P values < 0.05 are considered significant and are displayed bold.

Abbreviations: CR, complete remission; PD, progressive disease; PR, partial remission; SD, stable disease; WT, wild-type.

Efficacy of cetuximab in tumors with high AREG expression according to sidedness

122 patients with RAS, BRAF, and PIK3CA WT tumors (right-sided: n = 26; left-sided: n = 96) received cetuximab-containing regimens. In left-sided tumors with high AREG expression, OS was more pronounced and PFS was significantly longer. In tumors of the right colon, high AREG expression was significantly associated with higher OS [24.5 vs. 16.1 months, HR = 0.32 (95% CI, 0.11–0.87); P = 0.03] and numerically higher PFS (Table 3).

To our knowledge, this combined biomarker analysis represented the largest collective that was analyzed for AREG mRNA expression so far. High AREG expression was identified as a strong positive prognostic biomarker for OS, PFS, and ORR. Our analysis found a significant interaction of AREG expression with the use of cetuximab-containing regimens, whereas no effect was observed for bevacizumab or cytotoxic treatment alone. Finally, patients with right-sided RAS/BRAF WT tumors showed longer OS and PFS when treated with cetuximab in the presence of high AREG expression.

AREG and EREG are ligands of EGFR, which promote activation of the MAPK signaling pathway by ligand binding, which lead to proliferation and invasion of cancer cells (3, 4). Inhibition of EGFR signaling in RAS WT patients with high ligand levels resulted in significantly increased OS and PFS (9, 10, 12, 18, 19). Although a strong prognostic effect of high AREG expression was observed for treatment strategies with and without anti-EGFR antibodies (7, 13), we could confirm significant interaction only with cetuximab-containing regimens in our analysis for RAS WT tumors. However, effect sizes depended on the respective trial (CIOX or FIRE-3).

Biologically, autocrine loops were identified which promoted transcription and translation of EGFR ligands via MMP1, which itself was activated by ligand binding to EGFR (3). Tumors with elevated expression of EGFR ligands might represent a subgroup of tumors with an EGFR overstimulated activation of the MAPK pathway, for which EGFR signaling is vital for tumor growth. We therefore hypothesized that EGFR inhibition would substantially increase survival of patients with tumors showing high AREG expression.

In all patients, a major prognostic benefit on OS and PFS was observed irrespectively of treatment in tumors with high simultaneous AREG and EGFR expression compared to high AREG, but low EGFR expression. The subgroup of patients who had all-WT tumors and who were treated with cetuximab derived numerical benefit of high AREG expression irrespective of EGFR status, though. Thus, a prognostic, but not predictive, impact of EGFR expression in combination with AREG expression on treatment efficacy of cetuximab was observed in this analysis, analogous to previous results (20, 21). The validity of these findings might be biased by the limited sample size of 38 patients. Confirmation of the potential coherence of AREG and EGFR expression in a larger collective is therefore warranted and necessary.

Beyond activating mutations within the MAPK and PIK3CA-AKT signaling pathway, HER2/neu overexpression and amplification, respectively, were recognized as additional predictive biomarkers in the subset of patients with RAS WT tumors. HER2/neu-positive tumors occurred with a prevalence of 5% and were associated with worse response to anti-EGFR treatment than negative tumors (22–24). Our analysis did not adjust for this factor, as assessments of HER2/neu status across all trials were heterogeneous (FIRE-1: IHC; FIRE-3: next-generation sequencing; CIOX: no assessment). Therefore, HER2/neu expression and amplification could have additionally influenced outcome of patients treated with anti-EGFR antibodies.

EREG expression was associated with high variance and nonparametric data distribution in our analysis. In contrast to AREG expression, combined analysis of all trial collectives with consistent threshold values was not possible. However, we were able to reproduce improvements of OS and PFS in patients of FIRE-1 with high EREG expression (data not shown), confirming its prognostic role in patients treated with and without anti-EGFR antibodies (10–12, 25, 26). The choice of a consistent threshold value and its translation into clinical practice, however, might be impaired owing to the heterogeneity of mRNA expression in the respective collectives.

In addition to RAS status, primary tumor sidedness has been shown to be an important predictor of anti-EGFR treatment efficacy (1, 2, 27). Molecular characteristics such as higher prevalence of BRAF V600E mutations and high-grade microsatellite instability, higher tumor mutational burden, and enrichment of immune-like gene expression–based phenotypes accumulated in primary tumors of the right colon (28, 29). While RAS WT tumors in the left colon were associated with good response to anti-EGFR agents, survival of patients with right-sided RAS WT tumors was significantly worse (2, 27). An explanation for this phenomenon is still missing. Recommendations for the optimal treatment of RAS WT right-sided primary tumors differed across guidelines (ESMO: no stratification by sidedness; NCCN: no anti-EGFR antibodies in first-line treatment; ASCO: no anti-EGFR antibodies regardless of treatment line; refs. 30–32). Although the impact of primary tumor sidedness on anti-EGFR efficacy was retrospectively assessed in larger clinical trials, it should be respected for clinical decision-making in terms of the optimal first-line strategy in metastatic disease. Nevertheless, despite an accumulation of molecular characteristics, these tumors might still represent a heterogeneous subgroup, as some tumors of the right colon derived benefit from treatment with cetuximab (33). Further biomarkers are therefore needed to identify a subpopulation of RAS WT primary tumors in the right colon with susceptibility to anti-EGFR treatment.

Data on the distribution of EGFR ligands according to primary tumor side was limited owing to low sample sizes. However, elevated levels of EGFR ligand mRNA expression were observed more frequently in left- than right-sided tumors recently (8). We confirmed this observation by analyzing median expression per detailed primary tumor location. This finding might be associated with reported regulatory mechanisms of EGFR ligand expression, as hypermethylation, which is more frequently found in right-sided tumors (34), correlated with lower and demethylation with higher ligand expression, respectively (4–7). In our analysis, high AREG expression was evident in a small subset of ten patients with RAS WT right-sided primary tumors and was associated with significantly longer OS and numerically longer PFS, similar to left-sided primary tumors. Nevertheless, this analysis was formally not powered to investigate the hypothesis of different outcome according to the primary tumor side and AREG mRNA expression. Therefore, these findings should be interpreted with caution.

Taken together, we hypothesized that the favorable effect of high AREG expression on outcomes possibly depended on concomitant positivity of EGFR expression, but was independent of primary tumor side. Former inconsistencies in regard to the relevance of EGFR expression as a predictive biomarker for anti-EGFR treatment efficacy might be explained by differential EGFR ligand expression which may represent a missing link in this context (20, 21). Detailed characterization of MAPK activation by combined assessment of mutations, EGFR and AREG expression might therefore support the prediction of cetuximab efficacy regardless of the primary tumor location.

We were able to investigate a large collective of patients with mCRC for mRNA expression at different time points with minor methodologic bias. However, our results are limited by the fact that expression values could have been considered absolute (i.e., global minimal expression in FIRE-3 and high expression in CIOX and FIRE-1; Supplementary Fig. S2) and not trial-specific. This assumption, however, would have implicated major differences in mRNA expression between cohorts within the same entity without adjustment for methodologic bias such as RNA input. Second, we considered logarithmic values as real numbers for rescaling similar data of three trials to one scale. Although this method is vulnerable from a mathematical point of view, we showed comparable results for unnormalized data when trial-specific threshold values were taken into account. Therefore, data normalization produced at best minor variance of these results. Nevertheless, trial-specific data normalization failed for nonparametric data distribution such as EREG expression. Finally, baseline characteristics were slightly imbalanced for molecular parameters, as data for KRAS codon 59 and NRAS testing were missing in CIOX, and IHC EGFR expression was not assessed in FIRE-3.

In conclusion, we demonstrated high AREG mRNA expression at two threshold values as a strong prognostic biomarker for OS, PFS, and ORR in patients with mCRC. An interaction of AREG mRNA expression and anti-EGFR treatment was confirmed for OS and PFS in our analyses. Furthermore, positive prognostic effects of high AREG expression were observed in patients with IHC EGFR-positive tumors. The definition of patient populations with benefit from anti-EGFR treatment might be supported by the additional assessment of AREG and EGFR expression together with RAS/BRAF/PIK3CA mutational analyses. This approach may potentially be used to define patients with right-sided RAS WT tumors that benefit from EGFR antibodies. Prospective analyses are warranted to confirm the predictive role of high AREG expression (or surrogate biomarkers) for anti-EGFR treatment in mCRC.

A. Stahler reports grants from German Translational Research Consortium (DKTK; no. n/a), Weigand-Gravenhorst-Bohnewand-Fond (no. n/a, to D.P. Modest), Sanofi-Aventis (support of FIRE-1 trial), Pfizer (support of FIRE-1 and FIRE-3), and Merck KGaA Germany (support of CIOX and FIRE-3) during the conduct of the study; in addition, A. Stahler reports personal fees from Roche Pharma (travel reimbursement, honoraria for talks), AMGEN (travel reimbursement), and Merck (travel reimbursement) outside the submitted work. S. Stintzing reports personal fees from Amgen, Bayer, Pierre-Fabre, Lilly, Merck KGaA, Sanofi, Roche, MSD, Servier, Takeda, and Taiho outside the submitted work. D.P. Modest reports grants from Merck during the conduct of the study; in addition, D.P. Modest reports grants and personal fees from Merck, Amgen, Roche, and personal fees from Servier, Sanofi, BMS, MSD, PierreFabre, and Lilly outside the submitted work. I. Ricard reports personal fees from Roche outside the submitted work. C. Giessen-Jung reports personal fees from Roche (honoraria for talks and travel support) during the conduct of the study. A. Kiani reports personal fees from Merck (advisory board, honoraria) and Roche (advisory board, honoraria) outside the submitted work. T. Decker reports personal fees from Novartis (advisory boards) outside the submitted work. M. Moehler reports personal fees and nonfinancial support from Roche and Servier; grants and personal fees form MSD; grants, personal fees, and nonfinancial support from Merck Serono, BMS, and Bayer during the conduct of the study. A. Jung reports personal fees from Amgen, Pfizer, Archer DX, AstraZeneca, Biocartis, Boehringer Ingelheim, Bristol Myers Squibb (BMS), Merck-Serono, Merck Sharp & Dohme (MSD), Roche Pharma, Takeda, and Thermo Fisher outside the submitted work. V. Heinemann reports grants from Pfizer; grants and personal fees from Boehringer Ingelheim; personal fees from BMS, MSD, Novartis, Pierre Fabre, and Celgene; grants, personal fees and nonfinancial support from Merck, Servier, Bayer, Amgen, and Roche; grants and personal fees from Sanofi; and grants, personal fees, and nonfinancial support from Sirtex outside the submitted work. No potential conflicts of interest were disclosed by the other authors.

A. Stahler: Conceptualization, resources, data curation, formal analysis, investigation, visualization, methodology, writing-original draft, project administration, writing-review and editing. S. Stintzing: Conceptualization, resources, supervision, funding acquisition, project administration, writing-review and editing. D.P. Modest: Conceptualization, resources, supervision, funding acquisition, investigation, project administration, writing-review and editing. I. Ricard: Software, formal analysis, validation. C. Giessen-Jung: Resources, investigation, writing-review and editing. C. Kapaun: Resources, investigation, writing-review and editing. B. Ivanova: Resources, investigation, writing-review and editing. F. Kaiser: Resources, investigation, writing-review and editing. L. Fischer von Weikersthal: Resources, investigation, writing-review and editing. N. Moosmann: Resources, investigation, writing-review and editing. A. Schalhorn: Resources, investigation, writing-review and editing. M. Stauch: Resources, investigation, writing-review and editing. A. Kiani: Resources, investigation, writing-review and editing. S. Held: Data curation, software, formal analysis, validation, writing-review and editing. T. Decker: Resources, investigation, writing-review and editing. M. Moehler: Resources, investigation, writing-review and editing. J. Neumann: Resources, supervision, validation, investigation, methodology, writing-review and editing. T. Kirchner: Resources, supervision, investigation, project administration, writing-review and editing. A. Jung: Conceptualization, resources, supervision, investigation, methodology, writing-review and editing. V. Heinemann: Conceptualization, resources, supervision, funding acquisition, investigation, methodology, writing-original draft, project administration.

This work was supported by the German Translational Cancer Consortium (DKTK) and the Weigand-Bohnewand-Gravenhorst-Fonds (grant no. n/a, to D.P. Modest). Trials were financially supported by Sanofi-Aventis (FIRE-1); Pfizer (FIRE-1, FIRE-3); and Merck KGaA, Darmstadt, Germany (CIOX, FIRE-3). We thank Sabine Sagebiel-Kohler, Gertrud Lenz and Sebastian Heucke for supporting qRT-PCR analyses.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Holch
JW
,
Ricard
I
,
Stintzing
S
,
Modest
DP
,
Heinemann
V
. 
The relevance of primary tumour location in patients with metastatic colorectal cancer: a meta-analysis of first-line clinical trials
.
Eur J Cancer
2017
;
70
:
87
98
.
2.
Arnold
D
,
Lueza
B
,
Douillard
JY
,
Peeters
M
,
Lenz
HJ
,
Venook
A
, et al
Prognostic and predictive value of primary tumour side in patients with RAS wild-type metastatic colorectal cancer treated with chemotherapy and EGFR directed antibodies in six randomized trials
.
Ann Oncol
2017
;
28
:
1713
29
.
3.
Carpenter
BL
,
Chen
M
,
Knifley
T
,
Davis
KA
,
Harrison
SM
,
Stewart
RL
, et al
Integrin alpha6beta4 promotes autocrine epidermal growth factor receptor (EGFR) signaling to stimulate migration and invasion toward hepatocyte growth factor (HGF)
.
J Biol Chem
2015
;
290
:
27228
38
.
4.
Carpenter
BL
,
Liu
J
,
Qi
L
,
Wang
C
,
O'Connor
KL
. 
Integrin alpha6beta4 upregulates amphiregulin and epiregulin through base excision repair-mediated DNA demethylation and promotes genome-wide DNA hypomethylation
.
Sci Rep
2017
;
7
:
6174
.
5.
Lee
MS
,
McGuffey
EJ
,
Morris
JS
,
Manyam
G
,
Baladandayuthapani
V
,
Wei
W
, et al
Association of CpG island methylator phenotype and EREG/AREG methylation and expression in colorectal cancer
.
Br J Cancer
2016
;
114
:
1352
61
.
6.
Bormann
F
,
Stinzing
S
,
Tierling
S
,
Morkel
M
,
Markelova
MR
,
Walter
J
, et al
Epigenetic regulation of amphiregulin and epiregulin in colorectal cancer
.
Int J Cancer
2019
;
144
:
569
81
.
7.
Stintzing
S
,
Ivanova
B
,
Ricard
I
,
Jung
A
,
Kirchner
T
,
Tannapfel
A
, et al
Amphiregulin (AREG) and epiregulin (EREG) gene expression as predictor for overall survival (OS) in oxaliplatin/fluoropyrimidine plus bevacizumab treated mCRC patients-analysis of the phase III AIO KRK-0207 trial
.
Front Oncol
2018
;
8
:
474
.
8.
Kuramochi
H
,
Nakajima
GO
,
Hayashi
K
,
Araida
T
,
Yamamoto
M
. 
Amphiregulin/epiregulin mRNA expression and primary tumor location in colorectal cancer
.
Anticancer Res
2019
;
39
:
4729
36
.
9.
Khambata-Ford
S
,
Garrett
CR
,
Meropol
NJ
,
Basik
M
,
Harbison
CT
,
Wu
S
, et al
Expression of epiregulin and amphiregulin and K-ras mutation status predict disease control in metastatic colorectal cancer patients treated with cetuximab
.
J Clin Oncol
2007
;
25
:
3230
7
.
10.
Baker
JB
,
Dutta
D
,
Watson
D
,
Maddala
T
,
Munneke
BM
,
Shak
S
, et al
Tumour gene expression predicts response to cetuximab in patients with KRAS wild-type metastatic colorectal cancer
.
Br J Cancer
2011
;
104
:
488
95
.
11.
Jacobs
B
,
De Roock
W
,
Piessevaux
H
,
Van Oirbeek
R
,
Biesmans
B
,
De Schutter
J
, et al
Amphiregulin and epiregulin mRNA expression in primary tumors predicts outcome in metastatic colorectal cancer treated with cetuximab
.
J Clin Oncol
2009
;
27
:
5068
74
.
12.
Seligmann
JF
,
Elliott
F
,
Richman
SD
,
Jacobs
B
,
Hemmings
G
,
Brown
S
, et al
Combined epiregulin and amphiregulin expression levels as a predictive biomarker for panitumumab therapy benefit or lack of benefit in patients with RAS wild-type advanced colorectal cancer
.
JAMA Oncol
2016
;
2
:
633
42
.
13.
Stahler
A
,
Heinemann
V
,
Giessen-Jung
C
,
Crispin
A
,
Schalhorn
A
,
Stintzing
S
, et al
Influence of mRNA expression of epiregulin and amphiregulin on outcome of patients with metastatic colorectal cancer treated with 5-FU/LV plus irinotecan or irinotecan plus oxaliplatin as first-line treatment (FIRE 1-trial)
.
Int J Cancer
2016
;
138
:
739
46
.
14.
Fischer von Weikersthal
L
,
Schalhorn
A
,
Stauch
M
,
Quietzsch
D
,
Maubach
PA
,
Lambertz
H
, et al
Phase III trial of irinotecan plus infusional 5-fluorouracil/folinic acid versus irinotecan plus oxaliplatin as first-line treatment of advanced colorectal cancer
.
Eur J Cancer
2011
;
47
:
206
14
.
15.
Moosmann
N
,
von Weikersthal
LF
,
Vehling-Kaiser
U
,
Stauch
M
,
Hass
HG
,
Dietzfelbinger
H
, et al
Cetuximab plus capecitabine and irinotecan compared with cetuximab plus capecitabine and oxaliplatin as first-line treatment for patients with metastatic colorectal cancer: AIO KRK-0104–a randomized trial of the German AIO CRC study group
.
J Clin Oncol
2011
;
29
:
1050
8
.
16.
Heinemann
V
,
von Weikersthal
LF
,
Decker
T
,
Kiani
A
,
Vehling-Kaiser
U
,
Al-Batran
SE
, et al
FOLFIRI plus cetuximab versus FOLFIRI plus bevacizumab as first-line treatment for patients with metastatic colorectal cancer (FIRE-3): a randomised, open-label, phase 3 trial
.
Lancet Oncol
2014
;
15
:
1065
75
.
17.
Neumann
J
,
Wehweck
L
,
Maatz
S
,
Engel
J
,
Kirchner
T
,
Jung
A
. 
Alterations in the EGFR pathway coincide in colorectal cancer and impact on prognosis
.
Virchows Arch
2013
;
463
:
509
23
.
18.
Jonker
DJ
,
Karapetis
CS
,
Harbison
C
,
O'Callaghan
CJ
,
Tu
D
,
Simes
RJ
, et al
Epiregulin gene expression as a biomarker of benefit from cetuximab in the treatment of advanced colorectal cancer
.
Br J Cancer
2014
;
110
:
648
55
.
19.
Cushman
SM
,
Jiang
C
,
Hatch
AJ
,
Shterev
I
,
Sibley
AB
,
Niedzwiecki
D
, et al
Gene expression markers of efficacy and resistance to cetuximab treatment in metastatic colorectal cancer: results from CALGB 80203 (Alliance)
.
Clin Cancer Res
2015
;
21
:
1078
86
.
20.
Moroni
M
,
Veronese
S
,
Benvenuti
S
,
Marrapese
G
,
Sartore-Bianchi
A
,
Di Nicolantonio
F
, et al
Gene copy number for epidermal growth factor receptor (EGFR) and clinical response to antiEGFR treatment in colorectal cancer: a cohort study
.
Lancet Oncol
2005
;
6
:
279
86
.
21.
Sartore-Bianchi
A
,
Moroni
M
,
Veronese
S
,
Carnaghi
C
,
Bajetta
E
,
Luppi
G
, et al
Epidermal growth factor receptor gene copy number and clinical outcome of metastatic colorectal cancer treated with panitumumab
.
J Clin Oncol
2007
;
25
:
3238
45
.
22.
Sartore-Bianchi
A
,
Trusolino
L
,
Martino
C
,
Bencardino
K
,
Lonardi
S
,
Bergamo
F
, et al
Dual-targeted therapy with trastuzumab and lapatinib in treatment-refractory, KRAS codon 12/13 wild-type, HER2-positive metastatic colorectal cancer (HERACLES): a proof-of-concept, multicentre, open-label, phase 2 trial
.
Lancet Oncol
2016
;
17
:
738
46
.
23.
Meric-Bernstam
F
,
Hurwitz
H
,
Raghav
KPS
,
McWilliams
RR
,
Fakih
M
,
VanderWalde
A
, et al
Pertuzumab plus trastuzumab for HER2-amplified metastatic colorectal cancer (MyPathway): an updated report from a multicentre, open-label, phase 2a, multiple basket study
.
Lancet Oncol
2019
;
20
:
518
30
.
24.
Sartore-Bianchi
A
,
Amatu
A
,
Porcu
L
,
Ghezzi
S
,
Lonardi
S
,
Leone
F
, et al
HER2 positivity predicts unresponsiveness to EGFR-targeted treatment in metastatic colorectal cancer
.
Oncologist
2019
;
24
:
1395
402
.
25.
Pentheroudakis
G
,
Kotoula
V
,
De Roock
W
,
Kouvatseas
G
,
Papakostas
P
,
Makatsoris
T
, et al
Biomarkers of benefit from cetuximab-based therapy in metastatic colorectal cancer: interaction of EGFR ligand expression with RAS/RAF, PIK3CA genotypes
.
BMC Cancer
2013
;
13
:
49
.
26.
Llovet
P
,
Sastre
J
,
Ortega
JS
,
Bando
I
,
Ferrer
M
,
Garcia-Alfonso
P
, et al
Prognostic value of BRAF, PI3K, PTEN, EGFR copy number, amphiregulin and epiregulin status in patients with KRAS codon 12 wild-type metastatic colorectal cancer receiving first-line chemotherapy with anti-EGFR therapy
.
Mol Diagn Ther
2015
;
19
:
397
408
.
27.
Tejpar
S
,
Stintzing
S
,
Ciardiello
F
,
Tabernero
J
,
Van Cutsem
E
,
Beier
F
, et al
Prognostic and predictive relevance of primary tumor location in patients with RAS wild-type metastatic colorectal cancer: retrospective analyses of the CRYSTAL and FIRE-3 trials
.
JAMA Oncol
2017
;
3
:
194
201
.
28.
Salem
ME
,
Weinberg
BA
,
Xiu
J
,
El-Deiry
WS
,
Hwang
JJ
,
Gatalica
Z
, et al
Comparative molecular analyses of left-sided colon, right-sided colon, and rectal cancers
.
Oncotarget
2017
;
8
:
86356
68
.
29.
Guinney
J
,
Dienstmann
R
,
Wang
X
,
de Reynies
A
,
Schlicker
A
,
Soneson
C
, et al
The consensus molecular subtypes of colorectal cancer
.
Nat Med
2015
;
21
:
1350
6
.
30.
Van Cutsem
E
,
Cervantes
A
,
Adam
R
,
Sobrero
A
,
Van Krieken
JH
,
Aderka
D
, et al
ESMO consensus guidelines for the management of patients with metastatic colorectal cancer
.
Ann Oncol
2016
;
27
:
1386
422
.
31.
National Comprehensive Cancer Network
. 
Clinical practice guidelines in oncology (NCCN Guidelines) colon cancer version 4.2020
. 
2020
.
32.
Chiorean
EG
,
Nandakumar
G
,
Fadelu
T
,
Temin
S
,
Alarcon-Rozas
AE
,
Bejarano
S
, et al
Treatment of patients with late-stage colorectal cancer: ASCO Resource-Stratified Guideline
.
JCO Glob Oncol
2020
;
6
:
414
38
.
33.
Price
T
,
Shen
L
,
Ma
B
,
Esser
R
,
Chen
W
,
Gibbs
P
, et al
Phase II APEC trial: the impact of primary tumor side on outcomes of first-line cetuximab plus FOLFOX or FOLFIRI in patients with RAS wild-type metastatic colorectal cancer
.
Asia Pac J Clin Oncol
2019
;
15
:
225
30
.
34.
Weisenberger
DJ
,
Siegmund
KD
,
Campan
M
,
Young
J
,
Long
TI
,
Faasse
MA
, et al
CpG island methylator phenotype underlies sporadic microsatellite instability and is tightly associated with BRAF mutation in colorectal cancer
.
Nat Genet
2006
;
38
:
787
93
.

Supplementary data