Does anti-EGFR therapy improve outcome in advanced colorectal cancer? A systematic review and meta-analysis

Cancer Treatment Reviews 38 (2012) 618–625

Contents lists available at SciVerse ScienceDirect

Cancer Treatment Reviews journal homepage: www.elsevierhealth.com/journals/ctrv

Anti-Tumour Treatment

Does anti-EGFR therapy improve outcome in advanced colorectal cancer? A systematic review and meta-analysis Claire L. Vale a,⇑, Jayne F. Tierney a, David Fisher a, Richard A. Adams b, Richard Kaplan a, Timothy S. Maughan c, Mahesh K.B. Parmar a, Angela M. Meade a a b c

Medical Research Council Clinical Trials Unit, Aviation House, 125 Kingsway, London WC2B 6NH, UK Clinical Senior Lecturer/Honorary Consultant in Oncology, Cardiff University and Velindre Cancer Centre, Cardiff, UK Gray Institute for Radiation Oncology and Biology, University of Oxford, Oxford OX3 7DQ, UK

a r t i c l e

i n f o

Article history: Received 25 August 2011 Received in revised form 1 November 2011 Accepted 5 November 2011

Keywords: Colorectal cancer Anti-EGFR therapy Systematic review Meta-analysis

a b s t r a c t Background: Randomised controlled trials (RCTs) of anti-EGFR monoclonal antibodies (MAb) in patients with advanced colorectal cancer (aCRC) have reported conflicting results. Methods: A systematic review of RCTs comparing standard treatments ± anti-EGFR MAbs was conducted. Hazard ratios (HR) for progression-free (PFS) and overall survival (OS) were derived for patients with wild-type (WT) and mutant KRAS. Prespecified analyses were conducted for line of treatment, MAb used, chemotherapy regimen, and choice of fluouropyrimidine. Trials using bevacizumab on both arms were included in a sensitivity analysis. Results: Fourteen eligible RCTs were identified, with results by KRAS status available for ten RCTs. For third line treatment, the effect of anti-EGFR MAbs depended on KRAS status (interaction p < 0.00001), with a PFS benefit for patients with WT KRAS only (HR = 0.43, 95% CI 0.35–0.52, p < 0.00001). For first and second line treatment, the effect also appeared to depend on KRAS status (interaction p = 0.0003), again with the PFS benefit only for patients with WT KRAS (HR = 0.83, 95% CI 0.76–0.90, p < 0.0001). Differences between trial results (heterogeneity p = 0.02, I2 = 62%) were best explained by the fluouropyrimidine used, with PFS benefits confined to trials combining MAbs alongside 5FU-based chemotherapy (HR = 0.77, 95% CI 0.70–0.85, p < 0.00001). There was no evidence of a PFS benefit when MAbs were given with bevacizumab. Conclusions: For aCRC patients with WT KRAS, there are clear benefits of anti-EGFR MAbs in the third line and in the first and second line, when used alongside infusional 5FU-based regimens. However, there is no benefit for patients with KRAS mutations. Ó 2011 Elsevier Ltd. All rights reserved.

Introduction The monoclonal antibodies (MAb) cetuximab (Erbitux, ImClone, Bristol-Myers Squibb, Merck-Serono) and panitumumab (Vectibix, Amgen) are approved by the FDA for advanced colorectal cancer (aCRC). The results of prior reviews1,2 suggested that aCRC patients harbouring KRAS mutations at codons 12 and 13 do not benefit from anti-EGFR MAbs. However, these reviews were conducted prior to completion of pertinent randomised controlled trials (RCTs) and included data from retrospective or non-randomised studies. A more recent review3 did not include data from the most recent RCT4 and main results were based on all trials, irrespective of the line of treatment. No formal comparisons of the treatment effects by the MAb used or line of treatment were reported and results were limited to patients with WT KRAS. A meta-analysis based on the results of ⇑ Corresponding author. Tel.: +44 0 20 76704723. E-mail address: [email protected] (C.L. Vale). 0305-7372/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ctrv.2011.11.002

two RCTs has also been reported.5 Furthermore, none of these reviews included RCTs that did not report results by KRAS status or those that had included bevacizumab. We aimed to provide a comprehensive, unbiased synthesis of the effects of anti-EGFR MAbs for aCRC and to compare the effects of treatment in patients expressing WT KRAS with those expressing mutant KRAS oncogenes. Methods All methods were pre-specified in a protocol (available on request). Completed RCTs that aimed to randomise patients of any age with aCRC and compared anti-EGFR MAbs either alone, or combined with chemotherapy versus the same standard treatment alone were eligible. Trials that included bevacizumab on both arms were also eligible for inclusion in a separate sensitivity analysis. To identify all relevant RCTs, systematic searches6 were conducted of Medline, Cochrane CENTRAL, clinical trial registers (PDQ, ClinicalTrials.gov) and conference proceedings (ASCO, ASCO

619

C.L. Vale et al. / Cancer Treatment Reviews 38 (2012) 618–625

GI, World Gastrointestinal Congress, ESMO and ECCO). We also searched reference lists of the included RCTs and relevant clinical reviews and contacted experts in the field. The risk of bias of individual trials was assessed in terms of the randomisation sequence generation, allocation concealment, availability of complete outcome data or evidence of selective outcome reporting.6 We also investigated potential patient selection bias in the KRAS tested population by comparing the patient characteristics and results for this group of patients with those for all patients randomised. The primary outcome was progression free survival (PFS). Secondary outcomes were overall survival (OS) and best overall response rates (ORR). Data on design, treatment and patient characteristics were extracted from each trial report. Outcome data were extracted separately for patients with WT KRAS and KRAS mutations. Where insufficient data were available in trial publications and for unpublished trials, data were sought directly from investigators. Statistical analysis Main analyses Estimating effects within trials. For each outcome measure, we estimated effects separately for patients with WT KRAS and KRAS mutations. For PFS and OS, a hazard ratio (HR) was either extracted or estimated from the reported analyses.7–9 For PFS, the ratio of the HRs for KRAS WT and mutant subgroups were used to estimate the interaction between treatment effect and KRAS mutational status. For overall response rate, an odds ratio (OR) was calculated from the reported number of complete and partial responses and the total numbers of patients in each arm. Estimating effects across trials. For individual trials, HRs and ORs for the effect of treatment on each outcome measure and HRs for the interaction between treatment effect and KRAS status were combined across trials using the fixed effect model. Heterogeneity was assessed using v2 test and I2 statistic. The random effects model

was used to assess whether the results were robust to the choice of model. For trials in the first and second line, to explore whether trial characteristics impacted on the effect of anti-EGFR MAbs, we pre-specified analyses grouping trials according to the line of treatment (first or second line); the antibody used (cetuximab or panitumumab); whether patients were previously untreated or not (first or second line); the chemotherapy regimen (oxaliplatinbased or irinotecan based); and the choice of fluouropyrimidine (5FU or capecitabine). This sometimes necessitated dividing patients from within a trial into separate trial groups (Table 1). Differences between the HRs for trial groups were assessed using tests for interaction and F-ratios,10 focussing on the primary outcome of PFS but also carried out for OS to support or refute any patterns found. Interpretation was based on results from all trials together, or on trial groups if differences in the treatment effect between groups were found. Sensitivity analyses To assess whether the effects are consistent (or not) when bevacizumab is also used, we repeated the above analyses, including trials that compared chemotherapy (or supportive care) plus bevacizumab with or without anti-EGFR MAb. Results Included trials Searches identified 27 potentially eligible RCTs (Fig. A online) however, 13 of these did not meet the eligibility criteria or had not yet completed recruitment. Of the 14 eligible trials, three trials (SAKK4104,11 EPIC12 and EXPLORE13) have not reported by KRAS status (Table A online) and the investigators of one further trial14 were unable to supply results until it is published in full. Ten trials have reported results by KRAS status (Table 1), although two of

Table 1 Trial characteristics. Trial

Anti-EGFR antibodya

Chemotherapyb (dose per cycle)

First line trials comparing chemotherapy ± anti-EGFR MAb CRYSTAL19 Cetuximab IRI 180 mg/m2, 5FU 400 mg/m2 bolus + 2400 mg/m2 infusion, FA 400 mg/m2 OPUS20 Cetuximab OX 85 mg/m2, 5FU 400 mg/m2 bolus + 1200 mg/m2 infusion, FA 400 mg/m2 PRIME21 Panitumumab OX 85 mg/m2, 5FU 400 mg/m2 bolus + 1200 mg/m2 infusion, FA 400 mg/m2 COIN22 XELOX Cetuximab OX 130 mg/m2 d1, CAP 2000 mg/m2, d1–14 OxMdG Cetuximab OX 85 mg/m2, 5FU 400 mg/m2 bolus + 2400 mg/m2 infusion, FA 350 mg Nordic VII4 Cetuximab OX 85 mg/m2, 5FU bolus 1000 mg/m2, FA 120 mg/m2 Sub total Second line trials comparing chemotherapy ± anti-EGFR MAb Study 2005018123 Panitumumab IRI 180 mg/m2, 5FU 400 mg/m2 bolus + 2400 mg/m2 infusion, LV 400 mg/m2 (or 200 mg/m2 l-leucovorin) Sub total

CT cycle length (weeks)

Patients randomised (analysed)

2 2 2 3 2 2

1217 (1198) 344 (337) 1183 1076 554 379 4753 (4727)

2

1186 1186

Third line trials comparing best supportive care ± anti-EGFR MAb Cetuximab n/a NCIC CO1718 Study 2002040817 Panitumumab n/a Sub total Sensitivity analysis only First line trials comparing chemotherapy CAIRO215 Cetuximab PACCE16 Ox Panitumumab IRI Panitumumab Sub total Total a b

plus Bevacizumab ± anti-EGFR MAb OX 130 mg/m2, CAP 2000 mg/m2, d1–14, Bev 7.5 mg/kg OX/5FU/LV/Bev (various regimens) IRI/5FU/LV/Bev (various regimens)

Cetuximab: 400 mg/m2 d1, then 250 mg/m2 weekly; Panitumumab: 6.0 mg/kg every 2 weeks. Ox, oxaliplatin; Cap, capecitabine; 5FU, 5-fluourouracil; FA, folinic acid; IRI, irinotecan; Bev, bevacizumab; LV, leucovorin.

572 463 1035 (1035)

3 2 2

755 (736) 823 230 1808 (1789) 8782 (8737)

620

C.L. Vale et al. / Cancer Treatment Reviews 38 (2012) 618–625

these15,16 used bevacizumab in both arms and are therefore only included in the sensitivity analysis. Of the eight trials included in the main analyses, two gave MAbs in the third line setting.17,18 Six trials used MAbs either as first line4,19–22 or second line treatment23 in combination with chemotherapy regimens that included either oxaliplatin or irinotecan combined with 5FU or capecitabine (Table 1). Assessment of risk of bias Risk of bias was judged to be low or unclear for PFS for all 10 included trials (Table B online). Missing data largely resulted from unavailable tumour samples or inconclusive KRAS test results. Where reported, baseline characteristics for the subset of patients in whom KRAS status was assessed were similar to those for all randomised patients suggesting a low risk of patient selection bias in the KRAS tested population. Furthermore, PFS results for the two groups were comparable (Fig. B, online only). Included patients Kras status was available from 5996 patients (86% of the total randomised) from eight trials eligible for the main analyses and for 1385 (77% of the total randomised) from the two further trials eligible for the sensitivity analyses. The distribution of patients with WT KRAS and KRAS mutations was similar across all trials (Table 2). Overall around 60% of included patients were men with median age 61–63 years. The majority had good performance status (ECOG/WHO 0–1) and where data were available; around 20% had metastases confined to the liver only (Table 2). Trials of third line treatment Patients with wild type KRAS Large benefits were observed for third line treatment, with median PFS increasing from 2 to 4.7 months (HR = 0.43, 95% CI 0.35–0.52, p < 0.00001; heterogeneity p = 0.57, I2 = 0%; Fig. 1a). There was also evidence of an improvement in OS with MAbs (HR = 0.76, 95% CI 0.62–0.92, p = 0.006; Fig. 1b), albeit with differences between the results of the trials (heterogeneity p = 0.004,

I2 = 88%). The survival benefit was less clear when the random effects model was used (HR = 0.74, 95% CI 0.42–1.32, p = 0.31). Patients with KRAS mutations There was clear evidence that the effect of anti-EGFR MAbs on PFS depended on KRAS mutational status (interaction HR = 0.43, 95% CI 0.32–0.58, p < 0.00001). For patients with KRAS mutations, there was no evidence of improved PFS (HR = 0.99, 95% CI 0.80– 1.23, p = 0.93, Fig. 1a) or OS (HR = 1.00 (95% CI 0.80–1.26) p = 0.99, Fig. 1b) with anti-EGFR MAbs. Trials of first and second line treatment Patients with wild type KRAS There were benefits of anti-EGFR MAbs on PFS in trials of first or second line treatment (HR = 0.83, 95% CI 0.76–0.90, p < 0.0001), although there was evidence of differences in the effect of MAbs between these trials (heterogeneity p = 0.02, I2 = 62%). We found no evidence that these differences were explained by patients being previously treated or not, albeit that there was only one trial of second line treatment23 (first or second line treatment: interaction p = 0.21) or the MAb used (cetuximab or panitumumab; interaction p = 0.20). There is some suggestion that the differences in the effect maybe partially explained by the use of oxaliplatin-based or irinotecan-based chemotherapy (interaction p = 0.02; F-ratio = 2.50, p = 0.17, Table C online). However, the best explanation for these differences was the choice of fluouropyrimidine, (interaction p = 0.003, F-ratio = 5.84, p = 0.06, Fig. 2a, Table C online). Benefits appeared confined to trials using MAbs alongside 5-FU based chemotherapy (HR = 0.77, 95% CI 0.70–0.85, p < 0.00001). For these trials, there was an absolute improvement in median PFS from 7 to 9.5 months and no strong evidence of heterogeneity (p = 0.20, I2 = 32%). The capecitabine trial group comprised solely of a subset of patients from one trial.22 There was also an improvement in OS with MAbs (HR = 0.89, 95% CI 0.82–0.97, p = 0.01), with no strong evidence that the effect differed across trials (heterogeneity p = 0.25, I2 = 24%). There was no strong evidence that the treatment effect varied by the line of treatment (first or second line; interaction p = 0.60); the MAb used (cetuximab or panitumumab; interaction p = 0.34) or use of oxaliplatin or irinotecan chemotherapy (interaction p = 0.07).

Table 2 Patient characteristics for included trials. Trial

Patients randomised (n)

Patients with known KRAS status (% KRAS WT)

First line trials comparing chemotherapy ± anti-EGFR MAb CRYSTAL19 1217 1063 (63%) OPUS20 344 315 (57%) * PRIME21 1183 1096 (60%) COIN22 XELOX 1076 1294 (56%) OxMdG 554 Nordic VII26 379 324 (60%) Second line trials comparing chemotherapy ± anti-EGFR MAb * Study 1186 1083 (55%) 2005018123 Third line trials comparing best supportive care ± anti-EGFR MAb NCIC CO1718 572 394 (58%) Study 2002040817 463 427 (57%) Sensitivity analysis only First line trials comparing chemotherapy plus Bevacizumab ± anti-EGFR MAb CAIRO215 755 520 (60%) PACCE16 OX 823 865 (60%) IRI 230 *

Median age (range)

Sex (% male)

Performance status 0– 1 (%)

Liver only metastases (%)

61 61 62 63

(19–84) (24–82) (24–85) (22–87)

61% 54% 63% 66%

96% 91% 95% 92%

21% 27% 17% 23%

61 (24–75)

59%

96%

Unknown

61 (28–86)

61%

94%

17%

63 (27–88) 63 (27–83)

64% 64%

77% 86%

Unknown Unknown

62 (27–83) 61 (22–89)

60% 57%

100% 100%

Unknown Unknown

Patient characteristics based on KRAS tested population only (otherwise based on all randomised patients.

621

C.L. Vale et al. / Cancer Treatment Reviews 38 (2012) 618–625

Log Hazard Ratio KRAS Wild Type NCIC CO1718 Study 2002040817

-0.9163 -0.7985

SE

Hazard Ratio

0.1499 0.1406

Subtotal (95% CI)

KRAS Mutant NCIC CO1718 Study 2002040817

HR=0.43, p<0.00001 -0.01 -0.01

0.1568 0.1587

HR=0.99, p=0.93

Subtotal (95% CI)

Figure 1a

0.2

KRAS Wild type NCIC CO1718 Study 2002040817

-0.5978 -0.0101

0.5

1

5

0.1506 0.1384

HR=0.76, p=0.006

Subtotal (95% CI)

KRAS Mutant NCIC CO1718 Study 2002040817

2

-0.0202 0.0198

0.1713 0.1574

Subtotal (95% CI)

Figure 1b

HR=1.00, p=0.99 0.2 0.5 Favours treatment

1

2 5 Favours control

Trials of third line treatment: Figure 1a: Progression free survival KRAS wild type: HR=0.43 (95% CI 0.35-0.52) p<0.00001; Heterogeneity Chi² = 0.33, p = 0.57; I² = 0% KRAS mutant: HR=0.99 (95% CI 0.80-1.23) p=0.93; Heterogeneity Chi² = 0.00, p = 1.00; I² = 0%

Figure 1b: Overall survival KRAS wild type: HR=0.76 (95% CI 0.62-0.92) p=0.0006; Heterogeneity Chi² = 8.26, p = 0.004; I² = 88% KRAS mutant: HR=1.00 (95% CI 0.80-1.26) p=0.99; Heterogeneity Chi² = 0.03, p = 0.86; I² = 0%

Fig. 1. Trials of third line treatment: (a) progression free survival and (b) overall survival.

However, the survival benefit appeared confined to trials using 5FU (HR = 0.86, 95% CI 0.78–0.95, p = 0.002; heterogeneity p = 0.61, I2 = 0; interaction p = 0.04, Fig. 2b), albeit the second line treatment and capecitabine trial groups consisted of a single trial23 or a subset of patients from one trial.22 Data on best overall response were available from six trials, including 2996 patients with WT KRAS, with 788 complete or partial responses in the treatment arm (53%) and 611 in the control arm (40%). While there was a 77% improvement with MAbs (OR = 1.77, 95% CI 1.52–2.06, p < 0.00001), effects varied considerably across trials (heterogeneity p < 0.0001, I2 = 81%). Patients with KRAS mutations There was evidence that the effect of anti-EGFR MAbs on PFS depended on KRAS mutational status (interaction HR = 0.78, 95% CI 0.68–0.89, p = 0.0003), however this relationship was inconsistent across trials (heterogeneity p = 0.001, I2 = 73%). There was no evidence of a PFS benefit with MAbs in the first or second line (Fig. 3a, HR = 1.06, 95% CI 0.96–1.17, p = 0.25) however there was considerable variation between the trial results (heterogeneity p = 0.01, I2 = 63%). Whilst this seemed to be best explained by the line of treatment, there was only one trial of second line treatment (first line or second line; interaction p = 0.03). There was no OS benefit in patients with KRAS mutations (HR = 1.04, 95% CI 0.95–1.15, p = 0.39, Fig. 3b), or any suggestion of variation between the trial results (heterogeneity p = 0.59, I2 = 0%). There was a tendency towards poorer overall response rates with anti-EGFR MAbs in first and second line treatment (OR = 1.19, 95% CI 0.99–1.43) which was not conventionally significant (p = 0.06). There was no evidence of differences in response rates between trials (heterogeneity p = 0.67, I2 = 0%) nor any suggestion of differences in the effect between first and second line treatment (interaction p = 0.72).

Sensitivity analysis of bevacizumab in patients with WT KRAS Two trials, including 1808 randomised patients, used bevacizumab in both of their arms15,16 (Table 1). Including data on 833 patients with WT KRAS from these trials (Table 2), alongside the six trials of MAb without bevacizumab in first or second line treatment, led to more variation between the trial results for PFS (heterogeneity p < 0.00001, I2 = 75%). Furthermore, there was clear evidence of an interaction between the effect of anti-EGFR MAbs on PFS and use of bevacizumab (interaction p = 0.00002), with poorer PFS associated with the addition of bevacizumab (HR = 1.27, 95% CI 1.06–1.51, p = 0.008, Fig. 4a). Only one of the two trials that used bevacizumab16 reported sufficient data for inclusion in the meta-analysis for OS (Fig. 4b). Discussion We aimed to bring together the results of all eligible RCTs to estimate the effect of adding anti-EGFR MAb therapy to standard treatments in aCRC. Although 14 eligible trials were identified, one trial has no results available, three trials have not reported results by KRAS status and two further studies are only eligible for inclusion in our sensitivity analyses. Therefore our main findings for the effects of MAbs in patients with WT or mutant KRAS are limited to eight trials. Reassuringly, where data were available for all randomised patients and patients for whom KRAS was tested, we saw no evidence of selection bias for those patients receiving KRAS tests (Fig. B, online only). Distribution of randomised patients with WT and mutant KRAS in each study was similar, such that in total, 58% of patients had WT KRAS. Overall there was a difference in the effect of treatment on PFS between patients with WT and mutant KRAS, with benefits of MAbs confined to patients with WT KRAS, as anticipated.

622

C.L. Vale et al. / Cancer Treatment Reviews 38 (2012) 618–625

Log Hazard Ratio Trials using 5FU Crystal19 OPUS20 PRIME21 Study 2005018123 COIN OxMdG22 Nordic VII 4 Subtotal (95% CI)

Hazard Ratio

SE

-0.3624 -0.5621 -0.2231 -0.3147 -0.2634 0.0677

0.1124 0.2366 0.0982 0.1077 0.1379 0.1549

HR=0.77, p<0.00001

Trials using Capecitabine COIN Xelox 22 Subtotal (95% CI)

0.0561

0.0961

HR=1.06, p=0.56 HR=0.83, p<0.0001

Total (95% CI) 0.2

Figure 2a Trials using 5FU Crystal19 OPUS20 PRIME21 Study 2005018123 COIN OxMdG22 Nordic VII 4 Subtotal (95% CI)

-0.2282 -0.1567 -0.1863 -0.1625 -0.0758 0.131

0.5

1

2

5

0.088 0.1813 0.1072 0.101 0.1516 0.1784

HR=0.86, p=0.001

Trials using Capecitabine COIN Xelox 22

0.0901

0.109

Subtotal (95% CI)

HR=1.09, p=0.41

Total (95% CI)

HR=0.89, p=0.01

Figure 2b

0.2 0.5 Favours treatment

1

2 Favours control

5

Trials of first and second line treatment (KRAS wild type patients only): Figure 2a: Progression free survival Trials using 5FU: HR=0.77 (95% CI 0.70 – 0.85) p<0.00001 ; Test for heterogeneity: Chi² = 7.33, p= 0.20, I² = 32% Trials using capecitabine: HR=1.06 (95% CI 0.88 – 1.28) p=0.56 Overall: HR=0.83 (95% CI 0.76 – 0.90) p<0.0001; Te st for heterogeneity: Chi² = 15.89, p = 0.01, I² = 62%

Figure 2b: Overall survival Trials using 5FU: HR=0.86 (95% CI 0.78 – 0.94) p=0. 001; Test for heterogeneity: Chi² = 3.63, p= 0.60), I² = 0% Trials using capecitabine: HR=1.09 (95% CI 0.88 – 1.35) p=0.41 Overall: HR=0.89 (95% CI 0.82 – 0.97) p=0.01; Te st for heterogeneity: Chi² = 7.88, p = 0.25, I² = 24%

Fig. 2. Trials of first and second line treatment (KRAS wild type patients only): (a) progression free survival and (b) overall survival.

Log Hazard Ratio

SE

0.1579 0.6043 0.2546 -0.1625 0.0439 0.0751 -0.3425

0.1414 0.2618 0.1131 0.1133 0.1535 0.1051 0.1844

Hazard Ratio

Progression free survival Crystal19 OPUS20 PRIME21 Study 2005018123 COIN OxMdG22 COIN Xelox 22 Nordic VII 4

HR=1.06, p=0.25

Total (95% CI)

Figure 3a

0.2

0.5

1

2

5

Overall Survival Crystal19 OPUS20 PRIME21 Study 2005018123 COIN OxMdG22 COIN Xelox 22 Nordic VII 4

0.0344 0.2546 0.2151 -0.0619 0.0144 -0.0329 0.0296

0.1101 0.1992 0.1202 0.1057 0.1659 0.1152 0.2135

HR=1.04, p=0.39

Total (95% CI)

Figure 3b

0.2 0.5 Favours treatment

1

2 Favours control

5

Patients with KRAS mutations: Figure 3a: Progression free survival HR= 1.06 (95% CI 0.96 – 1.17) p=0.25; Test for heterogeneity: Chi² = 16.41, p = 0.01, I² = 63.4%

Figure 3b: Overall survival HR=1.04 (95% CI 0.95 – 1.15) p=0.39; Test for heterogeneity Chi² = 4.64, p = 0.59; I² = 0%

Fig. 3. Patients with KRAS mutations: (a) progression free survival and (b) overall survival.

623

C.L. Vale et al. / Cancer Treatment Reviews 38 (2012) 618–625 Log Hazard Ratio Trials not using bevacizumab Crystal19 OPUS20 PRIME21 Study 2005018123 COIN OxMdG22 COIN Xelox22 Nordic VII4

-0.3624 -0.5621 -0.2231 -0.3147 -0.2634 0.0561 0.0677

Hazard Ratio

SE 0.1124 0.2366 0.0982 0.1077 0.1379 0.0961 0.1549

HR=0.83, p<0.0001

Subtotal (95% CI) Trials using bevacizumab CAIRO215 PACCE Iri16 PACCE Ox16

0.1458 0.4055 0.3075

0.129 0.3096 0.1357

Figure 4a Trials not using bevacizumab Crystal19 OPUS20 PRIME21 Study 2005018123 COIN OxMdG22 COIN Xelox22 Nordic VII4

HR=1.27, p=0.008 0.2

-0.2282 -0.1567 -0.1863 -0.1625 -0.0758 0.0901 0.131

0.5

1

2

0.088 0.1813 0.1072 0.101 0.1516 0.109 0.1784 HR=0.89, p=0.010

Subtotal (95% CI) Trials using bevacizumab PACCE Iri16 PACCE Ox16

5

0.2469 0.6366

0.4775 0.5652

HR=1.51, p=0.26

Subtotal (95% CI)

Figure 4b

0.2 0.5 Favours treatment

1

2 Favours control

5

Sensitivity analysis of trials including bevacizumab for patients with wild type KRAS Figure 4a: Progression free survival Trials not using bevacizumab: HR=0.83 (95% CI 0.76 – 0.90) p<0.0001; Test for heterogeneity: Chi² = 15.89, p= 0.01, I² = 62% Trials using bevacizumab: HR=1.27 (95% CI 1.06 – 1.51) p=0.008; Test for heterogeneity: Chi² = 1.07, p= 0.59, I² = 0% Test for subgroup differences: Interaction Chi² = 18.37, p< 0.0001

Figure 4b: Overall Survival Trials not using bevacizumab: HR=0.89 (95% CI 0.82 – 0.97) p=0.010; Test for heterogeneity: Chi² = 7.88, p= 0.25, I² = 24% Trials using bevacizumab: HR=1.51(95% CI 0.74 – 3.08) p=0.26; Test for heterogeneity: Chi² = 0.28 p= 0.60, I² = 0% Test for subgroup differences: Interaction Chi² = 2.04, p =0.15

Fig. 4. Sensitivity analysis of trials including bevacizumab for patients with wild type KRAS: (a) progression free survival and (b) overall Survival.

However, the relationship between this benefit and KRAS status varied between trials. Within the group of patients with WT KRAS, the largest relative benefits were observed in third line treatment, although the reasons for this are unclear. The absolute improvement in median PFS from 2 to 4.7 months may be considered worthwhile by many clinicians and patients. For trials of MAbs in earlier lines of treatment, the treatment effects were inconsistent across trials. Variations were explained by the fluouropyrimidine used and to a lesser degree, use of oxaliplatin or irinotecan based chemotherapy. Moreover, when we adjusted for the choice of fluoropyrimidine in an exploratory multivariate Cox regression, there was no significant difference in effect between trials using oxaliplatin or irinotecan (data not shown). Thus, the benefits for PFS and OS were limited to trials combining MAbs with any 5FU-based chemotherapy, with an absolute improvement in PFS from 7 to 9.5 months. It remains unclear from our results whether MAbs are effective when combined with capecitabine-based chemotherapy, since only a subset of patients from one trial in this review used capecitabine.22 Although there has been some suggestion that capecitabine may be less effective than 5FU in first line treatment of aCRC24–26 most previous RCTs and meta-analyses have reported similar outcomes with both fluouropyrimidines.27–32 Interestingly, results of one trial4 in this review that used bolus (rather than infusional) 5FU were more akin to the trial using capecitabine.22 Whilst it is tempting to speculate,

the reasons for these similarities maybe unrelated to the chemotherapy employed. Two trials that used bevacizumab and chemotherapy15,16 showed no benefit of adding anti-EGFR MAbs in patients with WT KRAS. Theoretically, targeting both EGFR and VEGF pathways should increase anti-tumour activity,33 however, our results may suggest this is potentially detrimental for patients. One hypothesis is that receiving the two antibodies together increases toxicity, potentially causing dose delays, therefore reducing dose intensity and potentially increasing mortality.16 In the trials where MAbs were given as third-line treatment, there was also a survival benefit for patients with WT KRAS, but effects differed between the trials. This may be a result of differing extents of patient crossover on relapse. In one of these trials,18 only 13 patients on the control arm (<7%) crossed over to receive antiEGFR MAb whereas in the other trial,17 the majority of patients with WT KRAS (76%) crossed over. Results of the latter trial17 perhaps suggest similar survival benefits when anti-EGFR MAbs are given on relapse, although other explanations cannot be ruled out. By contrast, the extent of patient crossover in the trials of first or second line treatment is unclear, however, it may be that the effect of MAbs on survival in these trials is diluted because some patients on the control arms received MAbs on relapse. For ORR, whilst there is a large and significant benefit observed overall in patients with WT KRAS, there is large variation in the results of

624

C.L. Vale et al. / Cancer Treatment Reviews 38 (2012) 618–625

individual trials. This may reflect differences in the classification of ORR between trials. Whilst we have data on KRAS mutational status for almost 6000 randomised patients in total, it was not possible to assess the impact of anti-EGFR therapy by KRAS status for all eligible trials. Three trials11–13 that randomised 1474 patients (Table A, online only) reported results based on all randomised patients only, presumably due to a lack of tumour sample availability. Reassuringly, PFS results for all randomised patients from these trials are in keeping with those for patients with WT KRAS from our main analysis (Fig. C, online only). Therefore we are fairly confident that KRAS subgroup data from these three trials would be unlikely to alter our conclusions regarding anti-EGFR MAbs in patients with KRAS WT. However, KRAS subgroup data from two of these trials12,13 of second line treatment would have enabled us to be more confident about our findings in the second line setting, currently limited to the results of a single trial.23 Final results from one further study of anti-EGFR MAb combined with chemotherapy are as yet unreported.14 However, it closed early due to poor recruitment and is underpowered to detect differences in survival or progression-free survival. Consequently, it is likely to have only minimal impact on the results of this analysis. Whilst it has not been possible to explore the potential predictive role of specific KRAS point mutations in terms of treatment benefits from the use of anti-EGFR MAbs, the identification of additional point mutations in KRAS and other key signaling molecules downstream of the EGFR is likely to reduce the proportion of patients with aCRC found to benefit from these treatments.34 Accurate and reliable identification of those unlikely to benefit from these treatments is vital and novel therapeutic approaches should be sought for them. It would be necessary to collect individual patient data from all relevant trials to further explore the impact of these additional mutations on treatment effect in the context of a meta-analysis. For aCRC patients with WT KRAS, our results have demonstrated significant and worthwhile benefits from the use of antiEGFR MAbs in third line therapy and clear, albeit potentially more modest, benefits when MAbs are given alongside 5FU-based chemotherapy in first and second line treatment. Some uncertainties remain as to whether this benefit applies when therapy does not include an infusional 5FU component. However, when anti-EGFR MAbs are combined with bevacizumab and chemotherapy as 1st or 2nd line treatment, PFS is reduced. Furthermore, this systematic review clearly demonstrates that the benefits of anti-EGFR MAbs are confined to aCRC patients with WT KRAS, with no such evidence of benefits for patients with KRAS mutations.

Role of the funding source The UK Medical Research Council (MRC) funded this research. It had no input in writing this manuscript or in the decision to submit this work for publication.

Authors and contributors AMM and MKBP had the original idea for this review, which was developed by CV and JT, who also conducted the literature searches and devised the protocol, with input and advice from all authors. CV, JT and DF conducted the analyses. All authors were involved in the interpretation of the results. CV and JT drafted the manuscript with input and advice from all authors.

Conflicts of interest None of the authors have reported financial conflicts of interest in relation to this submission. RAA received consultancy fees, speakers payments and travel expenses from both Roche and Merck Serono in relation to activities outside this submission. TM received grants for his institution from Cancer Research UK, Merck Serono and Immatics, as well as personal payments for consultancy, travel and speakers payments from Merck Serono relating to the MRC COIN trial as well as in relation to activities outside this submission.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ctrv.2011.11.002.

References 1. Linardou H, Dahabreh IJ, Kanaloupiti D, et al. Assessment of somatic k-RAS mutations as a mechanism associated with resistance to EGFR-targeted agents: a systematic review and meta-analysis of studies in advanced non-small-cell lung cancer and metastatic colorectal cancer. Lancet Oncol 2008;9(10):962–72. 2. 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 antiepidermal growth factor receptor monoclonal antibody therapy. J Clin Oncol 2009;27(12):2091–6. 3. Petrelli F, Borgonovo K, Cabiddu M, Ghilardi M, Barni S. Cetuximab and panitumumab in KRAS wild-type colorectal cancer: a meta-analysis. Int J Colorectal Dis 2011; doi: 10.1007/s00384-011-1149-0. 4. Tveit K, Guren T, Glimelius B, et al. Randomized phase III study of 5fluorouracil/folinate/oxaliplatin given continuously or intermittently with or without cetuximab as first-line treatment of metastatic colorectal cancer: the Nordic VII study (NCT00145314) by the Nordic Colorectal Cancer Biomodulation Group. Ann Oncol 2010;21(8(supplement)):LBA20. 5. Bokemeyer C, Kohne CH, Rougier P, Stroh C, Schlichting M, Van Cutsem E. Cetuximab with chemotherapy (CT) as first-line treatment for metastatic colorectal cancer (mCRC): analysis of the CRYSTAL and OPUS studies according to KRAS and BRAF mutation status. J Clini Oncol 2010;28(15S):3506. 6. Higgins JPT, Green SJ, editors. Cochrane handbook for systematic reviews of interventions. Chichester: John Wiley & Sons Ltd.; 2008. 7. Parmar MKB, Torri V, Stewart L. Extracting summary statistics to perform metaanalyses of the published literature for survival endpoints. Stat Med 1998;17:2815–34. 8. Tierney JF, Stewart LA, Ghersi D, Burdett S, Sydes MR. Practical methods for incorporating summary time-to-event data into meta-analysis. Trials 2007;8(1):16. 9. Williamson PR, Tudur Smith C, Hutton JL, Marson AG. Aggregate data metaanalysis with time-to-event outcomes. Stat Med 2002;21:3337–51. 10. The International Collaborative Ovarian Neoplasm (ICON) Group. Paclitaxel plus carboplatin versus standard chemotherapy with either single-agent carboplatin or cyclophosphomide, doxorubicin, and cisplatin in women with ovarian cancer: the ICON3 randomised trial. Lancet 2002;360(9332):505–15. 11. Borner M, Koeberle D, Von Moos R, et al. Adding cetuximab to capecitabine plus oxaliplatin (XELOX) in first-line treatment of metastatic colorectal cancer: a randomized phase II trial of the Swiss Group for Clinical Cancer Research SAKK. Ann Oncol 2008;19(7):1288–92. 12. Sobrero AF, Maurel J, Fehrenbacher L, et al. EPIC: phase III trial of cetuximab plus irinotecan after fluoropyrimidine and oxaliplatin failure in patients with metastatic colorectal cancer. J Clin Oncol 2008;26(14):2311–9. 13. Polikoff J, Mitchell EP, Badarinath S, et al. Erbitux (cetuximab) plus FOLFOX for colorectal cancer (EXPLORE): preliminary efficacy analysis of a randomized phase III trial. J Clin Oncol 2005;23(16S):3574. 14. Venook A, Niedzwiecki D, Hollis D, et al. Phase III study of irinotecan/5FU/LV (FOLFIRI) or oxaliplatin/5FU/LV (FOLFOX) ± cetuximab for patients with untreated metastatic adenocarcinoma of the colon or rectum (MCRC): CALGB 80203 preliminary results. J Clin Oncol 2006;24(18S):3509. 15. Tol J, Koopman M, Cats A, et al. Chemotherapy, bevacizumab, and cetuximab in metastatic colorectal cancer. N Engl J Med 2009;360(6):563–72. 16. Hecht JR, Mitchell E, Chidiac T, et al. A randomized phase IIIB trial of chemotherapy, bevacizumab, and panitumumab compared with chemotherapy and bevacizumab alone for metastatic colorectal cancer. J Clin Oncol 2009;27(5):672–80. 17. Van Cutsem E, Peeters M, Siena S, et al. Open-label phase III trial of panitumumab plus best supportive care compared with best supportive care alone in patients with chemotherapy-refractory metastatic colorectal cancer. J Clin Oncol 2007;25(13):1658–64.

C.L. Vale et al. / Cancer Treatment Reviews 38 (2012) 618–625 18. 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(17):1757–65. 19. Van Cutsem E, Lang I, D’haens G, et al. KRAS status and efficacy in the first line treatment of patients with metastatic colorectal cancer (MCRC) treated with FOLFIRI with or without cetuximab: the CRYSTAL experience. J Clin Oncol 2008;26(15S):2. 20. Bokemeyer C, Bondarenko I, Hartmann JT, et al. KRAS stsus and efficacy of first line treatment of patients with metastatic colorectal cancer (MCRC) with FOLFOX with or without cetuximab: the OPUS experience. J Clin Oncol 2008;26(15S):4000. 21. Douillard JY, Siena S, Cassidy J, et al. Randomized phase 3 study of panitumumab with FOLFOX4 compared to FOLFOX4 alone as 1st-line treatment (tx) for metastatic colorectal cancer (mCRC): the PRIME trial. Eur J Cancer Suppl 2009;7(3):6. 22. Maughan TS, Adams R, Smith CG, et al. Addition of cetuximab to oxaliplatinbased first-line combination chemotherapy for treatment of advanced colorectal cancer: results of the randomised phase 3 MRC COIN trial. Lancet 2011;377(9783):2103–14. 23. Peeters M, Price T, Hotko Y, et al. Randomized phase III study of panitumumab with fluorouracil, leucovorin, and irinotecan (FOLFIRI) compared with FOLFIRI alone as second-line treatment in patients with metastatic colorectal cancer. J Clin Oncol 2010;28(31):4706–13. 24. Diaz-Rubio E, Tabernero J, Gomez-Espana A, et al. Phase III study of capecitabine plus oxaliplatin compared with continuous-infusion fluorouracil plus oxaliplatin as first-line therapy in metastatic colorectal cancer: final report of the Spanish Cooperative Group for the Treatment of Digestive Tumors Trial. J Clin Oncol 2007;25(27):4224–30. 25. Porschen R, Arkenau HT, Kubicka S, et al. Phase III study of capecitabine plus oxaliplatin compared with fluorouracil and leucovorin plus oxaliplatin in metastatic colorectal cancer: a final report of the AIO Colorectal Study Group. J Clin Oncol 2007;25(27):4217–23. 26. Fuchs CS, Marshall J, Mitchell E, et al. Randomized, controlled trial of irinotecan plus infusional, bolus, or oral fluoropyrimidines in first-line treatment of

27.

28.

29.

30.

31.

32.

33. 34.

625

metastatic colorectal cancer: results from the BICC-C Study. J Clin Oncol 2007;25(30):4779–86. Arkenau HT, Arnold D, Cassidy J, et al. Efficacy of oxaliplatin plus capecitabine or infusional fluorouracil/leucovorin in patients with metastatic colorectal cancer: a pooled analysis of randomized trials. J Clin Oncol 2008;26(36):5910–7. Cao Y, Liao C, Tan A, Liu L, Mo Z, Gao F. Capecitabine plus oxaliplatin versus fluorouracil plus oxaliplatin as first line treatment for metastatic colorectal cancer: meta-analysis of six randomised trials. Colorectal Dis 2010;12(1):16–23. Comella P, Massidda B, Filippelli G, et al. Randomised trial comparing biweekly oxaliplatin plus oral capecitabine versus oxaliplatin plus i.v. bolus fluorouracil/ leucovorin in metastatic colorectal cancer patients: results of the Southern Italy Cooperative Oncology study 0401. J Cancer Res Clin Oncol 2009;135(2):217–26. Cassidy J, Clarke S, Diaz-Rubio E, et al. Randomized phase III study of capecitabine plus oxaliplatin compared with fluorouracil/folinic acid plus oxaliplatin as first-line therapy for metastatic colorectal cancer. J Clin Oncol 2008;26(12):2006–12. Martoni AA, Pinto C, Di Fabio F, et al. Capecitabine plus oxaliplatin (xelox) versus protracted 5-fluorouracil venous infusion plus oxaliplatin (pvifox) as first-line treatment in advanced colorectal cancer: a GOAM phase II randomised study (FOCA trial). Eur J Cancer 2006;42(18):3161–8. Rothenberg ML, Cox JV, Butts C, et al. Capecitabine plus oxaliplatin (XELOX) versus 5-fluorouracil/folinic acid plus oxaliplatin (FOLFOX-4) as second-line therapy in metastatic colorectal cancer: a randomized phase III noninferiority study. Ann Oncol 2008;19(10):1720–6. Tabernero J. The role of VEGF and EGFR inhibition: implications for combining anti-VEGF and anti-EGFR agents. Mol Cancer Res 2007;5(3):203–20. 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(8):753–62.