KRAS wild-type state predicts survival and is associated to early radiological response in metastatic colorectal cancer treated with cetuximab

Transcription

1 original article Annals of Oncology 19: , 2008 doi: /annonc/mdm496 Published online 12 November 2007 KRAS wild-type state predicts survival and is associated to early radiological response in metastatic colorectal cancer treated with cetuximab W. De Roock 1, H. Piessevaux 2, J. De Schutter 1, M. Janssens 3, G. De Hertogh 4, N. Personeni 5, B. Biesmans 1, J.-L. Van Laethem 6, M. Peeters 7, Y. Humblet 8, E. Van Cutsem 5 & S. Tejpar 1,5 * 1 Center for Human Genetics, Katholieke Universiteit Leuven, Leuven; 2 Service de Gastro-ente rologie, Cliniques universitaires Saint-Luc, Universite catholique de Louvain, Brussels; 3 Independent biostatistician, Mechelen; 4 Department of Morphology and Molecular Pathology University Hospital Gasthuisberg, Katholieke Universiteit Leuven, Leuven; 5 Digestive Oncology Unit, University Hospital Gasthuisberg, Katholieke Universiteit Leuven, Leuven; 6 Department of Gastroenterology, Gastro- Intestinal Cancer Unit, Erasme University Hospital, Universite Libre de Bruxelles, Brussels; 7 Digestive Oncology Unit, Ghent University Hospital, Ghent; 8 Clinique des Pathologies Tumorales du Coˆlon et du Rectum, Centre du Cancer, Cliniques universitaires Saint-Luc, Universite catholique de Louvain, Brussels, Belgium Received 26 June 2007; revised 12 September 2007; accepted 14 September 2007 original article Background: KRAS mutation status is a candidate marker for predicting survival in patients with metastatic colorectal cancer (mcrc) treated with cetuximab (CTX). Patients and methods: We studied the KRAS mutation status of 113 patients with irinotecan refractory mcrc treated with CTX in clinical trials. A predictive model for objective response (OR), progression-free survival (PFS) and overall survival (OS) was constructed using logistic and Cox regression. Results: OR was seen in 27 of 66 KRAS wild-type (WT) patients versus 0 of 42 in KRAS mutants. Median OS was significantly better in KRAS WT versus mutants (43.0 versus 27.3 weeks; P = 0.020). Decrease in tumor sizes was significantly larger at all time points in WT patients. KRAS WT patients with an initial relative decrease of tumor size >9.66% at week 6 had a significantly better median OS compared with all other patients (74.9 versus 30.6 weeks; P = ). Within KRAS WT patients OS was significantly better in patients with an initial decrease compared with those without [median OS: 74.9 versus 30.6 weeks (P = )]. Conclusions: KRAS WT status is associated to survival benefit in CTX treated mcrc. This benefit is even more pronounced in those patients with early radiological response. These characteristics may be exploited for response prediction. Key words: cetuximab, colorectal cancer, EGFR, KRAS, survival introduction A recent advance in oncology is the targeting of the epidermal growth factor receptor (EGFR) in the treatment of many tumor types. Cetuximab (CTX) (ErbituxÒ, Merck KGaA, Darmstadt, Germany) is a chimeric immunoglobulin G1 monoclonal antibody which binds the EGFR with high affinity and competitively inhibits ligand binding [1]. This prevents activation of downstream signalling pathways such as the PI3K/ Akt, RAS/Erk and STAT pathways, resulting in the inhibition of cellular proliferation and in the induction of apoptosis. However, for many tumor types including metastatic colorectal cancer (mcrc), it is not clear what proportion of tumors are dependent on EGFR signalling for their survival, nor how additional molecular alterations present in the tumor may *Correspondence to: Prof. S. Tejpar, Digestive Oncology Unit, University Hospital Gasthuisberg, Herestraat 49, 3000 Leuven, Belgium. Tel: ; Fax: ; sabine.tejpar@uz.kuleuven.be influence primary or secondary resistance to EGFR inhibitors. CTX is approved for irinotecan-resistant mcrc expressing EGFR by immunohistochemistry (IHC). Response rates in this group however only amount to 23% in combination with chemotherapy and about 10% in monotherapy [2, 3]. Increased response rates in the combination arm did not translate into increased survival. As yet no clinical or molecular markers are available to identify those patients with a longer overall survival (OS). Predictive markers of response and survival benefit after CTX are urgently required to allow the rational and effective use of these drugs. In a small series, Lièvre et al. [4] have shown objective response (OR) to CTX to be excluded in KRAS-mutated CRC and most importantly, showed an increase in OS for the KRAS wild-type (WT) patients. We retrospectively studied the ability of KRAS mutations and tumor regression on computerized tomography (CT) to predict OR, progression-free survival (PFS) and OS benefit in a large series of mcrc patients treated with CTX. ª The Author Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please journals.permissions@oxfordjournals.org

2 Annals of Oncology patients and methods patient characteristics We included 113 irinotecan refractory mcrc patients with positive IHC for EGFR derived from four clinical trials (EVEREST, BOND, SALVAGE and BABEL) from four Belgian centers [2, 5, 6] (Appendix 1; Table 1 and Table 2, the latter being available online). Patients were included according to availability of formalin-fixed paraffin-embedded (FFPE) tumor tissue. This retrospective study was carried out after approval of the Ethics Review Board. DNA extraction and mutation analysis Tumor DNA was extracted from FFPE sections after macrodissection. The presence of KRAS and BRAF mutations was determined by an allelic discrimination assay on a 7500HT Real Time PCR System (Applied original article Biosystems, Foster City, CA). KRAS mutations located within the codon 12 (n = 6) and 13 (n = 1) and the BRAF mutation at codon 600 were screened for. All mutations were confirmed by direct sequencing. Probes and conditions were developed together with P. Laurent-Puig and are available upon request [4]. evaluation of OR, PFS and OS Tumor response was evaluated by CT every 6 weeks until week 24, 30 or 36 and from then on every 12 weeks. Response Evaluation Criteria in Solid Tumors (RECIST) were used to classify tumor response in complete response (CR), partial response (PR), stable disease (SD) and progressive disease (PD) [7]. PFS was calculated from start of CTX to either progression of disease, death from any cause or last radiological assessment. OS was defined as the period from start of CTX until death. Table 1. Patient characteristics of 113 patients treated with CTX All KRAS wild type KRAS mutant Mean age (SD), t-test (NS) 59.6 (10.7) 58.8 (11.5) 60.7 (9.5) Sex n (column %), exact (NS) Male 70 (61.9) 44 (65.7) 26 (56.5) Female 43 (38.1) 23 (34.3) 20 (43.5) Number of previous lines of chemotherapy (column %), chi-square (NS) 1 15 (13.3) 7 (10.4) 8 (17.4) 2 55 (48.7) 30 (44.8) 25 (54.3) 3 32 (28.3) 23 (34.3) 9 (19.6) 4 8 (7.1) 4 (6.0) 4 (8.7) 5 3 (2.7) 3 (4.5) 0 (0) Median time since diagnosis of the primary (weeks) (range), Mann Whitney U (NS) Median time since diagnosis of the metastasis (weeks) (range), Mann Whitney U (NS) (25 534) 134 (50 534) (25 436) 94 (25 473) 92.1 ( ) 95.1 (25 311) Trial n (column %), chi-square (NS) BABEL 2 (1.8) 2 (3) 0 (0) BOND 44 (38.9) 27 (40.3) 17 (37) EVEREST 50 (44.2) 28 (41.8) 22 (47.8) SALVAGE 17 (15) 10 (14.9) 7 (15.2) Treatment n (column %), chi-square (NS) CTX monotherapy 30 (26.5) 18 (26.9) 12 (26.1) CTX + irinotecan 83 (73.5) 49 (73.1) 34 (73.9) CTX 250 mg/m 2 + irinotecan 68 (81.9) 37 (75.5) 31 (91.1) CTX 300 mg/m 2 + irinotecan 1 (1.2) 1 (2.0) 0 (0) CTX 350 mg/m 2 + irinotecan 5 (6.0) 4 (8.2) 1 (2.9) CTX 400 mg/m 2 + irinotecan 1 (1.2) 0 (0) 1 (2.9) CTX 500 mg/m 2 + irinotecan 8 (9.6) 7 (14.3) 1 (2.9) Skin toxicity n (column %), chi-square (NS) Grade 0 6 (5.3) 3 (4.5) 3 (6.5) Grade 1 65 (57.5) 36 (53.7) 29 (63) Grade 2 35 (31) 25 (37.3) 10 (21.7) Grade 3 7 (6.2) 3 (4.5) 4 (8.7) BABEL, BOND [2], EVEREST [6] and SALVAGE [5]. SD, standard deviation; NS, nonsignificant; CTX, cetuximab. Volume 19 No. 3 March 2008 doi: /annonc/mdm

3 original article Annals of Oncology changes in tumor size For 102 patients, tumor measurements carried out during the clinical trial were available at baseline, week 6 and every consecutive evaluation. The sum of the longest dimensions of the selected lesions in each patient was calculated, and the relative change of this sum from baseline was computed at every time point. skin toxicity Skin toxicity was graded according to National Cancer Institute Common Toxicity Criteria (version 2.0) [8]. The highest grade of skin toxicity observed was used. statistical analysis Differences in OR rates between patients with and without KRAS mutations were evaluated by means of a two-sided Fisher s exact test. Relative changes in tumor size at each time point between tumors with and without KRAS mutations were compared using the Student s t-test. A multiple logistic regression was carried out to determine the impact of covariates on OR. The PFS and the OS were estimated by the Kaplan Meier method and compared with use of the log-rank test. A Cox regression analysis with backward Wald stepwise elimination procedure for the PFS and the OS was built. The significance level for the removal of a covariate from the model was The assumption of normality (t-test) was checked using the Kolmogorov Smirnov test, the assumption of proportional hazards (Cox regression) via Schoenfeld residuals. A P value of <0.05 was considered to indicate statistical significance. Analyses were carried out using SAS version 8.2 and SPSS for Windows results patient characteristics The baseline and treatment characteristics of the patients are represented in Table 1. There was no imbalance between patients with and without a KRAS mutation. mutations KRAS mutations were detected in 46 of 113 (40.7%) tumors (Table 2, available online). The BRAF V600E mutation was found in 6 (5.6%) of 107 assessable patients. No patient had tumors carrying both BRAF and KRAS mutations (Fisher s exact test, P = 0.050) (Table 2, available online). mutation status versus OR In five patients, OR could not be assessed due to dropout before first evaluation. OR (CR + PR) was observed exclusively in the KRAS WT group. KRAS mutations were absent in the 27 OR patients and found in 42 of 81 (51.9%) in nonresponders (SD + PD) (Fisher s exact test, P = ). In SD and PD, KRAS mutations were present in 31 of 59 (52.5%) and 11 of 22 (50.0%) cases, respectively. OR rate was 27 of 66 (41%) in KRAS WT versus 0 of 42 (0%) in KRAS mutants (Table 3). The impact of the KRAS mutation state on OR rate was found both in the combination and the monotherapy group (Fisher s exact test: combination therapy, P = ; monotherapy, P = 0.126). The OR rate was not significantly different in BRAF mutants (1 of 6 (16.6%)) versus WT (26 of 98 (26.5%)) (Fisher s exact test, P = 1). Table 3. Responses to CTX in 108 patients according to KRAS mutation status KRAS status Total Wild type Mutant All patients, n (column %) Best response CR 1 (2) 0 1 (1) PR 26 (39) 0 26 (24) SD 28 (42) 31 (74) 59 (55) PD 11 (17) 11 (26) 22 (20) Total Monotherapy n (column %) Best response PR 5 (28) 0 5 (18) SD 9 (50) 9 (90) 18 (64) PD 4 (22) 1 (10) 5 (18) Total Combitherapy n (column %) Best response CR 1 (2) 0 1 (1) PR 21 (44) 0 21 (26) SD 19 (40) 22 (69) 41 (51) PD 7 (15) 10 (31) 17 (21) Total Details per treatment regimen are given; objective responses (CR + PR) were observed exclusively in the KRAS wild-type group. monotherapy, cetuximab alone; combitherapy, cetuximab with irinotecan. CTX, cetuximab; CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease. survival analysis In univariate analysis, the median PFS was different between KRAS WT and mutants but did not reach statistical significance in the whole study population {24 weeks [95% confidence interval (CI) ] in WT versus 12 weeks (95% CI ) in mutants (log-rank test, P = 0.074)} (Figure 1A, available online). However, median PFS in the combination therapy group was significantly different between WT [34 weeks (95% CI ) and mutants [12 weeks (95% CI )] (log-rank test, P = 0.016)] but not in the monotherapy group [WT 12 weeks (95% CI ) versus mutants 12 weeks (95% CI ) (log-rank test, P = 0.351)] (Figure 1B and C, available online). The median OS was significantly different between KRAS WT and mutants in the whole study population [43.0 weeks (95% CI ) versus 27.3 weeks (95% CI ) (log-rank test, P = 0.020)] (Figure 2A) and in the combination therapy group [44.7 weeks (95% CI ) versus 27.3 weeks (95% CI ) (log-rank test, P = 0.003)], but not in the monotherapy group [27 weeks (95% CI ) versus 25.3 weeks (95% CI ) (log-rank test, P = 0.330)] (Figure 2B and C) Table 4, available online). changes in tumor size All patients had to have documented radiological progression before starting CTX. For 102 patients, tumor measurements were available at baseline and at least at the first evaluation at 510 De Roock et al. Volume 19 No. 3 March 2008

4 Annals of Oncology original article 6 weeks. Four had no measurements available to us, two had measurements outside the predefined time points and five stopped the clinical trial before 6 weeks [death (n = 1), biochemical reasons (n = 1), bowel obstruction (n = 1), bowel ischemia (n = 1) or dyspnoea (n = 1)]. The decrease in tumor sizes was more pronounced at all time points in WT patients compared with mutants. The mean relative change in tumor size was: % (95% CI to 27.27) in KRAS WT versus 2.27% (95% CI to 6.41) in KRAS mutants (P = ) at week 6, 224.5% (95% CI to ) versus 1.99% (95% CI to 7.62) (P = ) at week 12 and 241.2% (95% CI to ) versus 6.09% (95% CI to 14.98) (P = ) at week 24, respectively (Figure 3). When we used bidimensional (World Health Organization) instead of unidimensional (RECIST) measurements, the difference remained equally significant (data not shown). Subgroup analysis restricted to SD patients, revealed a similar effect of KRAS status on tumor shrinkage. The differences were statistically significant at week 18: 28.84% (95% CI to 1.54) in WT versus 4.19% (95% CI to 12.55) (P = 0.044) in mutants and at week 24: % (95% CI to 26.63) in WT versus 6.09% (95% CI to 14.95) (P = ) in mutants. We carried out a sensitivity analysis to identify the best cut-off, to use the first radiological measurement at 6 weeks as a predictor variable for PFS and OS. OR patients had a median survival of weeks (95% CI ). We used the lower limit of the 95% CI of the OS to categorize the state variable and generated a Receiver Operating Characteristics (ROC) curve testing the relative decrease of the tumor size at 6 weeks. The area under the ROC curve was (95% CI ). Cut-off value 29.66% generated the optimal combination of sensitivity (75%) and specificity (74%). There is a clear association between KRAS mutation status and relative decrease of tumor size >9.66% at week 6 (Fisher s exact test, P = ) (Table 5, available online). We found a significant difference in median PFS and OS between patients with or without relative decrease of tumor size >9.66% at week 6 [median PFS: 36 weeks (95% CI ) in early responders versus 12 weeks (95% CI ) in early nonresponders (log-rank test, P = ); median OS: 65.9 weeks (95% CI )] in early responders versus 31.3 weeks (95% CI ) in early nonresponders (log-rank test, P = )]. Figure 2. Kaplan Meier plots of the overall survival (weeks) according to KRAS mutation status (A) whole study group (n = 112), (B) monotherapy patients (n = 30) and (C) combination therapy patients (n = 82). predictors of OR In the logistic regression analysis, using age, sex, KRAS mutation status, skin toxicity, the number of previous lines of chemotherapy and treatment regimen as covariates, we found KRAS status (hazard ratio unavailable because constant in responders) and skin toxicity [hazard ratio (per increase in one grade): 2.35 (95% CI ; P = 0.070)] being predictive of OR. predictors of survival A Cox regression analysis looked at the independent factors (age, sex, KRAS mutation status, skin toxicity, the number of Volume 19 No. 3 March 2008 doi: /annonc/mdm

5 original article Annals of Oncology not possible to elaborate due to important interaction between both variables. Figure 3. Individual curves of the relative change in tumor size in 102 patients during treatment, stratified according to KRAS mutation status. The sum of the longest dimensions of selected lesions in each patient was calculated, and the relative change of this sum from baseline was computed for each patient at every time point and upon progression. previous lines of chemotherapy and treatment regimen) influencing PFS and OS. PFS was found to be determined by skin toxicity [hazard ratio (per increase in one grade): (95% CI ); (P = )]. OS was determined by skin toxicity [hazard ratio (per increase in one grade): (95% CI ); (P = )] and KRAS mutation status [hazard ratio: (95% CI ); (P = 0.018)]. When the Cox regression analyses were carried out replacing KRAS mutation status with relative decrease of tumor size >9.66% at week 6, we found PFS being determined by skin toxicity [hazard ratio (per increase in one grade): (95% CI ); (P = 0.046) and by relative decrease of tumor size >9.66% at week 6 (hazard ratio: (95% CI ); (P = ) and OS by the same variables (hazard ratios: (95% CI ); (P = 0.004) and (95% CI ); (P = ), respectively]. A model combining KRAS mutation status and early radiological response, together with all the other clinical variables, was interaction between KRAS mutation status and early radiological changes in predicting outcome We analyzed the interaction of KRAS and early radiological response on tumor behavior. For each patient the smallest tumor size and the tumor size at progression were plotted against their time of occurrence (Figure 4). Patients were grouped according to their KRAS and early radiological response status. In patients with early radiological response, we observed that only KRAS WT patients (Figure 4 panel A versus C) showed further decrease in tumor size beyond 6 weeks and achieved the largest tumor size reductions. In addition, these are the patients that show delayed occurrence of progression. Within the KRAS WT patients, only those patients with an early radiological response (Figure 4 panel A versus B) continued to decrease over time and had delayed time to progression. In patients without regression at week 6, no subsequent tumor reduction or delayed occurrence of progression was observed even in KRAS WT patients. Lastly, in rare KRAS mutant patients delayed time to progression was observed, which was not associated to a tumor reduction. The interaction of KRAS and early radiological response was also analyzed on the OS. In KRAS WT patients OS was significantly better in patients with an initial decrease compared with those without [median OS: 74.9 weeks (95% CI ) versus 30.6 weeks (95% CI ) (log-rank test: P = )]. In contrast, in KRAS mutant patients, OS is similar, whether initial tumor decrease is observed or not [median OS: 17.6 weeks (95% CI ) versus 32 weeks (95% CI ) (log-rank test, P = 0.664)] (Figure 5). When comparing OS of patients with KRAS WT status and early tumor size reduction to all others we found a profound survival benefit [median OS: 74.9 weeks (95% CI ) versus 30.6 weeks (95% CI ) (log-rank test: P = )]. discussion The KRAS mutation state of the tumor has been proposed as a promising marker of response to CTX in mcrc [4]. In the setting of irinotecan refractory patients we confirm the clear relationship between KRAS WT state and OR, and demonstrate that KRAS WT is a strong predictor of a significant increase in PFS and OS in these patients. The ideal biomarker should identify responders and patients with improved survival with high sensitivity and specificity. However, the KRAS status of a tumor may still fall short in this respect, as not all WT patients respond or have improved survival and some mutant patients experience long-term disease control. Despite the encouraging hazard ratios for survival (0.62) seen in KRAS WT patients as a group, half of these patients will not benefit. In an attempt to further stratify these patients we looked at the additional predictive value of early tumor size reduction on survival. The time point of 6 weeks, at first evaluation, was chosen to avoid selection bias due to dropouts for progression. The early tumor size reduction was shown to 512 De Roock et al. Volume 19 No. 3 March 2008

6 Annals of Oncology original article Figure 4. Tumor behavior during CTX treatment. For every patient nadir in tumor size (filled inverted triangles) and the tumor size at progression (open circles) are depicted. The X-axis represents time on treatment and the Y-axis the relative change in tumor size. Patients are grouped according to their KRAS mutation status and radiological response at week 6. be a strong predictor of survival in a Cox regression analysis (hazard ratio: 0.42). There was a correlation between early tumor size reduction and the KRAS mutation status. Both parameters in conjunction were found to predict tumor behavior at later time points. This is demonstrated by the observation that early tumor size reduction will only translate in longlasting and more profound tumor control in KRAS WT patients and only KRAS WT patients with initial tumor decrease will have further and sustained tumor reduction. This translates into survival benefit as shown by the Kaplan Meier analysis where only KRAS WT patients with an early tumor size reduction had a significantly improved survival. We attempted to build a Cox regression model for predicting survival, including all the clinical variables, KRAS mutation status and early tumor size reduction. The magnitude of the interaction between the KRAS mutation status and early tumor size reduction precluded reliable modeling. Skin toxicity, a well-known predictor of response, was one of the predictor variables in our models. It was found to impact independently on all outcome parameters: OR, PFS and OS. In univariate log-rank analysis, we found a trend for an impact of skin toxicity on PFS and OS, only in patients with KRAS WT status and early radiological response (data not shown). We would like to point out that skin toxicity remains a difficult parameter to use as a response predictor in general, because of the lack of appropriate toxicity criteria specifically tailored to the effect of EGFR inhibitors. Also, the time point at which the skin toxicity should be graded to maximally reflect its predictive value has not been determined. These issues need to be specifically addressed in the future. Selecting patients with both KRAS WT status and early tumor size reduction identifies a subgroup with a median OS of 74.9 weeks in comparison to 38.7 weeks for the whole study population or 43.0 weeks for all KRAS WT patients. If the end point were to identify patients surviving longer than 64.5 weeks, our combined parameters would yield a positive predictive value of 57% and a negative predictive value of 88%. Volume 19 No. 3 March 2008 doi: /annonc/mdm

7 original article Figure 5. Kaplan Meier plots of the overall survival (weeks) in patients with or without a reduction of tumor size >9.66% at 6 weeks, stratified according to their KRAS mutation status (A) KRAS wild-type patients; (B) KRAS mutant patients. The false-positive patients however, still had a clinically relevant survival comparable to that of the whole KRAS WT group (median 43 weeks). The false-negative patients (8) consisted of one WT patient with an initial tumor size decrease of 9.63% whereas the cut-off was set at 9.66% and seven mutant patients. These patients, as shown in panel D of Figure 4, are characterized by the absence of tumor decrease and a long time to progression and OS. This suggests that in KRAS mutant tumors some of the proliferation may still be driven by EGFR signalling and inhibition will result in tumor stabilization. This idea was reinforced when we studied tumor growth rates in KRAS mutant patients before and following initiation of CTX treatment. A clear decrease in tumor growth rates was observed (data not shown). This suggests there may be a subgroup of KRAS mutant patients who may benefit from long-term disease stabilization upon CTX treatment. Lastly, there has been a report on two KRAS mutant patients with an OR on CTX [9, 10]. In our own series, we identified some Annals of Oncology KRAS mutant patients with tumor shrinkage approaching the 30% cut-off (Figure 3). The mutual exclusion of KRAS mutations and OR may therefore not stand in larger patient sets, however, we found that those KRAS mutant patients who had tumor size decrease had no increased survival (Figure 4C and D). To be considered for clinical use the molecular and radiological markers presented will need prospective validation. Our model necessitates the combination of a molecular marker and an early response marker that can only be assessed after 6 weeks of treatment. This is not common practice. However, there are clinical scenarios in which our data could already be put to use. In those patients where no tumor material is available, or determination of KRAS status is not feasible, we have shown that early response at 6 weeks is independently associated to longer survival. Alternatively, KRAS status alone could be used to select those patients in which to initiate treatment, but a stopping rule depending on the response at 6 weeks could be introduced. It is our opinion however that in the current paucity of therapeutic options after failing irinotecan- or oxaliplatin- based regimens, the median OS seen in KRAS mutants may be considered clinically relevant. This may represent the biggest barrier to clinical implementation. We show that the KRAS mutation state impacts significantly on three fundamental aspects of the mcrc treated with CTX: tumor decrease, PFS and OS. Our series consisted mostly of combination therapy patients and no valid conclusions could be drawn for the monotherapy patients due to the small sample size. Kathamba-Ford et al. [11], focussing solely on 80 monotherapy patients, showed a significant association between the KRAS status with both OR and PFS. However, in addition, our series shows the significant association between KRAS status and better OS. A similar trend observed in a smaller series of patients (n = 81) was recently presented [12]. Improved response rates do not always correlate with survival [13]. In our series, the correlation between PFS and OS was good both in the KRAS WT and mutant patients (data not shown), suggesting all the OS benefit derived from the response to therapy. Another recent paper, also found a significant difference in time to progression according to KRAS mutation status in 59 patients [14]. Purely prognostic effects of the KRAS status cannot be excluded without a randomized trial [15]. In two studies of resected liver metastasis no prognostic effect of the KRAS mutations status could be demonstrated nor was there a proliferative advantage for KRAS mutant over WT metastases [16, 17]. A strong prognostic effect of KRAS in mcrc is unlikely, as the RASCAL study could only demonstrate a significant prognostic effect of KRAS in Dukes C but not Dukes D patients [18]. In the current models, KRAS is regarded as the main effector of the EGFR signalling cascade and may represent a bottleneck in the signalling pathway. Mutations of KRAS will constitutively activate the pathway and render it resistant to upstream inhibitors. If this model holds true, KRAS mutations may be negative predictors of response even in the presence of supra-normal upstream activation such as seen with EGFR amplifications or increased ligand expression [11]. 514 De Roock et al. Volume 19 No. 3 March 2008

8 Annals of Oncology In conclusion, patients characterized by the WT state and initial tumor regression tumor represent a subgroup with the largest survival benefit. The challenge lies in identifying the molecular markers that type these subgroups, so they can be identified early on in the therapeutic approach. In this regard, the recently published data [11] on high EGFR ligand expression in tumors responding to CTX is promising. We believe our work and that of others represent significant steps towards the optimal use of EGFR-targeting therapies. funding Belgian Foundation against Cancer (SCIE ); Unrestricted grant, Merck KGaA, Darmstadt, Germany. acknowledgements The authors wish to thank Liesbeth Lemmens, Hilde Marsé, Martine Crabbé, Dimphna Van Reijen (clinical trial coordinators, University Hospital Gasthuisberg, Leuven, Belgium), Muriel De Cooman (clinical trial coordinator, Cliniques universitaires St-Luc, Brussels, Belgium) and Nancy Van Damme (Universiteit Gent) for the data collection. ST had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. ST and EVC are Senior Clinical Investigator of the Fund for Scientific Research Flanders (Belgium) (Fonds Wetenschappelijk Onderzoek Vlaanderen). The research sponsor (University Hospital Gasthuisberg) had no role in the design and conduct of the study; collection, management, analysis and interpretation of the data; or preparation, review, or approval of the manuscript. appendix 1 Centers having provided tumor samples and patient data: University Hospital Gasthuisberg, Leuven, Belgium; Ghent University Hospital, Ghent, Belgium; Cliniques universitaires St-Luc, Brussels, Belgium; Hôpital Erasme, Brussels, Belgium. supplementary data Figure 1 and Tables 2, 4 and 5 are available online to subscribers. references 1. Harding J, Burtness B. Cetuximab: an epidermal growth factor receptor chimeric human-murine monoclonal antibody. Drugs Today 2005; 41: 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: original article 3. Saltz LB, Meropol NJ, Loehrer PJ Sr 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: Lièvre A, Bachet JB, Le Corre D et al. KRAS mutation status is predictive of response to cetuximab therapy in colorectal cancer. Cancer Res 2006; 66: Lenz HJ, Van Cutsem E, Khambata-Ford S et al. Multicenter phase II and translational study of cetuximab in metastatic colorectal carcinoma refractory to irinotecan, oxaliplatin, and fluoropyrimidines. J Clin Oncol 2006; 24: Tejpar S, Peeters M, Humblet Y et al. Dose-escalation study using up to twice the standard dose of cetuximab in patients with metastatic colorectal cancer (mcrc) with no or slight skin reactions on cetuximab standard dose treatment (EVEREST study): preliminary data. J Clin Oncol 2006; 24: 159S (Abstr 3554). 7. Therasse P, Arbuck SG, Eisenhauer EA et al. New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 2000; 92: Common Toxicity Criteria v2.0 (CTC). Cancer Therapy Evaluation Program. National Cancer Institute 1999; (13 July 2006, date last accessed). 9. Benvenuti S, Sartore-Bianchi A, Di Nicolantonio F et al. Oncogenic activation of the RAS/RAF signalling pathway impairs the response of metastatic colorectal cancers to anti-epidermal growth factor receptor antibody therapies. Cancer Res 2007; 67: Moroni M, Veronese S, Benvenuti S 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: Khambata-Ford S, Garrett CR, Clark EA 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: Finocchiaro G, Cappuzzo F, Jänne PA et al. EGFR, HER2 and Kras as predictive factors for cetuximab sensitivity in colorectal cancer ASCO Annual Meeting Proceedings Part I. J Clin Oncol 25 (18 Suppl): 2007 (Abstr 4021). 13. Buyse M, Thirion P, Carlson RW et al. Relation between tumour response to firstline chemotherapy and survival in advanced colorectal cancer: a meta-analysis. Lancet 2000; 356: Di Fiore F, Blanchard F, Charbonnier F et al. Clinical relevance of KRAS mutation detection in metastatic colorectal cancer treated by cetuximab plus chemotherapy. Br J Cancer 2007; 96: Jonker DJ, Karapetis CS, Moore M et al. Randomized Phase III Trial of Cetuximab Monotherapy Plus Best Supportive Care (BSC) Versus BSC Alone in Patients with Pretreated Metastatic Epidermal Growth Factor Receptor (EGFR)-Positive Colorectal Carcinoma: A Trial of the National Cancer Institute of Canada Clinical Trials Group (NCIC CTG) and the Australasian Gastro-Intestinal Trials Group (AGITG) [Abstract]. Presented at the American Association for Cancer Research Annual Meeting, Los Angeles, CA, April Kastrinakis WV, Ramchurren N, Maggard M et al. K-ras status does not predict successful hepatic resection of colorectal cancer metastasis. Arch Surg 1995; 130: Petrowsky H, Sturm I, Graubitz O et al. Relevance of Ki-67 antigen expression and K-ras mutation in colorectal liver metastases. Eur J Surg Oncol 2001; 27: Andreyev HJ, Norman AR, Cunningham D et al. Kirsten ras mutations in patients with colorectal cancer: the multicenter RASCAL study. J Natl Cancer Inst 1998; 90: Volume 19 No. 3 March 2008 doi: /annonc/mdm