Efficacy of continued cetuximab for unresectable metastatic colorectal cancer after disease progression during first-line cetuximab-based chemotherapy: a retrospective cohort study

This study assessed second-line continued use of cetuximab for treatment of unresectable metastatic colorectal cancer (mCRC) after disease progression during first-line cetuximab-based therapy. Consecutive patients with wild-type KRAS exon 2 and unresectable mCRC were retrospectively enrolled after disease progression during first-line cetuximab-based chemotherapy. Second-line continued cetuximab plus changed chemotherapy (cetuximab continuation group, n = 102) was compared with changed chemotherapy only (chemotherapy only group, n = 96) with respect to treatment efficacy and safety endpoints. NRAS and other KRAS genotypes were also detected as a post hoc analysis. The cetuximab continuation group showed better progression-free survival (median, 6.3 vs. 4.5 months, P = 0.004), overall survival (median, 17.3 vs. 14.0 months, P < 0.001) and disease control rate (70.6% vs. 53.1%, P = 0.011), and a potentially better overall response rate (18.6% vs. 9.4%, P = 0.062) than the chemotherapy only group. These benefits were seen mainly in patients with all RAS wild-type and exhibiting first-line early tumor shrinkage (ETS). For patients with other RAS mutations or who did not achieve first-line ETS, there was no difference between the two groups. These findings suggest that for patients with all RAS wild-type and unresectable mCRC who had disease progression during first-line cetuximab-based treatment, second-line continued cetuximab is effective. Moreover, ETS during first-line cetuximab-based treatment may be predictive of the efficacy of second-line continued cetuximab.


INTRODUCTION
Metastatic colorectal cancer (mCRC) is a major healthcare problem globally [1].During the course of their disease, approximately half of patients with colorectal cancer will develop distant metastasis [2], which is the major cause of death.If feasible, radical resection is the ideal treatment for mCRC, but in the majority of patients, mCRC is unresectable, even after intensive treatment with targeted agents plus chemotherapy [3].
Cetuximab is a promising agent that targets epidermal growth factor receptor (EGFR) and has shown an impressive ability to improve the tumor response and increase progression-free survival (PFS) and overall survival (OS) among patients with all RAS wild-type mCRC [4][5][6].It also significantly increases the number of patients with inoperable metastases whose tumors become resectable after treatment [3].However, for patients who fail to respond to standard first-line cetuximabbased chemotherapy, it is unclear whether continuation

Treatment exposure
Before enrollment, the first-line treatment was balanced between the two groups, with no significant differences in PFS (P = 0.796), early tumor shrinkage (ETS) rate (P = 0.821), ORR (P = 0.951) or receiving maintenance treatment (P = 0.661) during the period of first-line treatment.For second-line treatment, the chemotherapy regimen was changed for all patients.After second-line disease progression, no significant difference was observed between the two groups whether they received TACE/TAI, radiofrequency ablation, radiotherapy or chemotherapy.However, there tended be more patients in the cetuximab continuation group receiving bevacizumab following treatment (P = 0.052).Details are shown in Table 2.

Efficacy of second-line treatment
All 198 patients included in this study had KRAS exon 2 wild-type.Among them 171 (86.4%) experienced disease progression during second-line treatment: 86 (84.3%) in the cetuximab continuation group and 85 (88.5%) in the chemotherapy only group.The median PFS from the start of second-line treatment was 6.3 months in the cetuximab continuation group, which was significantly better than the 4.5 months in the chemotherapy only group (hazard ratio = 0.646, P = 0.004).In terms of OS, a total of 131 (66.2%) deaths occurred by the end of follow-up: 70 (68.6%) in the cetuximab continuation group and 61 (63.5%) in the chemotherapy only group.All of these patients died of mCRC.The median OS from the start of second-line treatment was 17.3 months in the cetuximab continuation group, which was significantly better than the 14.0 months in the chemotherapy only group (hazard ratio = 0.503, P < 0.001).In addition, the cetuximab continuation group had significantly better OS from the start of first-line treatment than the chemotherapy only group (median, 30.4 vs. 27.0months, hazard ratio = 0.629, P = 0.010).In terms of tumor response, the cetuximab continuation group had significantly better DCR (70.6% vs. 53.1%,odds ratio = 2.118, P = 0.011) and potentially better ORR (18.6% vs. 9.4%, odds ratio = 2.213, P = 0.062) than the chemotherapy only group (Table 3 and Figure 2).
As a retrospective study without randomization, some potential imbalances between the two groups could interfere with analysis of the efficacy of second-line

Subgroup analysis for second-line treatment
Subgroup analysis was conducted to find more accurate predictors for second-line continued use of cetuximab after progression during first-line cetuximabbased treatment.All potential factors were included in the planned subgroup analysis of PFS from the start of second-line treatment (as shown in Figure 4).ETS in first-line cetuximab-based treatment was demonstrated to be predictive of the efficacy of second-line continued cetuximab, with significant interaction (P = 0.010).Among all patients with KRAS exon 2 wild-type included in this study, 85 (42.9%) achieved ETS during first-line treatment.From start of second-line treatment in these patients, continued cetuximab significantly improved PFS (median, 7.7 vs. 4.5 months, hazard ratio = 0.377, P < 0.001), OS (median, 21.1 vs. 14.3 months, hazard ratio = 0.258, P < 0.001) and DCR (86.0% vs. 54.8%,odds ratio = 5.094, P = 0.002), and potentially improved ORR (34.9% vs. 16.7%,odds ratio = 2.679, P = 0.055) as compared with chemotherapy alone.In patients who did not reach ETS during first-line treatment, from  start of second-line treatment no significant difference was observed between the two groups with respect to PFS (median, 5.0 vs. 4.1 months, hazard ratio = 0.925, P = 0.695), OS (median, 15.3 vs. 13.4 months, hazard ratio = 0.740, P = 0.168), DCR (59.3% vs. 51.9%,odds ratio = 1.354,P = 0.425) and ORR (6.8% vs. 3.7%, odds ratio = 1.891,P = 0.681) (Table 3 and Figure 5).
No other significant predictor was found among these factors (primary tumor site, organs with metastases, number of liver metastatic lesions, first-line chemotherapy regimen, etc.).

Efficacy of treatment after second-line disease progression
For all patients included in this study, univariate and multivariate Cox regression analyses of OS were conducted after disease progression during second-line treatment.The results showed that second-line continued cetuximab (hazard ratio = 0.663, P = 0.034 by Cox regression) and intensive chemotherapy after second-line disease progression (hazard ratio = 0.417, P = 0.008 by Cox regression) were independent protective factors for better OS after second-line progression.Bevacizumab administered after second-line progression was a potential protective factor (hazard ratio = 0.644, P = 0.084 by Cox regression) (Table 5).

Safety
For all patients included in this study, information on grade 3 or higher adverse events that occurred during the second-line treatment were collected and analyzed.In general, the observed toxicity was mostly mild in both groups, and no deaths were attributable to second-line treatment.The overall incidence of second-line grade 3 or higher adverse events was 27.5% in cetuximab continuation group and 22.9% in chemotherapy only group (P = 0.463).Continued cetuximab significantly increased incidence of newly occurring acne-like rash in patients (10.8% vs. 2.1%, P = 0.013).With the exception of the cetuximab-specific acne-like rash, no difference in adverse events was observed between the two groups (22.5% vs. 20.8%,P = 0.770).Three (2.9%) patients in the cetuximab continuation group and 2 (2.1%) patients in the chemotherapy only group converted to best supportive care because of they could not tolerate the intensive treatment (Table 6).

DISCUSSION
Although there have been considerable advances in the medications used to treat colorectal cancer, the array of anti-cancer agents currently available for clinical use is still very limited.Every kind of medicine that could prolong patients' survival is precious, especially for those with incurable metastases.Second-line treatment with anti-vascular endothelial growth factor (anti-VEGF) antibodies was demonstrated to be effective as a crossline treatment for mCRC [7,8].However, anti-EGFR antibodies were traditionally not considered suitable for continuing use.Several experimental studies reported resistance to anti-EGFR treatment once first-line disease cetuximab [13,14].In our study (Figure 4), patients with BRAF wild-type could benefit more from cetuximab than those with a BRAF mutation.Unfortunately, the small sample size of BRAF subgroup meant that there was no statistically significant interaction.Further large-scaled studies are needed for confirmation.
Results from earlier studies suggest ETS is an important predictor of prognosis [15,16].ETS reflects the treatment efficacy in terms of both speed and depth, and is always more accurate than the traditional ORR.In patients who received first-line cetuximab-based chemotherapy, ETS indicated higher sensitivity to treatment with both targeted agents and cytotoxic chemotherapy.For patients achieving first-line ETS, first-line disease progression was mainly due to resistance to the cytotoxic chemotherapy regimens.Tumor sensitivity to cetuximab was preserved.Indeed, Ekblad et al. [17] observed increased sensitivity to cetuximab in oxaliplatin-resistant cell lines.Thus, the continuation of cetuximab in second-line treatment was more suitable for patients who achieved ETS in first-line cetuximab-based chemotherapy.
For patients who did not achieve ETS during firstline cetuximab-based treatment, second-line continued cetuximab had no beneficial effect.This may reflect acquired resistance to cetuximab.Diaz et al. [10] and Misale et al. [9] reported an induced KRAS mutation rate of 38% to 60% after initial cetuximab-based treatment, which would result in the failure of second-line continued cetuximab.Another explanation could be primary resistance of cetuximab not detected prior to treatment.This could result from other RAS, BRAF, PIK3CA or HER-2 amplification, and from other gene mutations [13,14,18,19] or intratumoral heterogeneity [20].For this reason, more detailed gene analysis is needed before applying cetuximab.
As a retrospective study, there were potential imbalances between the two test groups.To reveal and reduce this interference, all patients included in this study received first-line and second-line treatment strictly according to the protocols developed by the MDT.A precise follow-up was done to avoid missing any data related to the treatment.All potential factors were listed and compared in a baseline analysis.Multivariate analysis was also conducted for second-line PFS and OS after second-line disease progression.These analyses proved that during first-line and second-line treatment, there was no significant imbalance between the two test groups.However, there were potential imbalances in the treatment after second-line disease progression.The efficacy of second-line continued cetuximab was reliable in both PFS and OS, and was not significantly interfered with by these imbalances.
In summary, despite several drawbacks, our study showed that for patients with all RAS wildtype and initially

Study eligibility
Consecutive patients presenting between January 2012 and January 2015 with unresectable mCRC were retrospectively identified from the colorectal cancer database in Zhongshan Hospital, Fudan University (Shanghai, China).For these patients, disease evaluations and treatment strategies were conducted at multidisciplinary team (MDT) meetings.
The enrollment was set at the beginning of secondline treatment.The inclusion criteria were as follows: aged from 18 to 75 years; Eastern Cooperative Oncology Group (ECOG) performance status of 0 to 2; pathologically confirmed primary colorectal cancer; unresectable mCRC based on clear radiographic evidence as determined by an MDT; wild-type KRAS exon 2 codons 12 and 13; received cetuximab plus mFOLFOX6 (fluorouracil, leucovorin, and oxaliplatin) or FOLFIRI (fluorouracil, leucovorin, and irinotecan) as first-line treatment for at least 4 cycles; first-line disease progression assessed by an MDT; received second-line treatment within 4 weeks after progression on first-line treatment; received a different chemotherapy regimen (first-line mFOLFOX6 converted to second-line FOLFIRI; first-line FOLFIRI converted to second-line mFOLFOX6) alone or with continued cetuximab as second-line treatment for at least 4 cycles.Patients were excluded if they exhibited tumor peritoneal dissemination, first-line treatment intolerance or other cancers within previous 5 years of the end of follow-up (with the exception of squamous cell carcinoma of skin and cervical cancer in situ).Patients were also excluded if they were unable to afford cetuximab as secondline treatment.The chemotherapy regimen CapeOX (capecitabine and oxaliplatin), bevacizumab and other targeted agents, transcatheter arterial chemoembolization (TACE), transcatheter arterial infusion (TAI), radiotherapy and radiofrequency ablation were not permitted until second-line disease progression.Eligible patients were divided into two groups according to their second-line treatment: patients receiving continued cetuximab plus changed chemotherapy were the cetuximab continuation group, and patients receiving only changed chemotherapy were the chemotherapy only group.
The resectability of liver metastases was determined based on the ability to obtain a complete resection (i.e., negative margins), preserve an adequate liver remnant (> 30% of healthy liver), and preserve adequate vascular inflow and outflow as well as biliary drainage [3].The resectability of lung metastases was determined based on the ability to obtain complete resection (i.e., negative margins) based on the anatomic location and the extent of disease and ability to maintain adequate lung function [21,22].Metastases were deemed unresectable if the above criteria were not met.Bone/brain metastases and peritoneal disease were defined as unresectable.
This study was approved by the institutional review board of Zhongshan Hospital, Fudan University.The investigators obtained informed consent from each patient.

RAS and BRAF mutation analysis
KRAS exon 2 (codon 12/13) and BRAF mutations were analyzed before the first-line treatment.Colonoscopic biopsy of the primary tumors was conducted.DNA was extracted from formalin-fixed paraffin-embedded (FFPE) tumor tissues obtained at biopsy.The mutation status of KRAS exon 2 (codon 12 and 13) was assessed using polymerase chain reaction clamping and pyrosequencing techniques.BRAF mutation (V600E) was assessed using a similar approach.
Other RAS gene detection was conducted as a post hoc analysis during or after the follow-up was over.Samples were FFPE specimens from colonoscopic biopsy or primary tumor resections.Mutation status of KRAS exons 3 (codon 59 and 61) and exon 4 (codon 117 and 146), NRAS exons 2 (codon 12 and 13), exon 3 (codon 59 and 61) and 4 (codon 117 and 146) were also assessed using approaches similar to those mentioned above.All gene detection protocols were performed by Gene Tech (Shanghai) Company Limited.

Chemotherapy regimen
In this study, chemotherapy (mFOLFOX6, FOLFIRI) and cetuximab were given as previously reported [3].Intensive combination therapy was sustained until disease progression or adverse events were intolerable.In patients whose tumors remained stable and who experienced no tumor-associated symptoms for more than 4 months (8 cycles), maintenance treatment became a possible alternative to intensive combination therapy.During the maintenance treatment, fluorouracil/leucovorin was given instead of mFOLFOX6 or FOLFIRI.In patients with adverse events were ≥ Grade 3, the intensive combination therapy was suspended, and symptomatic treatment was given.Maintenance treatment was also used in patients who were still unable to tolerate the intensive treatment after a rest.If the disease progressed during the maintenance treatment, the previous intensive treatment was attempted again.

Figure 2 :
Figure 2: Progression-free survival and overall survival among all patients with KRAS exon 2 wild-type.Kaplan-Meier curves for (A) progression-free survival from the start of second-line treatment, (B) overall survival from start of second-line treatment, and (C) overall survival from start of first-line treatment.The curves compare continued cetuximab plus changed chemotherapy and changed chemotherapy only as second-line treatment in all patients.Cet.: Cetuximab; Chemo.: chemotherapy.P values were determined using the log-rank test.

Figure 3 :
Figure 3: Progression-free survival and overall survival among patients receiving other RAS detection.Kaplan-Meier curves for (A) progression-free survival and (B) overall survival among patients with all RAS wild-type; and for (C) progression-free survival and (D) overall survival among patients with other RAS mutant type.The curves compare continued cetuximab plus changed chemotherapy and changed chemotherapy only as second-line treatment.Cet.: Cetuximab; Chemo.: chemotherapy.P values were determined using the log-rank test.

Figure 5 :
Figure 5: First-line early tumor shrinkage was predictive of the efficacy of second-line treatment in all patients with KRAS exon 2 wild-type.Kaplan-Meier curves for (A) progression-free survival and (B) overall survival among patients with KRAS exon 2 wild-type and who achieved early tumor shrinkage during first-line treatment; and for (C) progression-free survival and (D) overall survival among patients with KRAS exon 2 wild-type but who did not achieve early tumor shrinkage during first-line treatment.The curves compare continued cetuximab plus changed chemotherapy and changed chemotherapy only as second-line treatments.Cet.: Cetuximab; Chemo.: chemotherapy; ETS: early tumor shrinkage.P values were determined using the log-rank test.

Figure 6 :
Figure 6: First-line early tumor shrinkage was predictive of the efficacy of second-line treatment in patients with all RAS wild-type.Kaplan-Meier curves for (A) progression-free survival and (B) overall survival among patients with all RAS wild-type and who achieved early tumor shrinkage during first-line treatment; and for (C) progression-free survival and (D) overall survival among patients with all RAS wild-type but who did not achieve early tumor shrinkage during first-line treatment.The curves compare the continued cetuximab plus changed chemotherapy and changed chemotherapy only as second-line treatment.Cet.: Cetuximab; Chemo.: chemotherapy; ETS: early tumor shrinkage.P values were determined using the log-rank test.

Table 2 : Treatment exposure Cetuximab continuation (N = 102) Chemotherapy only (N = 96) P value
# : Not PD: not reaching progressive disease by the end of follow-up.§:Intensive chemotherapy: receiving at least one intravenous chemotherapy medicine.Bold font for P < 0.10.www.impactjournals.com/oncotargetcontinuedcetuximab.Therefore, multivariate Cox regression analysis was conducted for confirmation.The results showed that first-line ETS (hazard ratio = 0.682, P = 0.041) and second-line continued cetuximab (hazard ratio = 0.675, P = 0.016) were independent protective factors for second-line PFS; other RAS mutant type (hazard ratio = 2.141, P = 0.003) and BRAF mutant type (hazard ratio = 3.001, P < 0.001) were independent risk factors.No significance was detected for primary tumor site, organs with metastases and first-line chemotherapy regimen, among others (Supplementary TableS1).

Table 3 : Efficacy of second-line treatment among all patients with KRAS exon 2 wild-type Total patients First-line ETS First-line NOT ETS
Chemo.: chemotherapy; ETS: early tumor shrinkage (8 weeks, shrinkage ≥ 20% of the tumor); PFS: progression-free survival; OS: overall survival; CI: confidence interval; CR: complete response; PR: partial response; SD: stable disease; PD: progressive disease.Overall response rate = CR + PR; Disease control rate = CR + PR + SD.P values of "Overall response rate" and "Disease control rate" were calculated using two-sided Pearson's χ 2 tests or Fisher's exact test for any cell expected count less than 5 or sample size less than 40.Bold font for P < 0.10; * for P < 0.05.

Table 4 : Efficacy of second-line treatment among patients receiving other RAS detection Patients with all RAS wild-type Patients with other RAS mutant type All First-line ETS First-line NOT ETS
Chemo.: chemotherapy; ETS: early tumor shrinkage (8 weeks, shrinkage ≥ 20% of the tumor); PFS: progression-free survival; OS: overall survival; CI: confidence interval; CR: complete response; PR: partial response; SD: stable disease; PD: progressive disease; NE: not evaluable.Overall response rate = CR + PR; Disease control rate = CR + PR + SD.Other RAS including KRAS exons 3 (codon 59 and 61) and exon 4 (codon 117 and 146), NRAS exons 2 (codon 12 and 13), exon 3 (codon 59 and 61) and 4 (codon 117 and 146).P values of "Overall response rate" and "Disease control rate" were calculated using two-sided Pearson's χ 2 tests or Fisher's exact test for any cell expected count less than 5 or sample size less than 40.Bold font for P < 0.10; * for P < 0.05.

Table 6 : Grade 3 or higher adverse events during second-line therapy
P values were calculated using two-sided Pearson's χ 2 tests or Fisher's exact test for any cell expected count less than 5 or sample size less than 40.Bold font for P < 0.10; * for P < 0.05.www.impactjournals.com/oncotargetunresectable mCRC who experienced disease progression during standard first-line cetuximab-based treatment, continuation of cetuximab was effective and safe as second-line treatment.ETS in first-line cetuximab-based treatment could be a significant predictor of the efficacy of second-line continued cetuximab.Therefore, the indications for cetuximab might be prudently expanded, and a further large randomized clinical trial should be conducted to confirm these conclusions.