Background
The estimated incidence of stage 1–3 colorectal cancer in China for 2017 is 100,000 patients, of which almost 70% did develop metastatic colorectal cancer (MCC), which is partially related to common delays in diagnosis [
1‐
3]. Despite the lack of multi-centre, open-label, double-arm trials that compared the efficacy and safety of different therapeutic regimens in patients with MCC, 4 prospective studies comparing the efficacy and safety of bevacizumab(B) and capecitabine (CAP) had indicated a safe and effective regimen for the treatment of patients with late-stage colorectal cancer [
3‐
6]. However, evidence from recent studies suggests that MCC carrying extended KRAS exon 2 mutations predicted for the lack of activity of anti-epidermal growth factor receptor (EGFR) therapies and resulted in restriction to the use of anti-EGFR monoclonal antibodies, making these drugs suitable for the so-called “super” wild-type patients only [
7‐
9]. Approximately 70% patients with KRAS exon 2 wild-type(wt) MCC tend to have a poor prognosis, with a median overall survival (mOS) of 4–8 months and a median progression-free survival (mPFS) of 3–9 months [
10‐
12].
KRAS exon 2 wt MCC has historically demonstrated restricted response to intervention, and prospective trials with small samples have presented poor outcomes [
10,
11]. Despite recent developments in targeted therapies, experience with these agents in KRAS exon 2 wt MCC is still inadequate [
12]. Previous retrospective studies have either excluded or lacked definite details of the inclusion of KRAS exon 2 wt MCC [
13,
14]. At the time of study development, bevacizumab, a recombinant humanised monoclonal immunoglobulin G antibody, was considered a favourable antiangiogenic agent and was approved by the US Food and Drug Administration(UFDA) for the first-line treatment of KRAS exon 2 wt MCC [
15]. The combination regimen of bevacizumab and CAP in MCC has been reviewed extensively [
16]. In previous retrospective studies, the addition of bevacizumab was observed to significantly improve the patient’s mPFS and mOS, with 5–7 months and 5–18 months for bevacizumab in comparison to 3–6 months and 4–12 months for placebo, respectively [
17]. The results of a prior phase 3 CAIRO3 trial demonstrated that MCC patients with stable disease or better after 6-cycle induction treatment with capecitabine, oxaliplatin, and bevazicumab (CAPOXB) had a significant benefit from CAP-B maintenance treatment compared to observation [
18]. Although the trial was not premeditated to detect a difference in mOS between groups, an absolute mOS benefit of 3.5 months was detected, which was not statistically significant (HR0.89, 95% CI 0.73–1.07). In addition, in the phase 3 CAIRO3 trial, the analysis of this type of patients with untreated characterised KRAS exon 2 wt MCC is disregarded. A multi-center, single-arm, phase II trial (CCOG-0801) has been conducted to investigate targeted therapy alone, where either bevacizumab was unavailable or the toxicity profile of bevacizumab had been considered adverse [
19]. Their study included 47 patients from 2008 to 2010 with untreated characterised KRAS exon 2 wt MCC, and patients underwent 6-cycle CAPOXB induction therapy, followed by CAP-B or CAP maintenance treatment. The CAP maintenance treatment group described 2-year clinical outcomes that were encouraging, with mPFS of 54% and mOS of 63%. Nevertheless, with a median follow-up of 30 months, adverse events had occurred in 9 (19.1%) cases, which raises questions about whether CAP maintenance treatment was acceptable in the absence of bevacizumab, regardless of the cost-effectiveness of CAP-B. Furthermore, hormone-based therapy, a potential MCC preventive intervention, may influence carcinogenesis through cellular pathways involving the KRAS oncogene in postmenopausal women. Nevertheless, most existing literature tend to ignore hormones as an interference factor, which is likely to lead to a weakening of the power to draw conclusions.
The purpose of this study was to compare the safety and efficacy of CAP and CAP-B as maintenance treatment after 6-cycle-CAPOXB induction treatment for Chinese postmenopausal women with untreated characterised KRAS exon 2 wt MCC even after progression with these drugs during the treatment.
Discussion
The present study followed Chinese postmenopausal women with untreated KRAS exon 2 wt MCC for a mean of 2 years, and the most important finding was that CAP-B is a feasible maintenance treatment for these patients after 6-cycle CAPOX-B induction treatment compared with CAP. The superiority of CAP-B over CAP after 6-cycle CAPOX-B in Chinese postmenopausal women with untreated KRAS exon 2 wt MCC remains a matter of debate, which precludes any recommendations.
In most patients, in daily practice, KRAS mutational status is evaluated in samples originating from primary intestinal lesions at the time of diagnostic colonoscopy [
9,
12]. The rationale for the application of anti-EGFR monoclonal antibodies in KRAS exon 2 wt MCC cases depended on the appropriate concordance of mutational status between primary and metastatic tumours, as presented in previous literature [
22,
23]. Nevertheless, noteworthy differences in the incidence of KRAS exon 2 mutations among tumour locations have been reviewed [
8,
9,
24]. The superiority of CAP-B over CAP remains controversial, which precludes any recommendations [
2,
6,
7]. A growing but still very limited body of literature comparing the clinical efficacy of CAP-B and CAP in the management of Chinese postmenopausal women with untreated KRAS exon 2 wt MCC after 6-cycle CAPOX-B induction treatment demonstrated comparable outcomes [
10]. Chen et al. [
25] noticed a longer mPFS in postmenopausal women receiving CAP-B treatment than those recieving CAP treatment at a mean follow-up of 2 years. Our finding further expounded the significant differences in the mPFS between groups but were inconsistent with several prior retrospective reports that showed no significant differences in the mPFS [
14,
22]. Furthermore, a prospective study by Yamaguchi et al.[26]comprising 31 cases with untreated KRAS exon 2 wt MCC receiving CAP-B or CAP treatment after 6-cycle CAPOX-B induction treatment confirmed no significant difference in the mPFS.
As using chemotherapy alone in the current treatment only has a modest, if any, benefit, we wanted to evaluate whether CAP-B or CAP as maintenance treatment after 6-cycle CAPOX-B induction treatment could improve mPFS and/or mOS in untreated KRAS exon 2 wt MCC [
27]. Only a few 3 phase II trials comparing CAP-B with CAP in similar regimens showed no improvement in mPFS or mOS [
1,
26,
27]. Comparing with prior trials using the identical strategy with bevacizumab, the last study reported by Gervais et al.[18]failed to obtain benefit, although CAP-B, which had been investigated in a small population of 27 cases, had an extraordinary mOS of 2 years.
This study undoubtedly showed that Chinese postmenopausal women with untreated KRAS exon 2 wt MCC managed in the CAP-B or CAP setting have almost indistinguishable 2-year PFS and OS when treated using neoadjuvant chemotherapy first, followed by chemoradiation and then surgery [
16]. Moreover, compared to no addition, the addition of bevacizumab produced statistically significant benefit [
15]. The CAP-B or CAP maintenance treatment has been espoused by previous investigations [
1,
15,
28]. Based on the above facts, CAP-B or CAP has become the wide-reaching favoured approach for those women with untreated KRAS exon 2 wt MCC [
1]. The application of CAPOX-B administration rather than other regimens was based predominantly on a US Intergroup study indicating outstanding outcomes with CAP therapy [
1,
2]. However, the CAP strategy was observed to be relevant to outcomes inferior to those with the CAP-B strategy in Chinese postmenopausal women with untreated MCC. A previous trial evaluating CAP as a maintenance treatment failed to reach its primary endpoint of increasing the mPFS from 2 to 4 months [
18]. Hence, CAP showed no distinct benefit for the cohort. In addition, the end-points, as mentioned above, tended to be inferior to the outcomes from previous meta-analyses with bevacizumab [
29,
30]. The current study consequently seemed to support the treatment concept that combining bevacizumab with CAP in maintenance treatment was superior to CAP alone. Nevertheless, comparison of the results from different maintenance treatments was challenging [
15]. The attributes of research object were difficult to achieve relative consistency. Thus, the study only demonstrated that combining bevacizumab with CAP in these cases previously treated with 6-cycle CAPOX-B induction treatment tended to result in a benefit.
Despite the small sample of the study and the drawback of the retrospective study, the strength and consistency of the CAP-B treatment results corroborated the findings from previous multi-centre studies [
31,
32]. The current therapeutic regimen was clearly different from previous neoadjuvant chemotherapy regimens reported by several Asian researchers [
33,
34]. Given the positive mPFS and mOS of our results in the CAP-B setting in untreated KRAS exon 2 wt MCC, we have an optimistic opinion that the positive results were attributable to bevacizumab. However, despite the documented benefits of bevacizumab in the management of MCC, rare adverse reactions have been reported, i.e., thromboembolic events, diarrhea, wound healing abnormalities, irreversible leuco-encephalopathy syndrome [
29‐
31]. In addition, maintenance treatment with CAP-B does not appear to be cost-effective [
9].
This study should be interpreted in light of important limitations. First, the nature of the retrospective study with all the problems inherent with this methodology limits the level of confidence of our conclusion, and it is possible that every potential confounding variable, such as underlying diseases, failed to be addressed in our analyses. Second, limited sample size may have introduced bias. Nevertheless, the focus of our study is to assess an area that has not been studied extensively in the literature. Third, our analysis has the lack of generalisability because our study population included only postmenopausal women with KRAS exon 2 wild-type MCC. Fourth, every attempt was made to adjust for all potential confounders, but other unmeasured factors may also be relevant.