Unlike multicenter studies, this single-center study performed retrospective comparisons of the efficacy of chemo-targeted agents in WT RAS mCRC. KRAS mutations predict poor prognosis and response from cetuxamab and panitumumab therapies in CRC, hence we ruled out this as a prognostic factor and selected patients with WT RAS CRC as our candidates (18). In preliminary pathological gene profiles, only 38 patients (29.5%) had all WT RAS status, and the majority of patients (69, 53.5%) had WT KRAS/NRAS exons 2 and 3. Low prevalence of KRAS or NRAS exon 4 mutation was found in a retrospective cohort study in North Africa (19), however, potential KRAS or NRAS exon 4 mutations were not assessed comprehensively in our database.
In our real-world data, therapy with cetuximab/panitumumab led to better OS and PFS than therapy with bevacizumab, and the difference in OS even reached significance. Notably, OS and PFS were better in patients with left-sided primary tumors and on cetuximab/panitumumab than those receiving bevacizumab, but no significant difference was found. In the patients with right-sided primary tumors, our data showed better survival outcomes in those receiving cetuximab/panitumumab as the first-line agent, but no statistical difference was found. Primary tumor sidedness is a prognostic factor in patients with WT RAS mCRC, and right-sided primary tumors was associated with poor prognosis. A cohort study with 178 patients with mCRC showed similar results in that superior OS and PFS were observed when anti-EGFR was added to therapy for patients with for left-sided primary tumors (20).
A meta-analysis of RCTs with large sample sizes (21) showed that the ORR and OS (overall response) after anti-EGFR therapy were superior to those after anti-VEGF therapy in patients with WT RAS mCRC (odds ratio = 0.79, 95% CI = 0.68–0.92, p = 0.002) but not in patients with WT KRAS mCRC (odds ratio = 0.81, 95% CI = 0.64–1.01, p = 0.06). However, the exact mechanism has not been clarified yet. Early tumor shrinkage may be beneficial when anti-EGFR is added as first-line chemotherapy for WT RAS mCRC (22). Our study’s results are compatible with the abovementioned results. The reasons for the difference in PFS may be as follows: there was a relatively small number of patients in each group; there was a high rate of peritoneal metastasis in the bevacizumab group, which implied poor treatment response and prognosis (23); and a higher number of patients in the bevacizumab group (n = 9) compared with the cetuximab/panitumumab group (n = 5) regardless of the incomplete mutation data (24).
The poor prognoses of patients with mCRC are due to right-sided primary tumor. Wu et al. (25) recently performed a network meta-analysis of large RCTs and found that in patients with WT RAS disease, anti-EGFR is the most effective treatment for left-sided, and bevacizumab is effective for right-sided tumors. Our real-world data showed a similar effect, that is, using cetuximab/panitumumab provided nearly 4 months of survival benefit compared with using bevacizumab for patients with left-sided tumors, but no benefits to survival were observed when bevacizumab was used for patients with right-sided tumors. A few patients with right-sided tumors (n = 32) were present in the cohort, and the prevalence of RAS and BRAF mutations may have a prognostic impact. Our study ruled out RAS mutation possibility at initial enrollment. However, a previous review (26) revealed patients with right-sided tumors had significantly higher prevalence of BRAF mutations than patients with left-sided tumors (16.3% vs. 4.3%). Hence, we applied the current consensus of using anti-EGFR for patients with WT RAS/BRAF and left-sided tumor. With regard to patients with right-sided tumors, further discussions through MDT conference should be performed case by case.
In the early 2000s, only 10–25% of patients with CRC liver metastasis underwent liver metastasectomy according to the extent of their disease. Owing to inadequate liver remnant, hepatic surgery is unsuitable for many patients (27). Surgical intervention, including primary resection, secondary resection, and even third resection after systemic treatment, plays an important role in patients with mCRC. In the FIRE-3 central review analysis, patients with resectable disease and the best response who underwent resection had OS superior to those who did not undergo resection (51.3 vs. 30.8 months) (28). Another study reported that 22% of patients underwent surgery after treatment, which is higher than the percentage for other clinical trials (9, 11, 29). Meanwhile, a small trial with liver-limited mCRC (30) showed similar rates of resectability or conversion. The liver-limited mCRC conversion rate in our study was 34.1% in the cetuximab/panitumumab group and 23.8% in the bevacizumab group, and these values are similar to those of the above-mentioned trials. In a total of 82 patients with non-resectable liver metastases from CRC, no significant benefits on disease-free survival were shown for bevacizumab when added to chemotherapy (31). No comparison of conversion rate with additional anti-EGFR or anti-VEGF after first-line systemic treatment is available for WT RAS mCRC (31–33). Further studies or trials should be performed.
Study limitations. Firstly, a single-center population with a certain number of patients was used, but this limitation is unlikely to have influenced the results. Panitumumab was officially included in the Taiwanese national health payment guideline in June 2018 because of the US National Institutes of Health policy. Secondly, survival benefits that can be obtained from different targeted agents for WT RAS mCRC cannot be statistically compared because of the small number of patients involved. Further research is necessary to elucidate the association. Thirdly, perfectly matching cases for propensity score-matching analysis were not obtained. Thus, the results should be managed and interpreted with caution. Fourthly, our Pathological Department has started to use complete KRAS/NRAS panels for tissue genetic typing in recent years, and thus some patients had only KRAS exon 2 and 3 data or the deficiency mutation profile of the KRAS/NRAS exon 4.