Treatment outcome comparisons of first‐line targeted therapy in patients with KRAS wild‐type metastatic colorectal cancer: A nationwide database study

Abstract Background The first‐line systemic therapy for metastatic colorectal cancer (mCRC) is a combination of one targeted therapy agent and a chemotherapy doublet. Whether bevacizumab or anti‐epidermal growth factor receptor (anti‐EGFR) monoclonal antibody (mAb) is the more effective addition to a chemotherapy doublet as the first‐line treatment for inoperable KRAS wild‐type mCRC remains controversial in prior clinical trials. Moreover, the association between the sidedness of primary tumors and the efficacy of anti‐EGFR mAb needs to be addressed. Methods We established a cohort of patients with KRAS wild‐type mCRC who were treated with first‐line targeted therapy plus doublet chemotherapy between 2013 and 2018 using Taiwan's National Health Insurance Research Database. Secondary surgery was defined as either resection of primary tumors, liver metastases, lung metastases, or radiofrequency ablation. Results A total of 6482 patients were included; bevacizumab and anti‐EGFR mAb were the first‐line targeted therapies in 3334 (51.4%) and 3148 (48.6%) patients, respectively. Compared with those who received bevacizumab, patients who received anti‐EGFR mAb exhibited significantly longer overall survival (OS; median, 23.1 vs. 20.2 months, p = 0.012) and time to treatment failure (TTF; median, 11.3 vs. 10 months, p < 0.001). Among left‐sided primary tumors, the OS and TTF benefits of anti‐EGFR mAb remained. Among right‐sided primary tumors, the OS and TTF were similar regardless of the type of targeted therapy. In multivariate analyses, first‐line anti‐EGFR mAb therapy remained an independent predictor of longer OS and TTF for left‐sided primary tumors. Patients who received anti‐EGFR mAb were more likely to receive secondary surgery (29.6% vs. 22.6%, p < 0.0001) than patients who received bevacizumab. Conclusion For patients who received first‐line doublet chemotherapy for KRAS wild‐type mCRC, adding anti‐EGFR mAb was associated with significantly longer OS and TTF, especially for left‐sided primary tumors.


| INTRODUCTION
Colorectal cancer (CRC) is one of most common cancers globally. 1,2 For inoperable metastatic colorectal cancer (mCRC), systemic therapy is the main treatment. The main frontline systemic chemotherapy agents include irinotecan, oxaliplatin, and 5-fluorouracil. Doublet combination chemotherapy is usually recommended, such as irinotecan in combination with 5-FU (FOLFIRI) or oxaliplatin in combination with 5-FU (FOLFOX). The first-line systemic therapy for mCRC is a combination of one targeted therapy agent and a chemotherapy doublet, such as FOLFIRI or FOLFOX, or a chemotherapy triplet in highly selected patients. 3,4 For KRAS wild-type mCRC, if a chemotherapy doublet backbone is applied instead of a triplet, two types of targeted therapy agents can be considered in combination. One is antivascular endothelial growth factor (VEGF) monoclonal antibody (mAb) bevacizumab, and the other is anti-epidermal growth factor receptor (anti-EGFR) mAb such as cetuximab or panitumumab.
Regarding which targeted therapy is the more suitable companion of a chemotherapy doublet for KRAS wildtype mCRC, randomized clinical trials have reported conflicting results. In the CALGB 80405 study, overall survival (OS) and progression-free survival (PFS) were similar between patients who received bevacizumab or cetuximab in combination with a chemotherapy doublet. 5,6 In the FIRE-3 study, patients who received cetuximab plus FOLFIRI had similar PFS but significantly longer OS compared with patients who received bevacizumab plus FOLFIRI. 7,8 By contrast, in the PEAK study, patients who received panitumumab plus FOLFOX had significantly longer PFS and OS compared with patients who received panitumumab plus FOLFOX. [9][10][11] Furthermore, although the PFS was similar, patients who received panitumumab plus FOLFOX had significantly longer OS compared with patients who received panitumumab plus FOLFOX in the PARADIGM study. 12 Therefore, whether anti-VEGF mAb or anti-EGFR mAb is suitable for incorporation in this first-line systemic therapy remains unclear. Larger scale of clinical evidence is warranted to address this uncertainty on a realistic level.
Taiwan's National Health Insurance (NHI) program is a mandatory single-payer system covering 97%-98% of the population. 13 The associated database provided an opportunity to examine an unselected national population who had received fully reimbursed systemic anticancer therapies. Thus, we conducted this study to compare the treatment outcomes of first-line targeted therapy agents added to a chemotherapy doublet as first-line therapy for mCRC.

| Data source
In this retrospective population-based cohort study, we linked data from three national databases, namely Taiwan's National Health Insurance Research Database for all inpatient and outpatient prescriptions, Taiwan Cancer Registry for patient demographics and disease status, and National Death Registry for death records. We retrieved data from Health and Welfare Data Science Center, Ministry of Health and Welfare, Taiwan. Detailed processes have been described in our previous studies. 14,15 In brief, all personal data extracted from these databases were encrypted and analyzed anonymously. Any results that applied to less than 3% of the targeted group were not processed to ensure privacy. The study was approved by the Institutional Review Board of National Taiwan University Hospital.

| Study population and variables
All patients who fulfilled the following criteria were included in this study: (1) patients who received a diagnosis of histology-proven primary CRC (ICD-O-3 C180 to C189, C19, and C20); (2) patients who were diagnosed as KRAS wild-type mCRC; (3) patients who were aged 18 years or over; (4) patients who received first-line targeted therapy of either bevacizumab, cetuximab, or panitumumab between January 1, 2013 and December 31, 2018, as the first-line systemic therapy for KRAS wildtype mCRC; and (5) received the targeted therapy in combination with either irinotecan or oxaliplatin. The exclusion criteria were as follows: (1) patients who had missing clinical data; (2) patients who had received a diagnosis of hematological malignancies or Kaposi's sarcoma (ICD-O-3 morphology code 9140, 9590-9989); (3) patients who received irinotecan and oxaliplatin at the same time; and (4) patients who had received another targeted agent before.

K E Y W O R D S
bevacizumab, cetuximab, EGFR, metastatic colorectal cancer, panitumumab, secondary surgery According to the targeted therapy agents used in the first-line treatment regimen, patients were classified into the bevacizumab group and the anti-EGFR mAb group. The index date was the date of the first dose of targeted therapy and chemotherapy. Secondary surgery referred to resection of primary tumors, liver metastases, or lung metastases or radiofrequency ablation to liver metastases after the index date. Left-sided CRC was defined as the primary tumor origin in the rectum, sigmoid colon, descending colon, or splenic flexure. Right-sided CRC was defined as the primary tumor origin in the cecum, ascending colon, hepatic flexure colon, or transverse colon. Elderly patients were defined as age ≧70 years old. 16

| Statistical analysis
OS was calculated from the index date to the date of death; the data were censored if patients survived beyond the last follow-up date of December 31, 2019. The time to treatment failure (TTF) was calculated from the index date to the first dose of the next line of chemotherapy. Continuous variables expressed as mean were compared using the independent t test or Wilcoxon rank-sum test, whereas categorical variables expressed as frequency was compared using the chi-squared test or Fisher's exact test, as appropriate. OS and TTF were estimated using the Kaplan-Meier method, and the log-rank test was employed to compare between-group differences. The Cox proportional hazards model was applied to explore potential predictors of OS and TTF in multivariate analysis. Two-sided p values <0.05 were considered statistically significant. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).
Compared with those who received bevacizumab, the patients who received anti-EGFR mAb as part of their first-line systemic therapy were more likely to be men (p < 0.001), be aged 70 years or over (p < 0.001), undergo therapy at medical centers (p < 0.001), and have left-sided primary tumors (p < 0.001); by contrast, they were less likely to have initial stage IV disease (p < 0.001), have adenocarcinoma histology (p < 0.001), and receive irinotecan as a part of the combination chemotherapy (p < 0.001; Table 1).
We further analyzed the treatment outcomes according to the sidedness of the primary tumor. Among the patients with left-sided mCRC, those who received anti-EGFR mAb in first-line systemic therapy exhibited significantly longer OS (median, 24.9 vs. 22.9 months, p = 0.021; Figure 2A) and TTF (median, 12.2 vs. 10.4 months, p < 0.001; Figure 2B) compared with those who received bevacizumab. Among the patients with right-sided mCRC, those who received either anti-EGFR mAb or bevacizumab in their first-line systemic therapy exhibited similar OS (median, 15.6 vs. 16.1 months, p = 0.385; Figure 2C) and TTF (median, 7.6 vs. 8.8 months, p = 0.146; Figure 2D).

| Multivariate analysis
In the multivariate analyses of the entire cohort data with adjustment for age, gender, sidedness of the primary tumor, initial CRC stage, histology, combination chemotherapy, and hospital level, receiving anti-EGFR mAb or bevacizumab in combination with first-line chemotherapy was not associated with OS (hazard ratio [HR] = 0.96, p = 0.195). By contrast, the addition of anti-EGFR mAb as the first-line targeted therapy remained an independent predictor of longer TTF (HR = 0.91, p < 0.001; Table 2,  Table S1). In addition, right-sided primary tumor and initial stage IV disease were independent predictors of poor OS and TTF (Table S1).
Because the sidedness of the primary tumor was associated with the choice of first-line targeted therapy, OS, and TTF (Tables 1 and 2), we performed separate multivariate analyses for the patients with left-sided and rightsided primary tumors. For the patients with left-sided primary tumors, first-line anti-EGFR mAb was an independent predictor of OS (HR = 0.93, p = 0.030) and TTF (HR = 0.84, p < 0.001; Table 3). For the patients with rightsided primary tumors, first-line anti-EGFR mAb was not associated with OS (p = 0.270) but was an independent predictor of shorter TTF (HR = 1.12, p = 0.047; Table 3). Initial stage IV disease was an independent predictor of poor OS and TTF, regardless of the sidedness of primary tumors (Table S1).

| Secondary surgery
Among the 6482 patients in this study, 1685 (26.0%) received secondary surgery after the initiation of firstline systemic therapy. The patients who received secondary surgery exhibited significantly longer OS than the patients who did not (median, 37.9 vs. 17.2 months, p < 0.001; Figure 3A). The patients who received anti-EGFR had a significantly higher chance of undergoing secondary surgery (29.6% vs. 22.6%, p < 0.001) and a significantly shorter time to secondary surgery compared with the patients who received bevacizumab (p < 0.001, Figure 3B). Among the patients who received secondary surgery, the median time to secondary surgery was 5.7 months in the anti-EGFR mAb group and 7.1 months in the bevacizumab group.
For the patients with left-sided primary tumors, those who received anti-EGFR had a significantly higher chance of undergoing secondary surgery (31.1% vs. 23.6%, p < 0.001) and a significantly shorter time to secondary surgery compared with those who received bevacizumab ( Figure 3C). Among the patients with leftsided primary tumors who received secondary surgery, the median time to secondary surgery was 5.7 months in the anti-EGFR mAb group and 7.2 months in the bevacizumab group.
For the patients with right-sided primary tumors, those who received anti-EGFR had a higher chance of undergoing secondary surgery (23.4% vs. 19.9%, p = 0.113) and a significantly shorter time to secondary surgery compared with those who received bevacizumab ( Figure 3D). Among the patients with right-sided primary tumors who underwent secondary surgery, the median time to  Abbreviations: EGFR, epidermal growth factor receptor; mAb, monoclonal antibody; 5-FU, 5-fluorourocil.
Note: All data are presented as N (%) unless indicated otherwise. a Patients may shift between 5-FU and its analogs. b p < 0.05. secondary surgery was 5.8 months in the anti-EGFR mAb group and 6.7 months in the bevacizumab group.

| DISCUSSION
In this nationwide cohort study, we demonstrated that among patients who received first-line chemotherapy doublets for inoperable KRAS wild-type mCRC, the addition of anti-EGFR mAb led to significantly longer OS and TTF compared with the addition of bevacizumab. This benefit was mainly observed in the patients with left-sided primary tumors. In the multivariate analysis, anti-EGFR mAb treatment remained an independent predictor of longer OS and TTF among those with left-sided primary tumors. To the best of our knowledge, this is the largest (n = 6482) cohort study focusing on this topic. Our study results are compatible with those of the FIRE-3 and PARADIGM randomized clinical trial studies, which have demonstrated that patients who receive anti-EGFR mAb exhibit significantly longer OS than patients who receive bevacizumab in combination with a chemotherapy doublet. [7][8][9][10][11]17 However, the phase III CALGB 80405 study reported similar OS between patients who received either bevacizumab or cetuximab as the combination targeted therapy agents. 5,6,9 A pooled analysis of  these three studies indicated that anti-EGFR mAb as the first-line targeted agent for RAS wild-type mCRC significantly benefitted OS, especially in patients with left-sided mCRC. 18 In Taiwan, our recent treatment consensus also addressed the importance of first-line anti-EGFR mAb in combination with the chemotherapy doublet, especially for left-sided tumors and for therapy with curative intent. 19 The association between the sidedness of primary tumors and the efficacy of anti-EGFR mAb has been reported in many unplanned subgroup analyses of randomized studies. 9,20-22 A meta-analysis of 12 randomized studies with propensity score matching verified this association 23 ; however, this study did not include the aforementioned FIRE-3 and PEAK studies. Thus, it was not surprising that the study did not reveal anti-EGFR mAb to be more effective than bevacizumab as the companion targeted therapy in first-line systemic treatment for left-sided KRAS wildtype mCRC. In another meta-analysis of six main clinical studies including FIRE-3 and PEAK studies, the benefit of anti-EGFR mAb for left-sided KRAS wild-type mCRC was disclosed. 24 Currently, anti-EGFR mAb is suggested with the preferred regimen for left-sided KRAS wild-type mCRC in most updated guidelines. 25,26 We determined several possible reasons for the survival benefit of first-line anti-EGFR mAb, especially in patients with left-sided primary tumors, in a single population without the need for combining multiple cohorts. We studied an unselected nationwide cohort with a large sample size through multivariate analysis. In addition, over 90% of the patients received irinotecan as the companion chemotherapy instead of oxaliplatin. The results of several clinical trials revealed that irinotecan-based doublet chemotherapy in combination with targeted therapy led to longer PFS and OS, although these were nonsignificant, than oxaliplatin-based doublet chemotherapy. [27][28][29] These results imply that irinotecan might exert more synergistic effects in combination with targeted therapy than oxaliplatin. 30 Whether anti-EGFR mAb can be considered a firstline targeted therapy option for patients with rightsided primary tumors remains controversial. In the National Comprehensive Cancer Network guideline, its use is not favored in patients with expanded RAS wild-type right-sided mCRC. 31 Contrarily, in the pan-Asian adapted European Society for Medical Oncology guideline, anti-EGFR mAb plus a chemotherapy doublet is as an option for patients who were unsuitable for bevacizumab plus a chemotherapy triplet. 4 In our data, among the patients with right-sided primary tumors, those who received bevacizumab exhibited significantly longer TTF than those who received anti-EGFR mAb, but this difference in TTF did not translate to longer OS. Our data support the use of anti-EGFR mAb as one of the first-line options for targeted therapy for right-sided mCRC, which is in line with a consensus guideline in Taiwan. 19 Secondary surgery after first-line treatment for mCRC provides a crucial chance of cure. 8,32 When combined with doublet chemotherapy, anti-EGFR mAb resulted in more frequent early tumor shrinkage and a deeper objective tumor response than bevacizumab, 33 which could lead to a higher secondary resection rate. However, in the PEAK study, the secondary resection rate was 14.7% and 16.2% in the panitumumab and bevacizumab arms, respectively. 34 In the DEEPER trial, cetuximab in combination with triplet chemotherapy induced significantly deeper tumor responses compared with bevacizumab plus a chemotherapy triplet 35 ; however, the secondary resection rate was similar between the two arms. Furthermore, in the PARADIGM study, the secondary resection rate was 16.5% and 10.9% in the panitumumab and bevacizumab arms, respectively. 17 Our study indicated that anti-EGFR mAb plus chemotherapy led to significantly higher secondary resection rates, regardless of the primary tumor location, which was compatible with the results in the PARADIGM study. Because of the retrospective nature of our study, we could not adjust for other critical confounding factors, such as tumor size and metastatic site. Whether anti-EGFR mAb results in a higher secondary surgery rate must be further explored.
Our current study had some limitations. First, the data of prescriptions not reimbursed through NHI were unavailable, including those of second-line bevacizumab treatment or immunotherapy paid for by patients themselves. However, our data reflected the actual scenario when second-line targeted therapy was not reimbursed. Second, information on genetic aberrations other than KRAS mutations, such as BRAF and DNA mismatch repair status, was unavailable. Instead, our data implied that even in the population that exhibited the NRAS or BRAF mutation, bevacizumab as a first-line targeted therapy did not offer greater survival benefits over anti-EGFR F I G U R E 3 (A) Kaplan-Meier plots illustrating the overall survival of patients according to whether they received secondary surgery or radiofrequency ablation (RFA); (B-D) Kaplan-Meier plots illustrating time to secondary surgery or RFA according to the treatment regimens of the (B) whole population, (C) patients with left-sided primary tumors, and (D) patients with right-sided primary tumors. p values were determined using the log-rank test. OS, overall survival; anti-EGFR, anti-epidermal growth factor; 2nd, secondary. mAb. Immunotherapy was proven to be beneficial for microsatellite instability-high colorectal cancer in multiple clinical trials, [36][37][38] but the enrollment period of our study was before the publication of these results. Many genetic alterations were found to be predictive of the efficacy of specific therapies, such as HER2 amplification and tumor mutation burden. 39 However, the evidence of these genetic alterations were for salvage therapy rather than first-line therapy. Finally, as mentioned, we could not verify whether the disease extent before systemic therapy differed between the anti-EGFR mAb group and bevacizumab group.
In conclusion, for the patients who received first-line doublet chemotherapy for KRAS wild-type mCRC, the addition of anti-EGFR mAb was associated with significantly longer OS and TTF than the addition of bevacizumab, especially for those with left-sided primary tumors.