Background
Colorectal cancer (CRC) remains one of the most common causes of morbidity and mortality in the world and is the second most common cause of cancer death in the United States, with approximately 53,200 deaths annually.1 Advances in molecular biology and genomic profiling have led to the identification of several genetic aberrations and tumor characteristics that allow for distinct subgroups of metastatic CRC (mCRC) with both prognostic and predictive implications. One of these subgroups is BRAF-mutated (BRAFmt) CRC. BRAF mutations are present in roughly 10% of patients with mCRC2 and portend a particularly unfavorable prognosis, with a median survival for patients with metastatic disease of <12 months.3,4 BRAFmt mCRC tumors are typically right-sided, poorly differentiated, and T4 staged, and they characteristically affect older women.5
BRAF mutations can be predictive of response to targeted therapy, as shown in the recently published BEACON trial.6 In this study, 655 patients with BRAFmt mCRC whose disease progressed on at least one prior line of therapy were randomized to cetuximab + the investigator’s choice of either FOLFIRI or irinotecan versus encorafenib (BRAF inhibitor) + cetuximab (anti–endothelial growth factor receptor [EGFR]) or encorafenib + cetuximab + binimetinib (MEK inhibitor). Objective responses were significantly higher and overall survival (OS) was significantly improved in the doublet and triplet arms compared with chemotherapy.6 Based on these results, encorafenib gained regulatory approval by the FDA, and the doublet of encorafenib + anti-EGFR therapy has become the second-line standard of care in this patient population.
Optimal first-line (1L) treatment of patients with BRAF-positive mCRC remains to be defined. Given the poor prognosis of BRAFmt mCRC, it has been hypothesized that giving FOLFOXIRI up front may give patients the best chance to be exposed to the most active agents in this disease. Indeed, a phase II trial with FOLFOXIRI in BRAFmt mCRC resulted in a median OS of 24.1 months, which compared favorably with historical control patients.7 A randomized phase III study recently evaluated 1L FOLFOXIRI + bevacizumab versus FOLFIRI + bevacizumab in patients with mCRC.8 The triplet arm resulted in improved OS (29.9 vs 25.8 months; hazard ratio [HR], 0.80; 95% CI, 0.65–0.98) at the expense of increased toxicity. The survival benefit seemed to be independent of BRAF mutations. However, the study only enrolled a small number of patients with BRAF mutations (n=28). Median OS in the BRAF and RAS wild-type (wt) group was 37.1 months compared with 13.4 months for the BRAFmt group (HR, 2.79; 95% CI, 1.75–4.46; P<.0001). These results again highlight the poor outcomes associated with BRAFmt mCRC.
Although potential improvement in survival with FOLFOXIRI has been suggested, current NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) for Colon Cancer (Version 4.2020) still do not give preference to one specific 1L cytotoxic regimen for patients with BRAFmt mCRC.9 As such, in this study we sought to use real-world data from a large cohort of patients in the United States to assess changes in BRAF testing rate over time, to determine the prevalence of FOLFOXIRI use, and to compare survival outcomes in mCRC stratified by BRAF mutation status and 1L therapy.
Methods
Data Source
This study used the nationwide Flatiron Health electronic health record (EHR)–derived deidentified database. During the study period, the deidentified data originated from a geographically diverse group of approximately 280 US cancer clinics (∼800 centers of care), representing a mixture of academic and community care centers. Deidentified patient-level data in the longitudinal Flatiron Health database include structured data, such as laboratory values, and unstructured data extracted from patient charts using a technology-enabled abstraction process.10,11 University of Utah Institutional Review Board approval of the study protocol was obtained before study conduct and included a waiver of informed consent.
Study Population
We reviewed the EHR-derived deidentified database for patients diagnosed with mCRC between January 1, 2013, and December 31, 2018. Patients who had documented BRAF mutation testing at any time within 6 months of metastatic diagnosis and received standard 1L chemotherapy were included for analysis. Standard 1L chemotherapy was defined as FOLFOX, CAPEOX, FOLFIRI, or FOLFOXIRI. Patients who did not have a visit or medication order within 90 days of metastatic diagnosis were excluded to ensure that patients were engaged with care at the data-providing institution.
Statistical Analysis
Statistical differences between groups were determined using chi-square tests for categorical variables and Wilcoxon rank sum tests for quantitative variables. Kaplan-Meier estimates with corresponding log-rank tests and Cox proportional hazards modeling were used to compare survival outcomes stratified by BRAF mutation status and 1L therapy received. Treatment comparative effectiveness was assessed for BRAFmt tumors via a matching weighted Cox proportional hazards model.12,13 Matching weights were constructed via treatment propensity scores estimated via a random forest14 based on age at metastatic diagnosis, sex, metastatic diagnosis year, mismatch repair (MMR) status, ECOG performance status at 1L initiation, and de novo versus non–de novo metastatic disease. Proportional hazards assumptions for multivariate Cox proportional hazards models were assessed using Schoenfeld tests and graphical inspection for no pattern in the scaled Schoenfeld residuals across time. An additional multivariate analysis controlled for the same potential confounders. The primary outcome of interest was OS, defined as the time from the date of metastatic diagnosis to the date of death. Statistical analysis was performed using R version 3.5.2 (R Foundation for Statistical Computing).15
Results
Study Cohort and Patient Characteristics
We selected 20,333 patients diagnosed with mCRC between 2013 and 2018. Within this group, we included 9,948 patients who received standard 1L chemotherapy and had a recorded visit or medication order within 90 days of diagnosis. Of these patients, 4,457 had documented BRAF testing results, with 466 (10.5%) listed as having tumors harboring a BRAFmt (Figure 1). Of the 423 BRAFmt tumors included in our multivariate analysis, 311 (73.5%) were V600E, 3 (0.7%) were V600K, 52 (12%) were “other,” and 57 (13%) were unknown.
Patient cohort selection flowchart.
Abbreviations: 1L, first-line; mCRC, metastatic colorectal cancer.
Citation: Journal of the National Comprehensive Cancer Network 20, 2; 10.6004/jnccn.2021.7059
Patient cohort selection flowchart.
Abbreviations: 1L, first-line; mCRC, metastatic colorectal cancer.
Citation: Journal of the National Comprehensive Cancer Network 20, 2; 10.6004/jnccn.2021.7059
Patient cohort selection flowchart.
Abbreviations: 1L, first-line; mCRC, metastatic colorectal cancer.
Citation: Journal of the National Comprehensive Cancer Network 20, 2; 10.6004/jnccn.2021.7059
Our analysis of the 466 BRAFmt mCRC tumors suggested that patients with these tumors are more likely to be female (55% vs 42% for patients with BRAFwt; P<.0001). BRAFmt mCRC tends to present at a later age, with a mean age at metastatic diagnosis of 64 years for BRAFmt versus 60 years for BRAFwt (P<.0001). We found that 7% of BRAFmt tumors had concurrent mutations in KRAS; 29% of these (10/34) were V600E, and the remaining were either unknown or “other.” In patients who were tested for microsatellite instability (MSI)/MMR, we found that 16% of BRAFmt tumors were MMR-deficient, compared with 3% of BRAFwt tumors (P<.0001). Demographic and clinical characteristics of the patients included in the study are summarized in Table 1.
Patient Characteristics
Uptake of BRAF Testing
Testing at any time in the patient disease course for the presence of the BRAF mutation was documented in only 14.1% of our study patients diagnosed with mCRC in 2013 (Figure 2). Although the proportion of patients with documented BRAF testing increased with each subsequent year, even in the most recent year (2018), only 56% of patients in this population had documented testing for a BRAF mutation (supplemental eTable 1, available with this article at JNCCN.org). In sum, of all the patients who met our inclusion criteria, only 34.9% had a documented test for the BRAF mutation.
Documented testing for BRAF mutation within 6 months of metastatic colorectal cancer diagnosis in patients receiving standard first-line treatment between 2013 and 2018.
aMay not have 6 months of follow-up in the testing data for those diagnosed in 2018.
Citation: Journal of the National Comprehensive Cancer Network 20, 2; 10.6004/jnccn.2021.7059
Documented testing for BRAF mutation within 6 months of metastatic colorectal cancer diagnosis in patients receiving standard first-line treatment between 2013 and 2018.
aMay not have 6 months of follow-up in the testing data for those diagnosed in 2018.
Citation: Journal of the National Comprehensive Cancer Network 20, 2; 10.6004/jnccn.2021.7059
Documented testing for BRAF mutation within 6 months of metastatic colorectal cancer diagnosis in patients receiving standard first-line treatment between 2013 and 2018.
aMay not have 6 months of follow-up in the testing data for those diagnosed in 2018.
Citation: Journal of the National Comprehensive Cancer Network 20, 2; 10.6004/jnccn.2021.7059
Prognosis of BRAF Mutations and Uptake of FOLFOXIRI
The median OS was 15.4 months in the BRAFmt group compared with 28.1 months in the BRAFwt group (Figure 3). A multivariate analysis adjusting for potential confounders, including 1L treatment, also showed the negative prognostic implications of a mutation in BRAF (HR, 0.48; 95% CI, 0.41–0.56; P<.001; supplemental eTable 2).
Kaplan-Meier curve showing overall survival of patients with metastatic colorectal cancer stratified by BRAF mutational status.
Citation: Journal of the National Comprehensive Cancer Network 20, 2; 10.6004/jnccn.2021.7059
Kaplan-Meier curve showing overall survival of patients with metastatic colorectal cancer stratified by BRAF mutational status.
Citation: Journal of the National Comprehensive Cancer Network 20, 2; 10.6004/jnccn.2021.7059
Kaplan-Meier curve showing overall survival of patients with metastatic colorectal cancer stratified by BRAF mutational status.
Citation: Journal of the National Comprehensive Cancer Network 20, 2; 10.6004/jnccn.2021.7059
Only 3% (n=16) of patients with BRAFmt tumors received 1L FOLFOXIRI, with a median OS of 13.8 months compared with 15.5 months in those receiving a chemotherapy doublet (FOLFOX, CAPEOX, or FOLFIRI; 95% CI, 4.9–not reached vs 14.3–19.0, respectively; P=.38; Figure 4).
Kaplan-Meier curve showing overall survival of patients with BRAF mutation receiving doublet chemotherapy (FOLFOX, CAPEOX, or FOLFIRI) versus triplet chemotherapy (FOLFOXIRI).
Citation: Journal of the National Comprehensive Cancer Network 20, 2; 10.6004/jnccn.2021.7059
Kaplan-Meier curve showing overall survival of patients with BRAF mutation receiving doublet chemotherapy (FOLFOX, CAPEOX, or FOLFIRI) versus triplet chemotherapy (FOLFOXIRI).
Citation: Journal of the National Comprehensive Cancer Network 20, 2; 10.6004/jnccn.2021.7059
Kaplan-Meier curve showing overall survival of patients with BRAF mutation receiving doublet chemotherapy (FOLFOX, CAPEOX, or FOLFIRI) versus triplet chemotherapy (FOLFOXIRI).
Citation: Journal of the National Comprehensive Cancer Network 20, 2; 10.6004/jnccn.2021.7059
In addition, in BRAFmt tumors, matching weighted analysis did not detect a significant improvement in OS with the use of FOLFIRI + bevacizumab (HR, 0.90; 95% CI, 0.58–1.39; P=.63) or FOLFOX/CAPEOX + bevacizumab (HR, 0.81; 95% CI, 0.52–1.26; P=.35) when compared with a chemotherapy doublet alone. Notably, matching weights achieved well-balanced comparison groups on age at metastatic diagnosis, gender, metastatic diagnosis year, MMR status, ECOG performance status at 1L initiation, and de novo versus non–de novo metastatic disease (all standardized mean differences <0.20). Results based on multivariate analysis were similar (Table 2).
Multivariate Survival Analysis of 1L Therapy Groups in Patients With BRAF Mutation
Discussion
There has been a recent surge of interest in the field of oncology to evaluate the efficacy of various treatment strategies in real-world populations. These analyses can help confirm whether the standard of care as determined in tightly controlled clinical trials is truly applicable to a more general population. This is the largest study evaluating real-world outcomes of BRAFmt mCRC with front-line therapies. There are several key findings from our study. First, we were able to confirm the negative prognostic implication of a BRAF mutation in mCRC in this large, real-world population. Second, we showed that although the documented BRAF testing rate in this US cohort has increased over time, in 2018 the testing rate still remained quite low at 56%. Third, we showed that the use of up-front triplet chemotherapy has not been widely adopted by oncologists in the United States for treatment of BRAFmt mCRC. Finally, our study did not detect a significant difference in OS between the most commonly used 1L treatment strategies for BRAFmt mCRC.
Previous studies have reported the clinicoepidemiologic characteristics of patients with BRAFmt CRC. Our analysis of real-world data is consistent with prior results and shows that patients with BRAF mutations are older, more likely to be female, and more likely to have MMR deficiency compared with patients with BRAFwt. Although we did not subanalyze OS based on BRAF and MSI/MMR status combined, it is known that BRAFmt/microsatellite stable subgroups tend to fare the worst, and therefore consideration of MSI/MMR status may be a useful additional tool for prognostication in BRAFmt mCRC.16 In addition, we were surprised at how many patients had concurrent KRAS and BRAF mutations (7%).
Prior literature has also highlighted the poor OS associated with mCRC harboring BRAF mutations.17 Our real-world data analysis supports this observation. We found that the median OS of patients with BRAFwt disease was almost double that of patients with BRAFmt tumors (28.1 vs 15.4 months). Multivariate analysis showed a similar pattern.
The TRIBE study showed numerically improved OS/progression-free survival in patients with BRAFmt mCRC treated with FOLFOXIRI + bevacizumab versus FOLFIRI + bevacizumab in the first-line setting. Although this finding was not statistically significant, the lower HR in this group compared with the RASwt group suggested a greater magnitude of benefit with the triplet in patients with BRAF mutations.8 However, the TRIBE2 trial did not confirm the benefit with triplet therapy in patients with BRAFmt mCRC (HR for second progression-free survival [PFS2], 1.23; 95% CI, 0.72–2.09, with PFS2 representing the time from randomization to objective tumor progression on next-line treatment, or death from any cause).18 A recent meta-analysis in patients with unresectable mCRC demonstrated improvement in OS with triplet chemotherapy + bevacizumab when compared with doublet chemotherapy + bevacizumab for mCRC in general.19 However, when looking specifically at BRAFmt tumors, that same benefit from triplet therapy was not demonstrated. Although only 3% of the patients in our analysis received 1L triplet chemotherapy with FOLFOXIRI + bevacizumab, our real-world data study showed no evidence of a survival benefit for triplet over doublet therapy (13.8 vs 15.5 months; P=.38).
Although adding antiangiogenics to cytotoxics in patients with BRAF mutations is often done in clinical practice, the role of antiangiogenics in the treatment of BRAFmt tumors remains to be definitively established.20–22 Post hoc analysis from studies in both the first- and second-line settings suggests that adding bevacizumab improves survival in patients with BRAFmt tumors. A limitation of this analysis is the small sample size. In the IFL trial, which evaluated the addition of bevacizumab to bolus fluorouracil + irinotecan in the 1L setting in mCRC, a trend toward improved survival was seen in the BRAFmt group (15.9 vs 7.9; HR, 0.11; 95% CI, 0.01–1.06). However, only 10 BRAFmt tumors were included in this analysis. The RAISE trial tested ramucirumab (a monoclonal antibody against VEGFR-2) in the second line in combination with FOLFIRI. In the BRAFmt subset (n=41), there was a nonstatistically significant trend toward improved OS in the ramucirumab arm (9 vs 4.2 months; HR, 0.54; 95% CI, 0.25–1.13). The VELOUR trial evaluated the benefit of adding aflibercept (a fusion protein that binds VEGF) to second-line FOLFIRI in patients with mCRC. The study showed a trend toward improved OS in the BRAFmt group (n=36) that received aflibercept (10.3 vs 5.5 months; 95% CI, 0.16–1.09).23 In our analysis (BRAFmt n=466), we were not able to show a survival benefit for patients with BRAF mutations who received any doublet therapy (FOLFOX/CAPEOX/FOLFIRI) + bevacizumab compared with those who did not receive bevacizumab. It is possible that the benefit seen in controlled study patients does not translate when this regimen is applied to all-comers with BRAFmt mCRC. However, additional efforts are needed to better understand the role of antiangiogenics in combination with cytotoxics or targeted therapy in BRAFmt tumors.
Of the 9,948 patients included in our study, we were intrigued to note that even in the year of highest testing (2018), only slightly more than half of the patients had a documented test for BRAF mutation (Figure 2). Importantly, of the 5,251 patients (44%) without documented BRAF testing, only 1,620 (30.7%) had a KRAS or NRAS mutation detected at any time. Thus, using the detection of a RAS mutation as an explanation for the absence of BRAF testing seems to be the case in only a fraction of patients. It is also important to note that this low BRAF testing rate was observed prior to publication of the BEACON trial results in 2019, which showed improved survival with targeted therapy compared with cytotoxics in the second line in V600E BRAF mCRC tumors.6 Ideally, these results should drive implementation of universal BRAF testing in this patient population, because encorafenib plus anti-EGFR has been established as a new standard of care in the second line in these patients, especially in the absence of MMR deficiency.
Our study has some limitations. We were able to include only patients who underwent palliative chemotherapy and had standard follow-up and documented BRAF testing. However, the characteristics of our BRAF patient population recapitulate what has been reported in other series and validate our data. The specific BRAF mutation was “unknown” for approximately 25% of our patient population, which may lead to some dilution of our analysis in patients with non-V600E mutations. In our study, only 16 patients received triplet chemotherapy in the front line, which precludes any definitive conclusion regarding the benefit of triplet therapy in this patient population. In addition, because this is a nonrandomized trial, there is a possibility for selection bias regarding which patients were offered FOLFOXIRI. Concerning BRAF testing, it is possible that there are missing data from patient charts, because the patients may have been tested out of the network our data were drawn from. Consequently, we cannot extrapolate the testing rate to the entire US mCRC population. In addition, because our inclusion was BRAF testing known within 6 months of diagnosis, it is possible that we missed patient testing data from 2018 due to a patient being diagnosed late in the year without mutational testing yet resulted. Final limitations include the retrospective study design and the potential for inclusion bias and unmeasured confounders.
Conclusions
In this real-world data analysis, we did not find significant differences in the OS of patients with BRAFmt mCRC based on which type of 1L therapy they received. This analysis supports that the presence of BRAFmt portends poor survival outcomes in mCRC. Finally, despite the known negative prognostic impact from BRAF mutations, documented testing rates remain low, with only 56% of our study patients undergoing testing in 2018. The results of the BEACON trial, published in 2019,6 are likely to drive an increase in BRAF testing in the future.
References
- 1.↑
Siegel RL, Miller KD, Goding Sauer A, et al. Colorectal cancer statistics, 2020. CA Cancer J Clin 2020;70:145–164.
- 2.↑
Barras D, Missiaglia E, Wirapati P, et al. BRAFV600E mutant colorectal cancer subtypes based on gene expression. Clin Cancer Res 2017;23:104–115.
- 3.↑
Wang J, Shen J, Huang C, et al. Clinicopathological significance of BRAFV600E mutation in colorectal cancer: an updated meta-analysis. J Cancer 2019;10:2332–2341.
- 4.↑
Samowitz WS, Sweeney C, Herrick J, et al. Poor survival associated with the BRAF V600E mutation in microsatellite-stable colon cancers. Cancer Res 2005;65:6063–6069.
- 5.↑
Loupakis F, Moretto R, Aprile G, et al. Clinico-pathological nomogram for predicting BRAF mutational status of metastatic colorectal cancer. Br J Cancer 2016;114:30–36.
- 6.↑
Kopetz S, Grothey A, Yaeger R, et al. Encorafenib, binimetinib, and cetuximab in BRAF V600E–mutated colorectal cancer. N Engl J Med 2019;381:1632–1643.
- 7.↑
Loupakis F, Cremolini C, Salvatore L, et al. FOLFOXIRI plus bevacizumab as first-line treatment in BRAF mutant metastatic colorectal cancer. Eur J Cancer 2014;50:57–63.
- 8.↑
Cremolini C, Loupakis F, Antoniotti C, et al. FOLFOXIRI plus bevacizumab versus FOLFIRI plus bevacizumab as first-line treatment of patients with metastatic colorectal cancer: updated overall survival and molecular subgroup analyses of the open-label, phase 3 TRIBE study. Lancet Oncol 2015;16:1306–1315.
- 10.↑
Ma X, Long L, Moon S, et al. Comparison of population characteristics in real-world clinical oncology databases in the US: Flatiron Health, SEER, and NPCR [preprint published online May 30, 2020]. medRxiv, doi: 10.1101/2020.03.16.20037143
- 11.↑
Birnbaum B, Nussbaum N, Seidl-Rathkopf K, et al. Model-assisted cohort selection with bias analysis for generating large-scale cohorts from the EHR for oncology research [preprint published online January 13, 2020]. ArXiv, arXiv:2001.09765v1
- 12.↑
Li L, Greene T. A weighting analogue to pair matching in propensity score analysis. Int J Biostat 2013;9:215–234.
- 13.↑
Yoshida K, Hernández-Díaz S, Solomon DH, et al. Matching weights to simultaneously compare three treatment groups: comparison to three-way matching. Epidemiology 2017;28:387–395.
- 15.↑
R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2013.
- 16.↑
Lochhead P, Kuchiba A, Imamura Y, et al. Microsatellite instability and BRAF mutation testing in colorectal cancer prognostication. J Natl Cancer Inst 2013;105:1151–1156.
- 17.↑
Modest DP, Ricard I, Heinemann V, et al. Outcome according to KRAS-, NRAS- and BRAF-mutation as well as KRAS mutation variants: pooled analysis of five randomized trials in metastatic colorectal cancer by the AIO colorectal cancer study group. Ann Oncol 2016;27:1746–1753.
- 18.↑
Cremolini C, Antoniotti C, Rossini D, et al. Upfront FOLFOXIRI plus bevacizumab and reintroduction after progression versus mFOLFOX6 plus bevacizumab followed by FOLFIRI plus bevacizumab in the treatment of patients with metastatic colorectal cancer (TRIBE2): a multicentre, open-label, phase 3, randomised, controlled trial. Lancet Oncol 2020;21:497–507.
- 19.↑
Cremolini C, Antoniotti C, Stein A, et al. Individual patient data meta-analysis of FOLFOXIRI plus bevacizumab versus doublets plus bevacizumab as initial therapy of unresectable metastatic colorectal cancer [published online August 20, 2020]. J Clin Oncol, doi: 10.1200/JCO.20.01225
- 20.↑
Heinemann V, von Weikersthal LF, Decker T, et al. FOLFIRI plus cetuximab versus FOLFIRI plus bevacizumab as first-line treatment for patients with metastatic colorectal cancer (FIRE-3): a randomised, open-label, phase 3 trial. Lancet Oncol 2014;15:1065–1075.
- 21.↑
Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004;350:2335–2342.
- 22.↑
Van Cutsem E, Tabernero J, Lakomy R, et al. Addition of aflibercept to fluorouracil, leucovorin, and irinotecan improves survival in a phase III randomized trial in patients with metastatic colorectal cancer previously treated with an oxaliplatin-based regimen. J Clin Oncol 2012;30:3499–3506.
- 23.↑
Wirapati P, Pomella V, Vandenbosch B, et al. VELOUR trial biomarkers update: impact of RAS, BRAF, and sidedness on aflibercept activity [abstract]. J Clin Oncol 2017;35(Suppl):Abstract 3538.
