Abstract
Background
Encorafenib-cetuximab has been approved for pretreated BRAFV600E-mutated metastatic colorectal cancer (mCRC) patients based on efficacy demonstrated in the randomized phase III BEACON trial. The aim of this real-world effectiveness study is to improve knowledge on the generalizability of trial results.
Methods
This population-based real-world study includes all mCRC patients in the Netherlands treated with encorafenib-cetuximab since approval. Individual patient data and pathology reports were collected. Overall survival (OS) was compared to BEACON and subgroup analyses were conducted for patients who would have been eligible and ineligible for BEACON.
Results
166 patients were included with a median follow-up time of 14.5 months. Median OS was 6.7 months (95% CI:6.0–8.3) and differed from BEACON (9.3 months; 95% CI:8.0–11.3, p-value 0.002). Thirty-six percent of real-world patients would have been ineligible for the BEACON trial. Trial ineligible subgroups with symptomatic brain metastases and WHO performance status ≥2 had the poorest median OS of 5.0 months (95% CI:4.0-NR) and 3.9 months (95% CI:2.4-NR).
Conclusion
This real-world cohort of mCRC patients treated with encorafenib-cetuximab showed a clinically relevant efficacy-effectiveness gap for OS. The chance of survival benefit from encorafenib-cetuximab in patients with brain metastases and/or WHO performance status ≥2 is negligible as neither efficacy nor effectiveness has been demonstrated.
Subject terms: Cancer therapy, Colorectal cancer
Background
Metastatic colorectal cancer (mCRC) is a heterogeneous disease characterized by molecularly distinct subgroups with different prognosis and response to treatment [1]. The BRAFV600E mutation occurs in approximately 10% of mCRC patients treated with systemic treatment within clinical trials while reaching up to 20% of mCRC patients in real-world settings, and is associated with an unfavorable prognosis [2, 3]. The randomized phase III BEACON trial showed a median overall survival (OS) benefit of 3.4 months for the experimental group receiving encorafenib-cetuximab treatment versus the control group receiving irinotecan-based chemotherapy with cetuximab in pretreated BRAFV600E-mutated mCRC patients [4, 5]. Consequently, encorafenib-cetuximab combination treatment received regulatory approval by the European Medicine Agency (EMA) and the Food and Drug Administration (FDA) and was included as treatment option in (inter)national guidelines for pretreated patients with BRAFV600E-mutated mCRC [6–8].
Randomized controlled trials (RCT’s) - such as the BEACON trial - are essential to prove treatment efficacy and lead to approval and registration of new oncological therapies. However, trials are conducted under strictly controlled conditions and patients are enrolled based on increasingly stringent eligibility criteria, resulting in a highly selected homogeneous population that is not representative of the whole patient population in daily clinical practice [9, 10]. This can lead to an efficacy-effectiveness gap; efficacy describes how an intervention performs under ideal circumstances, while effectiveness describes how it performs under conditions of real-world clinical practice where biological and behavioral variability is often more prevalent [11]. Population-based real-world studies which include all patients that received a certain treatment are well suited to improve knowledge on the generalizability of trial results for daily clinical practice. These studies can identify predictors for survival benefit and support decision making in everyday clinical practice. Our aim is to provide insight in the survival of patients with BRAFV600E-mutated mCRC treated with encorafenib-cetuximab using a population-based real-world cohort from the Netherlands.
Methods
Study design and data collection
This population-based real-world cohort study includes all mCRC patients in the Netherlands treated with encorafenib-cetuximab since approval in October 2020 until June 2022. Follow-up was available until February 2023. Patients were identified by the Dutch Hospital Data foundation (DHD). Hospitals in the Netherlands are required by law to provide monthly data to DHD on all dispensed (expensive) medication irrespective of billing status. Hence, this method ensures inclusion of all patients, without selection, who were treated with encorafenib-cetuximab for mCRC. After identification, individual patient data was collected from electronic patient records by qualified data managers of the Netherlands Comprehensive Cancer Organization (IKNL). IKNL collects data in the Netherlands Cancer Registry (NCR) from all patients with cancer and uses an opt-out approach to consent for this registry. All data were pseudonymized prior to the transfer from the NCR to the researchers. Additionally, original pathology reports were obtained for each patient by linkage with the The Dutch Nationwide Pathology Databank (PALGA) which covers pathology results of the entire Dutch population. Mutation analysis techniques and results were retrieved from these reports.
The Dutch mCRC treatment guideline recommends encorafenib-cetuximab for patients with BRAFV600E-mutated mCRC after progression on first-line treatment [12]. First-line treatment options include either doublet chemotherapy (a fluoropyrimidine plus oxaliplatin or irinotecan) with bevacizumab, or triplet chemotherapy (a fluoropyrimidine plus oxaliplatin and irinotecan) with bevacizumab, based on tumor-location, patient age, and performance status for patients with proficient mismatch repair (pMMR) or microsatellite stable (MSS) tumors [8]. Pembrolizumab is recommended for mCRC patients with tumors harboring deficient mismatch repair (dMMR) or microsatellite instability (MSI). Trifluridine-tipiracil is to be considered as a last treatment line in relatively fit patients. Monotherapy with epidermal growth factor receptor (EGFR) inhibitors is not recommended for patients with BRAF V600E-mutated tumors. Ramucirumab, aflibercept, and regorafenib are not reimbursed in the Netherlands.
Population-based real-world cohort
All patients with mCRC and administration of at least one dose of encorafenib-cetuximab were included. None of the included patients were treated with encorafenib-cetuximab as part of a clinical trial or a nominal use program. Subgroup analyses were conducted for patients that would have been eligible and ineligible for the BEACON trial. Categorization was based on key eligibility criteria of the BEACON trial and included: BRAFV600E mutation, progression of disease after 1 or 2 prior regimens in the metastatic setting, world health organization (WHO) performance status of 0 or 1 at start of encorafenib-cetuximab treatment, neutrophil count of ≥1.5 × 109/L, absence of symptomatic brain or leptomeningeal metastasis, no prior treatment with any RAF, MEK or EGFR inhibitor and no concurrent or previous invasive malignancy within five years of encorafenib-cetuximab treatment initiation. Patients were categorized as ineligible if one or more key eligibility criteria were not met. In case of missing data on key eligibility criteria patients were excluded from this analysis, however a sensitivity analysis with imputed data, including all patients, was performed.
Statistical analysis
Patient characteristics were compared using chi-square test for categorical variables and independent t-test for continuous variables. Primary endpoint was OS, defined as the length of time from initiation of encorafenib-cetuximab treatment to death by any cause. OS was estimated using the Kaplan–Meier method with censoring on the date of last follow-up if patients were still alive. The log-rank test was used to compare OS between subgroups. Median follow-up time was estimated by reversed Kaplan–Meier analyses.
To compare OS estimated in the population-based real-world cohort with OS estimated in the BEACON trial, the Kaplan–Meier curve of patients treated with encorafenib-cetuximab in the BEACON trial was reconstructed applying the validated method of Wei and Royston [13]. First, WebPlotDigitizer (version 4.6) was used to extract the data points of survival probability of the published updated survival curves of BEACON [4]. These extracted data points and the information on the number of patients at risk were then used to reconstruct the Kaplan–Meier curves and estimate 1-year OS. A univariable and multivariable Cox regression model was derived including relevant prognostic factors based on literature and expert opinion [14]. Categorical variables with less than 5 patients per group were excluded for multivariable analysis. Multiple imputation by a substantive model compatible with fully conditional specification was used for missing data for multivariable analysis [15]. Proportional hazard assumption was statistically tested with Schoenfeld residuals, and was not violated after stratifying for the variable ‘presence of other malignancies’. A p-value of <0.05 was considered statistically significant. All analyses were performed in R version 3.5.1 (packages ‘gtsummary’, ‘reconstructKM’, ‘prodlim’, ‘survminer’, ‘survival’, ‘smfcs’ and ’Table 1’ were used) [16].
Table 1.
Patient and tumor characteristics.
| Real-world cohort (N = 166) | Trial eligible real-world cohorta (N = 74) | Trial ineligible real-world cohorta (N = 41) | p value (comparison of trial eligible and trial ineligible real-world cohort) | |
|---|---|---|---|---|
| Age | ||||
| Mean age in years (SD) | 63.5 (10.3) | 62.4 (10.6) | 65.9 (8.7) | 0.1 |
| Median age in years (range) | 64 (29–85) | 63 (29–81) | 67 (45–81) | |
| Sex, no. of male | 72 (43%) | 35 (47%) | 19 (46%) | 0.9 |
| Time to metastases, synchronous | 113 (68%) | 48 (65%) | 25 (61%) | 0.7 |
| Primary tumor resected | 98 (59%) | 44 (59%) | 29 (71%) | 0.2 |
| WHO performance status | ||||
| 0 | 24 (23%) | 21 (28%) | 3 (10%) | (eligibility criterium) |
| 1 | 65 (62%) | 53 (72%) | 12 (39%) | |
| ≥2 | 16 (15%) | 0 (0%) | 16 (52%) | |
| Missing | 61 | – | 10 | |
| Location primary tumor | ||||
| Right-sided | 113 (68%) | 55 (74%) | 26 (63%) | 0.1 |
| Left-sided, including rectum | 51 (31%) | 17 (23%) | 15 (37%) | |
| Otherb | 2 (1%) | 2 (3%) | 0 (0%) | |
| Involvement of ≥3 metastatic sites | 83 (50%) | 33 (45%) | 25 (61%) | 0.1 |
| Presence of liver metastases | 101 (61%) | 44 (59%) | 26 (63%) | 0.7 |
| Presence of peritoneal metastases | 97 (58%) | 46 (62%) | 26 (63%) | 0.9 |
| Previous lines of therapy | ||||
| 1 | 110 (66%) | 49 (66%) | 15 (37%) | (eligibility criterium) |
| 2 | 46 (28%) | 25 (34%) | 16 (39%) | |
| ≥3 | 10 (6%) | 0 (0%) | 10 (24%) | |
| Prior oxaliplatin exposure | 149 (90%) | 71 (96%) | 34 (83%) | 0.018 |
| dMMR/MSI | 8 (5%) | 3 (4%) | 1 (3%) | 0.7 |
| Baseline CEA level >5 µg/L | 114 (87%) | 57 (77%) | 33 (89%) | 0.7 |
CEA carcinoembryonic antigen, dMMR mismatch repair deficiency, MSI microsatellite instability, SD standard deviation, WHO world health organization.
aIf all key eligibility criteria were met, but WHO performance status was missing, trial eligibility was considered unknown (N = 51). If one or more key eligibility criteria was not met and WHO performance status was missing patients were categorized as ineligible.
bOther is left- and right-sided due to a second primary malignancy or unknown.
Results
Patient population
A total of 166 mCRC patients were treated with encorafenib-cetuximab combination therapy in the Netherlands since approval. The median age was 64 years, 43% were male and 85% had a WHO performance score of 0-1 at start of encorafenib-cetuximab and 15% of ≥2 (Table 1). Patients with a WHO performance status of ≥2, included 15 patients with a WHO performance status of 2 and one patient with a WHO performance status of 4. There were no significant differences in patient and treatment characteristics between patients with an unknown WHO performance status (N = 61) compared to known WHO performance status (N = 105) except for the proportion of patients with only 1 prior treatment line which is 82% for patients with unknown WHO PS versus 58% in patients with known WHO PS (Supplementary Table 1).
Almost all patients were previously exposed to fluoropyrimidines (N = 163, 98%), and the large majority to oxaliplatin (N = 149, 90%) and anti-vascular endothelial growth factor (anti-VEGF, N = 124, 75%). Three patients were treated with another BRAF inhibitor within a clinical trial: one patient before and two patients after encorafenib-cetuximab treatment. BRAF mutation status was available for all patients. Mutation analysis was found to be performed on either primary tumor and/or metastatic tissue by next generation sequencing (NGS) in 134 cases (81%), Idylla in 20 cases (12%), Sanger sequencing in 3 cases (2%) and an unknown method in 9 cases (5%). BRAFV600E mutation was not confirmed in all patients; four patients had tumors harboring a BRAFnon-V600E mutation: p.D594N, p.A594G, p.K601N, and p.V600K/R. Mismatch repair status and/or microsatellite stability was available for 163 patients (98%) and was determined by immunohistochemistry in 152 cases (93%), by Idylla in 3 cases (2%), by polymerase chain reaction (PCR) in 1 case (1%) and an unknown method in 7 cases (4%). Three out of eight patients with a dMMR/MSI tumor were treated with mono-immunotherapy: two patients before start of encorafenib-cetuximab and one patient after progression on encorafenib-cetuximab (Supplementary Table 2).
The real-world patient population was divided in two subgroups: BEACON trial eligible (N = 74, 64%), and BEACON trial ineligible patients (N = 41, 36%). Patients with missing data on WHO performance status (N = 61) could partly be categorized due to ineligibility on other key eligibility criteria (N = 10/61). The remaining patients (N = 51) with missing data on a key eligibility criterium were excluded from this analysis. Ineligibility was based on one or more of the following reasons: no BRAFV600E mutation (N = 4), more than 2 prior regimens (N = 10), WHO performance status ≥2 (N = 16), neutrophil count of ≤1.5 × 109/L (N = 3), presence of symptomatic brain or leptomeningeal metastasis (N = 7), prior treatment with RAF, MEK or EGFR inhibitor (N = 2) and concurrent other invasive malignancies (N = 9). Trial eligible patients and trial ineligible patients did not differ significantly with regards to tumor and patient characteristics except for previous oxaliplatin exposure (96 versus 83%, Table 1).
Comparison with the BEACON trial
Patient characteristics of our real-world cohort (N = 166) differed from the 220 patients treated with encorafenib-cetuximab in the BEACON trial in median age (64 versus 61 years), primary tumor sidedness (68 versus 50% right-sided), WHO performance status (23 versus 51% WHO 0), previous lines of therapy (28 versus 34% with 2 prior lines and 6 versus 0% with ≥3 prior lines), previous oxaliplatin exposure (90 versus 95%) and baseline carcinoembryonic antigen (CEA) > 5 µg/L (87 versus 70%, Supplementary Table 3). Median duration of exposure to encorafenib-cetuximab treatment was 17 weeks in our cohort compared to 19 weeks in the BEACON trial. Characteristics of the 74 trial eligible cohort patients (64%) were most comparable to the BEACON population treated with encorafenib-cetuximab with regard to median age (63 years versus 61 years), previous lines of therapy (one previous line 66 versus 66%) and prior oxaliplatin exposure (96 versus 95%), although there were differences in WHO performance (WHO 0 was 28 versus 51%) and location of primary tumor (right-sided 74 versus 50%) (Supplementary Table 3). The 41 trial ineligible cohort patients (36%) were less comparable with a higher median age (67 years), less frequently received one previous line of therapy (37%) and were less often previously exposed to oxaliplatin (83%) (Supplementary Table 3).
Overall survival
Median follow-up time of the total real-world cohort was 14.5 months (95% CI: 12.6–23.6, range: 1.5–27.3 months), 128/166 (77%) patients died during follow-up time. Median OS was 6.7 months (95% CI: 6.0–8.3) which is a significant difference from the encorafenib-cetuximab arm in the BEACON trial (9.3 months; 95% CI: 8.0–11.3, p-value = 0.002, Fig. 1) with a hazard ratio (HR) of 1.47 (95% CI: 1.15–1.88). The subgroup of trial ineligible patients showed an inferior median OS of 6.0 months (95% CI: 4.9–9.2) compared to 7.3 months (95% CI: 6.4–10.5) for trial eligible cohort patients (p = 0.040) with a HR of 1.63 (95% CI: 1.05–2.53). Patients with unknown WHO performance status showed similar OS compared to patients with known WHO performance status (Supplementary Fig. 1A) and a sensitivity analysis with imputed data to categorize patients with unknown trial eligibility as either trial eligible or trial ineligible, demonstrated similar results (Supplementary Fig. 1B). Compared to BEACON, median OS was significantly shorter for trial ineligible cohort patients (p < 0.001), however not for trial eligible cohort patients (9.3 versus 7.3 months, p = 0.15) (Fig. 1b). 1-year OS was 14% for trial ineligible cohort patients and 30% for trial eligible cohort patients versus 42% for patients treated with encorafenib-cetuximab in the BEACON trial. Survival was especially short for patients who were trial ineligible due to symptomatic brain metastases (median OS 5.0 months; 95%: 4.0-NR), and due to WHO performance status ≥2 (median OS 3.9 months; 95% CI: 2.4-NR) (Fig. 1). Survival for patients with a BRAFnon-V600E mutation (N = 4) was heterogeneous, ranging from 4.2 months to 11.4 months. In multivariable analysis, age (HR 0.98, 95% CI: 0.95–0.99, p = 0.024), involvement of ≥3 metastatic sites (HR 1.61, 95% CI: 1.04–2.50, p = 0.034) and WHO performance status of ≥2 (HR 2.67, 95% CI: 1.32–5.40, p = 0.007) were independently associated with shorter OS (Table 2).
Fig. 1. Kaplan–Meier estimates of the overall survival of mCRC patients treated with encorafenib-cetuximab.
Patients treated in the BEACON trial versus the real-world cohort (a, b) and stratified by brain metastases and WHO performance status in the real-world cohort (c, d).
Table 2.
Cox regression model of the association of prognostic factors on OS in the real-world cohort.
| Univariable regression | Multivariable regressiona | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Variable | Category | Reference | N | HR | 95% CI | p-value | HR | 95% CI | p-value |
| Patient and tumor characteristics | |||||||||
| Age | – | – | 166 | 0.98 | 0.96, 0.99 | 0.009 | 0.98 | 0.95, 0.99 | 0.024 |
| Sex | Female | Male | 166 | 1.01 | 0.72, 1.44 | 0.9 | 0.90 | 0.60, 1.33 | 0.6 |
| Sidedness | Left-sided, including rectum | Right-sided | 164 | 1.29 | 0.89, 1.86 | 0.2 | 1.13 | 0.73, 1.76 | 0.6 |
| Time to metastases | Synchronous | Metachronous | 166 | 1.32 | 0.91, 1.92 | 0.1 | 0.90 | 0.52, 1.54 | 0.7 |
| Primary tumor resection | Resection | No resection | 166 | 0.61 | 0.43, 0.87 | 0.006 | 0.65 | 0.39, 1.09 | 0.1 |
| Mismatch repair status | dMMR/MSI | pMMR/MSS | 163 | 0.99 | 0.46, 2.14 | 0.9 | 1.65 | 0.60, 4.53 | 0.3 |
| Metastatic site involvement | Involvement ≥3 organs | 1-2 organs | 166 | 1.71 | 1.20, 2.43 | 0.003 | 1.61 | 1.04, 2.50 | 0.034 |
| Liver metastases | Involvement | No involvement | 166 | 1.56 | 1.09, 2.25 | 0.016 | 1.14 | 0.73, 1.80 | 0.6 |
| Peritoneal metastases | Involvement | No involvement | 166 | 1.12 | 0.79, 1.59 | 0.5 | 1.03 | 0.65, 1.62 | 0.9 |
| Prior oxaliplatin | Yes | No | 166 | 1.48 | 0.78, 2.83 | 0.2 | 1.60 | 0.73, 3.52 | 0.2 |
| Eligibility criteria | |||||||||
| WHO performance status | ≥2 | 0-1 | 105 | 2.38 | 1.29, 4.38 | 0.005 | 2.67 | 1.32, 5.40 | 0.007 |
| Prior regimens | ≥2 | 1 | 166 | 0.79 | 0.55, 1.15 | 0.2 | 0.84 | 0.54, 1.31 | 0.4 |
| Brain metastases | Present | Absent | 166 | 2.03 | 0.94, 4.38 | 0.070 | 1.30 | 0.52, 3.22 | 0.6 |
Bold values indicate statistical significance p < 0.05.
CI confidence interval, dMMR deficient mismatch repair, HR hazard ratio, MSI microsatellite instability, MSS microsatellite stability, OS overall survival, pMMR proficient mismatch repair, WHO world health organization.
aMultivariable regression includes both patients and tumor characteristics as eligibility criteria.
Discussion
In this population-based real-world cohort of BRAFV600E-mutated mCRC patients treated with encorafenib-cetuximab, we show a clinically relevant efficacy-effectiveness gap. Median OS of patients treated in daily clinical practice (real world) was lower compared to patients enrolled in the BEACON trial (6.7 months versus 9.3 months, HR = 1.47 (95% CI: 1.15–1.88)) [3, 4]. In daily clinical practice encorafenib-cetuximab treatment is provided to a more general population compared to the stringent criteria used in clinical trials. Regarding patients that could be categorized into eligible or ineligible, more than a third of patients in our real-world cohort (36%) would have been ineligible for the BEACON trial. Patients that were trial eligible showed a median OS of 7.3 months with no significant difference to the BEACON trial (p = 0.15), demonstrating that patients in daily clinical practice that obtain comparable characteristics to BEACON trial patients show no significant different OS outcome. While especially the trial ineligible patients demonstrated a poor median OS of 6.0 months with a significant difference to the BEACON trial (p < 0.01). Specific trial ineligible subgroups that demonstrated the poorest median OS were patients with symptomatic brain metastases and WHO performance status ≥2 (5.0 months and 3.9 months, respectively), although no significant difference (p = 0.065) was demonstrated for patients with symptomatic brain metastases, likely due to the small number of patients (N = 7).
Treatment exposure was comparable with a median of 17 weeks in our real-world cohort and 19 weeks in the BEACON trial and follow-up was sufficient with 128/166 events for OS. However, we found our real-world cohort patients to be more heavily pretreated, and with a poorer WHO performance status compared to patients in the BEACON trial. In multivariable analysis, we found WHO performance status ≥2 to be an independent poor prognostic factor for OS, which could partly explain the difference in OS between the BEACON trial and our real-world cohort. Although it should be pointed out that WHO performance status was missing for 61/155 patients and missing values had to be imputed in multivariable analysis. Another difference to the BEACON trial was that real-world cohort patients with symptomatic brain metastases were exposed to encorafenib-cetuximab, which were excluded from the BEACON trial. Lastly, our cohort included four patients with tumors that harbored a BRAF mutation other than BRAFV600E. Treatment indication for encorafenib-cetuximab is stated by EMA for ‘adult patients with metastatic colorectal carcinoma (mCRC) with the BRAFV600E mutation who had received prior systemic therapy’ [6]. Since registration of encorafenib-cetuximab is limited to mCRC harboring a BRAFV600E mutation, these patients had in fact no treatment indication. Since it is known from pharmacodynamic studies that RAF inhibitors as encorafenib treatment would lead to no (clinical) response in patients with tumors without a class I mutation, meaning without a BRAFV600E mutation [17]. This stresses the importance of focusing on the specific type of BRAF mutation to ensure adequate patient selection for targeted treatment and emphasizes the importance for continuing medical education and feedback on guideline adherence [18]. All other patients in our dataset (162/166) had a treatment indication according to EMA.
To better understand the difference in findings between our real-world cohort and the BEACON trial, we categorized our cohort into trial eligible and trial ineligible patients based on the selection of inclusion and exclusion criteria applied in the BEACON trial. Not all in- and exclusion criteria of the BEACON trial were available in the real-world cohort and only key eligibility criteria were used. A limitation for this analysis was missing data on key eligibility criteria, especially WHO performance status. This restricted our analyses of trial eligible and ineligible patients to a subgroup of 115 out of 166 patients. We did however show that both patient characteristics and survival outcomes between patients with unknown and known WHO performance status were highly comparable and the sensitivity analysis with imputed data demonstrated similar results (Supplementary Fig. 1).
The trial ineligible patients are not represented in the BEACON trial, meaning that no evidence of efficacy for encorafenib-cetuximab treatment has been provided for these patients. In our cohort, these patients demonstrated a 1-year OS of only 14% versus the estimated 42% in the BEACON trial. Although we made use of reconstructed data of the BEACON trial, of which we estimated 1-year OS, the method used is well validated and expected to provide almost identical data [13]. Of the ineligible patients especially patients with symptomatic brain metastases and patients with a WHO performance status ≥2 had a very poor OS with a 6-month OS of only 29 and 25%, respectively. This illustrates that prescription of encorafenib-cetuximab to a more heterogeneous population in daily clinical practice leads to shorter OS compared to the homogenous population of the strictly monitored BEACON trial. This supports the existence of an efficacy-effectiveness gap [11]. The shorter OS in our population-based cohort in comparison to the BEACON trial could be explained by additional factors besides the differences in patient population. Due to the need for referral and trial enrollment procedures, patients with very rapid progressive disease would likely not have initiated treatment within a clinical trial such as BEACON. Furthermore, the choice and availability of systemic treatment after progression on encorafenib-cetuximab is likely to vary between countries and could potentially also lead to differences in OS. Both factors could also play a role in the non-significant OS difference of the trial eligible real-world cohort and BEACON trial.
A recently published Italian real-world cohort study described a median OS of 7.2 months for 97 patients treated with encorafenib-cetuximab within a nominal use program [19]. This is comparable with the median OS of 6.7 months of our cohort. The marginally shorter OS in our cohort could be due to availability of encorafenib-cetuximab treatment in every hospital in the Netherlands, and no written informed consent was required, hence ensuring a population-based selection of all patients who are treated in daily clinical care after treatment reimbursement. This is in contrast to inclusion of patients in a specific number of (academic) hospitals in Italy in the study of Boccaccino et al. [19]. The added value of our study is that we have an unselected cohort of patients receiving encorafenib-cetuximab. This means it is not based on any inclusion or exclusion criteria, making our data completely representative of daily clinical practice. Notably, in both the Italian real-world cohort study as in our own real-world cohort we found that older age did not lead to inferior OS. As the median OS is relatively short in the studied population it could be that age has less impact compared to other factors.
The observational nature of our cohort, the absence of data regarding adverse events and lack of a control group limits us from drawing firm conclusions regarding benefit and harm. However, we strongly doubt that patients with symptomatic brain metastases and/or a WHO performance status of ≥2 derive substantial survival benefit from encorafenib-cetuximab treatment given the very short survival times for these patients. WHO performance status ≥2 was an independent negative prognostic marker for patients treated with encorafenib-cetuximab in both the Italian and this Dutch cohort [19]. International treatment guidelines currently recommend encorafenib-cetuximab for pretreated patients with BRAFV600E-mutated mCRC, without elaborating on specific criteria for treatment eligibility. In general, only patients with WHO performance status score 0–2 and adequate organ functioning are eligible for systemic treatment [20]. These efficacy-based guidelines should be complemented and adapted with effectiveness evidence. Based on our results, encorafenib-cetuximab treatment should be discouraged in the current guidelines for pretreated BRAFV600E-mutated mCRC patients with brain metastases and/or a WHO performance status of ≥2. Additional data will be provided in the future from ongoing non-interventional prospective studies such as the BERING-CRC study (NCT04673955) and the Prospective Dutch Colorectal Cancer cohort [21].
When new therapy options are approved by EMA and FDA and included in clinical guidelines, the recommended target population is often generalized, and the recommendation is not as selective as the strict inclusion and exclusion criteria of the randomized phase III trial that demonstrated efficacy. Our study shows that in daily clinical practice indeed encorafenib-cetuximab treatment is given to a more general population than the population included in the BEACON trial. Data from the real world make it possible to better estimate prognosis of patients in daily clinical practice, including subgroups that were not included in trials, and counsel patients in a more personalized manner. Additionally, collection of high-quality real-world data regarding quality of life, adverse events, and clinical outcome will allow evaluation of effectiveness of novel agents in the real world to estimate generalizability of the trial results. We believe implementation of standardized reporting in electronic health records - including the currently often missing WHO performance status - is essential to achieve high-quality real-world data. This will result in actionable real-world evidence that is needed to complement efficacy-based guidelines and guide policy and decision making. It should be evident within the guideline which recommendations are based on efficacy and which on effectiveness results.
Conclusion
This population-based real-world cohort of mCRC patients treated with encorafenib-cetuximab showed a clinically relevant efficacy-effectiveness gap with a median OS of 6.7 months in our real-world cohort versus 9.3 months in the BEACON trial. The difference in survival is driven due to treatment administration in a more heterogenous population in the real world. 36% of patients treated in daily clinical practice were not represented in the BEACON trial. While there is no significant difference between trial eligible real-world patients and the BEACON trial, the chance of survival benefit from encorafenib-cetuximab in specific subgroups of trial ineligible real-world patients, such as patients with brain metastases and/or WHO performance status ≥2, is negligible; these patients were not included in the BEACON trial and show very poor OS in the real world, therefore lacking proof of efficacy and effectiveness. Our results highlight the importance of analyzing real-world effectiveness of novel agents in high-quality registry data to gain insights in the efficacy-effectiveness gap, inform medical oncologists of outcomes in daily clinical practice to improve personalized counseling of patients and complement prescription criteria in efficacy-based guidelines.
Supplementary information
Acknowledgements
The authors thank the registration team of the Netherlands Comprehensive Cancer Organization (IKNL) for the collection of data for the Netherlands Cancer Registry as well as IKNL staff for scientific advice. The authors also thank the registration team of the nationwide network and registry of histo- and cytopathology in the Netherlands (PALGA) for collection of mutation status in the Dutch cohort.
Author contributions
KZ: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Data curation, Writing – original draft, Writing – review & editing, Visualization. SN: Conceptualization, Investigation, Data curation, Writing – original draft, Writing – review & editing. FB: Methodology, Writing – review & editing, Supervision. MK: Conceptualization, Writing – review & editing, Supervision. PS: Writing – review & editing, Supervision. AG: Software, Writing – review & editing. GV: Conceptualization, Writing – review & editing, Project administration. JR: Conceptualization, Writing – review & editing, Supervision.
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Competing interests
KZ: Received research funding via institution from Bristol-Myers Squibb. SN: Received research funding via institution from Pierre-Fabre. Had travel, accommodations, or other expenses paid or reimbursed via institution by Servier. FB: Received research funding via institution from Bristol-Myers Squibb. MK: Received research funding via institution from Bayer, Bristol-Myers Squibb, Merck-Serono, Pierre Fabre, Servier, Roche, Sanofi, and Personal Genome Diagnostics. PS: Consulting or advisory role for Bayer, Bristol-Myers Squibb, MEDtalks via institution. AG: No competing interests. GV: Consulting or advisory role for Merck via institution. Received research funding via institution from Merck, Bayer, Personal Genome Diagnostics, Delphi Diagnostics, Bristol-Myers Squibb, Sirtex Medical. JR: Consulting or advisory role for Bayer, Bristol-Myers Squibb, Merck-Serono, Pierre Fabre, Servier via institution. Received research funding via institution from Bayer, Bristol-Myers Squibb, Merck-Serono, Pierre Fabre, Servier, HUB 4 organoids, and Cleara Biotech. Had travel, accommodations, or other expenses paid or reimbursed via institution by Servier.
Ethics approval and consent to participate
According to the Central Committee on Research involving Human Subjects, this type of registry-based study does not require approval from an ethics committee in the Netherlands. The study was approved by the Privacy Review Board and the scientific council of the Netherlands Comprehensive Cancer Organization (IKNL) which collects and guards the data for the Netherlands Cancer Registry (NCR). All data were pseudonymized prior to the transfer from IKNL to the researchers. The NCR uses an opt-out approach to consent. The study was performed in accordance with the Declaration of Helsinki.
Footnotes
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
The online version contains supplementary material available at 10.1038/s41416-024-02711-w.
References
- 1.Punt CJA, Koopman M, Vermeulen L. From tumour heterogeneity to advances in precision treatment of colorectal cancer. Nat Rev Clin Oncol. 2017;14:235–46. doi: 10.1038/nrclinonc.2016.171. [DOI] [PubMed] [Google Scholar]
- 2.Modest DP, Ricard I, Heinemann V, Hegewisch-Becker S, Schmiegel W, Porschen R, 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;27:1746–53. doi: 10.1093/annonc/mdw261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Sorbye H, Dragomir A, Sundström M, Pfeiffer P, Thunberg U, Bergfors M, et al. High BRAF Mutation Frequency and Marked Survival Differences in Subgroups According to KRAS/BRAF Mutation Status and Tumor Tissue Availability in a Prospective Population-Based Metastatic Colorectal Cancer Cohort. PloS One. 2015;10:e0131046. doi: 10.1371/journal.pone.0131046. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kopetz S, Grothey A, Yaeger R, Van Cutsem E, Desai J, Yoshino T, et al. Encorafenib, Binimetinib, and Cetuximab in BRAF V600E–Mutated Colorectal Cancer. N Engl J Med. 2019;381:1632–43. doi: 10.1056/NEJMoa1908075. [DOI] [PubMed] [Google Scholar]
- 5.Tabernero J, Grothey A, Van Cutsem E, Yaeger R, Wasan H, Yoshino T, et al. Encorafenib Plus Cetuximab as a New Standard of Care for Previously Treated BRAF V600E-Mutant Metastatic Colorectal Cancer: Updated Survival Results and Subgroup Analyses from the BEACON Study. J Clin Oncol. 2021;39:273–84. doi: 10.1200/JCO.20.02088. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Trullas A, Delgado J, Koenig J, Fuerstenau U, Dedorath J, Hausmann S, et al. The EMA assessment of encorafenib in combination with cetuximab for the treatment of adult patients with metastatic colorectal carcinoma harbouring the BRAFV600E mutation who have received prior therapy. ESMO Open. 2021;6:100031. doi: 10.1016/j.esmoop.2020.100031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Morris VK, Kennedy EB, Baxter NN, Benson AB, Cercek A, Cho M, et al. Treatment of Metastatic Colorectal Cancer: ASCO Guideline. J Clin Oncol. 2023;41:678–700. doi: 10.1200/JCO.22.01690. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Cervantes A, Adam R, Roselló S, Arnold D, Normanno N, Taïeb J, et al. Metastatic colorectal cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol. 2023;34:10–32. doi: 10.1016/j.annonc.2022.10.003. [DOI] [PubMed] [Google Scholar]
- 9.Phillips CM, Parmar A, Guo H, Schwartz D, Isaranuwatchai W, Beca J, et al. Assessing the efficacy-effectiveness gap for cancer therapies: A comparison of overall survival and toxicity between clinical trial and population-based, real-world data for contemporary parenteral cancer therapeutics. Cancer. 2020;126:1717–26. doi: 10.1002/cncr.32697. [DOI] [PubMed] [Google Scholar]
- 10.Srikanthan A, Vera-Badillo F, Ethier J, Goldstein R, Templeton AJ, Ocana A, et al. Evolution in the eligibility criteria of randomized controlled trials for systemic cancer therapies. Cancer Treat Rev. 2016;43:67–73. doi: 10.1016/j.ctrv.2015.12.006. [DOI] [PubMed] [Google Scholar]
- 11.Eichler H-G, Abadie E, Breckenridge A, Flamion B, Gustafsson LL, Leufkens H, et al. Bridging the efficacy-effectiveness gap: a regulator’s perspective on addressing variability of drug response. Nat Rev Drug Discov. 2011;10:495–506. doi: 10.1038/nrd3501. [DOI] [PubMed] [Google Scholar]
- 12.Dutch Colorectal Cancer Guideline. Accessed March 25, 2024. https://richtlijnendatabase.nl/richtlijn/colorectaal_carcinoom_crc/gemetastaseerd_colorectaalcarcinoom_crc/systemische_therapie_bij_niet_lokaal_behandelbare_metastasen_bij_crc.html.
- 13.Wei Y, Royston P. Reconstructing time-to-event data from published Kaplan–Meier curves. Stata J. 2017;17:786–802. doi: 10.1177/1536867X1801700402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Goey KKH, Sørbye H, Glimelius B, Adams RA, André T, Arnold D, et al. Consensus statement on essential patient characteristics in systemic treatment trials for metastatic colorectal cancer: Supported by the ARCAD Group. Eur J Cancer. 2018;100:35–45. doi: 10.1016/j.ejca.2018.05.010. [DOI] [PubMed] [Google Scholar]
- 15.Bartlett JW, Seaman SR, White IR, Carpenter JR. Multiple imputation of covariates by fully conditional specification: Accommodating the substantive model. Stat Methods Med Res. 2015;24:462–87. doi: 10.1177/0962280214521348. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.R Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing. 2023. https://www.R-project.org/.
- 17.Yaeger R, Corcoran RB. Targeting Alterations in the RAF-MEK Pathway. Cancer Discov. 2019;9:329–41. doi: 10.1158/2159-8290.CD-18-1321. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Larkin J, Del Vecchio M, Ascierto PA, Krajsova I, Schachter J, Neyns B, et al. Vemurafenib in patients with BRAF(V600) mutated metastatic melanoma: an open-label, multicentre, safety study. Lancet Oncol. 2014;15:436–44. doi: 10.1016/S1470-2045(14)70051-8. [DOI] [PubMed] [Google Scholar]
- 19.Boccaccino A, Borelli B, Intini R, Antista M, Bensi M, Rossini D, et al. Encorafenib plus cetuximab with or without binimetinib in patients with BRAF V600E-mutated metastatic colorectal cancer: real-life data from an Italian multicenter experience. ESMO Open. 2022;7:100506. doi: 10.1016/j.esmoop.2022.100506. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Schnipper LE, Smith TJ, Raghavan D, Blayney DW, Ganz PA, Mulvey TM, et al. American Society of Clinical Oncology identifies five key opportunities to improve care and reduce costs: the top five list for oncology. J Clin Oncol. 2012;30:1715–24. doi: 10.1200/JCO.2012.42.8375. [DOI] [PubMed] [Google Scholar]
- 21.Derksen JWG, Vink GR, Elferink MAG, Roodhart JML, Verkooijen HM, van Grevenstein WMU, et al. The Prospective Dutch Colorectal Cancer (PLCRC) cohort: real-world data facilitating research and clinical care. Sci Rep. 2021;11:3923. doi: 10.1038/s41598-020-79890-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.