STRATEGIC-1: multiple-line, randomized, open-label GERCOR-PRODIGE-39 phase III trial in unresectable RAS/BRAF wild-type metastatic colorectal cancer.

Introduction
The management of unresectable metastatic colorectal cancer (mCRC) is a comprehensive treatment strategy involving several therapy lines, maintenance periods, locoregional modalities (surgery/radiofrequency/radiotherapy), and treatment-free intervals. Standard chemotherapies include fluoropyrimidines, irinotecan, and oxaliplatin.1–9 In recent decades, substantial progress has been made in treating mCRC through targeted therapies, including anti-epidermal growth factor receptor (EGFR; cetuximab, panitumumab) or anti-vascular endothelial growth factor (VEGF; bevacizumab, aflibercept) monoclonal antibodies (mAbs). These therapies become integral components of treatment schedules, successfully improving clinical outcomes.1–10 After their first approval, anti-EGFR mAbs showed efficacy only in patients with wild-type KRAS/NRAS tumors.3,11 Despite substantial therapeutic advances, mCRC remains a leading cause of cancer-related mortality worldwide, underscoring the need for continual optimization of treatment strategies. First-line therapy for RAS wild-type mCRC combines a targeted agent (antiangiogenic/anti-EGFR) with doublet chemotherapy (irinotecan/fluoropyrimidine, FOLFIRI, oxaliplatin/fluoropyrimidine, FOLFOX, or XELOX).12 Tumor sidedness (right versus left) is an important prognostic factor and predictor of sensitivity to anti-EGFR agents in RAS wild-type mCRC, with left-sided tumors showing better sensitivity.13–17 In first-line treatment, bevacizumab with chemotherapy improves overall survival (OS) without tumor molecular selection, while anti-EGFR mAbs plus chemotherapy benefit RAS wild-type mCRC. Second-line regimen choice depends on prior first-line treatment, efficacy, tolerance, and adverse events (AEs). Effective therapies after progression on oxaliplatin-based first-line chemotherapy include irinotecan-based chemotherapy with antiangiogenics (bevacizumab.18/aflibercept 19) or anti-EGFR mAbs (cetuximab.20,21/panitumumab.22). The most active second-line regimen for patients progressing on irinotecan-based chemotherapy is FOLFOX-bevacizumab.23 Continuation of antiangiogenics benefits OS beyond first-line progression, whereas anti-EGFR mAb-based regimens did not show the same benefit in second-line therapy. After failure of two prior treatment lines without anti-EGFR agents, cetuximab or panitumumab has been shown to improve OS and can be administered as third-line therapy for RAS wild-type mCRC, either alone or with irinotecan.24 For refractory mCRC, regorafenib and trifluridine/tipiracil are approved, regardless of RAS status.25,26
While the therapeutic armamentarium for mCRC has expanded considerably, this progress has paradoxically introduced a new challenge: how best to sequence these treatments to maximize long-term patient benefit. The absence of clear evidence-based guidance on treatment sequencing has led to heterogeneous clinical practices and uncertainty regarding optimal decision-making across lines of therapy. Importantly, ineffective sequencing strategies may prevent patients from receiving all active agents, reduce cumulative treatment efficacy, and negatively affect OS. As mCRC is increasingly managed as a chronic disease, therapeutic sequencing has become a key determinant of patient outcomes, underscoring the need for dedicated strategy trials.
Patients’ treatment options should be viewed as a long-term plan, optimizing exposure to all active agents. After decades of development, tremendous progress in CRC-targeted drugs has improved treatment compliance, reduced toxicity, and more personalized treatment strategies. However, key unresolved questions remain about the optimal use and sequencing of these agents across lines: How should they be combined? How should they be sequenced? Which upfront combination therapy should be used? Should administration be continuous or intermittent, with maintenance and/or predefined break? Ultimately, randomized strategy trials are needed to define the optimal use and sequencing. Addressing this gap requires prospective studies designed to compare predefined treatment sequences, with the goal of maximizing cumulative drug exposure and long-term clinical benefit.
The STRATEGIC-1 trial was designed to address these critical unanswered questions by directly comparing treatment sequences in patients with unresectable RAS/BRAFV600E wild-type mCRC. STRATEGIC-1 evaluates comprehensive therapeutic strategies, incorporating first- and second-line regimens to determine the most effective and thus optimal sequence of targeted therapies. By focusing on a molecularly selected population and assessing outcomes across multiple lines of therapy, STRATEGIC-1 aimed to refine decision-making in mCRC, offering a strategic framework for optimizing long-term outcomes in an era of increasingly individualized cancer care.

Results

Patients
Between October 2013 and May 2019, 424 patients were randomized (212 in each arm). Of these, 263 with RAS/BRAFV600E wild-type mCRC received treatment (131 in arm A and 132 in arm B), while 161 (81 in arm A, 80 in arm B) with mutated or unknown RAS/BRAFV600E status were excluded (Fig. 1).
Treatment arms were well balanced in terms of demographic and clinical characteristics, including primary tumor location, with a similar distribution of left-sided tumors between the two arms (Table 1). The median age was 63.2 years (range, 40-85). Most patients had an ECOG PS of 0 (56.3%) and synchronous metastases (79.1%).

Treatments
At the time of the data cutoff date (July 13, 2023), 238 patients ended the study strategy (90.5%; 117 in arm A, 121 in arm B), while 25 patients (9.5%) remained on study treatment (Fig. 1). The main reasons for ending the strategy were PD after completing all planned treatments (n = 104), death (n = 47), and receiving therapy outside the original strategy (n = 40). Notably, the initial first-line regimen was reintroduced in 46 patients (35.1%) in arm A (authorized although unplanned) and 60 patients (45.5%) in arm B. A total of 152 patients (57.8%) received second-line therapy (83/131 in arm A and 69/132 in arm B; Supplementary Table S1). During the first sequence, irinotecan was stopped earlier than in the planned strategy in 43 patients (93.5%), cetuximab in 12 (26.5%) and 5-fluorouracil in 7 patients (15.2%). Post-strategy treatments were well balanced between arms.

Efficacy
In the whole population, after a median follow-up of 68.4 months (95% CI, 76.45-97.97) and 195 DDC events (96 in arm A, 99 in arm B), the median DDC was 22.8 months (95% CI, 20.4-28.8) in arm A and 23.5 months (95% CI, 17.9-26.3) in arm B (HRDDC 1.01, 95% CI, 0.76-1.34; log-rank P = 0.945; Fig. 2; Supplementary Fig. 1; Table 2). The median OS was 40.4 months (95% CI, 32.4-51.1) in arm A and 34.4 months (95% CI, 27.5-42.2) in arm B (HROS 1.30, 95% CI, 0.99-1.72; Supplementary Fig. 1; Table 2). The median TFS was 26.0 months (95% CI, 21.6-29.9) in arm A and 24.4 months (95% CI, 19.4-30.1) in arm B (HRTFS 1.00, 95% CI, 0.75-1.32; Supplementary Fig. 2; Table 2).
In the first-line treatment (n = 263), the median PFS was 11.9 months (95% CI, 10.6-13.8) in arm A and 11.9 months (95% CI, 10.2-12.6) in arm B (HRPFS-L1 1.07, 95% CI, 0.82-1.39; Supplementary Fig. 2. Table 2). The ORR was 82.4% (103 of 125 patients; 95% CI, 74.6-88.6) in arm A and 65.4% (85 of 130 patients; 95% CI, 56.5-73.5) in arm B (odds ratio 0.40, 95% CI, 0.23-0.72; Table 2). For patients who had a reintroduction of the full starting regimen, the ORR was 58.7% (27 of 46 patients; 95% CI, 38.7-85.4) in arm A and 31.7% (19 of 60 patients; 95% CI, 19.1-49.5) in arm B (odds ratio 0.33, 95% CI 0.15-0.73). In the second-line group (n = 152, 57.8%), there was no difference between arms in ORR, PFS, or OS (Supplementary Fig. 2; Table 2). In the third-line group, for arm B (n = 50, 37.9%), the median PFS was 4.4 months (95% CI, 3.4-5.6), and the ORR was 12.5% (six of 48 patients, 95% CI, 4.7-25.3; Table 2).
Of note, delaying the introduction of the targeted agent (cetuximab in arm A, bevacizumab in arm B) did not statistically affect PFS, DDC, or OS.

Primary tumor sidedness
Regardless of the treatment arm, the median OS was 37.0 months (95% CI, 32.6-44.0) in 208 patients (79.1%) with left-sided tumors and 29.8 months (95% CI, 18.0-47.7) in 36 patients (13.7%) with right-sided tumors (HROS 1.32, 95% CI, 0.90-1.94).
For patients with left-sided tumors, the median OS was 44.2 months (95% CI, 32.4-55.1) in arm A and 35.8 months (95% CI, 27.9-42.7) in arm B (HROS 1.37 (95% CI, 0.99-1.89). In 202 evaluable patients (97.1%), the ORR was 89.1% (82 of 92 patients; 95% CI, 80.9-94.7) in arm A and 66.4% (73 of 110 patients; 95% CI, 57.3-75.7) in arm B (odds ratio 0.25; 95% CI, 0.12-0.53).
For patients with right-sided tumors, the median OS was 34.9 months (95% CI, 15.7-59.9) in arm A and 20.9 months (95% CI, 13.7-54.9) in arm B (HROS 1.63, 95% CI, 0.76-3.49).

Surgery for metastasis
Surgery for metastasis, with or without radiofrequency ablation, was performed in 36 patients (27.9%) in arm A and 28 patients (21.2%) in arm B. The median OS from surgery from metastasis was 60.3 months (95% CI, 44.7-89.5) in arm A and 63.4 months (95% CI, 52.1-74.2) in arm B (HR 1.05, 95% CI, 0.56-1.96).

Safety
Across the entire treatment strategy, any grade AEs occurred in 129 patients (99.2%) in arm A and in 129 patients (97.7%) in arm B (Supplementary Table 2). Grade 3-4 AEs occurred in 94 patients (72.3%) in arm A and in 96 patients (72.7%) in arm B. Serious treatment-related AEs occurred in 20 patients (33.9%) in arm A and in 36 patients (50.7%) in arm B.
In the first-line treatment, the most frequent (≥10%) grade 3-4 AEs were alopecia (grade 1-2, 38.3%), acne-like syndrome (18.6%), neutropenia (12.6%), and diarrhea (10.2%) in arm A and alopecia (grade 1-2, 18.5%), neutropenia (13.0%), and hypertension (10.0%) in arm B (Supplementary Table 2).
In the second-line treatment, the most frequent (≥10%) grade 3-4 AEs were alopecia (grade 1-2, 12.5%), neuropathy (12.5%) and neutropenia (12.4%) in arm A and alopecia (grade 1-2, 30.3%), neutropenia (16.4%), and fatigue (13.4%) in arm B (Supplementary Table 2).
In third-line (arm B only), the most frequent (≥10%) grade 3-4 AEs were alopecia (grade 1-2, 15.9%) and acne-like syndrome (13.4%; Supplementary Table 2).

Health-related quality of life
Among 220 patients (83.6%) evaluable for HRQoL analysis (110 patients in each arm), the median TUDD of global health status was 28.0 months (95% CI, 12.0-NE) in arm A and 23.9 months (95% CI, 12.5-78.2) in arm B (HRTUDD 1.00, 95% CI, 0.66-1.52; Supplementary Fig. 2).

Discussion
STRATEGIC-1, the first randomized phase III study comparing multiline standard treatment strategies in patients with RAS/BRAFV600E wild-type mCRC, did not meet its primary endpoint (DDC) of demonstrating the superiority of starting oxaliplatin-based chemotherapy with bevacizumab over irinotecan-based chemotherapy with cetuximab. While the DDC results aligned with the hypothesis in arm B (23.5 versus 24 months), arm A exceeded expectations (22.8 versus 16.1 months). Optimizing sequencing therapy and determining the best approach for first-line and subsequent treatments is critical for improving patient outcomes.27 All secondary findings were uncorrected, exploratory, and suitable only for hypothesis generation.
In our study, starting with FOLFIRI-cetuximab followed by FOLFOX-bevacizumab showed better-than-expected efficacy (ORR 82%, median DDC 23 months, median OS 40 months). Previous phase III trials with FOLFIRI-cetuximab as first-line therapy (CRYSTAL,9 FIRE-313,28) reported a lower ORR (66%) in RAS wild-type mCRC, although BRAF mutants were not excluded, unlike our study. Our molecular selection, including BRAF testing in addition to RAS mutation analysis, may partly explain our findings. Additionally, 35% of patients had EGFR-based reintroduction during first-line treatment, with high efficacy (ORR 58%). While reintroducing the full initial regimen was not planned, it was authorized. However, 18% of patients in arm A showed no first-line response (most had PD), indicating primary or acquired resistance. Second-line studies using FOLFOX-bevacizumab after first-line FOLFIRI yielded similar outcomes (ORR 20% to 27%, median PFS 6.4-7.8 months, and median OS 12.9-19.6 months).23,29,30
Arm B, starting with FOLFOX-bevacizumab followed by FOLFIRI while maintaining bevacizumab and subsequently anti-EGFR mAb, achieved the expected DDC. Similar efficacy was observed in previous first- (FOLFOX-bevacizumab) and second-line (FOLFIRI-bevacizumab) studies (Supplementary Table 3). However, only 38% of arm B patients received anti-EGFR mAb as third-line therapy, with lower efficacy (ORR 12.5%, median PFS 4.4 months). This contrasts with a previous trial where third-line panitumumab-irinotecan in RAS/BRAF wild-type mCRC showed a higher ORR (46.0%) and median PFS (8.7 months).31 However, that trial enrollment was closed before standardizing antiangiogenic agent use beyond first-line progression, and 19% of patients did not receive prior bevacizumab. In contrast, all arm B patients in our study had prolonged bevacizumab exposure before anti-EGFR mAb treatment. Prior anti-VEGF therapy may reduce the efficacy of anti-EGFR therapy.32,33 by increasing circulating VEGFA levels.
Adverse events were consistent with the well-established safety profiles of the treatment regimens. Additionally, patients in both arms had similar use of poststudy drugs (TAS-102, regorafenib, EGFR mAbs rechallenge).
While the CALGB/SWOG 80405 trial.34 showed no difference in OS between first-line cetuximab and bevacizumab when used in combination with either oxaliplatin-based or irinotecan-based regimens, the FIRE-3.13,28 and PARADIGM.35 trials demonstrated an OS benefit for the anti-EGFR mAb and chemotherapy combination compared to bevacizumab plus chemotherapy. In our series, tumor molecular selection included the KRAS, NRAS and BRAF genes, and there was no difference in DDC between the treatment arms, but we observed significantly higher response rates in first-line treatment (both induction and reintroduction) and a numerically better OS when starting with FOLFIRI plus cetuximab. The first-line median PFS of approximately one year aligns with prior randomized phase III trials (FIRE-3,13,28 CALGB/SWOG 80405,34 PARADIGM,35 CAIRO5.36–38), which compared antiangiogenics and anti-EGFR agents in first-line therapy, with no difference between treatment arms. The median OS of 40.4 months in the cetuximab group was numerically higher than that found in previous trials (32.0-36.2 months).13,28,34,35 This difference may be partly explained by the exclusion of patients with BRAFV600E-mutated tumors in our study. Additionally, we observed an ORR benefit with first-line FOLFIRI plus cetuximab compared to mFOLFOX7 plus bevacizumab. The ORR of 82.4% in the immediate anti-EGFR group was higher than that in the trials mentioned above (57.8%-74.9%), while the ORR of 65.4% in the anti-VEGF group was within the reported ranges (53.5%-67.3%).13,28,34,35 In the CRYSTAL study,9 59 of 625 (9.4%) patients with KRAS wild-type mCRC harbored a BRAF mutation. Similarly, in the FIRE-3 study,13,28 48 of 400 (12%) patients with RAS wild-type mCRC had a BRAF mutation. In the FIRE-4.5 study dedicated to BRAF V600E mutant mCRC, patients were randomized to FOLFIRINOX with either cetuximab or bevacizumab.39,40 The response rate was 42% in the cetuximab arm and 57% in the bevacizumab arm, with an absolute difference of 15%.
Modest et al. 41 suggested that first-line treatment with anti-EGFR may represent a favorable condition for effective subsequent therapy, including antiangiogenic agents. Bennouna et al. 18 demonstrated that continuation of bevacizumab with chemotherapy as second-line therapy after PD on first-line bevacizumab plus chemotherapy significantly prolonged OS and PFS compared to chemotherapy alone. A meta-analysis reported by Wu et al. 42 of the pooled data from randomized clinical trials and retrospective studies comparing the efficacy of bevacizumab, cetuximab, and panitumumab combined with chemotherapy as a sequence of therapies for mCRC patients, found that the optimal sequence for first-to-second-line therapy in RAS wild-type mCRC is cetuximab-based therapy first, followed by a bevacizumab-based regimen.
Despite improved molecular selection of patients for EGFR inhibitors by extending genomic testing (KRAS, NRAS, BRAF), the RAS/BRAF wild-type population remains clinically and molecularly heterogeneous. Subgroup analyses of randomized trials have demonstrated that primary tumor location affects clinical and molecular features in mCRC in the first-line setting.43,44 Meta-analyses of unresectable RAS wild-type mCRC have shown that left-sided tumors treated with anti-EGFR plus chemotherapy exhibit significant survival benefits in ORR and OS but not PFS compared with chemotherapy alone or with anti-VEGF.14,24,25 In our series, 79% of patients with left-sided tumors and those treated with anti-EGFR had higher ORR and OS than those receiving bevacizumab (89.1% versus 66.4% and 44.2 versus 35.8 months, respectively; Supplementary Table 4). However, no significant differences were observed in PFS (13.8 versus 12.0 months), DDC (26.0 versus 24.2 months), or TFS (27.6 versus 25.1 months). Of note, OS was numerically higher in a limited subgroup of 36 patients with right-sided tumors starting with FOLFIRI-cetuximab.
Most RAS/BRAFV600E wild-type mCRC patients develop PD within a year of starting anti-EGFR therapy due to primary or secondary resistance, often linked to downstream activation of MAPK (EGFR/RAS/BRAF/MEK/ERK) and PIK3CA/AKT/mTOR pathways.45–47 Panels including several uncommon genomic alterations (e.g., HER2, MET amplification, NTRKs/ROS1/ALK fusions, PTEN/PIK3CA, MAPK mutations) of primary resistance to anti-EGFR mAb may optimize patient selection for anti-EGFR therapy.48,49 Future trials should incorporate extended molecular profiling, in addition to RAS/BRAF testing and tumor sidedness, to refine mCRC stratification. Liquid biopsy offers a noninvasive alternative or complementary to tissue biopsy, enhancing molecular screening and facilitating tailored treatment with anti-EGFR drugs. It aids in predicting responses, monitoring anti-EGFR administration/readministration, and optimizing individualized treatment sequencing in mCRC patients.
This study has several limitations. First, the sample size may limit the statistical power and generalizability of the findings. Second, MMR testing was not mandatory, resulting in a low testing rate. However, given the rarity of dMMR/MSI-high tumors in the RAS and BRAF wild-type population, this is unlikely to have significantly impacted the overall results. Additionally, some patients did not have complete RAS and BRAF status available at therapy initiation. Nevertheless, as previously observed in the FIRE-3 study,39 the delayed introduction of the targeted agents, cetuximab in arm A and bevacizumab in arm B, did not influence PFS, DDC, or OS, suggesting minimal effect on outcome measures.
In conclusion, the primary endpoint, DDC, was not met, and the STRATEGIC study was inconclusive to identify the optimal treatment sequence for patients with unresectable wild-type RAS/BRAFV600E mCRC. DDC was not an appropriate endpoint for hypothesis generation or for the estimation of clinical outcomes within the context of this strategy trial. The safety profiles aligned with the established safety profiles of these regimens.

Patients and methods

Study design and patients
STRATEGIC-1 was a prospective, open-label, multinational, randomized phase III study conducted at 72 centers in France, Israel, and Ireland.50 The trial was conducted according to the Declaration of Helsinki, International Conference on Harmonization Guidelines for Good Clinical Practice, and CONSORT guidelines. The study protocol (Supplementary Data 1) was approved for all participating centers by the French health authorities (the Agence Nationale de Sécurité du Médicament et des Produits de Santé [ANSM] on June 24, 2013 and the Independent Ethics Committee “Ile de France Paris VI” La Pitié Salpêtrière on April 12, 2013), and all patients provided informed consent.
The main eligibility criteria included histologically proven colorectal carcinoma; KRAS (exons 2-4), NRAS (exons 2-4), and BRAFV600E wild-type status (local assessment); age ≥18 years; an ECOG PS of 0-2; unresectable metastatic disease; and adequate bone marrow, hepatic, and renal functions. The KRAS, NRAS, and BRAF mutational status could be assessed after randomization within the first 2 cycles of first-line therapy.

Random assignment and study treatment
Patients were randomized (1:1) to first-line FOLFIRI-cetuximab followed by second-line mFOLFOX6-bevacizumab (control arm A); those two regimens were given continuously until disease progression (PD) or limiting toxicity or to first-line induction therapy with mFOLFOX7-bevacizumab for 3 months (6 cycles) followed by maintenance therapy with fluoropyrimidine plus bevacizumab and then oxaliplatin reintroduction at PD,51 followed by second-line therapy with FOLFIRI-bevacizumab followed by third-line therapy with anti-EGFR monoclonal antibody (cetuximab/panitumumab) with/without irinotecan (experimental arm B). Randomization was stratified by center, GERCOR prognostic score (based on ECOG PS and lactate dehydrogenase [LDH] level52), prior oxaliplatin use (yes/no), and metastatic disease extension (liver-only/other).

Study endpoints
The primary endpoint was the duration of disease control (DDC), defined as the sum of progression-free survival (PFS) of each active treatment course planned in the strategy. The DDC excluded intervals between PD and treatment reinitiation and PFS of inactive treatment if PD occurred at the first evaluation post reinitiation (reintroduction in a stop-and-go strategy or the subsequent treatment regimen in the multiple-line strategy). Censoring rules for DDC included study end without PD and use of unplanned therapeutic agent. Patients undergoing surgery for metastasis were not censored.53
Secondary endpoints included time-to-failure of strategy (TFS), OS, PFS, overall response rate (ORR) per RECIST 1.1 criteria, salvage surgery rate, safety (NCI-CTCAE v4.0), and health-related quality of life (HRQoL) using time until definitive deterioration (TUDD).

Statistical analysis
This trial aimed to demonstrate the superiority of arm B over arm A, assuming a 33% risk reduction in event risk (hazard ratio [HR] = 0.67) in arm B, based on an absolute gain of 8 months in median DDC from 16.1 months in arm A to 24.0 months in arm B. A total of 269 evaluable patients were needed to observe 226 events for a two-sided log-rank test (alpha 5% and beta 15%; Software: EAST 5.3). Including an anticipated 33% of patients without RAS/BRAFV600 status testing at randomization and a 5% drop-out rate, 423 patients were needed.
Efficacy was analyzed on modified intent-to-treat (mITT) population, including all randomized patients with locally assessed RAS/BRAFV600E wild-type status. Patients without this status after 2 cycles were excluded. The safety population included patients who received at least one dose of any study treatment. The HRQoL population consisted of mITT patients with at least one baseline HRQoL questionnaire. The primary objective used a two-sided type-I error of 5%. All other P values were exploratory.
Event-free rates were analyzed using the Kaplan‒Meier method, with survival medians and 95% CIs compared via the log-rank test. The treatment effects were evaluated by HR with 95% CI from the Cox proportional hazards model. Median follow-up was performed using the reverse Kaplan‒Meier method. All analyses were performed using SAS v9.4 and R software v4.1.

Supplementary information