Skip to main content
NCBI home page
JAMA Network logoLink to JAMA Network
. 2024 Nov 21;11(1):36–45. doi: 10.1001/jamaoncol.2024.5174

First-Line Systemic Treatment for Initially Unresectable Colorectal Liver Metastases

Post Hoc Analysis of the CAIRO5 Randomized Clinical Trial

Marinde J G Bond 1, Karen Bolhuis 2, Olaf J L Loosveld 3, Jan Willem B de Groot 4, Helga Droogendijk 5, Helgi H Helgason 6, Mathijs P Hendriks 7, Joost M Klaase 8, Geert Kazemier 9,10, Mike S L Liem 11, Arjen M Rijken 12, Cornelis Verhoef 13, Johannes HW de Wilt 14, Koert P de Jong 8, Michael F Gerhards 15, Martinus J van Amerongen 16, Marc RW Engelbrecht 17, Krijn P van Lienden 18, John J Hermans 19, I Quintus Molenaar 20, Dirk J Grünhagen 13, Bart de Valk 21, Brigitte C M Haberkorn 22, Emile D Kerver 23, Frans Erdkamp 24, Robbert J van Alphen 25, Daniëlle Mathijssen-van Stein 26, Aysun Komurcu 27, Anne M May 1, Rutger-Jan Swijnenburg 10,28, Cornelis J A Punt 1,29,✉, for the Dutch Colorectal Cancer Group
PMCID: PMC11583021  PMID: 39570583

Key Points

Question

What is the optimal induction systemic treatment for patients with colorectal cancer and unresectable liver-only metastases?

Findings

In this post hoc analysis of a multicenter randomized clinical trial of 530 patients with colorectal cancer with liver-only metastases, there was no difference in overall survival between treatment with folinic acid, fluorouracil, and oxaliplatin (FOLFOX) plus irinotecan plus bevacizumab (FOLFOXIRI) and FOLFOX/folinic acid, fluorouracil, and irinotecan (FOLFIRI) plus bevacizumab in patients with RAS/BRAFV600E-variant and/or right-sided tumors, nor between adding panitumumab vs bevacizumab to FOLFOX/FOLFIRI in patients with RAS/BRAFV600E wild-type and left-sided tumors.

Meaning

The trial results support the use of FOLFOX/FOLFIRI-bevacizumab for patients with initially unresectable colorectal cancer with liver-only metastases irrespective of RAS/BRAFV600E variant status and tumor sidedness.

Abstract

Importance

In patients with colorectal cancer and unresectable liver-only metastases (CRLM), treatment with folinic acid, fluorouracil, and oxaliplatin (FOLFOX) plus irinotecan (FOLFOXIRI) and bevacizumab vs FOLFOX/folinic acid, fluorouracil, and irinotecan (FOLFIRI) plus bevacizumab increased progression-free survival, response, and R0/R1 resection/ablation rates, as well as toxic effects in RAS/BRAFV600E-variant and/or right-sided tumors. FOLFOX/FOLFIRI–panitumumab vs FOLFOX/FOLFIRI–bevacizumab increased response at the cost of more toxic effects in RAS/BRAFV600E wild-type, left-sided tumors.

Objective

To present long-term outcomes of treatment with FOLFOXIRI plus bevacizumab vs FOLFOX/FOLFIRI plus bevacizumab and FOLFOX/FOLFIRI plus panitumumab vs FOLFOX/FOLFIRI + bevacizumab.

Design, Setting, and Participants

The randomized phase 3 CAIRO5 trial included patients with initially unresectable CRLM in 46 Dutch centers and 1 Belgian center between November 2014 and January 2022. A liver expert panel repeatedly evaluated resectability.

Intervention

Patients with RAS/BRAFV600E-variant and/or right-sided tumors randomly received FOLFOX/FOLFIRI–bevacizumab (group 1) or FOLFOXIRI-bevacizumab (group 2), and those with RAS/BRAFV600E wild-type, left-sided tumors received FOLFOX/FOLFIRI–bevacizumab (group 3) or FOLFOX/FOLFIRI–panitumumab (group 4). Adjuvant chemotherapy (ACT) after complete local treatment was recommended but not standard.

Main Outcomes and Measures

Overall survival (OS) was analyzed as a secondary outcome. Other outcomes were post hoc analyses.

Results

A total of 530 patients (327 male [62%] and 203 female individuals [38%]; median age, 62 [IQR, 54–69] years) were randomized: 148 in group 1, 146 in group 2, 118 in group 3, and 118 in group 4. The median OS in group 1 was 23.6 (95% CI, 20.1-27.5) vs 24.1 (95% CI, 21.0-30.9) months in group 2 (hazard ratio [HR], 0.90; 95% CI, 0.70-1.17; P = .44), and 39.9 (95% CI, 30.7-44.6) in group 3 vs 38.3 (95% CI, 35.3-51.3) months in group 4 (HR, 0.95; 95% CI, 0.68-1.32; P = .75). OS was longest after complete local treatment without early (≤6 months) recurrence (64.3 months; 95% CI, 57.6 to not reached) and salvage local treatment options after early recurrence (58.9; 95% CI, 47.3 to not reached), followed by patients without salvage local treatment after early recurrence (30.5; 95% CI, 24.4-33.4) and with incomplete local treatment (28.7; 95% CI, 25.9-38.3), and worst in patients with continued unresectability (18.3; 95% CI, 15.7-20.0). After confounder adjustment, ACT was associated with longer OS (HR, 0.66; 95% CI, 0.44-0.98) and relapse-free survival (HR, 0.65; 95% CI, 0.48-0.88) and less early recurrence without salvage local treatment (odds ratio, 0.46; 95% CI, 0.25-0.85).

Conclusions and Relevance

These results support using FOLFOX/FOLFIRI–bevacizumab for patients with initially unresectable CRLM irrespective of RAS/BRAFV600E status and tumor sidedness. Patients with complete local liver treatment with salvage local treatment in case of early recurrence had the longest OS. ACT might be considered for these patients.

Trial Registration

ClinicalTrials.gov NCT02162563

This randomized clinical trial examines different treatments for patients with unresectable colorectal cancer with liver-only metastases.

Introduction

Patients with initially unresectable colorectal cancer liver metastases (CRLM) may convert to receiving curative-intent local liver treatment with resection or thermal ablation (ie, local treatment) after systemic induction treatment. The randomized phase 3 CAIRO5 clinical trial aimed to find the optimal induction regimen in a population for whom the unresectability of CRLM at baseline was strictly defined.1 Previously, we showed that for patients with RAS/BRAFV600E-variant and/or right-sided primary tumors, treatment with folinic acid, fluorouracil, and oxaliplatin (FOLFOX) plus irinotecan (FOLFOXIRI) plus bevacizumab was associated with a small but significantly longer median progression-free survival (PFS), higher response rate, and higher complete local treatment rate (R0/R1 resection and/or ablation of all CRLM) at the cost of more toxic effects compared with FOLFOX/folinic acid, fluorouracil, and irinotecan (FOLFIRI) plus bevacizumab.1 In patients with RAS/BRAFV600E wild-type, left-sided tumors, adding panitumumab vs bevacizumab to FOLFOX/FOLFIRI showed no difference in median PFS, and the significantly higher response rate did not translate into a higher local treatment rate, whereas more toxic effects were was observed.1

In the case of conversion to local treatment after induction treatment, only 19% of patients remained recurrence-free.2 Thus, there is a rationale for adjuvant chemotherapy (ACT).3 For patients with upfront resectable CRLM, phase 3 randomized clinical trials (RCTs) have shown that perioperative chemotherapy or ACT may postpone recurrence, but to our knowledge no overall survival (OS) benefit has been demonstrated.4,5,6,7,8,9 In patients with initially unresectable CRLM, ACT efficacy has not been established. In this article, we present long-term outcomes, including those for OS and ACT efficacy.

Methods

Study Design and Participants

The open-label, phase 3 CAIRO5 RCT of the Dutch Colorectal Cancer Group was performed at 46 Dutch centers and 1 Belgian center between November 13, 2014, and January 31, 2022 (Supplement 1; eTable 1 in Supplement 2). The protocol has been previously published.1,10 In summary, eligible patients were 18 years or older, had histologically proven colorectal cancer with known RAS/BRAFV600E variant status, previously untreated and unresectable CRLM (as centrally assessed by a liver expert panel), a World Health Organization performance status of 0 to 1, and no contraindications for local or systemic treatment. The study was conducted according to the Declaration of Helsinki. The study was approved by the medical ethical committee of the Amsterdam University Medical Centers. All patients provided written informed consent. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guidelines.

Randomization

Patients with RAS/BRAFV600E-variant and/or right-sided tumors randomly received FOLFOX or FOLFIRI plus bevacizumab (group 1) or FOLFOXIRI plus bevacizumab (group 2). Patients with RAS/BRAFV600E wild-type and left-sided tumors randomly received FOLFOX/FOLFIRI–bevacizumab (group 3) or FOLFOX/FOLFIRI–panitumumab (group 4). The local investigator chose between FOLFOX/FOLFIRI based on patient preference. Randomization was stratified by CRLM resectability (potentially resectable vs permanently unresectable according to the panel), serum lactate dehydrogenase levels (normal vs abnormal), hospital, choice of irinotecan/oxaliplatin, and BRAFV600E variant status (only for groups 1 and 2). Randomization was done centrally by the Netherlands Comprehensive Cancer Organisation according to the Pocock minimization technique via a masked web-based allocation procedure (ALEA, version 17.1; FormsVision). Patients were enrolled by their treating physician. Investigators, physicians, and participants were not masked to treatment group allocation. The panel surgeons and radiologists were masked to treatment allocation. As of February 2017, patients with RAS/BRAFV600E-variant and/or right-sided tumors were assigned to groups 1 or 2, and those with RAS/BRAFV600E wild-type, left-sided tumors to groups 3 or 4, because data showed that anti–epidermal growth factor receptor (EGFR) treatment was not beneficial for right-sided or BRAFV600E-variant tumors.11,12

Procedures

Drugs were administered per standard regimens as previously described, twice weekly for up to 12 cycles.1 Bevacizumab was discontinued 5 weeks or less before planned surgery. Additional chemotherapy was allowed during this period. Continuing chemotherapy without bevacizumab/panitumumab as ACT after complete local treatment was not standard, but it was recommended to receive ACT 12 weeks or less of last local liver treatment to complete the planned 12 cycles. The decision for ACT was made by shared decision-making between patients and local investigators. For patients who did not receive local treatment, maintenance treatment with fluorouracil, folinic acid, and bevacizumab/panitumumab was recommended after 12 cycles.

The liver expert panel has been previously described.1 Resectability was evaluated at baseline and twice monthly thereafter. CRLMs were considered unresectable at baseline if no R0 resection could be achieved with 1-stage surgical resection. Resectability criteria during follow-up allowed all established local treatments to reach an R0 resection (ie, ablation, 2-stage hepatectomies, and portal vein embolization). After radiologic evaluation, 3 randomly selected panel surgeons independently evaluated computed tomography (and, if available, magnetic resonance imaging) scans and CRLM were categorized patients as having tumors that were (1) resectable, (2) potentially resectable, or (3) permanently unresectable. If no consensus (ie, same category selected) was obtained, 2 additional surgeons were consulted, and the majority vote was accepted as the final vote. If there was no majority, the panel chair determined the final vote.

Outcomes

OS was defined as time from randomization to death and censored if a patient was alive during their last clinical visit. Relapse-free survival was calculated from last local liver treatment until recurrence or death and censored on the last visit. Panel resectability and response assessments were masked. For evaluations of CT scans outside the scheduled panel evaluations, the local investigator’s assessment was used. Complete local treatment was defined as R0/R1 resection or ablation of all CRLM. R0 was defined as the absence of microscopic tumor invasion of the resection margin. Early recurrence was defined as disease progression or death 6 months or less after complete local CRLM treatment.13,14 Death and palliative or no curative-intent local treatment 6 months or less after recurrence were scored as an event for early recurrence without salvage local treatment.

Statistical Analysis

The sample size calculation was based on the primary outcome PFS. To detect a hazard ratio (HR) of 0.70 with 80% power at a 2-sided 5% significance level, 257 events were required for groups 1 and 2 and 256 events for groups 3 and 4. In March 2022, the steering group of the study followed the advice of the data and safety monitoring board to discontinue accrual in groups 3 and 4 due to futility. Results of the primary outcome were published previously.1

Analyses were based on a modified intention-to-treat analysis that excluded patients who withdrew consent before initiating study treatment or who did not meet major entry criteria. The length of observation time was measured with median follow-up, which was calculated with the reverse Kaplan-Meier method, and person-years of follow-up, which was calculated by multiplying the number of people in a study by the time each person spends. OS was prespecified as a secondary outcome, and the other outcomes were post hoc analyses. OS curves were estimated with the Kaplan-Meier method and compared with the 2-sided stratified log-rank test, adjusting for the stratification factors used in randomization. A per-protocol analysis of patients with RAS/BRAFV600E wild-type and left-sided tumors, excluding patients with BRAFV600E-variant and/or right-sided tumors, was done for groups 3 and 4. HRs and 95% CIs were calculated with Cox proportional hazards analysis. To compare OS between treatment groups 1 and 2 and groups 3 and 4, a stratified Cox proportional hazards analysis was used to adjust for the stratification factors used in randomization. The proportional hazards assumption for comparing OS between treatment groups 1 and 2 and 3 and 4 was met in both comparisons according to a test for independence between scaled Schoenfeld residuals and time.15 Visual inspection confirmed this. OS according to local treatment was analyzed in all groups and was also reported based on local treatment combined with early recurrence status. To account for potential immortal time bias, a sensitivity analysis with follow-up starting 6 months postoperatively was performed to compare OS between the early recurrence outcomes. Subgroup analyses were performed for potentially resectable vs permanently unresectable metastases (panel decision), and in groups 1 and 2 RAS vs BRAFV600E variant status, to evaluate the predictive and prognostic effects. OS was also compared between patients who underwent R0 vs R1 resection and RAS and BRAFV600E variants in groups 1 and 2. The comparison of OS between R0 and R1 resections was also adjusted for potential confounders and prognostic factors to determine whether the difference was independent. Postoperative outcomes, including margin status and recurrence outcomes after complete local treatment, were reported per group. Categorical outcomes were compared using the Fisher exact test.

Efficacy of ACT in terms of overall and relapse-free survival and early recurrence without salvage local treatment was analyzed by stabilized inverse probability treatment weighting (sIPTW) to adjust for potential confounding in patients who underwent complete local treatment. Patients who received 12 or more induction cycles were excluded since ACT was not recommended for these patients, and patients who died within 30 days postoperatively were excluded since ACT is usually not discussed with these patients. Baseline characteristics were compared with standardized mean differences (SMDs). A weighted sample was constructed with sIPTW that balanced patient characteristics according to ACT status.16 Propensity scores were calculated using multivariable logistic regression, with ACT as a dependent variable and prognostically important determinants or potential confounders as covariates.16 The following covariates were used: age, sex, RAS/BRAFV600E status, primary tumor location, synchronous (ie, metastases diagnosed ≤6 months after primary tumor17) vs metachronous metastases, induction regimen, number of induction cycles, World Health Organization performance status, number of involved liver segments, number, largest size, and distribution of CRLM (uni- or bilobar), carcinoembryonic antigen, lactate dehydrogenase levels, response, 1-stage vs 2-stage surgery, and Clavien-Dindo complications. The last known preoperative values were used. The covariate balance in the weighted samples was assessed with SMDs for continuous and categorical variables. Differences of 0.10 or less were considered negligible. Treatment effects were calculated by including stabilized weights in a Cox and logistic regression model. Statistical analyses were conducted using R (version 4.2.2; R Foundation). The data cutoff was February 7, 2024.

Results

A total of 530 patients (327 male [62%] and 203 female individuals [38%]; median age, 62 years [IQR 54–69 years]) were randomly assigned: 148 patients to group 1, 146 patients to group 2, 118 patients to group 3, and 118 patients to group 4 (Figure 1). After excluding 9 ineligible patients, 147 patients were allocated to group 1, 144 to group 2, 114 to group 3, and 116 to group 4 for the modified intention-to-treat analyses. Baseline characteristics were well balanced (eTable 2 in the Supplement). The median number of CRLM in all groups was 12 (IQR, 7-22), denoting a population with a high disease burden. Before the protocol amendment, 14 patients with a right-sided and/or BRAFV600E-variant tumor were randomly assigned to group 3 or 4.

Figure 1. Trial Profile.

Figure 1.

Open in a new tab

CRC indicates colorectal cancer; FOLFIRI, folinic acid, fluorouracil, and irinotecan; FOLFOX, folinic acid, fluorouracil, and oxaliplatin; mCRC, metastatic colorectal cancer.

All patients in groups 1, 3, and 4 received their assigned treatment regimens; 2 patients in group 2 received FOLFOX instead of FOLFOXIRI (Figure 1). In case of continued unresectability, 22 of 79 patients (28%) in group 1, 28 of 62 patients (45%) in group 2, 18 of 36 patients (50%) in group 3, and 8 of 36 patients (22%) in group 4 received maintenance treatment. After complete local treatment, 23 of 53 patients (43%) in group 1, 35 of 74 patients (47%) in group 2, 26 of 66 patients (39%) in group 3, and 33 of 67 patients (49%) in group 4 received ACT.

Median follow-up was 62.4 months (95% CI, 59.8-64.9) in groups 1 and 2 and 60.8 months (95% CI, 56.8-66.6) in groups 3 and 4. Person-years of follow-up were 340, 353, 354, and 348, in groups 1, 2, 3, and 4, respectively. With 241 observed events (83%), median OS was 23.6 months (95% CI, 20.1-27.5) in group 1 vs 24.1 months (95% CI, 21.0-30.9) in group 2 (stratified HR, 0.90; 95% CI, 0.70-1.17; stratified log-rank P = .44; Figure 2A). With 149 events (65%), median OS was 39.9 months (95% CI, 30.7-44.6) in group 3 vs 38.3 months (95% CI, 35.3-51.3) in group 4 (stratified HR, 0.95; 95% CI, 0.68-1.32; stratified log-rank P = .75; Figure 2B). In the per-protocol analysis of 216 patients with left-sided, RAS/BRAFV600E wild-type tumors, median OS was 41.2 months (95% CI, 31.3-48.4) in group 3 and 40.7 months (95% CI, 35.4-51.3) in group 4 (stratified HR, 0.96; 95% CI, 0.68-1.36; stratified log-rank P = .83) with 135 events (63%). Five-year survival was 14% in group 1, 17% in group 2, 27% in group 3, and 28% in group 4.

Figure 2. Overall Survival in the Modified Intention-to-Treat Analysis.

Figure 2.

Open in a new tab

A, Patients with right-sided and/or RAS or BRAFV600E-mutated primary tumors randomized to receive folinic acid, fluorouracil, and oxaliplatin (FOLFOX) or folinic acid, fluorouracil, and irinotecan (FOLFIRI) plus bevacizumab and FOLFOX plus irinotecan (FOLFOXIRI) and bevacizumab. B, Patients with left-sided and RAS and BRAFV600E wild-type primary tumors randomized to receive FOLFOX or FOLFIRI plus bevacizumab and FOLFOX or FOLFIRI plus panitumumab. HR indicates hazard ratio.

OS was significantly longer in patients with local treatment vs no or incomplete local treatment in all groups (eFigure 1 in Supplement 2). No interaction was observed between baseline panel evaluation (potentially resectable or permanently unresectable) or RAS/BRAFV600E variant status (for groups 1 and 2) and OS (eFigure 2 in Supplement 2). Median OS in patients with a RAS/BRAFV600E wild-type but right-sided tumor was significantly longer compared with patients with a BRAFV600E variant (HR, 3.07; 95% CI, 1.50-6.27) and not significantly different compared with patients with an RAS variant (HR, 1.68; 95% CI, 0.94-3.01; eFigure 3 in Supplement 2). In the overall study population, median OS was 29.6 months (95% CI, 27.5-32.3) with 390 events (75%). The best OS was observed after complete local treatment without early recurrence and salvage local treatment after early recurrence, followed by patients with incomplete local treatment and without salvage local treatment options after early recurrence, and was worst in patients without local treatment (Figure 3; eTable 3 in Supplement 2). OS was significantly longer after R0 vs R1 resections (HR, 1.69; 95% CI, 1.14-2.49; P < .001; eFigure 4 in Supplement 2). This remained after covariate adjustment in the Cox model for age, sex, variant status, sidedness, synchronous or metachronous metastases, preoperative number of liver segments involved, and preoperative number, size, and distribution of CRLM (HR, 1.64; 95% CI, 1.05-2.55; P = .03). There were no differences in margin status and recurrence outcomes between patients with complete local treatment in groups 1/2 and 3/4 (Table 1).

Figure 3. Overall Survival According to Extended Local Treatment Status in All Groups.

Figure 3.

Open in a new tab

HR indicates hazard ratio; NA, not applicable.

Table 1. Outcomes After Local Treatment.

Characteristic No. (%)
Patients with a right-sided and/or RAS or BRAFV600E-variant tumor Patients with a left-sided and RAS and BRAFV600E wild-type tumor
FOLFOX or FOLFIRI plus bevacizumab (n = 147) FOLFOXIRI plus bevacizumab (n = 144) P value FOLFOX or FOLFIRI plus bevacizumab (n = 114) FOLFOX or FOLFIRI plus panitumumab (n = 116) P value
Local treatment (resection/ablation) 68 (46) 82 (57) .08 78 (68) 80 (69) >.99
Complete local treatmenta 53 (36) 74 (51) <.01 66 (58) 67 (58) >.99
Outcomes of patients who received complete local treatment
Primary tumor resection
Before first local liver treatment 22 (42) 24 (32) NA 31 (47) 22 (33% NA
During local liver treatment 5 (9) 5 (7) 3 (5) 2 (3)
After local liver treatment 17 (32) 34 (46) 22 (33) 32 (48)
No primary tumor resection 9 (17% 11 (15) 10 (15) 11 (16)
Margin status
Ablation only 4 (8) 1 (1) .11 4 (6) 3 (4) >.99
R0 37 (70) 61 (82) 51 (77) 52 (78)
R1 12 (23) 12 (16) 11 (17) 11 (16)
R uncertain 0 1 (1)
Early recurrence within 6 mob 26 (49) 29 (39) .28 28 (42) 26 (39) .73
Early recurrence within 6 mo without salvage local treatment 18 (34) 24 (32) >.99 20 (30) 18 (27) .70
Location of recurrenceb
Liver only 20 (43) 28 (47) .75 38 (68) 45 (73) .62
Lung only 10 (21) 7 (12) 6 (11) 4 (6)
Lymph node 2 (4) 3 (5) 1 (2) 2 (3)
Peritoneal only 2 (4) 3 (5) 2 (4) 0
Bone only 1 (2) 1 (2) 0 0
Primary tumor/colon 1 (2) 0 0 0
Other 10 (21) 17 (29) 9 (16) 11 (18)
Unknown 1 (2) 0 0 0
Open in a new tab

Abbreviations: FOLFIRI, folinic acid, fluorouracil, and irinotecan; FOLFOX, folinic acid, fluorouracil, and oxaliplatin; NA, not applicable.

a

All liver metastases treated with an R0/R1 resection and/or ablation.

b

Three patients in arm 2 and 1 patient in arm 4 had death as an early recurrence event.

In patients who underwent no or incomplete local treatment, systemic treatment after first progression was administered for 77 of 89 patients (87%) in group 1 vs 47 of 60 patients (78%) in group 2 (P = .26), as well as 42 of 46 patients (91%) in group 3 vs 34 of 43 patients (79%) in group 4 (P = .14; Table 2). In patients who underwent complete local treatment and had recurrence or progression of a primary tumor when still in situ, subsequent systemic treatment was administered for 31 of 47 patients (66%) in group 1 vs 37 of 59 patients (63%) in group 2 (P = .84), as well as 43 of 56 patients (77%) in group 3 vs 47 of 62 patients (76%) in group 4 (P > .99). Surgery, radiotherapy, ablation, chemoembolization, and radioembolization for primary tumors or metastases in follow-up, without distinguishing between palliative and curative intent, was administered for 50 patients (34%) in group 1 vs 48 patients (33%) in group 2 (P > .99), as well as 47 patients (41%) in group 3 vs 69 (59%) in group 4 (P < .001; eTable 4 in Supplement 2).

Table 2. Systemic Treatment on First Progression or Recurrence.

Outcome No. (%)
Patients with a right-sided and/or RAS or BRAFV600E-variant tumor Patients with a left-sided and RAS and BRAFV600E wild-type tumor
FOLFOX or FOLFIRI plus bevacizumab (n = 147) FOLFOXIRI plus bevacizumab (n = 144) FOLFOX or FOLFIRI plus bevacizumab (n = 114) FOLFOX or FOLFIRI plus panitumumab (n = 116)
Prior no or incomplete local treatment (n = 94) Prior complete local treatment (n = 53) Prior no or incomplete local treatment (n = 70) Prior complete local treatment (n = 74) Prior no or incomplete local treatment (n = 48) Prior complete local treatment (n = 66) Prior no or incomplete local treatment (n = 49) Prior complete local treatment (n = 67)
Alive without progression or recurrence 1 (1) 6 (11) 0 12 (16) 1 (2) 10 (15) 1 (2) 4 (6)
Death without progression or recurrence 4 (4) 0 10 (14) 3 (4) 1 (2) 0 5 (10) 1 (1)
Progression or recurrence 89 (95) 47 (89) 60 (86) 59 (80) 46 (96) 56 (85) 43 (88) 62 (93)
Any systemic treatment on progression or recurrence 77 (87) 31 (66) 47 (78) 37 (63) 42 (91) 43 (77) 34 (79) 47 (76)
Use of drugs in patients on progression or recurrence
Fluoropyrimidine 61 (69) 26 (55) 40 (67) 32 (54) 36 (78) 38 (68) 27 (63) 44 (71)
Oxaliplatin 24 (27) 13 (28) 18 (30) 22 (37) 19 (41) 17 (30) 12 (28) 25 (40)
Irinotecan 62 (70) 19 (40) 22 (37) 21 (36) 34 (74) 30 (54) 29 (67) 35 (56)
Trifluridin/tipiracil 19 (21) 10 (21) 18 (30) 12 (20) 10 (22) 10 (18) 10 (23) 12 (19)
Anti-EGFR treatment 4 (4) 2 (4) 5 (8) 2 () 33 (72) 30 (54) 15 (35) 29 (47)
Bevacizumab 32 (36) 14 (30) 27 (45) 18 (31%) 14 (30) 18 (32) 25 (58) 31 (50)
Other systemic treatment 5 (6) 4 (9) 3 (5) 7 (12) 4 (9) 3 (5) 2 (5) 3 (5)
Open in a new tab

Abbreviations: FOLFIRI, folinic acid, fluorouracil, and irinotecan; FOLFOX, folinic acid, fluorouracil, and oxaliplatin; EGFR, epidermal growth factor receptor.

In the sIPTW analysis, 226 patients were included (eFigure 5 in Supplement 2). ACT was administered with a median of 5 cycles (IQR, 3-8). Patients who received ACT had less extensive disease, fewer induction cycles, and fewer Clavien-Dindo grade 3 or greater complications, with well-balanced characteristics (SMDs ≤0.10) after sIPTW (eTable 5 in Supplement 2). Clavien-Dindo grade 3 or greater complications are reported in eTable 6 in Supplement 2. Median follow-up was 60.5 months (95% CI, 55.5-63.6) in patients receiving ACT and 63.1 months (95% CI, 61.9-65.2) in patients without ACT. Person-years of follow-up was 416 in patients receiving ACT and 392 in patients without ACT. After adjusting for confounders with sIPTW, ACT was associated with an OS benefit (HR, 0.66; 95% CI, 0.44-0.98; P = .04). The estimated 5-year survival was 50% in patients who received ACT and 35% in patients not treated with ACT. After sIPTW, ACT was associated with a relapse-free survival benefit (HR, 0.65; 95% CI, 0.48-0.88; P < .001) and less early recurrence without salvage local treatment (odds ratio, 0.46; 95% CI, 0.25-0.85; P = .01).

Discussion

In this RCT, we observed no OS benefit for FOLFOXIRI-bevacizumab vs FOLFOX/FOLFIRI–bevacizumab in patients with initially unresectable CRLM with RAS/BRAFV600E-variant and/or right-sided primary tumors. This may not be unexpected given the previously reported small, although statistically significant, PFS benefit with FOLFOXIRI-bevacizumab.1 In patients with RAS/BRAFV600E wild-type and left-sided tumors, FOLFOX/FOLFIRI–panitumumab vs FOLFOX/FOLFIRI–bevacizumab also showed no OS difference. Between these treatments, PFS and complete local treatment rates were not different, although panitumumab was associated with a higher response rate and depth of response, but also with more toxic effects.1 Because of these differences in quality of response, it has been argued that PFS may not be an optimal end point, especially in studies with anti-EGFR antibodies.18 However, our results did not confirm the OS benefit for anti-EGFR–based vs anti-vascular endothelial growth factor–based treatment, as shown by others in the overall metastatic colorectal cancer (mCRC) population.19,20 Regarding the finding that the higher response rate with FOLFOX/FOLFIRI–panitumumab did not translate into more complete local treatments, we hypothesized that in this population with a poor prognosis with a median of 12 CRLMs, the observed response rate of 80% exceeds the threshold of converting CRLM to a resectable status. For example, anatomical locations could limit resectability.

Although OS analysis was underpowered as a secondary end point, it seems unlikely that a larger sample would have resulted in different outcomes given the shape of the Kaplan-Meier curves. The primary end point of PFS was met for FOLFOXIRI-bevacizumab in patients with RAS/BRAFV600E-variant and/or right-sided tumors. Nevertheless, we believe that this marginal benefit does not outweigh the increased toxic effects, particularly when considering that no OS benefit was observed.

Comparing our results with previous prospective studies in patients with initially unresectable CRLM was hampered because RAS/BRAFV600E status and primary tumor location were not prospectively considered when selecting treatment.21,22,23 Additionally, these studies lacked repeated, well-structured resectability assessments by a liver expert panel, possibly leading to suboptimal patient selection for local treatment, as shown previously.21 In non–liver-limited studies, an OS benefit has been observed for FOLFOXIRI-bevacizumab vs FOLFOX/FOLFIRI–bevacizumab.24,25 An individual patient data meta-analysis suggested that the HRs for comparing OS between FOLFOXIRI-bevacizumab and FOLFOX/FOLFIRI–bevacizumab were similar in patients with and without liver-only metastases, although it was only significant for patients with non–liver-limited metastases.25 Nevertheless, due to previously mentioned reasons, these results cannot be directly compared with our results. In previous studies in the overall population with mCRC, FOLFOX/FOLFIRI–panitumumab vs FOLFOX/FOLFIRI–bevacizumab demonstrated an OS benefit.19,20,26 However, this was lost in patients with CRLM and RAS wild-type and left-sided tumors according to subgroup analyses from 2 studies.26,27 Together these data suggest that outcomes in patients with mCRC may not be extrapolated to the subgroup of patients with CRLM. Differences in tumor biology and behavior of hepatic vs extrahepatic metastases may explain the disparity in findings between our study and other studies with non–liver-limited disease. Further research on this topic is warranted. A recent study found no benefit for FOLFOXIRI-panitumumab in patients with RAS/BRAFV600E wild-type tumors.28

Despite the frequent occurrence of early recurrence, we observed a significant OS benefit for local CRLM treatment in all groups. In line with other studies, OS was not impaired in patients with early recurrence after local treatment if salvage local treatment was performed.29 Our observation that more complete local treatments after FOLFOXIRI-bevacizumab did not affect OS implies some patients might not benefit from local treatment. This is also reflected by about one-third of patients having recurrence 6 months or less after complete local treatment that was not amenable to salvage local treatment and was associated with worse OS. However, recurrence outcomes were not different between treatment groups. In a previous analysis, we showed that clinically available parameters cannot sufficiently predict early recurrence.30 Therefore, we recommend offering local treatment to all patients for whom this is technically feasible. Research is ongoing to identify patients who are technically suitable to receive local treatment but who derive no benefit from this. Given the large intersurgeon variability regarding resectability evaluations, decision-making is preferably based on a liver expert panel rather than a single-surgeon decision.31

Strengths and Limitations

International guidelines recommend limiting induction and adjuvant systemic treatment for initially unresectable CRLM to 6 months without distinguishing between preoperative and postoperative duration.32 In this study, ACT was administered to 39% to 49% of patients after complete local treatment. We observed a significant benefit for ACT in overall and relapse-free survival. Some patients may not have received ACT due to earlier toxic effects or insufficient conditions. However, this decision may also be attributed to the lack of an OS benefit in phase 3 studies in patients with upfront resectable CRLM4,5,6,7,8,9 and the lack of data in patients with initially unresectable CRLM. To our knowledge, this is the first study that evaluated efficacy of ACT and demonstrated its benefits in a clearly defined cohort of patients with initially unresectable CRLM. Limitations concerned its post hoc character, small sample, and potentially residual confounding despite the attempt to balance confounders through sIPTW. However, in the absence of RCT data, which are challenging to perform, our results potentially justify discussing the option of ACT in patients who underwent complete local treatment after the downsizing of initially unresectable CRLM.

Conclusions

The long-term results of the CAIRO5 randomized clinical trial support using FOLFOX/FOLFIRI plus bevacizumab for patients with initially unresectable CRLM. This finding was irrespective of RAS/BRAFV600E variant status and tumor sidedness.

Supplement 1.

Trial protocol

jamaoncol-e245174-s001.pdf (677.6KB, pdf)
Supplement 2.

eTable 1. Centres and principal investigators

eTable 2. Baseline characteristics

eFigure 1. Overall survival according to local treatment status

eFigure 2. Subgroup analyses for overall survival

eFigure 3. Overall survival according to RAS/BRAFV600E mutation status in group A and B

eTable 3. Sensitivity analysis for comparing overall survival between early recurrence outcomes after complete local treatment

eFigure 4. Overall survival in R0 versus R1 resected patients

eTable 4. Local treatment in follow-up

eFigure 5. Flowchart eligible patients for stabilised inverse probability treatment weighting

eTable 5. Characteristics of patients included in the stabilised inverse probability treatment weighting analysis

eTable 6. Clavien-Dindo grade 3 or higher postoperative complications

jamaoncol-e245174-s002.pdf (1MB, pdf)
Supplement 3.

Nonauthor collaborators

jamaoncol-e245174-s003.pdf (110KB, pdf)
Supplement 4.

Data sharing statement

jamaoncol-e245174-s004.pdf (15.8KB, pdf)

References

  • 1.Bond MJG, Bolhuis K, Loosveld OJL, et al. ; Dutch Colorectal Cancer Study Group . First-line systemic treatment strategies in patients with initially unresectable colorectal cancer liver metastases (CAIRO5): an open-label, multicentre, randomised, controlled, phase 3 study from the Dutch Colorectal Cancer Group. Lancet Oncol. 2023;24(7):757-771. doi: 10.1016/S1470-2045(23)00219-X [DOI] [PubMed] [Google Scholar]
  • 2.Lam VWT, Spiro C, Laurence JM, et al. A systematic review of clinical response and survival outcomes of downsizing systemic chemotherapy and rescue liver surgery in patients with initially unresectable colorectal liver metastases. Ann Surg Oncol. 2012;19(4):1292-1301. doi: 10.1245/s10434-011-2061-0 [DOI] [PubMed] [Google Scholar]
  • 3.Petrelli NJ. Perioperative or adjuvant therapy for resectable colorectal hepatic metastases. J Clin Oncol. 2008;26(30):4862-4863. doi: 10.1200/JCO.2008.18.5868 [DOI] [PubMed] [Google Scholar]
  • 4.Portier G, Elias D, Bouche O, et al. Multicenter randomized trial of adjuvant fluorouracil and folinic acid compared with surgery alone after resection of colorectal liver metastases: FFCD ACHBTH AURC 9002 trial. J Clin Oncol. 2006;24(31):4976-4982. doi: 10.1200/JCO.2006.06.8353 [DOI] [PubMed] [Google Scholar]
  • 5.Mitry E, Fields ALA, Bleiberg H, et al. Adjuvant chemotherapy after potentially curative resection of metastases from colorectal cancer: a pooled analysis of two randomized trials. J Clin Oncol. 2008;26(30):4906-4911. doi: 10.1200/JCO.2008.17.3781 [DOI] [PubMed] [Google Scholar]
  • 6.Kobayashi A, Hasegawa K, Saiura A, et al. A randomized controlled trial evaluating efficacy of adjuvant oral uracil-tegafur (UFT) with leucovorin (LV) after resection of colorectal cancer liver metastases: The UFT/LV study. J Clin Oncol. 2014;32(suppl 15):1. doi: 10.1200/jco.2014.32.15_suppl.358424276780 [DOI] [Google Scholar]
  • 7.Hasegawa K, Saiura A, Takayama T, et al. Adjuvant oral uracil-tegafur with leucovorin for colorectal cancer liver metastases: a randomized controlled trial. PLoS One. 2016;11(9):e0162400. doi: 10.1371/journal.pone.0162400 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Kanemitsu Y, Shimizu Y, Mizusawa J, et al. ; JCOG Colorectal Cancer Study Group . Hepatectomy followed by mFOLFOX6 versus hepatectomy alone for liver-only metastatic colorectal cancer (JCOG0603): a phase II or III randomized controlled trial. J Clin Oncol. 2021;39(34):3789-3799. doi: 10.1200/JCO.21.01032 [DOI] [PubMed] [Google Scholar]
  • 9.Nordlinger B, Sorbye H, Glimelius B, et al. ; EORTC Gastro-Intestinal Tract Cancer Group; Cancer Research UK; Arbeitsgruppe Lebermetastasen und–tumoren in der Chirurgischen Arbeitsgemeinschaft Onkologie; Australasian Gastro-Intestinal Trials Group; Fédération Francophone de Cancérologie Digestive . Perioperative FOLFOX4 chemotherapy and surgery versus surgery alone for resectable liver metastases from colorectal cancer (EORTC 40983): long-term results of a randomised, controlled, phase 3 trial. Lancet Oncol. 2013;14(12):1208-1215. doi: 10.1016/S1470-2045(13)70447-9 [DOI] [PubMed] [Google Scholar]
  • 10.Huiskens J, van Gulik TM, van Lienden KP, et al. Treatment strategies in colorectal cancer patients with initially unresectable liver-only metastases, a study protocol of the randomised phase 3 CAIRO5 study of the Dutch Colorectal Cancer Group (DCCG). BMC Cancer. 2015;15:365-365. doi: 10.1186/s12885-015-1323-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Pietrantonio F, Petrelli F, Coinu A, et al. Predictive role of BRAF mutations in patients with advanced colorectal cancer receiving cetuximab and panitumumab: a meta-analysis. Eur J Cancer. 2015;51(5):587-594. doi: 10.1016/j.ejca.2015.01.054 [DOI] [PubMed] [Google Scholar]
  • 12.Holch JW, Ricard I, Stintzing S, Modest DP, Heinemann V. The relevance of primary tumour location in patients with metastatic colorectal cancer: a meta-analysis of first-line clinical trials. Eur J Cancer. 2017;70:87-98. doi: 10.1016/j.ejca.2016.10.007 [DOI] [PubMed] [Google Scholar]
  • 13.Viganò L, Capussotti L, Lapointe R, et al. Early recurrence after liver resection for colorectal metastases: risk factors, prognosis, and treatment: a LiverMetSurvey-based study of 6,025 patients. Ann Surg Oncol. 2014;21(4):1276-1286. doi: 10.1245/s10434-013-3421-8 [DOI] [PubMed] [Google Scholar]
  • 14.Imai K, Allard MA, Benitez CC, et al. Early recurrence after hepatectomy for colorectal liver metastases: what optimal definition and what predictive factors? Oncologist. 2016;21(7):887-894. doi: 10.1634/theoncologist.2015-0468 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Grambsch PM, Therneau TM. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika. 1994;81(3):515-526. doi: 10.1093/biomet/81.3.515 [DOI] [Google Scholar]
  • 16.Austin PC, Stuart EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Stat Med. 2015;34(28):3661-3679. doi: 10.1002/sim.6607 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Mekenkamp LJM, Koopman M, Teerenstra S, et al. Clinicopathological features and outcome in advanced colorectal cancer patients with synchronous vs metachronous metastases. Br J Cancer. 2010;103(2):159-164. doi: 10.1038/sj.bjc.6605737 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Heinemann V, Stintzing S, Modest DP, Giessen-Jung C, Michl M, Mansmann UR. Early tumour shrinkage (ETS) and depth of response (DpR) in the treatment of patients with metastatic colorectal cancer (mCRC). Eur J Cancer. 2015;51(14):1927-1936. doi: 10.1016/j.ejca.2015.06.116 [DOI] [PubMed] [Google Scholar]
  • 19.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(10):1065-1075. doi: 10.1016/S1470-2045(14)70330-4 [DOI] [PubMed] [Google Scholar]
  • 20.Venook AP, Niedzwiecki D, Innocenti F, et al. Impact of primary (1o) tumor location on overall survival (OS) and progression-free survival (PFS) in patients (pts) with metastatic colorectal cancer (mCRC): analysis of CALGB/SWOG 80405 (Alliance). J Clin Oncol. 2016;34(15)(suppl):3504-3504. doi: 10.1200/JCO.2016.34.15_suppl.3504 [DOI] [Google Scholar]
  • 21.Folprecht G, Gruenberger T, Bechstein WO, et al. Tumour response and secondary resectability of colorectal liver metastases following neoadjuvant chemotherapy with cetuximab: the CELIM randomised phase 2 trial. Lancet Oncol. 2010;11(1):38-47. doi: 10.1016/S1470-2045(09)70330-4 [DOI] [PubMed] [Google Scholar]
  • 22.Ychou M, Rivoire M, Thezenas S, et al. A randomized phase II trial of three intensified chemotherapy regimens in first-line treatment of colorectal cancer patients with initially unresectable or not optimally resectable liver metastases. The METHEP trial. Ann Surg Oncol. 2013;20(13):4289-4297. doi: 10.1245/s10434-013-3217-x [DOI] [PubMed] [Google Scholar]
  • 23.Gruenberger T, Bridgewater J, Chau I, et al. Bevacizumab plus mFOLFOX-6 or FOLFOXIRI in patients with initially unresectable liver metastases from colorectal cancer: the OLIVIA multinational randomised phase II trial. Ann Oncol. 2015;26(4):702-708. doi: 10.1093/annonc/mdu580 [DOI] [PubMed] [Google Scholar]
  • 24.Cremolini C, Antoniotti C, Rossini D, et al. ; GONO Foundation Investigators . 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(4):497-507. doi: 10.1016/S1470-2045(19)30862-9 [DOI] [PubMed] [Google Scholar]
  • 25.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. J Clin Oncol. 2020;38(28):JCO2001225. doi: 10.1200/JCO.20.01225 [DOI] [PubMed] [Google Scholar]
  • 26.Watanabe J, Muro K, Shitara K, et al. Panitumumab vs bevacizumab added to standard first-line chemotherapy and overall survival among patients with RAS wild-type, left-sided metastatic colorectal cancer: a randomized clinical trial. JAMA. 2023;329(15):1271-1282. doi: 10.1001/jama.2023.4428 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Holch JW, Ohnmacht A, Stintzing S, et al. Refining first-line treatment decision in RAS wildtype (RAS-WT) metastatic colorectal cancer (mCRC) by combining clinical biomarkers: results of the randomized phase 3 trial FIRE-3 (AIO KRK0306). J Clin Oncol. 2024;42(3)(suppl):13-13. doi: 10.1200/JCO.2024.42.3_suppl.1337883739 [DOI] [Google Scholar]
  • 28.Rossini D, Antoniotti C, Lonardi S, et al. Upfront modified fluorouracil, leucovorin, oxaliplatin, and irinotecan plus panitumumab versus fluorouracil, leucovorin, and oxaliplatin plus panitumumab for patients with RAS/BRAF wild-type metastatic colorectal cancer: the phase III TRIPLETE study by GONO. J Clin Oncol. 2022;40(25):2878-2888. doi: 10.1200/JCO.22.00839 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Hellingman T, de Swart ME, Heymans MW, Jansma EP, van der Vliet HJ, Kazemier G. Repeat hepatectomy justified in patients with early recurrence of colorectal cancer liver metastases: a systematic review and meta-analysis. Cancer Epidemiol. 2021;74:101977. doi: 10.1016/j.canep.2021.101977 [DOI] [PubMed] [Google Scholar]
  • 30.Bolhuis K, Bond MJG, Van Amerongen MJ, et al. ; Dutch Colorectal Cancer Group Liver Expert Panel . The role of tumour biological factors in technical anatomical resectability assessment of colorectal liver metastases following induction systemic treatment: an analysis of the Dutch CAIRO5 trial. Eur J Cancer. 2023;183:49-59. doi: 10.1016/j.ejca.2023.01.013 [DOI] [PubMed] [Google Scholar]
  • 31.Bond MJG, Kuiper BI, Bolhuis K, et al. Intersurgeon variability in local treatment planning for patients with initially unresectable colorectal cancer liver metastases: analysis of the Liver Expert Panel of the Dutch Colorectal Cancer Group. Ann Surg Oncol. 2023;30(9):5376-5385. doi: 10.1245/s10434-023-13510-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Cervantes A, Adam R, Roselló S, et al. ; ESMO Guidelines Committee . Metastatic colorectal cancer: ESMO clinical practice guideline for diagnosis, treatment and follow-up. Ann Oncol. 2023;34(1):10-32. doi: 10.1016/j.annonc.2022.10.003 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement 1.

Trial protocol

jamaoncol-e245174-s001.pdf (677.6KB, pdf)
Supplement 2.

eTable 1. Centres and principal investigators

eTable 2. Baseline characteristics

eFigure 1. Overall survival according to local treatment status

eFigure 2. Subgroup analyses for overall survival

eFigure 3. Overall survival according to RAS/BRAFV600E mutation status in group A and B

eTable 3. Sensitivity analysis for comparing overall survival between early recurrence outcomes after complete local treatment

eFigure 4. Overall survival in R0 versus R1 resected patients

eTable 4. Local treatment in follow-up

eFigure 5. Flowchart eligible patients for stabilised inverse probability treatment weighting

eTable 5. Characteristics of patients included in the stabilised inverse probability treatment weighting analysis

eTable 6. Clavien-Dindo grade 3 or higher postoperative complications

jamaoncol-e245174-s002.pdf (1MB, pdf)
Supplement 3.

Nonauthor collaborators

jamaoncol-e245174-s003.pdf (110KB, pdf)
Supplement 4.

Data sharing statement

jamaoncol-e245174-s004.pdf (15.8KB, pdf)

Articles from JAMA Oncology are provided here courtesy of American Medical Association

RESOURCES