FOLFOXIRI + bevacizumab versus FOLFOX/FOLFIRI + bevacizumab in patients with initially unresectable colorectal liver metastases (CRLM) and right-sided and/or RAS/BRAFV600E-mutated primary tumor: Phase III CAIRO5 study of the Dutch Colorectal Cancer Group.

LBA3506 Background: Patients (pts) with initially unresectable CRLM may qualify for curative-intent local therapy after downsizing by induction systemic therapy. The CAIRO5 study aims to find the optimal induction regimen. We present the results of pts with right-sided and/or RAS/ BRAFV600E mutated primary tumors. Accrual in pts with left-sided and RAS/ BRAFV600E wildtype tumors is ongoing. Methods: Pts were randomized between FOLFOX or FOLFIRI + bevacizumab (B) (arm A) and FOLFOXIRI-B (arm B), in both arms up to 12 cycles and followed by 5FU/LV/B maintenance. Prior systemic or local therapy for metastases was not allowed. Unresectability of CRLM at baseline was assessed by an online liver expert panel of surgeons and radiologists based on predefined criteria, and resectability every 2 months thereafter based on panel majority vote. The primary endpoint was progression-free survival (PFS). Secondary endpoints were R0/1 resection, overall survival, overall response rate (ORR), toxicity, pathologic response, postoperative morbidity and correlation of panel evaluations with outcome. Pts were stratified by potentially resectable vs permanently unresectable CRLM, serum LDH (normal/abnormal), BRAFV600E mutation, sidedness, choice of irinotecan vs oxaliplatin and institute. To detect a hazard ratio (HR) of 0.70 for PFS with 80% power and a 2-sided log-rank test at 5%, 257 events were required, assuming a median PFS of 8.7 months for arm A. Results: From December 2014 until March 2021, 294 pts were randomized, 148 in arm A and 146 in arm B in 43 Dutch and 1 Belgian sites. 3 ineligible pts were excluded. Median follow up was 40 months. Pts received a median of 10 vs 9 induction cycles in arm A vs B. Main characteristics were (arm A/B): median age 61/65 years, male 63.9/60.4%, right-sided primary tumor 39.5/41.7%, RAS mutant 85.7/86.1%, BRAFV600E mutant 6.8/8.3%, synchronous disease 86.4/89.6%, prior adjuvant chemotherapy 4.8/4.9%, median number of CRLM 12/12. With 259 events, median PFS in arm A vs B was 9.0 vs 10.6 months (stratified HR 0.74, 95% CI 0.57-0.96, p=0.02). ORR was 32.0% vs 52.1% (p<0.001), any grade ≥3 adverse events occurred in 58.5% vs 75.0% (p=0.003), the most frequent were neutropenia (12.9/38.2%, p<0.001), hypertension (14.3/13.9%, p=1) and diarrhea (3.4/19.4%, p<0.001) in arm A vs B, respectively. R0/1 resection ± ablation rates were 37.4% vs 51.4% (p=0.02) with 2-stage procedures in 16.4% vs 32.4% (p=0.04), postoperative complications occurred in 38.2% vs 51.2% (p=0.14), Clavien Dindo grade ≥3 in 14.7% vs 26.8% (p=0.08). Conclusions: FOLFOXIRI-B vs FOLFOX/FOLFIRI-B significantly increases PFS, ORR and R0/1 resections at the cost of increased toxicity in pts with initially unresectable CRLM and right-sided and/or RAS/ BRAFV600E mutated primary tumor. Clinical trial information: NCT02162563.