603P - A phase Ib study of the safety and efficacy of atezolizumab (atezo) + bevacizumab (bev) + cobimetinib (cobi) in patients (pts) with metastatic colorectal cancer (mCRC)

mCRC is a non-immunogenic tumour type with poor response to checkpoint inhibitors, hypothetically due to low tumour T-cell levels and incomplete blockade of MAPK-mediated growth factor signalling. Combining the anti-tumour immunity effects of the anti–PD-L1 Ab (atezo), MEK inhibitor (cobi) and anti-VEGF Ab (bev) in pts with mCRC may improve response. Atezo enhances T-cell priming/activation, cobi decreases tumour growth and increases tumour antigen presentation via MHC I and bev increases T-cell infiltration. Therefore, bev + atezo + cobi may enhance immune recognition of mCRC. This report is the primary analysis of a Phase Ib study of this combination. This study assessed the safety and activity of atezo (840mg q2w IV) + bev (5mg/kg q2w IV) + cobi (60mg qd 21 on/7 off oral) in pts with mCRC who progressed on≥1 prior line of treatment (tx) containing 5-FU and oxaliplatin/irinotecan. After safety run-in, dose-expansion and biopsy cohorts were enrolled. The primary objective was to assess safety. In the primary analysis (Sep 21, 2018, cutoff), 49 pts were evaluated. Median safety follow-up was 7.8mo (range, 1.4-25.4mo). Median prior lines of tx was 2 (range, 1-7). 33 pts (67%) had RAS-mutant tumours. 45 pts (92%) had microsatellite stable disease; 4 pts (8%) had unknown microsatellite status. 48 pts (98%) had tx-related AEs (TRAEs), 30 pts (61%) had G3-5 TRAEs and 1pt (2%) had bev-related G5 GI necrosis. AEs led to tx withdrawal in 9 pts (18%). Most common AEs were diarrhoea (78%), acneiform dermatitis (51%), fatigue (49%), rash (41%) and CPK increase (35%). Confirmed ORR was 8% in the overall population. See Table for activity by tx line and RAS status. Atezo + bev + cobi had an acceptable safety profile with manageable AEs with improved activity over that reported with SOC in 2L+ mCRC. A greater benefit trend was seen in pts with a RAS mutation vs wild-type tumours.Table603PTable2L therapy (n=25)3L+ therapy (n=24)Overall (N=49)Confirmed response per RECIST 1.1, n (%)a2 (8%)2 (8%)4 (8%)mPFS (range), moa5.5 (3.6-9.4)5.8(2.4-9.0)5.5 (3.6-7.5)mOS (range), moNE(11.7-NE)11.8 (8.7-21.8)11.8 (9.0-21.8)RAS mutant (n=33)RAS wild-type/ unknown (n=16)Pb HR (95% CI)DCR, n (%)c (95% CI)12 (36) (20, 55)5 (31) (11, 59)–mPFS, mo (95% CI)5.8 (3.6, 11.6)3.6 (2.6, 7.3)0.05 2.2 (1.1, 4.5)mOS, mo (95% CI)21.8 (9.0, NE)9.1 (3.6, 14.5)0.34 1.8 (0.6, 5.5)Ab, antibody; DCR, disease control rate; HR, hazard ratio; mPFS, median progression-free survival; mOS, median overall survival;