607 Background: To date, no biomarker of efficacy of nivolumab+/-ipilimumab (N+/-I) or anti-VEGFR TKI has been prospectively validated in mRCC. The BIONIKK trial showed promising objective response rate (ORR) and progression-free survival (PFS) with these treatments in first line (L1) after selection by tumour molecular group. We report OS and efficacy results of the second-line (L2) treatment. Methods: BIONIKK is a French multicentre non-comparative phase II trial, randomising 199 mRCC pts to receive N (58), NI (101) or TKI (40) in L1 according to four molecular groups (ccrcc1-4). ORR and PFS were already reported. With an additional follow-up of ≥20 months, we report OS from randomization and from the start of L2, as well as ORR and PFS with a TKI in L2 by molecular group. Results: With a median follow-up of 42.1 months (40.5-45.2), 86 (43%) patients died: 27/58 (46.5%), 39/101 (39%) and 20/40 (50%) in the N, NI, and TKI arm, respectively. Median OS were 43.4 months (95%CI=31.4-NR) with N, 52.7 months (95%CI=46-NR) with NI and 38.1 months (95%CI=33.2-NR) with TKI (table). 175 (88%) patients discontinued first-line treatment, including 20 deaths, and 129 (74%) received a L2, 38/58 (65.5%), 64/101 (63%), and 27/40 (67.5%) after N, NI and TKI, respectively. The most frequent L2 received after N+/-I was a TKI in 96/102 (94%) pts, including cabozantinib in 49, sunitinib/pazopanib in 32, axitinib in 13, and lenvatinib in 2. N was the most frequent L2 after TKI, 20/27 (74%). ORR with TKI in L2 was 28.5% (10/35) after N, 39% (24/61) after NI and 80% (4/5) after TKI, with marked benefit in ccrcc2 pts (table). The mPFS with TKI in L2 was 8.2 (95%CI=6.9-19.3) after N, 11.4 (95%CI= 8.9-16.8) after NI, and 12.1 (95%CI =11.4-NR) months after TKI, with a higher benefit in ccrcc2 pts (vs. ccrcc1+4, p=0.04). Conversely, ORR and mPFS with N after TKI in ccrcc2-pts were 12.5% (2/16) and 5.4 (2.6-NR) months, respectively. Median OS L2 was reported in the table. The updated ORR and PFS in L1 will presented at the Meeting, as well as PFS2 and efficacy by TKI type in L2. Conclusions: We report for the first-time OS and L2 efficacy results by molecular group in a randomized trial. Molecular selection also has an impact on treatment efficacy in L2. These results, together with those reported in L1, can inform clinicians on the best treatment sequence in L1-2. Clinical trial information: NCT02960906 . [Table: see text]