505 Background: The addition of novel mechanisms of action to established treatment strategies may further improve outcomes for patients with previously treated advanced ccRCC. We present updated results from substudy 03B (NCT04626518) of KEYMAKER-U03, a phase I/II study evaluating novel regimens in ccRCC. Methods: Adults with confirmed ccRCC and disease progression on or after PD-(L)1 inhibitor and VEGF-TKI were randomly assigned to open arms: pembrolizumab (pembro)/quavonlimab (qmab; anti–CTLA-4), pembro/favezelimab (fave; anti-LAG3), pembro + MK-4830 (anti-ILT4), pembro + belzutifan (bel; HIF-2α inhibitor), bel + lenvatinib (lenva; VEGF-TKI), and pembro + lenva. Primary end points were safety and confirmed ORR per RECIST v1.1 by BICR. Secondary end points included clinical benefit rate (CBR; CR + PR + SD ≥6 mo), duration of response (DOR) and progression-free survival (PFS) per RECIST v1.1 by BICR, and overall survival (OS). No formal comparisons were made across arms. Enrollment was planned for 50 participants (pts) in each arm, although enrollment would be stopped if the 6-mo PFS rate was ≤40% among the first 20 enrolled pts per protocol design. Results: Overall, 263 pts were enrolled (Table). Median study follow-up was ≥22.9 mo in all arms. Efficacy is reported in the table. Grade 3-5 treatment-related adverse events (TRAEs) occurred in 25%, 11%, 9%, 45%, 63%, and 52% of treated pts in the pembro/qmab, pembro/fave, pembro + MK-4830, pembro + bel, bel + lenva, and pembro + lenva arms, respectively. TRAEs led to death in 2 pts in the bel + lenva arm (cerebral hemorrhage and intracranial hemorrhage) and 1 pt in the pembro + lenva arm (esophageal perforation). Conclusions: Bel + lenva continued to demonstrate the highest antitumor activity followed by pembro + lenva. Pembro + bel activity was similar to historical bel monotherapy. In the small pembro/qmab cohort, median OS was not reached. No responses were observed in the anti-LAG3 or anti-ILT4 combo arms. The safety profile of each treatment was manageable. The HIF/VEGF axis continues to represent a key therapeutic target in previously treated ccRCC. These data support further evaluation of bel + lenva or other VEGF-TKIs. Clinical trial information: NCT04626518 . Pembro + Qmabn = 20 Pembro + Fave n = 20 Pembro + MK-4830 n = 24 Pembro + beln = 62 Bel + lenva n = 64 Pembro + lenvan = 73 ORR (95% CI), % 30 (12-54) 0 (0-17) 0 (0-14) 23 (13-35) 50 (37-63) 44 (32-56) CR, n (%) 0 0 0 2 (3) 3 (5) 1 (1) PR, n (%) 6 (30) 0 0 12 (19) 29 (45) 31 (42) CBR (95% CI), % 35 (15-59) 5 (0-25) 0 (0-14) 35 (24-49) 61 (48-73) 60 (48-72) DOR, median (range), mo NR(2.7-24.4+) – – 23.6 (2.8- 38.6+) NR (2.9-38.4+) 16.6 (2.5+-25.6+) PFS, median (95% CI), mo 5.4 (1.5-17.8) 1.6 (1.4-2.8) 1.5 (1.3-1.6) 5.5 (2.8-8.4) 15.0 (5.9-26.3) 9.7 (7.0-15.2) 12-mo PFS rate, % 34 7 NR 31 53 41 OS, median (95% CI), mo NR (7.8-NR) 22.1 (9.4-NR) 11.0 (6.5-18.0) 22.9 (14.2-NR) 36.4 (25.8-NR) 35.9 (21.9-NR) 12-mo OS rate, % 75 61 <jats:td colspan="1" rowspan="
Albigès et al. (2026) studied this question.