545 Background: The LITESPARK-013 study (NCT04489771) showed similar efficacy and safety between 200 mg and 120 mg (approved dose) doses of the HIF-2α inhibitor bel in pts with pretreated advanced ccRCC. We report exploratory biomarker analyses (including HIF-2α IHC, RNAseq, and WES) from this study. Methods: Pts with advanced ccRCC whose disease progressed after 1-3 prior systemic therapies (including an anti–PD-L1 therapy) were randomly assigned 1:1 to bel 200 mg or 120 mg QD. Treatment arms were pooled for biomarker analyses. Association of HIF-2α tumor proportion score (TPS; IHC D6T8V) and H-score (sum of each percentage multiplied by its corresponding intensity 0, 1+, 2+, 3+) with clinical outcomes (ORR, PFS, OS) were tested at prespecified multiplicity-adjusted α = 0.05. Associations of RNAseq-based HIF-2α co-expression and HIF-2α cancer cell scores, glycolysis, angiogenesis, and other HIF-2α–related genes and signatures were tested at prespecified multiplicity-adjusted α = 0.10. Associations for molecular subtypes (Motzer et al. Cancer Cell . 2020;38:803) were tested at nominal α = 0.05. DNA mutations were evaluated by WES. Results: Of 154 pts, HIF-2α IHC was evaluable in 136, RNAseq in 109, and WES in 113 pts. HIF-2α TPS and H-score were positively associated with PFS ( P = 0.016 and 0.012, respectively), but not ORR or OS. HIF-2α cancer cell score was positively associated with ORR ( P = 0.085), but not PFS or OS. No associations with clinical outcome were observed for HIF-2α co-expression score. Glycolysis and angiogenesis were positively associated with ORR ( P = 0.036) and PFS ( P = 0.070), respectively, but these were no longer significant after adjustment of other signatures. ORR was consistent across molecular subtypes and by DNA mutation status (Table), with notably higher ORR in the BAP1 mutant vs wild-type group. Conclusions: In this first RNA and WES biomarker analysis of bel, HIF-2α TPS and H-score were potentially associated with PFS, and BAP1 mutation was potentially predictive of improved ORR in pts with RCC. Molecular subtypes, other RCC-related signatures, and VHL mutation showed no associations or trends with ORR, PFS, or OS. Prospective biomarker evaluations are needed to confirm these findings. Clinical trial information: NCT02853344 . n ORR, % (95% CI) All RNAseq evaluable 109 26.6 (18.6-35.9) Molecular subtypes Angiogenic 12 33.3 (9.9-65.1) Angiogenic/stromal 24 25.0 (9.8, 46.7) Immune/proliferative 22 27.3 (10.7-50.2) Stromal/proliferative 17 23.5 (6.8-49.9) Proliferative 17 23.5 (6.8-49.9) Other (unassignable to a subtype) 17 29.4 (10.3-56.0) VHL Mut 80 25.0 (16.0-35.9) WT 33 24.2 (11.1-42.3) PBRM1 Mut 59 23.7 (13.6-36.3) WT 54 25.9 (15.0-39.7) SETD2 Mut <jats:td colspan="1" rowspan=
Agarwal et al. (2026) studied this question.