ESMO Congress 2026

Daraxonrasib (RMC-6236) is an oral, non-covalent RAS(ON) multi-selective tri-complex inhibitor that blocks downstream signaling by forming a tri-complex with the intracellular chaperone CypA and the active, GTP-bound RAS ^[1]. Its evidence base is anchored by the published phase 3 RASolute 302 trial: 500 patients with previously treated mPDAC were randomized, 91.8% harboring RAS G12 mutations, with OS and PFS in the G12 population as dual primary endpoints ^[2]; the earlier phase 1-2 study in 168 previously treated RAS-mutated PDAC patients established 300 mg as the phase 3 dose, with treatment-related adverse events of any grade in 96% and grade 3 or higher in 30% ^[6]. On resistance, paired-sample analysis identified recurrent acquired alterations such as RAS Y64, Y71 and kinase-dead/hypoactive BRAF in 18 of 40 patients ^[5].

At ESMO26 (October 23), Zev Wainberg presented subgroup analyses of RASolute 302: OS 13.2 vs 6.7 months (HR 0.40), with OS HRs of 0.49, 0.32 and 0.61 for G12D, G12V and G12R respectively, PFS HR 0.55 by BICR, and ORR 32% vs 11%; the OS benefit extended to patients with no identified tumor RAS mutation, and commentators acknowledged consistent benefit across subgroups while cautioning about statistical power in small subgroups. A parallel storyline was the Tango Therapeutics combination: daraxonrasib plus vopimetostat showed a confirmed ORR of 81.8% (18/22) in advanced pancreatic cancer — 85.7% (6/7) in 2L and 84.6% (11/13) in 3L — with 100% disease control and 9-month PFS estimated at 88.4%, using a lower daraxonrasib dose; the on-site zoldonrasib plus vopimetostat data showed a confirmed ORR of 54.2% (19/35), including one CR.

Two steps forward versus existing knowledge. First, while the NEJM primary paper established phase 3 evidence on dual primary endpoints in the G12 population ^[2], the subgroup analyses refine the benefit down to each genotype and even to patients without identified mutations, echoing the trial design in which the overall population included patients with no RAS mutation identified ^[2], though small samples warrant caution. Second, the Tango combination follows the preclinical logic that combined blockade of downstream, upstream and orthogonal pathways can prevent RAS inhibitor resistance and yield durable regressions ^[4], consistent with evidence that RAS(ON) multi-selective inhibition reshapes the tumor immune microenvironment and creates room for combinations ^[3]. Against the FDA exploratory monotherapy ORRs of 50% in G12V, 24% in G12D and 7% in G12R, the 81.8% is striking, but attribution is debated: the daraxonrasib cohort was predominantly G12V/G12R with no G12D, vopimetostat raised exposure approximately 1.5–2.4-fold, and the all-treated recalculation of 62.1% (18/29) is not a formal ITT/mITT figure.

The limitations are equally clear: subgroups are small with limited statistical power; the 81.8% pools two dose levels while the proposed phase 3 regimen uses the lower dose, whose confirmed ORR and durability are undisclosed; the 7 PDAC patients excluded by the 14-week evaluability rule have no reported follow-up update; at the proposed phase 3 dose there was no grade 3 rash but grade 3 treatment-related adverse events reached 40%, leaving open whether toxicity is genuinely reduced or redistributed. Outside the hall, a Spanish post claimed the EU has already approved daraxonrasib and urged faster national access, underscoring access and reimbursement tensions beyond the data.