Literature Substrate · Public Sources · 2026-05-21
RAS / MAPK as a Drug-Target Landscape
A reading of 152 peer-reviewed and preprint records across three threads: active-state RAS pharmacology, the KRAS G12C-OFF comparator class, and the Phase 3 record in metastatic pancreatic ductal adenocarcinoma.
The published RAS / MAPK literature now describes a target whose mechanism is replicated outside the sponsor, whose comparator class has delivered reproducible PFS without OS conversion at primary analysis, and whose disease floor in metastatic pancreatic cancer sits at one positive Phase 3 in twenty-five years with a 0.33-month median-OS magnitude. Three independent academic groups — Cuevas-Navarro (MSKCC/Lito), Pfaff & Shokat (UCSF), and Alexander (NCI/Frederick) — have confirmed the active-state mechanism with orthogonal methods, and Sang et al. Cell 2026 has now reported the first peer-reviewed clinical resistance profile for a tri-complex inhibitor. The page reads the public substrate against the next pivotal readouts and names the gaps the literature has not yet closed.
Records carry T1–T4 evidence tiers: T1 mechanistic (RCT, controlled trial, replicated proof), T2 functional (single-arm trial, validated model, biochemistry), T3 association (observational cohort, screen hit), T4 mention (review, preprint, conference abstract). 152 unique records: 65 / 53 / 9 / 25.
Generated 2026-05-21 · Evidence cutoff 2026-05-21 · 152 unique records (209 raw) · 7 databases · 12 sponsor-authored flagged
§1 · The Phase 3 floor in metastatic pancreatic cancer
The Phase 3 record in metastatic pancreatic cancer defines the floor any new agent reads against. Of roughly fourteen pathway-directed randomized Phase 3 studies in metastatic PDAC since 1997, one — erlotinib plus gemcitabine in PA.3 (Moore JCO 2007) — produced a statistically significant overall-survival improvement, and the magnitude was 0.33 months (HR 0.82). The current second-line chemotherapy control mOS is 6.1 months (NAPOLI-1).
The forest plot below shows the published peer-reviewed record. The final row carries the most recent peer-reviewed active-state RAS inhibitor publication for comparison: Wolpin NEJM May 2026, single-arm Phase 1/2, n=26 G12-mutant 2L 300 mg subgroup — not a randomized result.
The plot shows published data only. The Phase 3 RASolute 302 result sits at corporate-disclosure tier (2026-04-13 topline) outside the peer-reviewed substrate; see methodology footnote.
Every subsequent section reads against this floor.
§2 · Active-state mechanism, replicated outside the sponsor
The active-state inhibitor binds cytosolic cyclophilin A (CYPA / PPIA) and the GTP-bound RAS conformation simultaneously, forming a ternary complex with neomorphic contact geometry distinct from the switch-II / Helix-3 pocket that covalent KRAS G12C-OFF inhibitors exploit. Two RVMD-authored anchor papers describe the mechanism: Schulze et al. Science 2023 (PMID 37590355) and Holderfield et al. Nature 2024 (PMID 38589574).
Three non-RVMD academic groups have replicated or extended the mechanism with orthogonal methods. Cuevas-Navarro et al. Nature2025 (PMID 39476862, MSKCC / Lito lab) demonstrated that drug-bound CYPA stimulates GTP hydrolysis in a mutation-specific manner and predicted a hierarchy of allele sensitivities subsequently borne out clinically. Pfaff & Shokat 2026 (PMID 42030031, UCSF) solved a K-Ras·GDP·AlF₃ transition-state structure that resembles endogenous GTPase–GAP complexes, framing the tri-complex as a pharmacologic GAP mimetic rather than a passive effector-binding blocker. Alexander et al. JBC 2025 (PMID 40473215, NCI Frederick) used SPR and X-ray across an allele panel to confirm switch-II pocket conformation dependence and to show why switch-II accessibility kinetically constrains ON-state inhibitors.
The chemistry generalizes. The same scaffold has advanced to elironrasib (G12C-ON; Cregg et al. J Med Chem 2025, PMID 39993169) and zoldonrasib (G12D-ON; Weller et al. Science 2025). Mozzarelli, Cuevas-Navarro et al. Cancer Research 2025 (PMID 40644578, Castel lab) extended cyclophilin-templated chemistry to RIT1, a RAS-family GTPase implicated in lung adenocarcinoma and Noonan syndrome — evidence that the modality is a platform rather than a one-off compound.
At the search date, no competitor pharma program used cyclophilin recruitment chemistry to target RAS or any other GTPase. BridgeBio Oncology Therapeutics' BBO-11818 (PMID 41790032), Boehringer Ingelheim's BI-2865 / BI-2493 (PMID 40709733), and the Mirati / BMS and Hengrui G12D programs all target the switch-II / Helix-3 pocket directly. At the published-literature tier, the technical position is defensible.
One disclosure on lab continuity: the same MSK / Lito lab that confirmed the mechanism via Cuevas-Navarro Nature 2025 has also published the first peer-reviewed clinical resistance profile (Sang et al. Cell 2026, cited in §4). The two papers report different findings from different cohorts. The page treats Lito as one of three independent academic groups confirming the mechanism, and separately as the source of the clinical resistance cohort. Continuity of work, not contradiction.
§3 · The class battle
Comparator class, combinations, and competitor mechanism nodes form one argument, not three. The KRAS G12C-OFF class is the cautionary tale; the combinations record names the only peer-reviewed horizontal extension to date; the cross-read matrix integrates both into a single reading frame for the next pivotal readouts.
The G12C-OFF comparator class: PFS without OS at primary analysis
The KRAS G12C-OFF class reproduces its PFS signal across two pivotal trials and has not converted it to a statistically significant overall-survival advantage. Sotorasib in CodeBreaK 200 (de Langen Lancet 2023, PMID 36764316, n=345) raised median progression-free survival from 4.5 to 5.6 months versus docetaxel in pretreated KRAS G12C NSCLC (HR 0.66, 95% CI 0.51–0.86, p=0.0017). Adagrasib in KRYSTAL-12 (Barlesi Lancet2025, PMID 40783289, n=453) raised median PFS from 3.8 to 5.5 months (HR 0.58, 95% CI 0.45–0.76, p<0.0001). Neither trial produced a statistically significant overall-survival improvement at primary analysis. The published explanations name three structural drivers: a modest PFS delta (1.1–1.7 months) at the limit of OS detection, post-progression cross-over from docetaxel to KRAS-targeted therapy, and access to subsequent therapies on both arms. Divarasib (Sacher NEJM 2023, PMID 37611121) marks the upper bound for the class at ORR 53.4% and mPFS 13.1 months in a Phase 1 dataset.
Combinations: EGFR-in-CRC validated, SHP2 / SOS1 not
At peer-review tier only one combination axis is horizontally validated; the rest sit at preclinical signal, terminated trial, or null clinical result. EGFR co-blockade in colorectal cancer is the only validated case. CodeBreaK 300 (Fakih NEJM 2023, PMID 37870968) reports sotorasib plus panitumumab versus standard-of-care chemotherapy in KRAS G12C mCRC at HR 0.49 for PFS (95% CI 0.30–0.80, p=0.006). KRYSTAL-1 (Yaeger NEJM 2023, PMID 36546659) reports adagrasib plus cetuximab at 46% ORR and 6.9-month mPFS. Amodio Cancer Discovery 2020 (PMID 32430388) explains the mechanism: RTK-feedback intensity in CRC runs an order of magnitude above NSCLC and depends on lineage, and EGFR drives the dominant feedback loop.
SHP2 co-targeting shows clean preclinical combination activity with KRAS G12C-OFF (Weiss Cancer Discovery 2022, PMID 35404998) and constrained clinical results. SHERPA (Kanhailal EJC 2026, PMID 42114409, RMC-4630 plus an ERK inhibitor) terminated at three of four dose levels for tolerability — dose-limiting thrombocytopenia, AKI, diarrhea, LVEF reduction — with no radiologic responses. The recommended Phase 2 dose remained undetermined. Independent SHP2 programs (BBP-398 across four NCTs) are discontinued.
SOS1 and MEK combinations carry published preclinical signals — MRTX0902 plus adagrasib in 8 of 12 PDX models (PMID 38904222); trametinib plus docetaxel in G12C NSCLC at ORR 28% — but no peer-reviewed clinical readouts that exceed monotherapy ORR. MRTX0902 Phase 1 (NCT05578092) terminated; no peer-reviewed efficacy publication exists.
RAS-pathway inhibition with checkpoint blockade rests on one positive peer-reviewed preclinical foundation. Orlen Cancer Discovery 2025 (PMID 40057911) demonstrates T-cell-dependent depth and duration of response in immunocompetent PDAC models that receive daraxonrasib plus immunotherapy. Kemp Cancer Discovery 2023 (PMID 36472553) generalises the principle for MRTX1133 and KRAS G12D. No peer-reviewed clinical efficacy of any active-state RAS inhibitor plus checkpoint blockade has appeared. Begum 2021 (PMID 34590053) reports a sotorasib-specific hepatotoxicity signal when sequenced with prior checkpoint exposure; real-world reports show the signal does not generalize to adagrasib and read as sotorasib-specific idiosyncrasy rather than a class effect.
Cross-read matrix: competitor mechanism nodes against outcome states
Each competitor mechanism node is read against three outcome states (clean hit, mixed, clean miss). The cell text describes what the published mechanism literature already records under each outcome state — not a directional position on outcomes.
KRAS G12C-OFF
State of evidence: Mature peer-reviewed evidence: CodeBreaK 200 PFS+/OS– at primary analysis (PMID 36764316); KRYSTAL-12 PFS+/OS– at primary analysis (PMID 40783289); divarasib ph1 ORR 53.4% (PMID 37611121).
Clean hit
Confirms RAS-pathway addiction in lung G12C tumours and validates the comparator class. Tanaka 2021 (PMID 33824136) reports that RM-018 — a tri-complex chemotype binding outside the switch-II pocket — retained activity in vitro against the Y96D switch-II resistance mutation; the published clinical resistance literature has not yet measured the same question for daraxonrasib.
Mixed
The currently published pattern. The literature names three structural drivers of unconverted OS — modest PFS magnitude, post-progression cross-over, and subsequent-therapy access — and notes that the same drivers act on any pivotal trial in this population.
Clean miss
Falsifies long-term benefit from G12C addiction in lung cancer at the comparator-class tier. The published literature treats G12D and pan-RAS biology in PDAC as mechanistically and clinically distinct, so a G12C miss leaves those theses intact.
KRAS G12D
State of evidence: Sparse peer-reviewed evidence. Park NEJM 2026 (PMID 41879829) reports setidegrasib / ASP3082 ph1 PDAC ORR 24%, mOS 10.3 mo. Zhou Cancer Cell 2024 (PMID 38942026) covers HRS-4642 preclinically; the clinical data extend no further than a one-sentence mention. Hallin Nature Medicine 2022 (PMID 36216931) reports MRTX1133 preclinical activity; the MRTX1133 Phase 1 (NCT05737706) terminated without a peer-reviewed clinical readout.
Clean hit
Establishes clinical tractability of direct G12D engagement in PDAC. The active-state mechanism literature (Schulze 2023, Holderfield 2024, Cuevas-Navarro 2025) describes a single active-state geometry that engages G12D alongside additional alleles; the peer-reviewed clinical literature has not yet measured allele-by-allele ORR for multi-allelic agents.
Mixed
The currently published state at peer-review tier. Setidegrasib’s 24% ORR / 3.0 mo mPFS sits at the lower-middle band of the comparator-class ORR range.
Clean miss
Implies the PDAC tumour microenvironment defeats direct KRAS engagement irrespective of allele or state. Pan-RAS active-state agents retain mechanism rationale in CRC and NSCLC under this outcome; the PDAC thesis at peer-review tier does not.
SHP2
State of evidence: Weiss Cancer Discovery 2022 (PMID 35404998) reports preclinical combination activity with G12C-OFF. SHERPA (RMC-4630 + ERK inhibitor) terminated at three of four dose levels for tolerability with no responses (PMID 42114409). No peer-reviewed positive randomized SHP2 + KRAS combination readout exists. BBP-398 (Navire/BridgeBio) is discontinued across four NCTs.
Clean hit
Validates RTK-feedback interruption as a depth axis. Holderfield 2024 implies redundancy with SHP2, because active-state engagement binds the same wild-type RAS-GTP pool that SHP2 inhibition releases.
Mixed
The currently published reality. The SHERPA tolerability failure shows the therapeutic-index compression that follows from stacking pathway inhibitors at clinical dose.
Clean miss
Falsifies upstream GEF blockade as a viable clinical axis. Direct RAS engagement (OFF or ON) remains the only validated approach under this outcome.
SOS1
State of evidence: MRTX0902 + adagrasib augments activity in 8 of 12 PDX models (PMID 38904222). MRTX0902 clinical Phase 1 (NCT05578092) terminated. No peer-reviewed clinical efficacy data exist.
Clean hit
Confirms upstream GEF interruption as biologically meaningful at clinical tier. The same active-state-RAS redundancy argument applies as for SHP2.
Mixed
A null state at peer-review tier.
Clean miss
Operationally identical to a SHP2 miss. Direct RAS engagement remains the only validated clinical approach under this outcome.
MEK
State of evidence: Extensive peer-reviewed evidence. Trametinib + docetaxel in G12C-NSCLC reports ORR 28% (Clin Cancer Res 2023). Trametinib + gemcitabine in PDAC (NCT01231581) reports HR 0.98 with no signal (PMID 24915778). Three MEK trials in PDAC have read negative overall.
Clean hit
Establishes downstream vertical inhibition as achievable. The peer-reviewed combination literature does not support expectation of ORR amplification beyond monotherapy from MEK addition.
Mixed
The default expectation at peer-review tier, given prior MEK-doublet performance in this population.
Clean miss
Falsifies the "deepen pathway suppression" combination strategy at clinical tier. Horizontal combinations (EGFR, checkpoint) remain the only validated combination class under this outcome.
§4 · The risks: resistance, absences, structural questions
Resistance splits asymmetrically across mechanism classes, and the asymmetry separates one class architecturally from another. The peer-reviewed RAS-pathway literature reports twenty-three resistance mechanisms across the G12C-OFF and active-state RAS classes through 2026-05-21, sorted by their relationship to the tri-complex binding face.
Six routes evade the active-state mechanism because they sit outside the tri-complex binding face — switch-II pocket mutations including Y96D, KRAS amplification, secondary KRAS mutations, NRAS/HRAS bypass, wild-type RAS reactivation, and RAS-dependent RTK feedback. Nine routes operate downstream of or independent of RAS engagement and apply to any RAS-pathway agent — kinase-dead BRAF, MEK1 activating mutations, MET amplification, PTEN loss, oncogenic fusions, YAP/TAZ lineage switching, KEAP1/STK11/TP53 co-mutations, histologic transformation, and drug-tolerant persister phenotype. Five routes are novel to the active-state class, anchored by the first peer-reviewed clinical resistance profile (Sang et al. Cell 2026, PMID 42092352, n=40 daraxonrasib patients): RAS Y64 binding-interface mutations (~5% of cohort, clinical), kinase-dead BRAF newly observed as a tri-complex escape (~10%, clinical), CYPA/PPIA disruption and RAS Y71 (preclinical only at search date), and MYC amplification with differential effect between RAS-ON and G12C-OFF (preclinical). Two partial routes (RTK adaptive feedback, NF1 loss) act lineage-dependently. One axis (CNS penetration) remains unknown at the published-literature tier.
▸ Expand 23-row resistance matrix▾ Collapse resistance matrix
- KRAS switch-II pocket mutations (Y96D/C, H95, R68)Tanaka CD 2021 · PMID 33824136 [T1]
- KRAS G12C allele amplificationAwad NEJM 2021 · PMID 34161704 [T1]
- Secondary KRAS mutations (G12D/V/R, Q61H)Awad 2021 · PMID 34161704 [T1]
- NRAS / HRAS / MRAS bypass mutationsAwad 2021, Zhao 2021 · PMID 34161704 / 34759319 [T1]
- Wild-type RAS amplification / reactivationRyan CCR 2020 · PMID 31776128 [T2]
- RAS-dependent RTK feedback (epithelial ERBB2/3)Xue Nature 2020 · PMID 31915379 [T1]
- BRAF kinase-dead / class-III mutationsSang et al. Cell 2026 · PMID 42092352 [T1]
- MAP2K1 (MEK1) activating mutationsAwad 2021 · PMID 34161704 [T1]
- MET amplificationAwad 2021 · PMID 34161704 [T1]
- PTEN lossAwad 2021 · PMID 34161704 [T1]
- Oncogenic fusions (ALK, RET, RAF1, FGFR3)Awad 2021 · PMID 34161704 [T1]
- YAP/TAZ activation, lineage switchingKapoor Cell 2014 · PMID 24954535 [T2]
- KEAP1 / STK11 / TP53 co-mutationsArbour CCR 2018 · PMID 29089357 [T3]
- Histologic transformation (squamous, SCLC)Awad 2021 · PMID 34161704 [T1]
- RTK adaptive feedback (mesenchymal FGFR1 / AXL)Solanki CCR 2021 · PMID 33619172 [T2]
- NF1 loss (WT-RAS activation arm)Awad 2021 · PMID 34161704 [T1]
- CYPA / PPIA mutation or lossFoth preprint 2025 [T4]
- RAS Y64 binding-interface mutationSang et al. Cell 2026 · PMID 42092352 [T1]
- Kinase-dead BRAF (clinically observed in Sang et al. Cell 2026)Sang et al. Cell 2026 · PMID 42092352 [T1]
- RAS Y71 (RAS-RAF affinity enhancement; preclinical only)Sang et al. Cell 2026 · PMID 42092352 [T1, supplement]
- MYC amplification (RAS-ON > G12C-OFF)Wasko Nature 2024 [T2]
- Daraxonrasib CNS penetration / sanctuaryabsence of peer-reviewed PK [—]
- Drug-tolerant persister phenotypeSolanki 2026 · PMID 41165456 [T2]
Classification reflects structural and clinical observations published through 2026-05-21. Sang et al. Cell 2026 reports the first peer-reviewed clinical resistance profile for any tri-complex inhibitor — n=40 daraxonrasib patients, 18 with recurrent acquired alterations — and anchors the novel column.
Three structural questions the literature flags
Cyclophilin A immunosuppression: no peer-reviewed assessment exists.
The active-state RAS literature does not measure whether tri-complex CYPA sequestration impairs the calcineurin / NFAT axis. The cyclosporin A precedent supplies a mechanistic analogy, not data; the literature reports sanglifehrin-derived chemistries as non-immunosuppressive but offers no peer-reviewed clinical safety study for the tri-complex assets. The literature cannot close this gap; only a sponsor disclosure or a peer-reviewed safety study can.
OS conversion is the structural feature the comparator class has yet to settle.
CodeBreaK 200 and KRYSTAL-12 did not resolve the PFS-without-OS pattern at primary analysis. The next published data points on this question are the OS follow-up publications and the ESMO 2026 OS readouts for CodeBreaK 200, KRYSTAL-12, and JDQ443.
Tri-complex-specific clinical resistance has been observed in the first reported cohort.
Sang et al. Cell 2026 identifies recurrent acquired alterations in 18 of 40 daraxonrasib-treated patients (mixed tumor types — 17 NSCLC, 15 CRC, 4 melanoma, 4 other) — chiefly RAS Y64 binding-interface mutations and kinase-dead BRAF. RAS Y71 and CYPA/PPIA alterations appear preclinically only (Y71 in 2 of 4 cell lines; PPIA in 0 of 6 WES-tested patients, the panel does not cover PPIA in clinical NGS routinely). Resistance-detection rates match the G12C-OFF comparator class at 45%, with daraxonrasib PFS modestly longer (8.5 mo 2L PDAC vs 6.5–6.8 mo NSCLC for sotorasib / adagrasib), so the durability question reads "novel routes exist, comparable rate, slightly longer time-to-progression" — not "tri-complex resistance arrives faster."
What the literature does not yet establish
The mechanism is replicated outside the sponsor, but the cyclophilin safety axis remains unmeasured at clinical tier — these are not the same finding, and the page does not let them blur. The twelve items below mark load-bearing absences: gaps the published literature has not closed at search date, each bounding the inferences the prior sections support.
- Cyclophilin A immunosuppression / NFAT-axis clinical safety
The active-state RAS inhibitor literature does not measure whether tri-complex CYPA sequestration impairs the calcineurin / NFAT axis. Only a sponsor disclosure or a peer-reviewed clinical safety study can close this gap; the literature cannot.
- Daraxonrasib RASolute 302 full Phase 3 published readout
Revolution Medicines disclosed the topline at corporate-disclosure tier on 2026-04-13 (outside the peer-reviewed substrate; see methodology). The full dataset reads out at the ASCO Plenary 2026-05-31. Peer-reviewed publication is anticipated 2026–2027.
- RMC-9805 (zoldonrasib) peer-reviewed publication
No peer-reviewed publication exists. The data sit at conference-disclosure tier (AACR 2026 Plenary, 2026-04-19); see conference-frontier panel.
- ERAS-4001 (Erasca pan-KRAS)
Zero peer-reviewed publications.
- BI-2865 / BI-2493 clinical status
The preclinical literature is substantial (PMID 37258666, Nature 2023). The published literature stays silent on clinical disposition.
- RMC-6236 + checkpoint human efficacy
No peer-reviewed clinical publication exists. Orlen Cancer Discovery 2025 (PMID 40057911) establishes the preclinical foundation; clinical efficacy sits at corporate-disclosure or absence-of-evidence tier.
- KRYSTAL-2 (TNO155 + adagrasib)
No peer-reviewed primary publication identified.
- KRYSTAL-7 (adagrasib + pembrolizumab)
No peer-reviewed primary publication. Only conference data plus hepatotoxicity case reports exist.
- Olomorasib (Eli Lilly LY3537982) peer-reviewed clinical efficacy
Phase 3 in 1L NSCLC + pembrolizumab is active per ClinicalTrials.gov; no peer-reviewed Phase 1/2 efficacy publication appeared at search date.
- BMS-986504 (Bristol Myers Squibb pan-RAS)
Phase 2 (NCT07492680) is active; no peer-reviewed publication identified.
- Daraxonrasib CNS pharmacokinetics
No published comparative data exist.
- CYPA expression heterogeneity in responder vs non-responder biopsies
No peer-reviewed biopsy biomarker study identified.
§5 · The catalyst window
The substrate stops at 2026-05-21. The catalysts below land after that date and are the next data points the substrate reads against. The page takes no position on outcomes; it names the dates, the venues, the priors the literature already records, and the structural questions a buy-side or development-strategy reader is likely to ask.
Catalyst calendar
- 2026-05-31 (post search date)
RASolute 302 full Phase 3 dataset
ASCO 2026 Plenary Session LBA5, Hall B1
Revolution Medicines announced the topline as corporate disclosure 2026-04-13; see methodology footnote. The full dataset (subgroups, safety, QoL) lands at conference / not-yet-peer-reviewed tier on presentation. Peer-reviewed publication is anticipated 2026–2027.
- 2026-06-01 (post search date)
KRAS G12D allele-selective triplet (oral session)
ASCO 2026 — Developmental Therapeutics oral, abstracts 3006, 3007, 3008
Three competing G12D inhibitors share one session. Cross-reads to the G12D node of the class-battle matrix in §3.
- 2026-05-30 (post search date)
TSN1611 (oral KRAS G12D) Phase 1/2 NSCLC
ASCO 2026 — Lung Cancer NSCLC Metastatic rapid oral, abstract 8516
The first KRAS G12D inhibitor to occupy a dedicated NSCLC oral session.
- 2026-06-01 (post search date)
Atebimetinib (MEK) + gem/nab-paclitaxel in 1L mPDAC
ASCO 2026 — GI Pancreatic rapid oral, abstract 4013
Calibrates the MEK + chemotherapy backbone in 1L mPDAC. Cross-reads to the MEK node of the class-battle matrix in §3.
- 2026-Q3 (post search date)
KRYSTAL-10 Phase 3 (adagrasib + cetuximab, G12C CRC)
Primary completion 2026-07-31 per ClinicalTrials.gov; venue TBD
The first randomized Phase 3 of a G12C + EGFR combination.
- 2026-October (post search date)
ESMO 2026 KRAS G12C OS readouts
CodeBreaK 200 OS follow-up, KRYSTAL-12, JDQ443 Phase 3 — ESMO 2026
Three Phase 3 OS reads on the KRAS G12C-OFF class in NSCLC. Read against the published PFS-without-OS pattern of the comparator class in §3.
Conference-frontier data already landed, not peer-reviewed yet (T4)
The numbers below come from conference disclosures, not peer-reviewed publications. Each row carries venue, date, and confidence interval where reported.
Zoldonrasib (RMC-9805, KRAS G12D active-state inhibitor)
Phase 1, previously-treated KRAS G12D-mutant NSCLC, post-IO and post-platinum
ORR 52% (95% CI 32–71); mPFS 11.1 mo; 12-month PFS 48%; 12-month OS 73%· n = 27
AACR 2026 Plenary · 2026-04-19 · GlobeNewswire press release 2026-04-19 (RVMD IR)
Pre-peer-review. The cohort is small; report the confidence interval whenever the ORR appears.
Setidegrasib (ASP3082, KRAS G12D protein degrader)
Phase 1, KRAS G12D-mutant NSCLC (separately from the peer-reviewed Park NEJM 2026 publication, which covers PDAC and NSCLC G12D)
ORR 35.8%
ELCC 2026 · 2026-03-25 · Conference press release; Park NEJM 2026 (PMID 41879829) is the peer-reviewed counterpart for the broader Phase 1
The Park NEJM 2026 publication supersedes the NSCLC G12D subgroup details.
ASCO 2026 Developmental Therapeutics oral session, KRAS G12D
Three competing G12D allele-selective inhibitors in a single oral session
Abstract bodies embargoed; data tables release 2026-05-22 17:00 ET· n = per-abstract; not yet public
ASCO 2026 (abstracts 3006, 3007, 3008) · 2026-06-01 (post search date) · ASCO program planner via Oncoletter; RVMD IR press release 2026-04-21
The substrate captures the catalyst slot only; numeric content remains non-public at search date.
Topline-to-Plenary precedent
A reader at the next pivotal readout will ask: when a sponsor has disclosed a positive pivotal-trial topline preceding an ASCO Plenary presentation, has the Plenary dataset reproduced the topline? The three comparators below are the cleanest recent precedents the published record offers. The substrate names what the precedent records — not what the next Plenary will produce.
- ADAURA (osimertinib, adjuvant EGFR-mutated NSCLC)
AstraZeneca
Topline 2020-04-10: Independent Data Monitoring Committee recommended early unblinding for overwhelming efficacy in DFS.
Plenary 2020-05-31: Full Plenary dataset reproduced the topline: HR 0.17 (95% CI 0.12–0.23, p<0.0001) for DFS in stage II–IIIA EGFR-mutated NSCLC. (ASCO 2020 Plenary Session, Abstract LBA5)
Peer review · NEJM 2020 (PMID 32955177); long-term OS NEJM 2023 (PMID 37272535).
- DESTINY-Breast04 (trastuzumab deruxtecan, HER2-low mBC)
Daiichi Sankyo / AstraZeneca
Topline 2022-02-21: Topline disclosed statistically significant and clinically meaningful PFS and OS benefit versus chemotherapy.
Plenary 2022-06-05: Full Plenary dataset reproduced the topline: 50% reduction in risk of progression or death; standing ovation. (ASCO 2022 Plenary Session, Abstract LBA3)
Peer review · NEJM 2022 (PMID 35665782); long-term follow-up Nature Medicine 2025.
- TALAPRO-3 (talazoparib + enzalutamide, HRR-mutated mCSPC)
Pfizer
Topline 2026-03-29: Topline reported a statistically significant rPFS improvement exceeding the pre-specified target HR of 0.63.
Plenary 2026-06-02: Late-breaking presentation reproduced the topline rPFS magnitude with consistent subgroup effects and a strong OS trend. (ASCO 2026 — Late-Breaking Oral Abstract Session)
Peer review · Peer-reviewed publication pending at search date.
The three comparators reproduced their toplines at Plenary presentation. The published record offers no clean recent counter-example of a positive pivotal topline that diminished materially at full Plenary dataset; the precedent is small and selection-biased toward sponsor disclosure, and a fourth-case search at the search date did not surface an independently verified clean counter-example.
Competitive timing
The competitive window the substrate names: can a multi-allelic active-state agent enroll, manufacture, and read out its next set of pivotal trials before the G12D-selective field crowds in?
- G12D field convergence
Three competing G12D allele-selective inhibitors share one ASCO 2026 Developmental Therapeutics oral session (abstracts 3006, 3007, 3008). TSN1611 (oral KRAS G12D) has its first dedicated NSCLC oral session at ASCO 2026 (abstract 8516). Park NEJM 2026 (PMID 41879829) sets the published G12D bar at ORR 24% / mOS 10.3 mo for setidegrasib in PDAC.
- Sponsor cash and runway
Per Q1 2026 10-Q (Appendix): $1,908M cash + ST securities at 2026-03-31; trailing-twelve-month R&D expense $1,036M; implied runway ~5.5 quarters at Q1 2026 burn before the Royalty Pharma optionality tranches.
- Breakthrough Therapy Designations granted
Three FDA Breakthrough Therapy Designations (Appendix): Daraxonrasib 2L+ mPDAC G12 (2025-06-23); Elironrasib G12C NSCLC (2025-07-23); Zoldonrasib G12D NSCLC (2026-01-08) — the first BTD for any KRAS G12D asset in NSCLC.
The substrate stops at structural question. Whether the competitive window opens or closes is observable only as later readouts land; the page does not take a position.
§6 · How this was made
The methodology is the product. Every claim on this page is two clicks from a primary source; the page documents the six audit gates that produced the substrate.
- Gate 1 · DiscoveryForward and backward citation expansion via EuropePMC on eight anchor papers (Schulze 2023, Holderfield 2024, Awad 2021, Tanaka 2021, CodeBreaK 100, KRYSTAL-1, NAPOLI-1, PA.3). Seven databases queried — PubMed, EuropePMC, OpenAlex, Crossref, Semantic Scholar, BioRxiv, MedRxiv — through the ToolUniverse MCP layer. 209 raw records deduplicated to 152.
- Gate 2 · PMID primary-source verificationEvery PMID independently re-checked against EuropePMC by the pre-build verification agent: 27 of 27 pre-build verification anchors confirmed clean. A subsequent audit (2026-05-22) of the full ≥57 unique PMIDs rendered on the page found 3 citation errors (ADAURA long-term OS, JANUS-1/2, CCTG PA.7 first author), all corrected in commit shipping this audit. Full-text and supplements consulted per the Tier 0 full-text policy when peer-review-tier interpretation depended on it (resistance categories, allele-specific numerics, supplement-only data points such as Sang et al. Cell 2026 RMC-4791 and RAS Y71).
- Gate 3 · Editorial chainWilliams-pass editorial review for voice, claim-against-data alignment, and redundancy. Williamson-pass editorial review for skeptical numeric audit and tier-boundary discipline. Williams pass 2 audit confirmed source-of-truth coverage after content revisions.
- Gate 4 · Boundary disciplinePeer-reviewed evidence, T4 conference disclosures, and corporate-disclosure events occupy distinct tiers and never blur. Each conference row carries venue, date, sponsor, n, and confidence interval. Corporate-disclosure events (the 2026-04-13 RASolute 302 topline) appear only in the boundary footnote below and the catalyst calendar in §5 — never as a body number.
- Gate 5 · Post-build re-verification14 / 14 load-bearing claims re-verified against the rendered HTML after build, to confirm no numeric drift between source data and rendered prose.
- Gate 6 · Citation correctionsFive citation corrections logged below: journal of record, primary-publication vs OS update, related-trial disambiguation. The corrections survive because the audit gates surface them; future readers can verify any one of them in one click.
Search date · 2026-05-21. Evidence cutoff matches search date.
Databases queried
- PubMedPubMed_search_articles10 req/s (NCBI_API_KEY)
- EuropePMCEuropePMC_search_articlespublic
- OpenAlexopenalex_literature_searchpublic
- CrossrefCrossref_search_workspublic
- Semantic ScholarSemanticScholar_search_papersAPI key
- BioRxivBioRxiv_get_preprint / list_recent_preprintspublic
- MedRxivMedRxiv_get_preprintpublic
Expansion depth · Forward citations via EuropePMC_get_citations and backward via EuropePMC_get_references on anchor papers (Schulze 2023, Holderfield 2024, Awad 2021, Tanaka 2021, CodeBreaK 100, KRYSTAL-1, NAPOLI-1, PA.3). Depth-2 expansion not performed.
Exclusions
- · Trial scheduling — out of scope
- · NCT-by-NCT enumeration — out of scope
- · USPTO continuity / patent-family analysis — out of scope
- · ASCO abstract slotting — out of scope
- · Sponsor financial disclosures (cash, runway, royalty deals, valuation) — out of scope; not literature
Corporate-disclosure boundary
The page reads the peer-reviewed and preprint literature through 2026-05-21. Corporate disclosures filed since the search date — notably the RASolute 302 Phase 3 topline announced 2026-04-13 via 8-K and GlobeNewswire — sit outside that scope and appear only in this methodology footnote. The ASCO 2026 Plenary presents the full dataset 2026-05-31; peer-reviewed publication is anticipated 2026–2027 and will be incorporated then. Conference disclosures (AACR 2026, ELCC 2026, ASCO 2026 regular abstracts) appear separately as T4 conference-frontier rows, each tagged with venue, date, and confidence interval.
Citation corrections
- · Schulze 2023 appears in Science (PMID 37590355), not Nature as in some secondary references.
- · Fakih CodeBreaK 100 CRC appears in Lancet Oncology (PMID 34919824), not NEJM. The NEJM Fakih 2023 publication is CodeBreaK 300 (PMID 37870968) — a distinct trial.
- · The CodeBreaK 200 OS evidence cited here draws on the 2023 Lancet primary publication (PMID 36764316, primary endpoint PFS) and subsequent OS follow-up reporting. Statements about OS at primary analysis refer to the published primary analysis, not later updates.
- · MRTX1133: Hallin Nature Medicine 2022 (PMID 36216931) publishes preclinical activity; the clinical Phase 1 (NCT05737706) terminated without a peer-reviewed efficacy publication.
- · Daraxonrasib peer-reviewed efficacy figures (35% ORR, 8.5 mo mPFS, 13.1 mo mOS) refer to the 26-patient G12-mutant 300 mg 2L subgroup of Wolpin NEJM 2026 (PMID 42090791). The broader pooled G12/G13/Q61 group reports distinct numbers (ORR 29%, mOS 15.6 mo, n=38) — never blurred on this page.
Appendix · Sponsor context — primary-source filings
The figures below come from SEC EDGAR (10-Q, 8-K) and FDA Breakthrough Therapy Designation press releases — not peer-reviewed literature. They appear here as orientation; the page draws no inference from them.
Financials · as of 2026-05-21
- Cash and short-term marketable securities
- $1,908MCash $440.9M + short-term marketable securities $1,467.1M as of 2026-03-31
- Q1 2026 R&D expense
- $343.97MUp 67% year over year from $205.75M in Q1 2025
- Q1 2026 G&A expense
- $101.3MUp from $35.0M Q1 2025; consistent with pre-launch infrastructure investment
- Trailing-twelve-month R&D expense
- $1,036M
- Implied runway at Q1 2026 burn
- ~5.5 quarters (16–17 months)On current cash + ST securities, before Royalty Pharma tranches
$2.0B tiered Royalty Purchase Agreement (Royalty Pharma, executed 2025-06-23)
Initial $250M synthetic-royalty Tranche 1 already drawn. Four additional optionality tranches gated on clinical and commercial milestones; total undrawn optionality up to $1.25B. Tranche 2 ($250M) unlocks on positive RASolute 302 readout meeting agreed endpoint plus NDA election; Tranche 3 ($250M) gates on FDA approval of RMC-6236 for 2L mPDAC prior to 2028-07-01. Royalty rates: 4.55% of net sales up to $2B base tier (escalating to 7.80% if all five tranches drawn) stepping down to 2.50% / 1.00% / 0.0% at higher revenue tiers.
Source · Q1 2026 10-Q Note 21; Royalty Pharma 8-K 2025-06-23
FDA designations
- Daraxonrasib (RMC-6236)
Breakthrough Therapy Designation · 2L+ metastatic PDAC with KRAS G12 mutations
Granted 2025-06-23 · GlobeNewswire / RVMD IR 2025-06-23
- Elironrasib (RMC-6291)
Breakthrough Therapy Designation · KRAS G12C NSCLC, post-chemo + IO, no prior KRAS G12C inhibitor
Granted 2025-07-23 · GlobeNewswire / RVMD IR 2025-07-23
- Zoldonrasib (RMC-9805)
Breakthrough Therapy Designation — first BTD for any KRAS G12D asset in NSCLC · KRAS G12D NSCLC, post anti-PD-(L)1 therapy
Granted 2026-01-08 · GlobeNewswire / RVMD IR 2026-01-08
RAS(ON) pipeline assets — per Q1 2026 10-Q named pipeline
- RMC-6236
Daraxonrasib· Pan-RAS(ON) multi-selective tri-complex
Phase 3 · 2L+ mPDAC (registration); 1L mPDAC; 2L NSCLC RAS G12X excl G12C; adjuvant PDAC
- RMC-6291
Elironrasib· KRAS G12C(ON)-selective tri-complex
Phase 1/2 · NSCLC G12C (post-IO and post-chemo); CRC G12C; PDAC G12C
- RMC-9805
Zoldonrasib· KRAS G12D(ON)-selective tri-complex
Phase 1/2 · NSCLC G12D (post-anti-PD-(L)1); PDAC G12D; CRC G12D
- RMC-5127
Not yet assigned· KRAS G12V(ON)-selective tri-complex
Preclinical / IND-pending · Solid tumours (KRAS G12V)
Primary-source disclosures from Revolution Medicines Q1 2026 10-Q (accession 0001193125-26-208969, filed 2026-05-06), Q1 2026 8-K (accession 0001193125-26-152039, filed 2026-04-13), Royalty Pharma 8-K (2025-06-23), and FDA Breakthrough Therapy Designation press releases. Not peer-reviewed literature; surfaced as orientation context.
Appendix · Key citations
The list below carries the twenty-four most load-bearing references in the substrate. The source repository preserves the full deduplicated bibliography of 152 unique records.
- 01Schulze et al. Science 2023[T1]Tri-complex remodelling of cyclophilin A to target active-state KRAS. RVMD-authored foundational paper.PMID 37590355
- 02Holderfield et al. Nature 2024[T1]RMC-7977 / pan-RAS active-state inhibitor; preclinical underwriter of daraxonrasib clinical activity. RVMD-authored.PMID 38589574
- 03Cuevas-Navarro et al. Nature 2025[T1]MSKCC/Lito lab independent confirmation — drug-bound CYPA stimulates GTP hydrolysis.PMID 39476862
- 04Pfaff & Shokat Biol Chem 2026[T1]UCSF independent structural confirmation — TCI mimics GTPase-GAP transition state.PMID 42030031
- 05Alexander et al. JBC 2025[T1]NCI Frederick — SPR/X-ray across allele panel confirms switch-II pocket conformation dependence.PMID 40473215
- 06Tanaka et al. Cancer Discovery 2021[T1]First clinical Y96D switch-II resistance to G12C-OFF; RM-018 retained activity.PMID 33824136
- 07Skoulidis et al. NEJM 2021[T1]CodeBreaK 100 — sotorasib NSCLC Phase 2 (ORR 37.1%, mOS 12.5 mo).PMID 34096690
- 08de Langen et al. Lancet 2023[T1]CodeBreaK 200 — sotorasib Phase 3 vs docetaxel (HR 0.66 PFS at primary analysis).PMID 36764316
- 09Jänne et al. NEJM 2022[T1]KRYSTAL-1 — adagrasib NSCLC Phase 2.PMID 35658005
- 10Barlesi et al. Lancet 2025[T1]KRYSTAL-12 — adagrasib Phase 3 (HR 0.58 PFS).PMID 40783289
- 11Sacher et al. NEJM 2023[T1]Divarasib Phase 1 — NSCLC ORR 53.4%, mPFS 13.1 mo.PMID 37611121
- 12Fakih et al. NEJM 2023[T1]CodeBreaK 300 — sotorasib + panitumumab in mCRC Phase 3 (HR 0.49 PFS).PMID 37870968
- 13Yaeger et al. NEJM 2023[T1]KRYSTAL-1 mCRC — adagrasib + cetuximab ORR 46%.PMID 36546659
- 14Awad et al. NEJM 2021[T1]Clinical resistance landscape of adagrasib — 17 of 38 patients with mechanisms identified.PMID 34161704
- 15Sang et al. Cell 2026[T1]First clinical resistance profile for daraxonrasib (n=40, mixed tumor types: 17 NSCLC + 15 CRC + 4 melanoma + 4 other). Clinically: RAS Y64 binding-interface and kinase-dead BRAF. Preclinically only: RAS Y71, CYPA/PPIA. Supplement names RMC-4791 next-gen TCI. MSK/Lito-led, RVMD co-authored.PMID 42092352
- 16Moore et al. JCO 2007[T1]PA.3 — erlotinib + gemcitabine in mPDAC; only positive Phase 3 to date (ΔmOS 0.33 mo).PMID 17452677
- 17Conroy et al. NEJM 2011[T1]PRODIGE 4 — FOLFIRINOX in mPDAC (mOS 11.1 mo).PMID 21561347
- 18Von Hoff et al. NEJM 2013[T1]MPACT — gemcitabine + nab-paclitaxel (mOS 8.5 mo).PMID 24131140
- 19Wang-Gillam et al. Lancet 2016[T1]NAPOLI-1 — NAL-IRI + 5-FU/LV 2L (mOS 6.1 mo); modern second-line standard.PMID 26615328
- 20Wainberg et al. Lancet 2023[T1]NAPOLI-3 — NALIRIFOX 1L (mOS 11.1 mo).PMID 37708904
- 21Wolpin et al. NEJM 2026[T2]Daraxonrasib Phase 1/2 in RAS-mutant mPDAC — 13.1 mo mOS in 26-patient G12 300 mg 2L subgroup (single-arm). RVMD-sponsored.PMID 42090791
- 22Park et al. NEJM 2026[T1]ASP3082 / setidegrasib G12D Phase 1 — PDAC ORR 24%, mOS 10.3 mo.PMID 41879829
- 23Orlen et al. Cancer Discovery 2025[T2]RAS(ON) + checkpoint combination, preclinical — T-cell-dependent depth and duration of response in immunocompetent PDAC. RVMD co-authored.PMID 40057911
- 24Kanhailal et al. EJC 2026[T2]SHERPA — RMC-4630 + ERK inhibitor; combination terminated at three of four dose levels for tolerability with no responses.PMID 42114409
Bibliography source · ~/code/dd-bigbio/reports/revolution_medicines_2026-05-21/bibliography/bibliography_master.json