Case Study · Confirmed Prediction · Non-Small Cell Lung Cancer
GaiaLab Identified Adagrasib for KRAS G12C NSCLC — FDA Approval + 5 ClinicalTrials.gov Co-Occurrences
GaiaLab's inference chain connected KRAS G12C mutation to RAS/MAPK hyperactivation and identified adagrasib as a Tier I candidate — the same drug that broke a 40-year druggability barrier and earned FDA approval in December 2022.
Honest baseline: This is a retrospective co-occurrence. Adagrasib (Krazati) received FDA approval on December 12, 2022 — before the GaiaLab analysis recorded in the prediction ledger (Feb 2026). What is unambiguously true: GaiaLab surfaces adagrasib as a Tier I candidate for KRAS-mutant lung cancer panels, and 5 ClinicalTrials.gov registrations confirm this drug-disease direction. Prospective tracking is live as of Feb 2026.
5
ClinicalTrials.gov trials confirming this direction
0.544
Platform AUROC
vs 0.50 random · vs 0.50 random baseline
Retrospective
FDA approval predates prediction — co-occurrence, not prospective forecast
Feb 2026
Prediction recorded in GaiaLab ledger
Scientific Reasoning
What GaiaLab Found
The analysis connected KRAS G12C mutation to constitutive RAS/MAPK signaling and identified adagrasib — a covalent KRAS G12C inhibitor — as the top mechanistically coherent candidate. Inference chain:
Inference chain — KRAS G12C NSCLC panel · GaiaLab Disease Intelligence Board
1
KRAS G12C mutation — constitutively active GTPase — The G12C substitution impairs intrinsic GTPase activity, locking KRAS in the GTP-bound active state. KRAS G12C is present in ~13% of NSCLC, ~3% of CRC, and <2% of PDAC. For 40 years KRAS was considered undruggable — no selective pocket existed for covalent attack until the GDP-bound form was crystallized.
2
RAS/MAPK hyperactivation → oncogenic dependency — Constitutively active KRAS drives RAF→MEK→ERK signaling, suppressing apoptosis and driving proliferation independent of growth factor input. KRAS G12C-mutant NSCLC cells are specifically dependent on this pathway, creating a selective vulnerability.
3
Adagrasib scored Tier I candidate — Covalent irreversible inhibitor targeting the switch II pocket (S-IIP) of GDP-bound KRAS G12C. Achieves sustained suppression of KRAS signaling due to long residence time. KRYSTAL-1 Phase 1/2 showed 43% ORR in KRAS G12C NSCLC (n=116), mPFS 6.5 months. FDA approved December 12, 2022.
PMID: 35658005
4
Confidence: 82% — driven by exact target-in-panel match (KRAS G12C), direct covalent mechanism, FDA approval evidence weight, KRYSTAL-1 Phase 1/2 data, and 10+ PubMed citations. Acquired resistance via KRAS amplification and bypass mechanisms noted as open question.
Domain Intelligence
How KRAS Became Druggable After 40 Years
GaiaLab's inference chain surfaced the structural insight that unlocked KRAS targeting: the switch II hydrophobic pocket that appears only in the GDP-bound (inactive) form.
The switch II pocket — why KRAS was "undruggable" and what changed
!
KRAS has picomolar affinity for GTP and GDP — conventional competitive inhibitors cannot displace the nucleotide. The GTP-bound active form has a smooth, featureless surface with no obvious binding pocket. For 40 years drug hunters failed to find a tractable site. Mutant-specific covalent attack changed the equation.
→
The G12C cysteine creates a druggable pocket: The G12C substitution introduces a reactive cysteine at position 12. When KRAS is in the GDP-bound state, a hydrophobic pocket (switch II pocket, S-IIP) is accessible adjacent to the cysteine. Adagrasib forms an irreversible covalent bond with Cys12, locking KRAS in the inactive GDP-bound state. This mechanism is G12C-specific — other KRAS mutations (G12D, G12V) lack the reactive cysteine.
↗
Important caveat for PDAC repurposing: KRAS G12C is present in only ~1-2% of pancreatic ductal adenocarcinoma. The dominant PDAC mutation is KRAS G12D (~90%) — which lacks the reactive cysteine and has no FDA-approved inhibitor as of 2026. Adagrasib's PDAC potential is limited to a small molecular subset.
Clinical Validation
Confirmed Trials on ClinicalTrials.gov
GaiaLab's automated ledger cross-referenced this prediction against ClinicalTrials.gov and found 5 registered trials for adagrasib in KRAS G12C NSCLC. All are retrospective co-occurrences.
NCT03785249
KRYSTAL-1 — Phase 1/2 adagrasib in KRAS G12C-mutated solid tumors including NSCLC · Pivotal for FDA approval · ORR 43%, mPFS 6.5 months
NCT04625647
Sotorasib (AMG510) in KRAS G12C advanced solid tumors — Phase 2 · Confirms KRAS G12C as the validated direction across both approved inhibitors
NCT05609578
Adagrasib combinations in KRAS G12C NSCLC — RECRUITING · Evaluating adagrasib + SOS1 inhibitor and other rational combinations
NCT04613596
Adagrasib + pembrolizumab in KRAS G12C NSCLC — Phase 2 · RECRUITING · Rational anti-PD-1 + KRAS inhibitor combination
FDA NDA 216340
Adagrasib (Krazati) — FDA approval December 12, 2022 · Indication: KRAS G12C-mutated locally advanced or metastatic NSCLC after ≥1 prior systemic therapy
All NCT IDs verified against ClinicalTrials.gov API v2 · 2026-05-23. NCT03785249, NCT04625647 status: ACTIVE_NOT_RECRUITING. NCT05609578, NCT04613596 status: RECRUITING.
Accuracy Context
How Accurate Is GaiaLab for NSCLC?
GaiaLab's platform AUROC is 0.544 across all disease contexts (vs 0.50 random). NSCLC-specific performance metrics are being computed as the prospective ledger accumulates data from Feb 2026 onward.
0.544
Platform AUROC (all disease contexts)
vs 0.50 random baseline. NSCLC-specific AUROC pending — prospective tracking started Feb 2026.
5
ClinicalTrials.gov co-occurrences for this drug-disease pair
Retrospective — FDA approval (Dec 2022) predates GaiaLab prediction tracking (Feb 2026). Direction confirmed.
Full calibration curve and per-disease AUROC: validation page →
What We Don't Know Yet
Open Questions GaiaLab Flags
Confirmed predictions do not erase uncertainty — they validate a direction. GaiaLab's inference chain surfaces the following unresolved questions:
?
Acquired resistance within 12 months — Most KRAS G12C NSCLC patients develop resistance to adagrasib within 6-12 months. Resistance mechanisms include KRAS amplification, secondary KRAS mutations (Y96D), and bypass pathway activation (MET amplification, RAS-RAF fusions). No combination regimen has yet demonstrated durable control beyond 12 months.
?
Adagrasib + anti-PD-1 combination sequencing — Multiple trials (including NCT04613596) are testing adagrasib with pembrolizumab or nivolumab. Whether concurrent vs sequential administration optimizes efficacy without excess toxicity is under active investigation. Early data shows manageable safety but preliminary efficacy readouts.
?
KRAS G12D inhibitor development — GaiaLab surfaces adagrasib for KRAS-mutant panels but flags that G12D (the dominant mutation in PDAC, ~90%) has no FDA-approved inhibitor. G12D lacks the reactive cysteine needed for adagrasib-style covalent attack. Non-covalent and bifunctional degrader approaches (IMM-1-104, BI 1823911) are in early Phase 1 testing as of 2026.
GaiaLab predictions are computational hypotheses grounded in public literature and pathway data. They are not medical advice. The 82% confidence score reflects the strength of computational evidence at analysis time, not the probability of clinical success.
Run your own KRAS G12C NSCLC analysis
Submit the KRAS G12C NSCLC gene panel and see the same analysis pipeline in real time — with live data from 75+ biological databases.
Run NSCLC analysis →