Judgements

Content in the run’s language (EN)

240 judgements

Page 1 of 5Next

Pair	Criterion	Decision	Judge	Confidence	Rationale	Missing evidence	Evidence
EGFR — ALK	Perturbational evidence	Abstained	gpt-5.6-luna	95%	The dossiers contain perturbational evidence for both targets, but neither provides sufficiently explicit, directly adjudicable evidence of target modulation reversing disease-relevant phenotypes in a Patient-Derived Xenograft. EGFR evidence mentions PDX use without a specific, interpretable perturbation result (C1-CAT-ONCO-BIO-010), while ALK evidence describes siRNA effects in ALK-positive cell models and patient-derived material (C2-EV_PERTURBATION_0001) and CRISPR resistance studies with in vivo validation (C2-EV_PERTURBATION_0002), but does not clearly establish a PDX efficacy experiment. Because absence of clearly specified PDX evidence cannot be used to rank either target lower, the appropriate verdict is ABSTAIN.	For EGFR: a clearly identified patient-derived xenograft study showing EGFR perturbation—preferably with genetic perturbation or a selective chemical probe—reverses a disease-relevant phenotype, with cohort size, comparator, and outcome data. For ALK: direct evidence that the reported ALK perturbation was performed in a patient-derived xenograft, rather than only in a patient-derived cell line or other in vivo model, with demonstrated reversal of tumor phenotype and ideally orthogonal perturbation confirmation.	–
ALK — MET	Human genetics and causal evidence	Abstained	gpt-5.6-luna	98%	The dossiers provide somatic tumor-genomic and functional driver evidence for ALK and MET, including recurrent fusions or splice alterations (C1-EV_GENETICS_0001, C1-EV_GENETICS_0002, C1-EV_GENETICS_0003, C2-EV_GENETICS_0002, C2-EV_GENETICS_0003, C2-EV_GENETICS_0004). However, the stated criterion specifically calls for human-genetic evidence such as GWAS/fine-mapping, L2G-style assignment, molecular-QTL colocalisation, rare-variant burden, Mendelian disease, or biobank phenotypes. Those evidence types, including ancestry-diverse germline replication and comparative allelic-series evidence, are absent for both candidates. Somatic driver evidence cannot substitute for the requested human-genetic evidence, so there is no admissible basis to rank the candidates.	Genome-wide association or fine-mapping evidence; variant-to-gene/L2G assignment; disease-relevant molecular-QTL colocalisation; rare-variant burden or Mendelian evidence; biobank phenotypes; ancestry-diverse germline replication for either ALK or MET in NSCLC.	–
MET — ALK	Combination potential	Abstained	gpt-5.6-luna	98%	Neither dossier provides direct evidence for combination with chemotherapy. MET evidence supports combinations with EGFR inhibition in resistance models, not chemotherapy, while ALK evidence describes resistance mechanisms and comparisons of ALK inhibition versus platinum-pemetrexed chemotherapy rather than concurrent combination. Single-agent toxicity descriptions do not establish whether either target can be safely dosed with chemotherapy. Under the guardrail, mechanistic rationale without combination-specific safety and preclinical evidence is insufficient for ranking.	Direct in vivo or translational preclinical studies testing MET plus chemotherapy and ALK plus chemotherapy, including pathway-complementarity and resistance-mechanism analyses in the specified NSCLC context. A comparative overlapping-toxicity assessment for each target combined with chemotherapy, including doseability and tolerability of the combination.	In untreated advanced ALK-positive NSCLC, biomarker-selected patients derived substantially greater benefit from ALK inhibition with crizotinib than from platinum-pemetrexed chemotherapy. ALK is clinically validated as a predictive therapeutic biomarker in advanced NSCLC by randomized phase 3 evidence showing superior efficacy of ALK inhibition over platinum-pemetrexed chemotherapy.
MET — EGFR	Novelty, crowding, and strategic optionality	Abstained	gpt-5.6-luna	95%	MET has direct evidence of substantial current crowding in METex14 NSCLC, including multiple approved or clinically active inhibitors and competition from an ADC modality (C1-EV_COMPETITIVE_0005, C1-EV_COMPETITIVE_0006). It also has documented expansion across biomarker contexts and modalities (C1-EV_TRACTABILITY_0005). EGFR has evidence of multiple approved small-molecule TKI generations and potential antibody/TKI approaches, plus a broad set of possible indications (C2-EV_TRACTABILITY_0001, C2-CAT-ONCO-BIO-023, C2-CAT-ONCO-BIO-046). However, the EGFR dossier does not provide a directly comparable NSCLC competitor pipeline by phase, indication-expansion evidence grounded in programs, or a patent landscape. Therefore, the dossiers do not support a reliable comparison of strategic optionality relative to competitive crowding. Crowding and optionality cannot be netted into a verdict without comparable evidence.	A directly comparable EGFR NSCLC competitor pipeline by phase and indication, including active and approved agents. Candidate-specific patent landscape or freedom-to-operate evidence for both targets. Comparable evidence on platform fit and indication-expansion programs, rather than general catalogue descriptions.	The METex14 NSCLC market is already competitive, with multiple approved or clinically active MET inhibitors rather than a single differentiated standard. MET-targeted competition now includes a non-TKI modality with a broader MET-overexpression biomarker strategy. MET is tractable with an antibody–drug conjugate modality for a distinct biomarker-defined NSCLC population characterized by high c-Met protein overexpression. Multiple approved TKI generations Gibt es potenzielle alternative Zielstrukturen für EGFR? Bei welchen Krankheitsindikationen spielt EGFR eine spezifische Rolle?
KRAS — ALK	Human genetics and causal evidence	B wins	gpt-5.6-luna	72%	SECOND has the stronger causal gene-assignment evidence: the EML4-ALK fusion is a directly identified somatic alteration, has an unambiguous fusion-gene assignment, shows a defined activating direction, and was demonstrated to transform cells and cause tumors in models (C2-EV_GENETICS_0001, C2-EV_GENETICS_0002, C2-EV_MECHANISM_0003). Its recurrence is also supported by independent human tumor-genomic evidence (C2-EV_GENETICS_0003, C2-EV_GENETICS_0005). FIRST has substantial human evidence, including a KRAS allelic series with concordant activating direction and replication across tumor cohorts (C1-EV_GENETICS_0001, C1-EV_GENETICS_0002, C1-EV_GENETICS_0003), but the supplied dossier does not provide GWAS, QTL-colocalisation, or ancestry-diverse evidence, and its human tumor-genomic associations are less directly tied to a causal variant-to-gene demonstration than the EML4-ALK fusion evidence. The margin is narrow because KRAS has broader recurrence and allelic-series support, while ALK lacks the requested population-genetic and QTL evidence.	Neither dossier provides GWAS/fine-mapping or L2G scores, disease-relevant molecular-QTL colocalisation, ancestry-diverse replication, or rare-variant burden analyses. For KRAS, direct human functional evidence linking specific NSCLC alleles to tumor initiation is not supplied beyond recurrent somatic mutation and allelic-series evidence. For ALK, ancestry-diverse replication and systematic comparison of fusion-variant-specific causal effects are not supplied.	Causal oncogenic fusions A recurrent EML4–ALK fusion is a transforming somatic genetic alteration in human NSCLC. The recurrent EML4-ALK fusion is a transforming oncogenic driver in NSCLC. Independent comprehensive tumor-genome profiling confirms ALK rearrangements as recurrent somatic driver events in lung adenocarcinoma. Recurrent activating mutations KRAS is a recurrent, activating somatic driver in human lung adenocarcinoma, supporting KRAS inhibition as a genetically defined therapeutic strategy in NSCLC. Independent prospective clinical sequencing confirms KRAS as one of the dominant oncogenic drivers in metastatic human lung adenocarcinoma. Authoritative germline ALK susceptibility is established for neuroblastic tumors, not NSCLC. A large NSCLC germline sequencing study found that inherited susceptibility explains only a minority of cases and did not establish ALK as an NSCLC predisposition gene.
MET — KRAS	Cell and tissue relevance	A wins	gpt-5.6-luna	72%	MET has stronger direct human NSCLC tissue evidence: tumor-cell MET protein was assessed by immunohistochemistry, including primary tumors and matched nodal metastases, and recurrent MET exon-14 alterations were identified as tumor genomic drivers. KRAS has strong human genetic-driver evidence, but the supplied cell/tissue evidence mainly describes broad or near-ubiquitous expression and does not establish disease-specific localisation or dysregulation in mucus-producing glandular cells. The MET evidence is limited by intratumoral heterogeneity and discordance between protein expression and genomic alteration, and neither dossier provides the required cell-resolved evidence for the specified mucus-producing population; therefore the advantage is narrow rather than clear.	Single-cell or spatial transcriptomic/proteomic data, or cell-resolved immunohistochemistry, specifically measuring MET and KRAS expression, localisation, and alteration-associated dysregulation in human NSCLC mucus-producing glandular cells while controlling for cell-composition shifts and disease stage.	In NSCLC tissue, MET protein is primarily demonstrated in neoplastic tumor cells, with the highest prevalence in adenocarcinoma-containing and sarcomatoid histologies. MET expression and gene-copy alterations occur in both primary NSCLC tumors and matched lymph-node metastases, with higher MET positivity in nodal metastases. TCGA-integrated human tumor sequencing identified recurrent somatic MET exon 14 skipping as a likely driver alteration in lung adenocarcinoma. Diverse somatic MET exon 14 splice-site alterations occur in human lung cancers and define a recurrent genomic driver class. In welchen krankheitsrelevanten Zelltypen wird KRAS exprimiert? Wird KRAS in krankheitsrelevanten Zelltypen überexprimiert? MET protein overexpression is a poor proxy for the underlying MET genomic alteration in lung adenocarcinoma. Even in MET exon 14-altered NSCLC, MET protein expression can be absent or heterogeneous, weakening the assumption that the alteration marks a uniformly MET-expressing tumor-cell population. MET biomarker classification is substantially affected by intratumoral heterogeneity and discordance between assays.
ALK — KRAS	Competitive differentiation	A wins	gpt-5.6-luna	84%	ALK offers greater evidenced potential for clinically meaningful differentiation over chemotherapy because randomized phase 3 evidence in previously untreated ALK-positive NSCLC showed superior progression-free survival, response rate, symptoms, and quality of life versus platinum-pemetrexed chemotherapy (C1-EV_BIOMARKER_0002, C1-EV_CLINICAL_0001). However, the ALK segment is already highly competitive, with several preferred first-line TKIs and limited direct differentiation among them (C1-EV_COMPETITIVE_0001, C1-EV_COMPETITIVE_0006), so the margin is narrow. The KRAS dossier establishes target tractability but does not provide comparable direct first-line clinical evidence versus chemotherapy or a sufficient competitor-landscape analysis. This conclusion addresses scientific/clinical differentiation evidence, not market attractiveness.	For KRAS: randomized first-line evidence directly comparing a KRAS-directed therapy with chemotherapy in newly diagnosed NSCLC, plus a defined competitor-pipeline and treatment-landscape analysis establishing differentiation over current first-line options. For both candidates: evidence for the specific proposed agent, rather than target-level evidence, demonstrating clinically meaningful differentiation over chemotherapy in the stated newly diagnosed population.	In untreated advanced ALK-positive NSCLC, biomarker-selected patients derived substantially greater benefit from ALK inhibition with crizotinib than from platinum-pemetrexed chemotherapy. ALK is clinically validated as a predictive therapeutic biomarker in advanced NSCLC by randomized phase 3 evidence showing superior efficacy of ALK inhibition over platinum-pemetrexed chemotherapy. ALK-positive NSCLC is a highly competitive targeted-therapy segment with three preferred first-line ALK TKIs—alectinib, brigatinib, and lorlatinib—and no direct randomized head-to-head trials establishing superiority among them. Direct head-to-head evidence did not show brigatinib was superior to alectinib after crizotinib failure, indicating limited differentiation among ALK inhibitors. Ensartinib adds a newly approved first-line competitor with clinically meaningful PFS improvement over crizotinib, but its differentiation is not yet established against preferred next-generation ALK TKIs. Competition is extending toward fourth-generation ALK inhibitors designed to address compound resistance mutations that can emerge after lorlatinib.
KRAS — ALK	Evidence robustness (meta-evidence)	B wins	gpt-5.6-luna	88%	SECOND is more robust on the stated meta-evidence criterion. The ALK dossier identifies multiple source families, including a discovery study, TCGA profiling, a large real-world prevalence cohort, tissue pathology cohorts, genome-wide perturbation screens, randomized clinical trials and regulatory sources. It also preserves adverse or contradictory findings: incomplete ALK dependence, variant-specific sensitivity, subclonality, nonproductive FISH-positive cases, bypass resistance and lack of unadjusted overall-survival benefit. This makes source dependence and failure modes more visible. KRAS has some independent human-genomics support from TCGA, prospective MSK-IMPACT and a large retrospective cohort, but much of the dossier consists of repeated catalogue claims with limited provenance, no explicit negative-result or replication-status framework, and few criterion-specific contradiction or publication-bias indicators. The conclusion is therefore based on metadata quality and explicit heterogeneity, not on the larger quantity of ALK evidence.	An explicit sample-overlap and provenance matrix for all studies in both dossiers, especially whether the KRAS TCGA, MSK-IMPACT and retrospective NSCLC cohorts are independent. A systematic assessment of unpublished negative results, selective reporting and publication bias for both targets. Direct replication-status metadata for the key KRAS perturbation and in vivo PDX experiments, comparable to the explicitly heterogeneous and contradictory ALK evidence.	A recurrent EML4–ALK fusion is a transforming somatic genetic alteration in human NSCLC. Independent comprehensive tumor-genome profiling confirms ALK rearrangements as recurrent somatic driver events in lung adenocarcinoma. ALK rearrangements occur in a reproducible molecular subset of advanced NSCLC, enriched in younger and never-smoking patients. ALK rearrangement does not uniformly establish ALK pathway dependence across NSCLC models. ALK fusion variants produce materially different pathway dependency and inhibitor sensitivity. Genome-wide CRISPR knockout screens identify bypass alterations that permit survival under ALK inhibition, demonstrating context-dependent ALK-TKI dependence. Genome-wide CRISPR knockout identified MIG6/ERRFI1 loss as a mechanism of resistance to ALK TKIs through EGFR bypass signaling. CRISPR-based in vivo perturbation shows that EML4-ALK variants are functionally heterogeneous rather than a uniform dependency context. ALK is clinically validated as a predictive therapeutic biomarker in advanced NSCLC by randomized phase 3 evidence showing superior efficacy of ALK inhibition over platinum-pemetrexed chemotherapy. Second-generation ALK inhibition provides superior systemic and intracranial disease control compared with first-generation therapy in untreated advanced ALK-positive NSCLC. Lorlatinib provides durable long-term clinical benefit in treatment-naive advanced ALK-positive NSCLC, with sustained systemic and intracranial disease control. The pivotal PROFILE 1014 trial did not demonstrate a statistically significant overall-survival benefit for first-line crizotinib in the unadjusted analysis. ALK rearrangement does not uniformly establish ALK pathway dependence across NSCLC models. ALK fusion variants produce materially different pathway dependency and inhibitor sensitivity. ALK rearrangement is frequently subclonal or unevenly distributed across NSCLC tumor cells rather than present uniformly throughout the tumor. ALK-rearranged lung carcinomas can contain mixed histologic and cellular differentiation, so morphology alone does not define a single ALK-associated tissue or cell type. ALK positivity does not ensure durable dependence on ALK signaling because resistance can arise through both on-target and off-target mechanisms. ALK FISH positivity can represent a nonproductive rearrangement rather than an expressed, targetable ALK fusion. The pivotal PROFILE 1014 trial did not demonstrate a statistically significant overall-survival benefit for first-line crizotinib in the unadjusted analysis. Gibt es SNPs, die mit KRAS assoziiert sind? Gibt es bekannte Zusammenhänge zwischen den identifizierten SNPs in KRAS und dem Risiko oder der Prognose von {Disease}?
EGFR — MET	Biomarker and target engagement	B wins	gpt-5.6-luna	97%	MET offers the more credible integrated clinical biomarker strategy. MET exon 14 skipping has prospective clinical selection using tissue RNA or plasma cfDNA/NGS, is linked to objective responses and regulatory-approved treatment, and has a reported on-treatment molecular response based on serial METex14 ctDNA depletion. A validated pMET:total-MET assay provides a direct target-engagement and pharmacodynamic readout. EGFR has strong clinical precedent for mutation-directed therapy, but the supplied evidence does not provide an equally explicit clinical target-engagement or pharmacodynamic assay strategy; the pEGFR item is presented as a general biological measurement and is not shown to be clinically validated. MET's limitations—heterogeneous expression, assay discordance, and failure of MET-IHC alone—are documented, but they reinforce use of the more specific METex14/genomic strategy rather than outweighing it.	For EGFR, an explicitly validated clinical target-engagement assay, on-treatment pharmacodynamic biomarker, and biomarker-linked early efficacy dataset are not supplied; the dossier describes pEGFR as a biological readout but does not establish it as a clinical assay supporting the required decision. For the stated newly diagnosed, chemotherapy-backbone, in-vivo PDX context, neither dossier supplies a directly matched clinical biomarker-validation dataset.	MET exon 14 skipping is a clinically validated predictive biomarker for selective MET inhibition in NSCLC. Serial depletion of METex14 ctDNA is an on-treatment pharmacodynamic biomarker associated with clinical response to tepotinib. Validated pharmacodynamic assays based on the pMET:total-MET ratio provide a direct biochemical readout of MET pathway inhibition and are suitable for target-engagement studies. MET exon 14 skipping is clinically validated as a predictive biomarker in metastatic NSCLC through durable responses to selective MET inhibitors. Tepotinib has traditional FDA approval for metastatic NSCLC harboring MET exon 14 skipping alterations, with efficacy confirmed in a 313-patient dataset. Capmatinib is FDA-indicated for metastatic NSCLC with a mutation leading to MET exon 14 skipping, selected using an FDA-approved test. MET immunohistochemistry did not identify patients who benefited from onartuzumab in the randomized phase III METLung trial. MET biomarker classification is substantially affected by intratumoral heterogeneity and discordance between assays. MET amplification has weaker and threshold-dependent clinical validation than MET exon 14 skipping in NSCLC. Was ist der primäre biologische Messwert in Zellen nach der Beeinflussung von EGFR?
KRAS — EGFR	Safety and essentiality	A wins	gpt-5.6-luna	82%	KRAS has a meaningful on-target liability because it is broadly expressed and physiologically important; developmental and hematopoietic knockout phenotypes support this risk. However, the dossier specifically reports that mutant-selective covalent KRAS inhibitors spare wild-type KRAS (C1-EV_SAFETY_0001), providing a modality- and mechanism-specific mitigation. EGFR also has broad normal epithelial functions, severe knockout phenotypes involving skin, lung, and gastrointestinal tissues, and an explicit class effect of rash and diarrhoea from wild-type EGFR inhibition (C2-EV_SAFETY_0001). EGFR inhibition is described as affecting normal EGFR-expressing epithelia and as only partially mutation-selective (C2-CAT-ONCO-BIO-064). Therefore KRAS has the more favourable anticipated on-target safety profile, but the margin is narrow because KRAS essentiality and broad expression still indicate mechanism-based risk and direct comparative human safety evidence is absent.	Direct human loss-of-function, pan-essentiality, PheWAS, and tissue-specific knockout data for both targets, particularly in lung mucus-producing glandular cells. Comparative clinical exposure-toxicity data for mutation-selective KRAS small-molecule inhibition versus EGFR small-molecule inhibition.	Mutant-selective inhibitors spare wild type In welchen krankheitsrelevanten Zelltypen wird KRAS exprimiert? Ist KRAS üblicherweise essenziell? Welche Gewebe oder Organsysteme sind bei der Maus nach dem Knockout des KRAS-Gens betroffen? On-target dermatologic and GI toxicity Ist die Beeinträchtigung der Funktion von EGFR selektiv? Welcher Phänotyp tritt auf, wenn EGFR deletiert wird? In welchen krankheitsrelevanten Zelltypen wird KRAS exprimiert? Ist KRAS üblicherweise essenziell? Welche Gewebe oder Organsysteme sind bei der Maus nach dem Knockout des KRAS-Gens betroffen? Welche Rolle spielt EGFR in gesunden Zellen/Geweben? Welche zellulären Prozesse werden durch EGFR in gesundem Gewebe im Vergleich zu erkranktem Gewebe reguliert?
MET — KRAS	Evidence robustness (meta-evidence)	A wins	gpt-5.6-luna	90%	MET has a substantially better documented meta-evidence structure: its dossier identifies source types and provenance, records explicit cohort and trial dependence, includes independent human-genomic cohorts, replicated clinical validation with different selective inhibitors, and multiple negative or contradictory findings such as METLung, ATTENTION, biomarker discordance, and limited activity in low-level amplification. KRAS has evidence of biological and clinical tractability, but its dossier is dominated by catalogue-style answers and lacks explicit source-family mapping, sample-overlap assessment, replication status, and systematic negative-result availability. This is not treated as evidence that KRAS is biologically weaker; it means its robustness cannot be established from the supplied meta-evidence. The documented dependence and contradiction handling for MET therefore make FIRST more robust on this criterion.	For KRAS: explicit provenance metadata, source-family structure, sample-overlap flags, replication status, and availability of negative results across the cited catalogue answers and clinical/preclinical evidence For KRAS: independent replicated studies and documented contradictory or null findings for the specific NSCLC context, mucus-producing glandular cells, and patient-derived xenograft setting	TCGA-integrated human tumor sequencing identified recurrent somatic MET exon 14 skipping as a likely driver alteration in lung adenocarcinoma. Diverse somatic MET exon 14 splice-site alterations occur in human lung cancers and define a recurrent genomic driver class. A large human lung-cancer profiling cohort confirmed that METex14 alterations are distributed across major NSCLC histologies and occur in 2.7% of profiled lung cancers. MET immunohistochemical positivity did not identify a clinically responsive NSCLC population in the phase III METLung trial. Adding the MET inhibitor tivantinib to erlotinib failed to improve overall survival in the phase III ATTENTION trial. MET immunohistochemistry did not identify patients who benefited from onartuzumab in the randomized phase III METLung trial. MET exon 14 skipping is clinically validated as a predictive biomarker in metastatic NSCLC through durable responses to selective MET inhibitors. Tepotinib has traditional FDA approval for metastatic NSCLC harboring MET exon 14 skipping alterations, with efficacy confirmed in a 313-patient dataset. Co-occurring RAS–MAPK pathway alterations are common in MET exon 14-altered NSCLC and can limit response to MET-directed therapy. MET alterations are uncommon in NSCLC and MET exon 14 mutations may be associated with worse prognosis rather than uniformly favorable biology. MET protein overexpression is a poor proxy for the underlying MET genomic alteration in lung adenocarcinoma. CRISPR MET knockout in A549 NSCLC cells did not materially impair intrinsic proliferation, despite reducing dissemination-related phenotypes. Mutant-selective inhibitors spare wild type
MET — ALK	Competitive differentiation	B wins	gpt-5.6-luna	95%	ALK offers greater potential for clinically meaningful differentiation over chemotherapy because untreated, biomarker-selected ALK-positive patients showed superior progression-free survival and response rate with crizotinib versus platinum-pemetrexed chemotherapy in randomized phase 3 evidence (C2-EV_BIOMARKER_0002; C2-EV_CLINICAL_0001). The ALK landscape also contains clinically demonstrated differentiation through superior CNS and long-term disease control with next-generation inhibitors, although the class is highly competitive (C2-EV_COMPETITIVE_0001; C2-EV_COMPETITIVE_0002). For MET, the dossier supports activity of selective inhibitors in treatment-naive METex14 disease, but explicitly states that chemotherapy remains a background option and provides no evidence of a comparative efficacy advantage over chemotherapy (C1-EV_COMPETITIVE_0003). MET is additionally a crowded competitive segment with multiple approved or clinically active agents and a non-TKI competitor (C1-EV_COMPETITIVE_0005; C1-EV_COMPETITIVE_0006). This conclusion addresses differentiation over chemotherapy, not overall scientific merit or market attractiveness.	A randomized, direct comparison of a selective MET inhibitor versus chemotherapy in newly diagnosed METex14-positive NSCLC would most change the answer for MET.	In untreated advanced ALK-positive NSCLC, biomarker-selected patients derived substantially greater benefit from ALK inhibition with crizotinib than from platinum-pemetrexed chemotherapy. ALK is clinically validated as a predictive therapeutic biomarker in advanced NSCLC by randomized phase 3 evidence showing superior efficacy of ALK inhibition over platinum-pemetrexed chemotherapy. ALK-positive NSCLC is a highly competitive targeted-therapy segment with three preferred first-line ALK TKIs—alectinib, brigatinib, and lorlatinib—and no direct randomized head-to-head trials establishing superiority among them. The principal differentiation opportunity for MET-directed therapy is against non-targeted NSCLC treatment, rather than a proven efficacy advantage among MET inhibitors. The pivotal PROFILE 1014 trial did not demonstrate a statistically significant overall-survival benefit for first-line crizotinib in the unadjusted analysis. Direct head-to-head evidence did not show brigatinib was superior to alectinib after crizotinib failure, indicating limited differentiation among ALK inhibitors. The METex14 NSCLC market is already competitive, with multiple approved or clinically active MET inhibitors rather than a single differentiated standard. MET-targeted competition now includes a non-TKI modality with a broader MET-overexpression biomarker strategy.
EGFR — ALK	Cell and tissue relevance	B wins	gpt-5.6-luna	78%	ALK has stronger criterion-matched evidence because human NSCLC tissue studies directly detect ALK rearrangements and ALK protein by IHC in tumor cells, with localization to the pulmonary adenocarcinoma epithelial compartment (C2-EV_TISSUE_0001, C2-EV_TISSUE_0002, C2-EV_TISSUE_0006). However, the evidence also shows substantial subclonality and heterogeneous histologic differentiation, limiting confidence that the signal is specific to mucus-producing glandular cells (C2-EV_TISSUE_0004, C2-EV_TISSUE_0005). EGFR evidence supports expression in malignant epithelial cells and general tumor-versus-normal dysregulation, but does not provide direct human pathology, single-cell, or spatial evidence for the specified mucus-producing glandular cells (C1-CAT-ONCO-BIO-047, C1-CAT-ONCO-BIO-058, C1-CAT-ONCO-BIO-097). Therefore ALK is favored, but only narrowly because neither dossier resolves the critical cell type or fully controls for composition shifts.	For both candidates: single-cell or spatially resolved human NSCLC data specifically identifying mucus-producing glandular cells, with cell-composition and disease-stage adjustment. For EGFR: direct human pathology or IHC/proteomic evidence localizing dysregulated EGFR specifically to mucus-producing glandular tumor cells. For ALK: direct evidence that ALK rearrangement, ALK protein expression, or pathway dysregulation is specifically present in mucus-producing glandular cells rather than the broader epithelial tumor compartment.	In NSCLC, oncogenic ALK is found in malignant lung epithelial tumor cells, predominantly as an EML4–ALK fusion in a molecular subset of tumors. ALK rearrangements are directly detectable in lung adenocarcinoma tumor tissue and are associated with ALK protein expression in the tumor cells. ALK-positive NSCLC is most strongly represented in the lung adenocarcinoma epithelial compartment, with frequent pulmonary adenocarcinoma markers and characteristic tumor-cell morphology. ALK rearrangement is frequently subclonal or unevenly distributed across NSCLC tumor cells rather than present uniformly throughout the tumor. ALK-rearranged lung carcinomas can contain mixed histologic and cellular differentiation, so morphology alone does not define a single ALK-associated tissue or cell type.
ALK — MET	Competitive differentiation	A wins	gpt-5.6-luna	93%	ALK offers stronger evidence for clinically meaningful differentiation over chemotherapy in newly diagnosed disease: randomized phase 3 evidence showed superior progression-free survival and response rate versus platinum-pemetrexed chemotherapy in biomarker-selected patients. Its differentiation is also supported by durable systemic and intracranial control, although the ALK segment is crowded and incremental differentiation among preferred ALK inhibitors is not established. MET has clinically meaningful treatment-naive response data in METex14-positive disease and a stated opportunity relative to non-targeted treatment, but the dossier does not provide a direct randomized comparison against chemotherapy. MET differentiation is further limited by multiple approved or clinically active MET inhibitors and competition from a non-TKI modality. This conclusion concerns evidence for scientific/clinical differentiation over chemotherapy, not market attractiveness.	A randomized first-line trial directly comparing a MET-directed therapy in METex14-positive newly diagnosed NSCLC with chemotherapy, with clinically meaningful efficacy and intracranial outcomes	ALK is clinically validated as a predictive therapeutic biomarker in advanced NSCLC by randomized phase 3 evidence showing superior efficacy of ALK inhibition over platinum-pemetrexed chemotherapy. In untreated advanced ALK-positive NSCLC, biomarker-selected patients derived substantially greater benefit from ALK inhibition with crizotinib than from platinum-pemetrexed chemotherapy. ALK-positive NSCLC is a highly competitive targeted-therapy segment with three preferred first-line ALK TKIs—alectinib, brigatinib, and lorlatinib—and no direct randomized head-to-head trials establishing superiority among them. Lorlatinib differentiates through exceptional long-term systemic and intracranial disease control in treatment-naive ALK-positive NSCLC. The principal differentiation opportunity for MET-directed therapy is against non-targeted NSCLC treatment, rather than a proven efficacy advantage among MET inhibitors. MET exon 14 skipping is clinically validated as a predictive biomarker in metastatic NSCLC through durable responses to selective MET inhibitors. Competition is extending toward fourth-generation ALK inhibitors designed to address compound resistance mutations that can emerge after lorlatinib. Direct head-to-head evidence did not show brigatinib was superior to alectinib after crizotinib failure, indicating limited differentiation among ALK inhibitors. The METex14 NSCLC market is already competitive, with multiple approved or clinically active MET inhibitors rather than a single differentiated standard. MET-targeted competition now includes a non-TKI modality with a broader MET-overexpression biomarker strategy. MET amplification has weaker and threshold-dependent clinical validation than MET exon 14 skipping in NSCLC.
EGFR — MET	Human genetics and causal evidence	B wins	gpt-5.6-luna	82%	MET has stronger supplied human genomic evidence for a causal NSCLC role: multiple independent lung-cancer profiling datasets identify recurrent MET exon-14 splice alterations, including a broad allelic series of 126 distinct variants, with cis-acting splice-site changes linked to exon skipping, constitutive activation, and oncogenic transformation (C2-EV_GENETICS_0002, C2-EV_GENETICS_0003, C2-EV_GENETICS_0004). A clinical series also reports responses associated with these alterations (C2-EV_GENETICS_0006). EGFR has strong causal evidence for canonical activating somatic mutations, but the supplied genetics evidence is less detailed for variant-to-gene assignment and replication, while its germline familial NSCLC evidence is rare and incompletely quantified (C1-EV_GENETICS_0001; C1-CAT-ONCO-BIO-016; C1-CAT-ONCO-BIO-056). MET evidence is nevertheless limited by the absence of GWAS, fine-mapping, QTL colocalisation, ancestry diversity, and formal germline analyses; therefore the margin is narrow rather than clear.	Neither dossier supplies GWAS or fine-mapping results, L2G-style variant-to-gene scores, disease-relevant molecular-QTL colocalisation, ancestry-stratified analyses, or formal rare-variant burden results. Direct comparative germline evidence for NSCLC causality is also missing, particularly for MET; EGFR familial and rare-germline evidence is described but not quantified or replicated in the supplied records.	TCGA-integrated human tumor sequencing identified recurrent somatic MET exon 14 skipping as a likely driver alteration in lung adenocarcinoma. Diverse somatic MET exon 14 splice-site alterations occur in human lung cancers and define a recurrent genomic driver class. A large human lung-cancer profiling cohort confirmed that METex14 alterations are distributed across major NSCLC histologies and occur in 2.7% of profiled lung cancers. MET exon 14 splice-site mutations are present in approximately 4% of lung adenocarcinomas and can be clinically actionable in humans. Causal activating mutations Co-occurring RAS–MAPK pathway alterations are common in MET exon 14-altered NSCLC and can limit response to MET-directed therapy. MET alterations are uncommon in NSCLC and MET exon 14 mutations may be associated with worse prognosis rather than uniformly favorable biology.
ALK — KRAS	Resistance and escape liability	B wins	gpt-5.6-luna	67%	The direction is reversed: the stronger target is the one least vulnerable to escape. SECOND (KRAS) is favored narrowly because the ALK dossier contains particularly direct and high-quality clinical and functional evidence of acquired on-target compound mutations, non-ALK-dependent resistance, and bypass through EGFR-MAPK/PI3K signaling (C1-EV_TRACTABILITY_0007; C1-EV_BIOMARKER_0004; C1-EV_PERTURBATION_0002; C1-EV_PERTURBATION_0005). KRAS also has documented feedback reactivation, bypass signaling, secondary binding-site mutations, allele switching, and amplification, so it is not low-risk (C2-CAT-ONCO-BIO-004; C2-CAT-ONCO-BIO-013; C2-CAT-ONCO-BIO-063). However, those KRAS resistance data are predominantly catalogue-level and focused on KRAS-G12C, whereas ALK resistance is supported by repeat-biopsy clinical and validated mechanistic studies. Combination strategies could mitigate risk for both: KRAS combinations targeting RTK-SHP2-SOS/MAPK feedback are explicitly indicated, while ALK bypass may be mitigated by combining with or switching to pathway-directed agents, and sequential next-generation inhibitors may address selected on-target mutations. Biomarker strategies also mitigate risk: allele-specific KRAS testing and resistance genotyping can select appropriate inhibitors, while ALK fusion, expressed-fusion, variant, and resistance-mutation testing can refine therapy selection. The evidence therefore supports only a narrow SECOND verdict, not an absence of KRAS escape liability.	A matched, direct clinical comparison of acquired resistance frequency, clonal evolution, and post-progression treatment options for ALK-directed versus KRAS-directed therapy in NSCLC, separately by ALK fusion variant and KRAS allele.	The ALK small-molecule modality is vulnerable to acquired on-target resistance, particularly after sequential inhibitor treatment. ALK positivity does not ensure durable dependence on ALK signaling because resistance can arise through both on-target and off-target mechanisms. Genome-wide CRISPR knockout screens identify bypass alterations that permit survival under ALK inhibition, demonstrating context-dependent ALK-TKI dependence. Genome-wide CRISPR knockout identified MIG6/ERRFI1 loss as a mechanism of resistance to ALK TKIs through EGFR bypass signaling. Gibt es kompensatorische Mechanismen, die die Auswirkungen der Hemmung von KRAS abmildern können? Gibt es biochemische Signalwege, die die Hemmung von KRAS umgehen und dadurch zu einer Resistenz führen können? Ist KRAS von Resistenzmutationen betroffen? Gibt es kompensatorische Mechanismen, die die Auswirkungen der Hemmung von KRAS abmildern können? Gibt es biochemische Signalwege, die die Hemmung von KRAS umgehen und dadurch zu einer Resistenz führen können? Ist KRAS von Resistenzmutationen betroffen? Welche spezifischen genetischen Veränderungen in KRAS verleihen eine Resistenz gegen zielgerichtete Therapien oder Chemotherapie?
EGFR — ALK	Resistance and escape liability	B wins	gpt-5.6-luna	72%	Applying the reversed direction, ALK (SECOND) is judged less vulnerable, but only narrowly. EGFR has documented pathway redundancy and escape through MET/HGF and other bypass routes, plus acquired T790M and C797S resistance and therapeutic-selection-driven clonal evolution (C1-CAT-ONCO-BIO-004, C1-CAT-ONCO-BIO-013, C1-CAT-ONCO-BIO-063, C1-CAT-ONCO-BIO-066, C1-CAT-ONCO-BIO-078). ALK also has substantial escape liability: EGFR bypass after MIG6/ERRFI1 loss, KEAP1-mediated resistance, non-ALK-dependent progression, and compound on-target mutations after sequential inhibitors (C2-EV_PERTURBATION_0002, C2-EV_PERTURBATION_0003, C2-EV_BIOMARKER_0004, C2-EV_TRACTABILITY_0007). However, the ALK dossier provides a stronger biomarker-and-sequencing mitigation framework: ALK positivity selects the target, resistance mutations can guide later-line inhibitor choice, and broader sequential ALK inhibition can address some single-mutant resistance states (C2-EV_BIOMARKER_0001, C2-EV_BIOMARKER_0003, C2-EV_COMPETITIVE_0004). For EGFR, combination strategies could mitigate bypass risk, particularly EGFR inhibition with MET-directed therapy when MET activation is identified, and serial genotyping could guide mutation-specific inhibitor selection; these strategies do not eliminate the documented heterogeneous escape routes. For ALK, combination or biomarker strategies could likewise mitigate risk, including EGFR blockade for EGFR-bypass resistance and resistance-mutation-guided selection of next-generation ALK inhibitors. The evidence therefore favors ALK as less affected, but the difference is narrow because both targets have well-established redundancy and acquired-resistance mechanisms.	Matched head-to-head longitudinal resistance analyses for EGFR- versus ALK-driven NSCLC under comparable first-line treatment, including frequency and durability of bypass, on-target, and polyclonal resistance mechanisms in the specified patient-derived xenograft context.	Gibt es kompensatorische Mechanismen, die die Auswirkungen der Hemmung von EGFR abmildern können? Gibt es biochemische Signalwege, die die Hemmung von EGFR umgehen und dadurch zu einer Resistenz führen können? Ist EGFR von Resistenzmutationen betroffen? Gibt es Hinweise auf eine klonale Evolution unter Beteiligung von Veränderungen an EGFR während des Krankheitsverlaufs? Welche spezifischen genetischen Veränderungen in EGFR verleihen eine Resistenz gegen zielgerichtete Therapien oder Chemotherapie? Genome-wide CRISPR knockout screens identify bypass alterations that permit survival under ALK inhibition, demonstrating context-dependent ALK-TKI dependence. CRISPR-Cas9 resistance screens show that KEAP1 loss modulates response to ALK inhibition in ALK-mutant lung cancer cells. ALK positivity does not ensure durable dependence on ALK signaling because resistance can arise through both on-target and off-target mechanisms. The ALK small-molecule modality is vulnerable to acquired on-target resistance, particularly after sequential inhibitor treatment. ALK rearrangement does not uniformly establish ALK pathway dependence across NSCLC models. ALK fusion variants produce materially different pathway dependency and inhibitor sensitivity. Genome-wide CRISPR knockout identified MIG6/ERRFI1 loss as a mechanism of resistance to ALK TKIs through EGFR bypass signaling. CRISPR-based in vivo perturbation shows that EML4-ALK variants are functionally heterogeneous rather than a uniform dependency context.
KRAS — ALK	Druggability and tractability	B wins	gpt-5.6-luna	93%	Both candidates have evidence supporting small-molecule tractability in lung cancer. KRAS has a clinically validated covalent small-molecule modality, but the strongest evidence is allele-specific and concentrated on KRAS-G12C, with additional resistance and bypass mechanisms documented. ALK has broader and more mature tractability evidence: multiple approved oral small-molecule TKIs, regulatory companion-diagnostic infrastructure, randomized clinical efficacy in NSCLC, and evidence of systemic and intracranial activity. These data directly support binding-site accessibility, modality feasibility, delivery to lung cancer, and therapeutic precedent. ALK therefore has the stronger overall small-molecule tractability dossier, while acknowledging that exact evidence in mucus-producing glandular cells is missing.	Direct target-expression, binding-site engagement, intracellular exposure, and pharmacodynamic delivery data for ALK and KRAS specifically in mucus-producing glandular cells of lung PDX tumors. Comparative lung tissue and cell-type distribution data for the candidate small molecules in the specified mucus-producing glandular-cell context. For KRAS, tractability evidence for the full stated KRAS scope, particularly G12D, rather than predominantly clinically validated G12C.	Several approved ALK TKIs ALK is clinically tractable with an established oral small-molecule kinase-inhibitor modality in metastatic NSCLC. First-line ALK inhibition produces strong systemic and intracranial clinical efficacy, supporting high tractability of ALK in NSCLC. Alectinib demonstrates that selective, CNS-penetrant ALK inhibition is an effective modality in treatment-naive advanced NSCLC. ALK-targeted therapy has an established companion-diagnostic infrastructure, supporting practical biomarker selection and deployment of the modality. ALK is clinically validated as a predictive therapeutic biomarker in advanced NSCLC by randomized phase 3 evidence showing superior efficacy of ALK inhibition over platinum-pemetrexed chemotherapy. Clinically validated tractability Weist KRAS Mutationen auf, die neue medikamentös adressierbare Bindungsstellen schaffen? Welche Toolverbindungen für KRAS sind verfügbar? The ALK small-molecule modality is vulnerable to acquired on-target resistance, particularly after sequential inhibitor treatment. ALK positivity does not ensure durable dependence on ALK signaling because resistance can arise through both on-target and off-target mechanisms. Gibt es biochemische Signalwege, die die Hemmung von KRAS umgehen und dadurch zu einer Resistenz führen können? Ist KRAS von Resistenzmutationen betroffen?
ALK — EGFR	Evidence robustness (meta-evidence)	A wins	gpt-5.6-luna	93%	ALK has a more auditable evidence structure: it includes distinct source families such as an original discovery study, TCGA profiling, a large real-world cohort, multicenter tissue pathology, functional perturbation studies, and regulatory or clinical sources. It also explicitly preserves contradictory findings, including variable model dependence and bypass resistance. These features support assessment of replication and publication-bias vulnerability, even though some ALK items share the original discovery study or resistance experiments. EGFR has relevant evidence, but its dossier is overwhelmingly composed of catalogue-answer summaries with sparse provenance and no documented sample-overlap, replication, or negative-result framework. Thus its apparent breadth cannot be treated as independent robustness under this meta-evidence criterion.	For EGFR: independently identified primary-study references with provenance, cohort/sample-overlap flags, explicit replication status, and available negative or null results; the current dossier is dominated by catalogue summaries. For both candidates: a systematic source-family and sample-overlap audit across all listed studies, especially for repeated cohorts and summaries of the same experiments. For EGFR: directly documented contradictory findings or failed critical experiments in the specified NSCLC context, rather than only general statements of context dependence.	A recurrent EML4–ALK fusion is a transforming somatic genetic alteration in human NSCLC. Independent comprehensive tumor-genome profiling confirms ALK rearrangements as recurrent somatic driver events in lung adenocarcinoma. ALK rearrangements occur in a reproducible molecular subset of advanced NSCLC, enriched in younger and never-smoking patients. ALK rearrangements are directly detectable in lung adenocarcinoma tumor tissue and are associated with ALK protein expression in the tumor cells. ALK rearrangement does not uniformly establish ALK pathway dependence across NSCLC models. ALK fusion variants produce materially different pathway dependency and inhibitor sensitivity. Genome-wide CRISPR knockout identified MIG6/ERRFI1 loss as a mechanism of resistance to ALK TKIs through EGFR bypass signaling. Causal activating mutations Ligand-independent kinase activation Multiple approved TKI generations On-target dermatologic and GI toxicity ALK rearrangement does not uniformly establish ALK pathway dependence across NSCLC models. ALK fusion variants produce materially different pathway dependency and inhibitor sensitivity. CRISPR loss of KEAP1 can produce resistance to ALK inhibition in ALK-mutant lung cancer models. Genome-wide CRISPR knockout identified MIG6/ERRFI1 loss as a mechanism of resistance to ALK TKIs through EGFR bypass signaling. On-target dermatologic and GI toxicity
ALK — MET	Disease biology and mechanism coherence	A wins	gpt-5.6-luna	87%	ALK has the more coherent causal chain: a recurrent EML4-ALK fusion is identified as a human oncogenic driver, produces constitutive kinase signaling, activates defined downstream survival circuitry, and is selectively required in ALK-rearranged NSCLC models by direct knockdown and inhibitor perturbation. MET also has strong mechanistic evidence, especially for MET exon 14 skipping and MET-amplified subsets, including impaired receptor turnover and selective dependency in amplified models. However, MET dependence is more alteration- and context-dependent, with substantial heterogeneity and evidence that MET perturbation does not generally impair intrinsic proliferation. ALK likewise has important variant-specific and bypass-resistance limitations, so the advantage is narrow rather than clear. This judgment concerns mechanistic coherence and perturbational support, not druggability.	Direct head-to-head mechanistic comparison of ALK and MET inhibition in matched NSCLC models, particularly patient-derived models of the specified mucus-producing glandular cell context, with genetic and pharmacologic perturbation concordance.	Causal oncogenic fusions A recurrent EML4–ALK fusion is a transforming somatic genetic alteration in human NSCLC. ALK rearrangements in NSCLC create oncogenic ALK fusion proteins whose constitutive signaling promotes tumor-cell proliferation and survival. The recurrent EML4-ALK fusion is a transforming oncogenic driver in NSCLC. RAS–RAF–MEK–ERK/MAPK is the primary downstream survival pathway in EML4-ALK-positive lung adenocarcinoma models. Direct ALK knockdown selectively impairs viability and clonogenic growth of EML4-ALK-positive NSCLC cells. ALK signaling is essential for proliferation and survival in at least a subset of EML4-ALK-positive NSCLC models. TCGA-integrated human tumor sequencing identified recurrent somatic MET exon 14 skipping as a likely driver alteration in lung adenocarcinoma. MET exon 14 skipping activates MET by removing the juxtamembrane CBL-binding region, impairing receptor ubiquitination and down-regulation. MET is an HGF-activated receptor tyrosine kinase whose oncogenic signaling promotes proliferation and survival through downstream pathways including MAPK and PI3K/AKT. MET knockdown selectively impairs growth and survival of MET-amplified NSCLC cell lines. Pharmacologic MET inhibition suppresses MET-dependent lung cancer models in vitro and in vivo, with sensitivity linked to MET signaling complexes. ALK rearrangement does not uniformly establish ALK pathway dependence across NSCLC models. ALK fusion variants produce materially different pathway dependency and inhibitor sensitivity. Genome-wide CRISPR knockout identified MIG6/ERRFI1 loss as a mechanism of resistance to ALK TKIs through EGFR bypass signaling. Co-occurring RAS–MAPK pathway alterations are common in MET exon 14-altered NSCLC and can limit response to MET-directed therapy. MET perturbational dependency in NSCLC is restricted to a molecularly defined subset rather than being general across lung cancer models. CRISPR MET knockout in A549 NSCLC cells did not materially impair intrinsic proliferation, despite reducing dissemination-related phenotypes. Even in MET exon 14-altered NSCLC, MET protein expression can be absent or heterogeneous, weakening the assumption that the alteration marks a uniformly MET-expressing tumor-cell population.
MET — ALK	Novelty, crowding, and strategic optionality	A wins	gpt-5.6-luna	72%	MET offers slightly greater demonstrated strategic optionality because the dossier supports both genotype-selected small-molecule inhibition and a distinct antibody-drug-conjugate strategy for a broader c-MET-overexpression population, creating more than one clinically actionable biomarker/modality path (C1-EV_TRACTABILITY_0005, C1-EV_CLINICAL_0004, C1-EV_CLINICAL_0005). However, this advantage is narrow: METex14 NSCLC is already crowded with multiple approved or clinically active inhibitors and now an additional non-TKI competitor (C1-EV_COMPETITIVE_0005, C1-EV_COMPETITIVE_0006), which raises biological confidence but reduces strategic value. ALK also has meaningful optionality through adjuvant treatment and an established companion-diagnostic infrastructure (C2-EV_TRACTABILITY_0005, C2-EV_TRACTABILITY_0006), plus ongoing fourth-generation inhibitor development for compound resistance (C2-EV_COMPETITIVE_0005). Its competitive crowding is especially substantial, with several preferred first-line TKIs and ongoing next-generation competition (C2-EV_COMPETITIVE_0001, C2-EV_COMPETITIVE_0005). Neither dossier supplies a patent landscape, so the IP component cannot distinguish them.	A comparative patent-landscape analysis, including patent breadth, remaining term, freedom-to-operate, and composition-of-matter or biomarker claims for MET versus ALK. Evidence of MET or ALK expansion into additional non-NSCLC indications or validated cross-tumor platform leverage. Direct evidence comparing strategic differentiation and commercial crowding between the two targets.	MET is tractable with an antibody–drug conjugate modality for a distinct biomarker-defined NSCLC population characterized by high c-Met protein overexpression. High c-MET protein overexpression represents a second clinically actionable MET biomarker context in nonsquamous NSCLC, although the regulatory evidence is based on accelerated approval. The LUMINOSITY phase 2 study provides clinical evidence for telisotuzumab vedotin activity in c-Met protein-overexpressing nonsquamous EGFR-wild-type NSCLC. The METex14 NSCLC market is already competitive, with multiple approved or clinically active MET inhibitors rather than a single differentiated standard. MET-targeted competition now includes a non-TKI modality with a broader MET-overexpression biomarker strategy. ALK-directed oral therapy is tractable beyond metastatic disease, including adjuvant treatment after complete resection. ALK-positive NSCLC is a highly competitive targeted-therapy segment with three preferred first-line ALK TKIs—alectinib, brigatinib, and lorlatinib—and no direct randomized head-to-head trials establishing superiority among them. Competition is extending toward fourth-generation ALK inhibitors designed to address compound resistance mutations that can emerge after lorlatinib. The METex14 NSCLC market is already competitive, with multiple approved or clinically active MET inhibitors rather than a single differentiated standard. MET-targeted competition now includes a non-TKI modality with a broader MET-overexpression biomarker strategy. ALK-positive NSCLC is a highly competitive targeted-therapy segment with three preferred first-line ALK TKIs—alectinib, brigatinib, and lorlatinib—and no direct randomized head-to-head trials establishing superiority among them. Competition is extending toward fourth-generation ALK inhibitors designed to address compound resistance mutations that can emerge after lorlatinib.
ALK — EGFR	Integrated therapeutic hypothesis	A wins	gpt-5.6-luna	94%	ALK represents the stronger integrated target–disease–modality hypothesis because the dossier connects a causal NSCLC alteration to a defined oncogenic mechanism, selective perturbational dependence, validated biomarker selection, established small-molecule tractability, and multiple randomized clinical demonstrations of benefit over chemotherapy, including durable systemic and intracranial activity in previously untreated disease. The hypothesis is not uniformly applicable: ALK dependence varies by model and fusion variant, resistance is common, toxicity is clinically relevant, and the segment is crowded. EGFR also has a coherent causal and mechanistic rationale and established small-molecule tractability, but its dossier is dominated by repetitive catalogue summaries and lacks comparable study-level randomized clinical evidence for the stated newly diagnosed context or a demonstrated differentiated medicine opportunity. The available EGFR evidence therefore supports plausibility but not an overall hypothesis as strongly as the ALK evidence.	Direct randomized clinical validation of EGFR inhibition versus chemotherapy in newly diagnosed EGFR-mutant NSCLC, with evidence of differentiated benefit against current EGFR-directed standards. Direct head-to-head evidence demonstrating that either target provides differentiated value in the specified mucus-producing glandular-cell, patient-derived xenograft, in-vivo proof-of-concept context. For ALK, evidence specifically in the stated mucus-producing glandular-cell PDX model and a differentiation strategy that overcomes the crowded ALK-TKI segment.	Causal oncogenic fusions ALK rearrangements in NSCLC create oncogenic ALK fusion proteins whose constitutive signaling promotes tumor-cell proliferation and survival. Direct ALK knockdown selectively impairs viability and clonogenic growth of EML4-ALK-positive NSCLC cells. ALK is clinically tractable with an established oral small-molecule kinase-inhibitor modality in metastatic NSCLC. In untreated advanced ALK-positive NSCLC, biomarker-selected patients derived substantially greater benefit from ALK inhibition with crizotinib than from platinum-pemetrexed chemotherapy. ALK is clinically validated as a predictive therapeutic biomarker in advanced NSCLC by randomized phase 3 evidence showing superior efficacy of ALK inhibition over platinum-pemetrexed chemotherapy. Lorlatinib provides durable long-term clinical benefit in treatment-naive advanced ALK-positive NSCLC, with sustained systemic and intracranial disease control. ALK rearrangement does not uniformly establish ALK pathway dependence across NSCLC models. ALK fusion variants produce materially different pathway dependency and inhibitor sensitivity. ALK-positive NSCLC is a highly competitive targeted-therapy segment with three preferred first-line ALK TKIs—alectinib, brigatinib, and lorlatinib—and no direct randomized head-to-head trials establishing superiority among them. Competition is extending toward fourth-generation ALK inhibitors designed to address compound resistance mutations that can emerge after lorlatinib. On-target dermatologic and GI toxicity Gibt es kompensatorische Mechanismen, die die Auswirkungen der Hemmung von EGFR abmildern können? Gibt es biochemische Signalwege, die die Hemmung von EGFR umgehen und dadurch zu einer Resistenz führen können?
EGFR — ALK	Reversibility and therapeutic window	B wins	gpt-5.6-luna	78%	ALK has stronger evidence for a therapeutic window: narrow adult expression and limited normal-tissue essentiality support separation from normal-tissue effects, while oral once-daily dosing supports interruption and titration. EGFR inhibition is explicitly associated with on-target rash and diarrhea, narrowing its window. However, ALK inhibitors also have clinically important organ, pulmonary, cardiac, neurologic, and metabolic toxicities, and the dossiers do not provide direct comparative half-life, exposure-response, or post-interruption recovery data. Therefore the advantage for ALK is narrow rather than clear.	Direct comparative exposure-response data, modality-specific half-life and washout data, and clinical evidence showing recovery after dose interruption for representative EGFR and ALK small-molecule inhibitors.	Favourable on-target safety Wild-type ALK appears to have limited essentiality for survival of adult normal tissues, although loss of ALK produces specific reproductive and neuroendocrine phenotypes in mice. ALK is clinically tractable with an established oral small-molecule kinase-inhibitor modality in metastatic NSCLC. On-target dermatologic and GI toxicity Welche Toolverbindungen für EGFR sind verfügbar? ALK inhibition has a generally manageable clinical safety profile but carries clinically important risks, including hepatotoxicity, interstitial lung disease/pneumonitis, QT prolongation, bradycardia, severe visual loss, and embryo-fetal toxicity. Alectinib has clinically important hepatotoxicity and other organ toxicities that complicate the safety case for ALK inhibition. First-line lorlatinib produces a high burden of metabolic, neurologic and other adverse events, with more grade 3–4 toxicity than crizotinib.
ALK — EGFR	Perturbational evidence	Abstained	gpt-5.6-luna	93%	Neither dossier provides sufficiently specific, reproducible evidence that direct modulation of the target reverses disease-relevant phenotypes in bona fide NSCLC patient-derived xenografts. ALK has direct knockdown evidence in NSCLC models and mentions a patient-derived context, but the supplied text does not establish a PDX tumor-response experiment; its in vivo perturbational evidence is primarily resistance or dependency-context evidence rather than demonstrated disease reversal (C1-EV_PERTURBATION_0001, C1-EV_PERTURBATION_0002, C1-EV_SAFETY_0002). EGFR includes a catalogue-level assertion that preclinical validation studies used xenograft and PDX models, but no specific PDX perturbation result, target-specific tumor phenotype, or reproducible study details are supplied (C2-CAT-ONCO-BIO-010, C2-CAT-ONCO-BIO-117b). Therefore there is no adequate basis to compare the candidates under this criterion.	Direct, reproducible ALK or EGFR perturbation in bona fide NSCLC patient-derived xenografts demonstrating reversal of tumor-relevant phenotypes, ideally with orthogonal perturbations and replicated models.	–
ALK — KRAS	Safety and essentiality	A wins	gpt-5.6-luna	74%	ALK has the more favourable anticipated on-target window, but only narrowly. The ALK knockout mouse was viable and fertile, with specific reproductive and neuroendocrine phenotypes rather than broad developmental or adult-organ failure, and the dossier reports narrow adult expression (C1-EV_SAFETY_0004, C1-EV_SAFETY_0001). By contrast, KRAS is broadly expressed across normal tissues and cell types (C2-CAT-ONCO-BIO-047), is not universally dispensable, and complete loss produces embryonic lethality and adult hematopoietic defects (C2-CAT-ONCO-BIO-052, C2-CAT-ONCO-BIO-054, C2-CAT-ONCO-BIO-114). Mutant-selective KRAS inhibitors can spare wild-type KRAS, which mitigates but does not eliminate this on-target concern (C2-EV_SAFETY_0001). ALK clinical adverse events are class-, molecule-, and modality-associated and therefore were not treated as direct evidence of intrinsic ALK essentiality (C1-EV_SAFETY_0005, C1-EV_SAFETY_0006, C1-EV_SAFETY_0007). The absence of human ALK LoF and quantitative comparative essentiality data limits confidence.	Human ALK loss-of-function phenotype data and quantitative normal-tissue ALK expression breadth; direct comparative human or pan-essentiality data for ALK versus KRAS under mutation-selective versus nonselective inhibition.	Wild-type ALK appears to have limited essentiality for survival of adult normal tissues, although loss of ALK produces specific reproductive and neuroendocrine phenotypes in mice. Favourable on-target safety In welchen krankheitsrelevanten Zelltypen wird KRAS exprimiert? Ist KRAS üblicherweise essenziell? Ist KRAS für das Überleben oder die Proliferation von Krebszellen im Vergleich zu normalen Zellen essenziell? Welche Gewebe oder Organsysteme sind bei der Maus nach dem Knockout des KRAS-Gens betroffen? Mutant-selective inhibitors spare wild type ALK inhibition has a generally manageable clinical safety profile but carries clinically important risks, including hepatotoxicity, interstitial lung disease/pneumonitis, QT prolongation, bradycardia, severe visual loss, and embryo-fetal toxicity. Alectinib has clinically important hepatotoxicity and other organ toxicities that complicate the safety case for ALK inhibition. First-line lorlatinib produces a high burden of metabolic, neurologic and other adverse events, with more grade 3–4 toxicity than crizotinib.
KRAS — EGFR	Evidence robustness (meta-evidence)	A wins	gpt-5.6-luna	72%	FIRST has the more robustly documented meta-evidence package: it includes distinct, provenance-labeled human genomics sources, including a primary untreated cohort, an independent prospective clinical-sequencing cohort, and a large retrospective cohort (C1-EV_GENETICS_0002, C1-EV_GENETICS_0003, C1-EV_GENETICS_0004), plus a separately labeled mechanistic evidence item (C1-EV_MECHANISM_0001). SECOND includes strong claims of replication for EGFR driver mutations (C2-EV_GENETICS_0001), but provides less underlying source-family and cohort metadata. The margin is narrow because neither dossier adequately documents sample overlap, negative-result availability, publication-bias assessment, or replication in the exact stated experimental context.	Explicit sample-overlap analysis across the human genomics cohorts, including confirmation that the TCGA, MSK-IMPACT, and large retrospective NSCLC cohorts are independent. A structured source-family map and study-level provenance for both candidates, including methods for assessing publication bias and availability of negative or null replication results. Directly comparable replication and contradiction audits for EGFR and KRAS in the specified lung, mucus-producing glandular-cell, PDX, chemotherapy context.	KRAS is a recurrent, activating somatic driver in human lung adenocarcinoma, supporting KRAS inhibition as a genetically defined therapeutic strategy in NSCLC. Independent prospective clinical sequencing confirms KRAS as one of the dominant oncogenic drivers in metastatic human lung adenocarcinoma. KRAS mutations are common in human NSCLC, particularly lung adenocarcinoma, with KRAS G12C representing the largest mutation subtype and a genetically defined inhibitor-sensitive population. Constitutive RAS-MAPK activation
KRAS — EGFR	Resistance and escape liability	A wins	gpt-5.6-luna	72%	Applying the reversed direction, KRAS is the less vulnerable candidate, but only narrowly. Both targets have evidence for pathway redundancy, feedback or compensatory signaling, clonal evolution, and acquired resistance. KRAS inhibition can be bypassed through RTK–SHP2–SOS–RAS reactivation and parallel survival pathways, and resistance can arise through secondary switch-II-pocket mutations, activating alterations in trans, or amplification of the inhibited allele (C1-CAT-ONCO-BIO-013; C1-CAT-ONCO-BIO-063). EGFR inhibition has a broader upstream receptor-network escape landscape, including MET/HGF and other bypass receptors, downstream reactivation, and recurrent acquired EGFR alterations such as T790M and C797S, with documented clonal evolution under treatment (C2-CAT-ONCO-BIO-004; C2-CAT-ONCO-BIO-013; C2-CAT-ONCO-BIO-063; C2-CAT-ONCO-BIO-066). Combination strategies could mitigate risk for both: KRAS combinations with SHP2, SOS1, RTK, or downstream-pathway inhibitors are supported in the dossier (C1-CAT-ONCO-BIO-015; C1-CAT-ONCO-BIO-021), while EGFR combinations could address MET/HER-family or other bypass signaling (C2-CAT-ONCO-BIO-004; C2-CAT-ONCO-BIO-088). Biomarker strategies would also mitigate risk by selecting the specific KRAS allele or EGFR activating/resistance genotype and using longitudinal molecular monitoring to identify emerging resistant clones. The conclusion is limited because the dossiers do not provide a direct comparative resistance-rate analysis.	A head-to-head, matched NSCLC clinical comparison of resistance frequency, time to resistance, and clonal escape mechanisms under equivalent KRAS- versus EGFR-directed treatment. Variant- and biomarker-stratified longitudinal resistance data in the specified patient-derived xenograft and mucus-producing glandular-cell context.	Gibt es biochemische Signalwege, die die Hemmung von KRAS umgehen und dadurch zu einer Resistenz führen können? Ist KRAS von Resistenzmutationen betroffen? Gibt es Hinweise auf eine klonale Evolution unter Beteiligung von Veränderungen an KRAS während des Krankheitsverlaufs? Gibt es kompensatorische Mechanismen, die die Auswirkungen der Hemmung von EGFR abmildern können? Gibt es biochemische Signalwege, die die Hemmung von EGFR umgehen und dadurch zu einer Resistenz führen können? Ist EGFR von Resistenzmutationen betroffen? Gibt es Hinweise auf eine klonale Evolution unter Beteiligung von Veränderungen an EGFR während des Krankheitsverlaufs?
EGFR — KRAS	Resistance and escape liability	B wins	gpt-5.6-luna	68%	SECOND (KRAS) is judged slightly less vulnerable, with narrow margin. EGFR has multiple documented escape routes, including MET/HGF and other bypass signalling, downstream pathway activation, acquired T790M or C797S resistance mutations, and therapy-driven clonal evolution (C1-CAT-ONCO-BIO-004; C1-CAT-ONCO-BIO-013; C1-CAT-ONCO-BIO-063; C1-CAT-ONCO-BIO-066). KRAS also has substantial vulnerability: RTK–SHP2–SOS–RAS feedback, pathway reactivation, secondary switch-II-pocket mutations, allele switching, and KRAS amplification are documented (C2-CAT-ONCO-BIO-004; C2-CAT-ONCO-BIO-013; C2-CAT-ONCO-BIO-063). However, KRAS inhibition directly targets a downstream oncogenic signal hub, so escape through upstream EGFR or other RTK redundancy is less directly sufficient unless it reactivates RAS signalling; this supports a narrow advantage for KRAS, not a clear one. Combination strategies would mitigate risk for both: EGFR blockade could be paired with bypass-pathway inhibitors such as MET-, HER3-, or downstream-pathway inhibitors, while KRAS inhibition could be combined with SHP2, SOS1, or RTK blockade (C1-CAT-ONCO-BIO-004; C2-CAT-ONCO-BIO-015). Biomarker strategies would also mitigate risk by selecting the relevant EGFR activating mutation or KRAS allele, and by serially monitoring emergent resistance alterations or subclones (C1-CAT-ONCO-BIO-063; C1-CAT-ONCO-BIO-066; C2-CAT-ONCO-BIO-063; C2-CAT-ONCO-BIO-067).	A head-to-head, same-context NSCLC comparison of escape frequency, durability, and resistance burden for EGFR versus KRAS inhibition, ideally stratified by biomarker and treatment combination.	Gibt es kompensatorische Mechanismen, die die Auswirkungen der Hemmung von EGFR abmildern können? Gibt es biochemische Signalwege, die die Hemmung von EGFR umgehen und dadurch zu einer Resistenz führen können? Ist EGFR von Resistenzmutationen betroffen? Gibt es Hinweise auf eine klonale Evolution unter Beteiligung von Veränderungen an EGFR während des Krankheitsverlaufs? Gibt es kompensatorische Mechanismen, die die Auswirkungen der Hemmung von KRAS abmildern können? Gibt es biochemische Signalwege, die die Hemmung von KRAS umgehen und dadurch zu einer Resistenz führen können? Ist KRAS von Resistenzmutationen betroffen? Bestehen Sicherheitenabhängigkeiten mit KRAS?
MET — EGFR	Translational model validity	A wins	gpt-5.6-luna	78%	MET has direct pharmacologic efficacy evidence in six MET-amplified NSCLC patient-derived xenograft models, plus an independent set of six EGFR-mutant/MET-amplified patient-derived lung-cancer models showing in vivo sensitivity to MET inhibition (C1-EV_PERTURBATION_0002, C1-EV_PERTURBATION_0003). These are human-derived in vivo models directly relevant to NSCLC. EGFR has evidence describing genetically engineered EGFR-mutant mouse models and broad claims that EGFR tool compounds have been tested in PDX and other models, but the supplied dossier provides less specific, less directly documented efficacy evidence in the relevant human-derived models (C2-CAT-ONCO-BIO-010, C2-CAT-ONCO-BIO-117b). Neither dossier establishes that the models represent newly diagnosed disease or the specified mucus-producing glandular cell context, so the advantage is narrow.	Direct head-to-head efficacy data in treatment-naive, newly diagnosed NSCLC patient-derived models, ideally involving mucus-producing glandular cells or matched patient organoids, for both MET and EGFR.	Pharmacologic MET inhibition suppresses MET-dependent lung cancer models in vitro and in vivo, with sensitivity linked to MET signaling complexes. A subset of EGFR-mutant, MET-amplified NSCLC models becomes dependent on MET and responds to MET inhibition alone. Gibt es veröffentlichte Studien, in denen die verfügbaren Tool-Substanzen zur Validierung von EGFR in präklinischen Modellen eingesetzt wurden? Welche Ergebnisse wurden in den einzelnen Tiermodellen mit EGFR erzielt?
ALK — MET	Resistance and escape liability	B wins	gpt-5.6-luna	84%	Direction is reversed: the stronger target is the one least affected by escape. SECOND (MET) is favored narrowly because the ALK dossier contains direct clinical and mechanistic evidence of substantial acquired resistance liability: sequential ALK-TKI treatment selects compound ALK kinase-domain mutations, and progression can occur through both on-target mutations and non-ALK bypass mechanisms (C1-EV_TRACTABILITY_0007; C1-EV_BIOMARKER_0004). Genome-wide screens further identify EGFR/MAPK and PI3K/AKT/mTOR bypass through MIG6/ERRFI1 loss (C1-EV_PERTURBATION_0002; C1-EV_PERTURBATION_0005). MET also has meaningful escape liabilities, including pre-existing RAS–MAPK alterations, assay and intratumoral heterogeneity, and pathway redundancy (C2-EV_GENETICS_0005; C2-EV_BIOMARKER_0006; C2-EV_MECHANISM_0004; C2-EV_MECHANISM_0005). Therefore this is narrow rather than clear, and is not based solely on absence of MET resistance data. Biomarker strategies would partially mitigate risk for both targets: ALK fusion/variant, resistance-mutation, and longitudinal ctDNA selection can improve matching, while METex14 and high-level amplification selection can restrict treatment to more dependent tumors. Combination strategies may also mitigate bypass, such as EGFR blockade for MIG6-deficient ALK models and combined EGFR/MET inhibition for MET-mediated EGFR resistance, but the supplied evidence does not establish that these approaches eliminate escape in the stated NSCLC context.	A matched, prospective clinical comparison of acquired on-target and bypass resistance rates for MET-directed versus ALK-directed therapy, including longitudinal clonal evolution under comparable treatment exposure.	The ALK small-molecule modality is vulnerable to acquired on-target resistance, particularly after sequential inhibitor treatment. ALK positivity does not ensure durable dependence on ALK signaling because resistance can arise through both on-target and off-target mechanisms. Genome-wide CRISPR knockout screens identify bypass alterations that permit survival under ALK inhibition, demonstrating context-dependent ALK-TKI dependence. Genome-wide CRISPR knockout identified MIG6/ERRFI1 loss as a mechanism of resistance to ALK TKIs through EGFR bypass signaling. Co-occurring RAS–MAPK pathway alterations are common in MET exon 14-altered NSCLC and can limit response to MET-directed therapy. MET biomarker classification is substantially affected by intratumoral heterogeneity and discordance between assays. Co-occurring RAS–MAPK pathway alterations are common in MET exon 14-altered NSCLC and can limit response to MET-directed therapy. MET amplification can provide a bypass resistance pathway to EGFR inhibition by activating ERBB3-dependent PI3K signaling. Both MET amplification and HGF-driven MET activation can mediate resistance to EGFR inhibitors through distinct downstream signaling routes. Even in MET exon 14-altered NSCLC, MET protein expression can be absent or heterogeneous, weakening the assumption that the alteration marks a uniformly MET-expressing tumor-cell population. MET perturbational dependency in NSCLC is restricted to a molecularly defined subset rather than being general across lung cancer models.
MET — KRAS	Human genetics and causal evidence	B wins	gpt-5.6-luna	82%	SECOND has stronger supplied human-genetic evidence for a causal NSCLC role. The dossier documents a recurrent activating allelic series, high prevalence in lung adenocarcinoma, and replication across a TCGA cohort, an independent prospective metastatic cohort, and a large clinical sequencing cohort (C2-EV_GENETICS_0001 through C2-EV_GENETICS_0004). MET is also supported by recurrent, directly characterized somatic driver alterations and multiple METex14 variants (C1-EV_GENETICS_0002 through C1-EV_GENETICS_0004), but its evidence is narrower and less frequent, with adverse prognostic heterogeneity and resistance-related caveats (C1-EV_GENETICS_0005, C1-EV_GENETICS_0007). Neither dossier supplies GWAS/fine-mapping, L2G scores, molecular-QTL colocalisation, ancestry-diverse replication, or biobank/rare-variant evidence, so the margin is narrow and the conclusion rests mainly on replicated somatic driver genomics and allelic-series evidence rather than the full stated rubric.	For both candidates: GWAS/fine-mapping or L2G variant-to-gene assignment, disease-relevant molecular-QTL colocalisation, ancestry-diverse replication, and population-level rare-variant or biobank evidence. For MET: an independent germline or population-genetic causal association beyond recurrent somatic tumor alterations. For KRAS: direct NSCLC-specific QTL colocalisation and ancestry-diverse causal association evidence.	Recurrent activating mutations KRAS is a recurrent, activating somatic driver in human lung adenocarcinoma, supporting KRAS inhibition as a genetically defined therapeutic strategy in NSCLC. Independent prospective clinical sequencing confirms KRAS as one of the dominant oncogenic drivers in metastatic human lung adenocarcinoma. KRAS mutations are common in human NSCLC, particularly lung adenocarcinoma, with KRAS G12C representing the largest mutation subtype and a genetically defined inhibitor-sensitive population. TCGA-integrated human tumor sequencing identified recurrent somatic MET exon 14 skipping as a likely driver alteration in lung adenocarcinoma. Diverse somatic MET exon 14 splice-site alterations occur in human lung cancers and define a recurrent genomic driver class. A large human lung-cancer profiling cohort confirmed that METex14 alterations are distributed across major NSCLC histologies and occur in 2.7% of profiled lung cancers. Co-occurring RAS–MAPK pathway alterations are common in MET exon 14-altered NSCLC and can limit response to MET-directed therapy. MET alterations are uncommon in NSCLC and MET exon 14 mutations may be associated with worse prognosis rather than uniformly favorable biology.
KRAS — MET	Druggability and tractability	Tie	gpt-5.6-luna	80%	Both targets have strong, modality-specific small-molecule tractability evidence in NSCLC. KRAS has a structurally characterized druggable switch-II pocket, mutation-specific covalent chemistry, available tool compounds, and clinically validated inhibitors (C1-EV_TRACTABILITY_0001; C1-CAT-ONCO-BIO-032; C1-CAT-ONCO-BIO-115). MET has an established small-molecule kinase-binding site, multiple approved oral MET TKIs, and direct clinical validation in genotype-selected MET exon 14-skipping NSCLC (C2-EV_TRACTABILITY_0001; C2-EV_TRACTABILITY_0002; C2-EV_TRACTABILITY_0003; C2-EV_TRACTABILITY_0004). MET tractability is less uniform for amplification and is limited by biomarker thresholding (C2-EV_TRACTABILITY_0007; C2-EV_CLINICAL_0006), while KRAS has allele-specific resistance and broader mutation-dependent requirements. The supplied evidence does not establish a decisive advantage for either target on lung delivery or the specified mucus-producing glandular cell type, so the supported conclusion is a positive, comparable tie rather than abstention.	Head-to-head experimental structures and quantitative pocket-quality metrics for KRAS versus MET kinase-domain binding sites. Small-molecule exposure, target engagement, and efficacy data specifically in mucus-producing glandular lung cells or the stated lung PDX context. Comparative data on selectivity, intracellular accessibility, protein turnover, and lung-tissue delivery for the two targets.	Clinically validated tractability Weist KRAS Mutationen auf, die neue medikamentös adressierbare Bindungsstellen schaffen? Welche Toolverbindungen für KRAS sind verfügbar? Approved MET TKIs MET is clinically tractable in NSCLC through genotype-selected, oral small-molecule kinase inhibition of MET exon 14–skipping tumors. Capmatinib produced substantial responses in MET exon 14–skipping NSCLC, providing direct clinical proof of target tractability for this biomarker-defined subgroup. A second selective oral MET inhibitor, tepotinib, has demonstrated sufficient efficacy for traditional approval in MET exon 14–skipping metastatic NSCLC. Broad MET amplification is less reliably tractable than MET exon 14 skipping and requires biomarker refinement, particularly regarding amplification level. MET amplification has weaker and threshold-dependent clinical validation than MET exon 14 skipping in NSCLC.
MET — ALK	Disease biology and mechanism coherence	B wins	gpt-5.6-luna	84%	Both targets have substantial evidence supporting an inhibition mechanism in NSCLC. ALK has the more internally coherent causal chain: a recurrent EML4–ALK fusion is shown to transform cells, generate constitutive kinase signaling, activate defined downstream pathways, and support tumor-cell proliferation and survival; direct ALK knockdown selectively reduces viability and clonogenic growth in fusion-positive models. MET also has a strong mechanism, particularly for MET exon 14 skipping, where loss of the CBL-binding region causes impaired receptor turnover, sustained signaling, and selective dependency in MET-amplified models and PDXs. However, MET dependency is more explicitly restricted to molecular subsets and is weakened by alteration- and expression-dependent heterogeneity. ALK likewise has variant and bypass-resistance limitations, so the advantage is narrow rather than clear.	Directly matched mechanistic perturbation and patient-derived xenograft evidence in mucus-producing glandular NSCLC cells, comparing MET and ALK inhibition under the stated biomarker context.	Causal oncogenic fusions A recurrent EML4–ALK fusion is a transforming somatic genetic alteration in human NSCLC. ALK rearrangements in NSCLC create oncogenic ALK fusion proteins whose constitutive signaling promotes tumor-cell proliferation and survival. The recurrent EML4-ALK fusion is a transforming oncogenic driver in NSCLC. RAS–RAF–MEK–ERK/MAPK is the primary downstream survival pathway in EML4-ALK-positive lung adenocarcinoma models. Direct ALK knockdown selectively impairs viability and clonogenic growth of EML4-ALK-positive NSCLC cells. TCGA-integrated human tumor sequencing identified recurrent somatic MET exon 14 skipping as a likely driver alteration in lung adenocarcinoma. MET exon 14 skipping activates MET by removing the juxtamembrane CBL-binding region, impairing receptor ubiquitination and down-regulation. MET knockdown selectively impairs growth and survival of MET-amplified NSCLC cell lines. Pharmacologic MET inhibition suppresses MET-dependent lung cancer models in vitro and in vivo, with sensitivity linked to MET signaling complexes. ALK rearrangement does not uniformly establish ALK pathway dependence across NSCLC models. ALK fusion variants produce materially different pathway dependency and inhibitor sensitivity. CRISPR-based in vivo perturbation shows that EML4-ALK variants are functionally heterogeneous rather than a uniform dependency context. Co-occurring RAS–MAPK pathway alterations are common in MET exon 14-altered NSCLC and can limit response to MET-directed therapy. MET perturbational dependency in NSCLC is restricted to a molecularly defined subset rather than being general across lung cancer models. CRISPR MET knockout in A549 NSCLC cells did not materially impair intrinsic proliferation, despite reducing dissemination-related phenotypes.
ALK — KRAS	Integrated therapeutic hypothesis	A wins	gpt-5.6-luna	90%	ALK represents the stronger integrated target–disease–modality hypothesis because the dossier directly connects ALK rearrangement to NSCLC biology, provides a validated predictive biomarker and established oral small-molecule modality, and includes multiple randomized first-line clinical studies showing superiority or durable benefit over chemotherapy or crizotinib (C1-EV_BIOMARKER_0001, C1-EV_BIOMARKER_0002, C1-EV_CLINICAL_0001, C1-EV_CLINICAL_0002, C1-EV_CLINICAL_0004). Its main limitation is differentiation in a crowded ALK-TKI segment, together with resistance and toxicity concerns. KRAS has greater mutation prevalence and clinically validated small-molecule tractability, particularly for G12C (C2-EV_GENETICS_0002, C2-EV_GENETICS_0004, C2-EV_TRACTABILITY_0001), but the supplied evidence is less directly relevant to newly diagnosed patients, chemotherapy comparison, the specified PDX setting, and a differentiated G12C/G12D medicine. Much of the KRAS support is catalogue-level, generic, or explicitly context-dependent. Therefore the direct clinical coherence of ALK outweighs KRAS's broader but less mature and less first-line-specific hypothesis.	Q1-Q19 criterion scores and the criterion-weighted composite required by the Q20 guardrail A randomized first-line NSCLC study directly testing a KRAS-mutant inhibitor against chemotherapy in newly diagnosed patients Direct in vivo PDX efficacy evidence for either target in the specified mucus-producing glandular-cell context	ALK is clinically validated as a predictive therapeutic biomarker in advanced NSCLC by randomized phase 3 evidence showing superior efficacy of ALK inhibition over platinum-pemetrexed chemotherapy. Second-generation ALK inhibition provides superior systemic and intracranial disease control compared with first-generation therapy in untreated advanced ALK-positive NSCLC. Lorlatinib provides durable long-term clinical benefit in treatment-naive advanced ALK-positive NSCLC, with sustained systemic and intracranial disease control. ALK rearrangement is a validated predictive biomarker for ALK-targeted therapy in metastatic NSCLC. In untreated advanced ALK-positive NSCLC, biomarker-selected patients derived substantially greater benefit from ALK inhibition with crizotinib than from platinum-pemetrexed chemotherapy. ALK is clinically tractable with an established oral small-molecule kinase-inhibitor modality in metastatic NSCLC. ALK-targeted therapy has an established companion-diagnostic infrastructure, supporting practical biomarker selection and deployment of the modality. Direct ALK knockdown selectively impairs viability and clonogenic growth of EML4-ALK-positive NSCLC cells. KRAS is a recurrent, activating somatic driver in human lung adenocarcinoma, supporting KRAS inhibition as a genetically defined therapeutic strategy in NSCLC. KRAS mutations are common in human NSCLC, particularly lung adenocarcinoma, with KRAS G12C representing the largest mutation subtype and a genetically defined inhibitor-sensitive population. Clinically validated tractability ALK-positive NSCLC is a highly competitive targeted-therapy segment with three preferred first-line ALK TKIs—alectinib, brigatinib, and lorlatinib—and no direct randomized head-to-head trials establishing superiority among them. Competition is extending toward fourth-generation ALK inhibitors designed to address compound resistance mutations that can emerge after lorlatinib. ALK inhibition has a generally manageable clinical safety profile but carries clinically important risks, including hepatotoxicity, interstitial lung disease/pneumonitis, QT prolongation, bradycardia, severe visual loss, and embryo-fetal toxicity. Alectinib has clinically important hepatotoxicity and other organ toxicities that complicate the safety case for ALK inhibition. ALK positivity does not ensure durable dependence on ALK signaling because resistance can arise through both on-target and off-target mechanisms. Mutant-selective inhibitors spare wild type Gibt es kompensatorische Mechanismen, die die Auswirkungen der Hemmung von KRAS abmildern können? Gibt es biochemische Signalwege, die die Hemmung von KRAS umgehen und dadurch zu einer Resistenz führen können? Ist KRAS üblicherweise essenziell? Ist KRAS selektiv für die Indikation {Disease}?
KRAS — ALK	Time to decisive experiment	Abstained	gpt-5.6-luna	98%	The dossiers contain evidence of reagent/modality availability and some preclinical model use for both candidates, including KRAS tool compounds and PDX-related studies (C1-CAT-ONCO-BIO-010, C1-CAT-ONCO-BIO-115) and approved ALK inhibitors with companion-diagnostic infrastructure (C2-EV_TRACTABILITY_0001, C2-EV_TRACTABILITY_0006). However, they do not provide organisation-specific capacity, historical cycle times, operational readiness for the specified PDX context, or comparative costs. Without those criterion-specific data, the faster and lower-cost target cannot be adjudicated.	Organisation-specific internal capability and capacity for KRAS versus ALK Historical cycle times from assay setup through PDX in vivo proof-of-concept readout Procurement, reagent, animal-model, and study costs Availability and readiness of PDX models matching mucus-producing glandular cells, newly diagnosed NSCLC, genomic biomarker status, and chemotherapy backbone	–
ALK — MET	Evidence robustness (meta-evidence)	A wins	gpt-5.6-luna	76%	ALK has evidence from multiple separately identified randomized clinical trials and includes both supportive and contradictory findings, including a negative unadjusted overall-survival result and documented resistance or nonuniform dependency. MET also has supportive and negative findings, but its dossier explicitly identifies extensive dependence among major clinical, biomarker, safety, and competitive evidence clusters, with several conclusions drawn from GEOMETRY mono-1, VISION, and LUMINOSITY. Therefore ALK has the more robust package against source dependence and contradiction, but the margin is narrow because explicit overlap and publication-bias metadata are incomplete for both candidates.	A complete, candidate-symmetric provenance and sample-overlap audit, including explicit cohort-sharing and publication-bias assessments for the ALK evidence family and independent replication details for each clinical result.	ALK is clinically validated as a predictive therapeutic biomarker in advanced NSCLC by randomized phase 3 evidence showing superior efficacy of ALK inhibition over platinum-pemetrexed chemotherapy. Second-generation ALK inhibition provides superior systemic and intracranial disease control compared with first-generation therapy in untreated advanced ALK-positive NSCLC. Brigatinib demonstrated superior first-line efficacy, including intracranial activity, versus crizotinib in ALK-positive advanced NSCLC. Lorlatinib provides durable long-term clinical benefit in treatment-naive advanced ALK-positive NSCLC, with sustained systemic and intracranial disease control. Ensartinib provides regulatory-confirmed clinical validation of ALK as a treatment target in previously untreated locally advanced or metastatic NSCLC. The pivotal PROFILE 1014 trial did not demonstrate a statistically significant overall-survival benefit for first-line crizotinib in the unadjusted analysis. ALK rearrangement does not uniformly establish ALK pathway dependence across NSCLC models. ALK fusion variants produce materially different pathway dependency and inhibitor sensitivity. MET exon 14 skipping is clinically validated as a predictive biomarker in metastatic NSCLC through durable responses to selective MET inhibitors. Tepotinib has traditional FDA approval for metastatic NSCLC harboring MET exon 14 skipping alterations, with efficacy confirmed in a 313-patient dataset. MET immunohistochemical positivity did not identify a clinically responsive NSCLC population in the phase III METLung trial. Adding the MET inhibitor tivantinib to erlotinib failed to improve overall survival in the phase III ATTENTION trial. The pivotal PROFILE 1014 trial did not demonstrate a statistically significant overall-survival benefit for first-line crizotinib in the unadjusted analysis. ALK rearrangement does not uniformly establish ALK pathway dependence across NSCLC models. ALK fusion variants produce materially different pathway dependency and inhibitor sensitivity. MET immunohistochemical positivity did not identify a clinically responsive NSCLC population in the phase III METLung trial. Adding the MET inhibitor tivantinib to erlotinib failed to improve overall survival in the phase III ATTENTION trial. MET amplification has weaker and threshold-dependent clinical validation than MET exon 14 skipping in NSCLC.
MET — ALK	Perturbational evidence	A wins	gpt-5.6-luna	91%	MET has explicit pharmacologic perturbation evidence in six MET-amplified lung-cancer patient-derived xenograft models, with MET-inhibition-associated signaling suppression and antitumor responses, plus concordant activity across additional patient-derived models in vivo (C1-EV_PERTURBATION_0002, C1-EV_PERTURBATION_0003). The ALK dossier contains direct knockdown evidence in cell models and CRISPR resistance-screen evidence, but does not provide comparably direct ALK modulation producing disease-relevant responses in multiple patient-derived xenografts. MET evidence is therefore stronger for this specific PDX perturbational criterion, despite limited orthogonal genetic validation in PDX.	Direct ALK knockdown or knockout in multiple ALK-positive patient-derived xenografts with tumor-regression or disease-control readouts, ideally with rescue or an orthogonal perturbation. Orthogonal genetic MET perturbation validated in MET-dependent patient-derived xenografts.	Pharmacologic MET inhibition suppresses MET-dependent lung cancer models in vitro and in vivo, with sensitivity linked to MET signaling complexes. A subset of EGFR-mutant, MET-amplified NSCLC models becomes dependent on MET and responds to MET inhibition alone. CRISPR MET knockout in A549 NSCLC cells did not materially impair intrinsic proliferation, despite reducing dissemination-related phenotypes. Genome-wide CRISPR knockout screens identify bypass alterations that permit survival under ALK inhibition, demonstrating context-dependent ALK-TKI dependence. Genome-wide CRISPR knockout identified MIG6/ERRFI1 loss as a mechanism of resistance to ALK TKIs through EGFR bypass signaling.
MET — KRAS	Safety and essentiality	B wins	gpt-5.6-luna	66%	KRAS has broad physiologic expression and documented essentiality during development and in hematopoietic compartments, which limits the margin and makes the advantage narrow rather than clear. However, the supplied KRAS safety evidence indicates that mutant-selective covalent inhibitors spare wild-type KRAS, providing a mechanistic basis for a more favourable anticipated on-target profile under small-molecule inhibition. MET has a documented clinical toxicity burden across selective inhibitors, including edema and serious pulmonary, hepatic, pancreatic, and embryo-fetal toxicities. Some of these effects may be molecule- or modality-specific rather than purely on-target, so they should not all be attributed to MET mechanism; nevertheless, the class-level clinical safety burden is more directly evidenced than for KRAS. The comparison therefore modestly favours KRAS, with substantial uncertainty due to incomplete comparative human safety and essentiality data.	Direct human LoF and pan-essentiality data for MET in normal adult tissues. Comparative clinical safety data for mutation-selective KRAS inhibitors, including allele-specific and tissue-specific toxicity. Direct evidence separating on-target KRAS toxicities from molecule-specific or modality-specific effects.	Mutant-selective inhibitors spare wild type In welchen krankheitsrelevanten Zelltypen wird KRAS exprimiert? Ist KRAS üblicherweise essenziell? Ist KRAS für das Überleben oder die Proliferation von Krebszellen im Vergleich zu normalen Zellen essenziell? Clinical targeting of MET has substantial but monitorable toxicity, including edema, interstitial lung disease, hepatotoxicity, pancreatic toxicity, and embryo-fetal toxicity. A second selective MET inhibitor confirms both target dependence and a clinically relevant toxicity burden in MET exon 14–skipping NSCLC. Ist KRAS üblicherweise essenziell? Ist KRAS für das Überleben oder die Proliferation von Krebszellen im Vergleich zu normalen Zellen essenziell? Tepotinib has clinically important, potentially fatal pulmonary toxicity in MET-altered NSCLC. Tepotinib safety includes hepatotoxicity, pancreatic enzyme elevations and embryo-fetal toxicity requiring monitoring or treatment modification.
KRAS — MET	Combination potential	Abstained	gpt-5.6-luna	96%	Neither dossier provides direct evidence for combination with chemotherapy. KRAS has preclinical combination and resistance evidence involving SOS1/SHP2 and related pathway feedback, while MET has in vivo combination evidence with EGFR inhibition and clinically documented toxicity; these are not chemotherapy combinations and cannot establish the requested rationale. The required chemotherapy-specific combination efficacy and overlapping-toxicity evidence are missing for both candidates, so ranking would exceed the dossier evidence.	Direct in vivo combination studies of each target inhibitor with chemotherapy in NSCLC, including efficacy, pathway complementarity, and resistance-mechanism characterization. A prospective or experimentally supported overlapping-toxicity assessment for each target inhibitor combined with the specified chemotherapy backbone.	Gibt es kompensatorische Mechanismen, die die Auswirkungen der Hemmung von KRAS abmildern können? Bestehen Sicherheitenabhängigkeiten mit KRAS? Both MET amplification and HGF-driven MET activation can mediate resistance to EGFR inhibitors through distinct downstream signaling routes. Clinical targeting of MET has substantial but monitorable toxicity, including edema, interstitial lung disease, hepatotoxicity, pancreatic toxicity, and embryo-fetal toxicity.
ALK — KRAS	Time to decisive experiment	Abstained	gpt-5.6-luna	98%	The dossiers establish that both targets have small-molecule tools and some diagnostic or preclinical model resources (C1-EV_TRACTABILITY_0001, C1-EV_TRACTABILITY_0006, C2-EV_TRACTABILITY_0001, C2-CAT-ONCO-BIO-010, C2-CAT-ONCO-BIO-115), but they do not provide the organisation-specific readiness, capacity, historical cycle times, or costs needed to determine which target reaches a high-information in vivo go/no-go decision faster and more cheaply. Evidence of biological or clinical tractability cannot substitute for this learning-efficiency criterion.	Comparative assay and reagent availability specifically for the required NSCLC mucus-producing glandular-cell and PDX in vivo proof-of-concept workflow Evidence that the organisation has internal capability, capacity, and operational readiness to execute the ALK versus KRAS PDX study Historical end-to-end cycle times, study throughput, and per-study costs for comparable in vivo PDX proof-of-concept experiments Evidence that the available ALK or KRAS tools produce a rapid, interpretable pharmacodynamic and efficacy readout in the specified model	–
ALK — KRAS	Combination potential	Abstained	gpt-5.6-luna	96%	Neither dossier supplies the criterion-critical evidence needed to adjudicate combination with chemotherapy: direct chemotherapy combination preclinical studies and an overlapping-toxicity or co-dosing assessment. ALK includes resistance-mechanism data and clinically important target-inhibitor toxicities, while KRAS includes general pathway-bypass and combination-vulnerability discussion, but these do not establish chemotherapy complementarity or safe combined dosing. Under the rubric and guardrail, the appropriate judgment is abstention rather than ranking one candidate on indirect evidence.	For both candidates: direct NSCLC combination studies pairing the target inhibitor with chemotherapy, preferably in PDX or other in vivo models, with efficacy, scheduling, and resistance analyses. For both candidates: formal overlapping-toxicity, tolerability, and pharmacokinetic/pharmacodynamic assessment for concurrent dosing with the specified chemotherapy backbone. For ALK: chemotherapy-specific evidence connecting documented bypass mechanisms such as EGFR–MAPK/PI3K signaling to improved combination activity; resistance evidence is present but not chemotherapy-combination evidence (C1-EV_PERTURBATION_0002, C1-EV_PERTURBATION_0005). For KRAS: chemotherapy-specific pathway-complementarity and in vivo combination evidence; the supplied combination-related material concerns KRAS-pathway partners rather than chemotherapy (C2-CAT-ONCO-BIO-004, C2-CAT-ONCO-BIO-013, C2-CAT-ONCO-BIO-015).	–
ALK — MET	Biomarker and target engagement	A wins	gpt-5.6-luna	80%	ALK offers the more credible overall clinical biomarker strategy because ALK rearrangement is an FDA-recognized predictive biomarker with an established companion-diagnostic infrastructure, including FDA-authorized testing, and randomized phase 3 evidence of superior efficacy over platinum-pemetrexed chemotherapy in biomarker-selected patients (C1-EV_BIOMARKER_0001, C1-EV_BIOMARKER_0002, C1-EV_TRACTABILITY_0006). Longitudinal plasma ALK mutation kinetics provide clinically linked pharmacodynamic evidence consistent with on-target inhibition (C1-EV_BIOMARKER_0005). METex14 also has strong clinical validation, FDA-approved selection, prospective tissue/plasma testing, and ctDNA molecular-response evidence (C2-EV_BIOMARKER_0001, C2-EV_BIOMARKER_0002, C2-EV_CLINICAL_0003), and its pMET:total-MET assay is a notable direct target-engagement strength (C2-EV_BIOMARKER_0005). However, MET's clinical evidence is primarily phase 2 and is accompanied by substantial assay and biomarker heterogeneity, while ALK has stronger regulatory precedent and randomized comparative efficacy evidence. The margin is narrow because MET has more explicit direct target-engagement assay evidence, whereas ALK selection can be weakened by nonproductive FISH-positive rearrangements (C1-EV_BIOMARKER_0006) and resistance-related loss of dependence (C1-EV_BIOMARKER_0004).	Direct comparative evidence in newly diagnosed NSCLC and in the specified mucus-producing glandular-cell patient-derived xenograft context showing which biomarker strategy best predicts target engagement and early in vivo efficacy. A prospectively validated ALK-specific pharmacodynamic assay directly measuring target inhibition comparable to the validated pMET:total-MET assay.	ALK rearrangement is a validated predictive biomarker for ALK-targeted therapy in metastatic NSCLC. In untreated advanced ALK-positive NSCLC, biomarker-selected patients derived substantially greater benefit from ALK inhibition with crizotinib than from platinum-pemetrexed chemotherapy. Longitudinal plasma ALK mutation kinetics provide pharmacodynamic evidence consistent with on-target ALK inhibition by lorlatinib. ALK-targeted therapy has an established companion-diagnostic infrastructure, supporting practical biomarker selection and deployment of the modality. MET exon 14 skipping is a clinically validated predictive biomarker for selective MET inhibition in NSCLC. Serial depletion of METex14 ctDNA is an on-treatment pharmacodynamic biomarker associated with clinical response to tepotinib. Validated pharmacodynamic assays based on the pMET:total-MET ratio provide a direct biochemical readout of MET pathway inhibition and are suitable for target-engagement studies. Capmatinib is FDA-indicated for metastatic NSCLC with a mutation leading to MET exon 14 skipping, selected using an FDA-approved test. ALK positivity does not ensure durable dependence on ALK signaling because resistance can arise through both on-target and off-target mechanisms. ALK FISH positivity can represent a nonproductive rearrangement rather than an expressed, targetable ALK fusion. MET immunohistochemistry did not identify patients who benefited from onartuzumab in the randomized phase III METLung trial. MET biomarker classification is substantially affected by intratumoral heterogeneity and discordance between assays.
KRAS — MET	Translational model validity	B wins	gpt-5.6-luna	90%	MET has direct efficacy evidence in patient-derived NSCLC xenograft models, including six MET-amplified lung-cancer PDX models and additional patient-derived EGFR-mutant/MET-amplified models showing in vivo sensitivity to MET inhibition (C2-EV_PERTURBATION_0002, C2-EV_PERTURBATION_0003). Its clinical relevance is further supported by durable responses to selective MET inhibitors in treatment-naive MET exon 14–skipping NSCLC, including capmatinib and tepotinib datasets (C2-EV_CLINICAL_0001, C2-EV_CLINICAL_0002). KRAS has strong human-genetic and clinical tractability evidence, but the dossier does not provide comparably specific, high-quality KRAS efficacy results from admissible translational models in NSCLC or newly diagnosed disease. The MET evidence is biomarker- and alteration-dependent, and the negative or weaker findings for unselected MET models do not negate the stronger evidence in the defined MET-dependent subsets.	Direct KRAS efficacy data from NSCLC patient-derived xenografts, organoids, primary patient cells, genetically engineered or humanized in vivo models, specifically in treatment-naive/newly diagnosed settings. Direct MET-versus-KRAS comparative efficacy data using the same validated translational model and newly diagnosed treatment context.	Pharmacologic MET inhibition suppresses MET-dependent lung cancer models in vitro and in vivo, with sensitivity linked to MET signaling complexes. A subset of EGFR-mutant, MET-amplified NSCLC models becomes dependent on MET and responds to MET inhibition alone. MET exon 14 skipping is clinically validated as a predictive biomarker in metastatic NSCLC through durable responses to selective MET inhibitors. Tepotinib has traditional FDA approval for metastatic NSCLC harboring MET exon 14 skipping alterations, with efficacy confirmed in a 313-patient dataset. MET perturbational dependency in NSCLC is restricted to a molecularly defined subset rather than being general across lung cancer models. CRISPR MET knockout in A549 NSCLC cells did not materially impair intrinsic proliferation, despite reducing dissemination-related phenotypes. MET amplification has weaker and threshold-dependent clinical validation than MET exon 14 skipping in NSCLC.
ALK — MET	Time to decisive experiment	Abstained	gpt-5.6-luna	98%	The dossiers establish that both targets have clinically tractable small-molecule inhibitors and some preclinical or patient-derived model evidence (for example, C1-EV_TRACTABILITY_0001, C1-EV_PERTURBATION_0001, C2-EV_TRACTABILITY_0001, and C2-EV_PERTURBATION_0002), but they do not provide the operational evidence required to compare time or cost to a decisive in vivo proof-of-concept decision. In particular, model readiness, internal capacity, procurement or assay timelines, historical cycle times, and study costs are absent. Biological validation and modality tractability cannot substitute for these learning-efficiency measures.	Candidate-specific assay and reagent inventory or procurement timelines; readiness and availability of the specified NSCLC patient-derived xenograft models, including mucus-producing glandular-cell relevance; internal in vivo capability, staffing, throughput, and capacity; historical cycle times and costs for comparable proof-of-concept studies; predefined go/no-go decision endpoints and expected information yield.	–
ALK — KRAS	Perturbational evidence	A wins	gpt-5.6-luna	80%	ALK has direct perturbational evidence from ALK knockdown showing selective loss of viability and clonogenic growth in ALK-rearranged NSCLC models, plus CRISPR-based evidence in a patient-derived NSCLC context and pharmacologic in vivo dependency evidence. KRAS has strong genetic and mechanistic support, but the supplied perturbational dossier does not provide a clearly documented, reproducible PDX disease-reversal result: its explicit xenograft example is a human cell-line xenograft, while the PDX references are general or catalogue-level claims. The ALK evidence is therefore stronger, but the margin is narrow because the supplied ALK PDX result is truncated and includes context-dependent resistance evidence.	A directly documented, reproducible KRAS perturbation study showing disease-relevant reversal in a genuine patient-derived xenograft, with matched KRAS-mutant and KRAS-wild-type controls. A complete report of the ALK patient-derived model in C1-EV_PERTURBATION_0001, including whether it was a PDX and whether in vivo tumor regression—not only viability loss—was demonstrated.	Direct ALK knockdown selectively impairs viability and clonogenic growth of EML4-ALK-positive NSCLC cells. Genome-wide CRISPR knockout screens identify bypass alterations that permit survival under ALK inhibition, demonstrating context-dependent ALK-TKI dependence. ALK signaling is essential for proliferation and survival in at least a subset of EML4-ALK-positive NSCLC models. CRISPR loss of KEAP1 can produce resistance to ALK inhibition in ALK-mutant lung cancer models. Genome-wide CRISPR knockout identified MIG6/ERRFI1 loss as a mechanism of resistance to ALK TKIs through EGFR bypass signaling. CRISPR-based in vivo perturbation shows that EML4-ALK variants are functionally heterogeneous rather than a uniform dependency context. ALK rearrangement does not uniformly establish ALK pathway dependence across NSCLC models. ALK fusion variants produce materially different pathway dependency and inhibitor sensitivity.
MET — ALK	Human genetics and causal evidence	B wins	gpt-5.6-luna	78%	ALK has slightly stronger causal human-genetic evidence because the EML4–ALK fusion is directly assigned to ALK, has an explicitly stated direction of effect, was identified in human NSCLC, and was shown to transform cells; independent tumor-genomic profiling also reproduced recurrent ALK rearrangements (C2-EV_GENETICS_0001, C2-EV_GENETICS_0002, C2-EV_GENETICS_0003, C2-EV_MECHANISM_0003). MET also has strong somatic causal evidence: recurrent exon-14 splice alterations are gene-specific, mechanistically activate MET, and are replicated across profiling cohorts (C1-EV_GENETICS_0002, C1-EV_GENETICS_0003). The margin is narrow because neither dossier provides the GWAS, QTL colocalisation, rare-variant, biobank, or ancestry-diverse evidence explicitly requested, and the available evidence is primarily somatic tumor-genomic and functional rather than population human genetics. The ALK germline evidence concerns neuroblastic tumors rather than NSCLC and does not undermine the somatic NSCLC fusion evidence (C2-EV_GENETICS_0004, C2-EV_GENETICS_0007).	No GWAS or fine-mapping results, L2G-style variant-to-gene scores, disease-relevant molecular-QTL colocalisation, rare-variant burden analyses, biobank phenotype analyses, or ancestry-diverse replication are supplied for either candidate. Direct comparative evidence using matched NSCLC cohorts and explicitly independent replication of the causal gene assignment is not supplied.	Causal oncogenic fusions A recurrent EML4–ALK fusion is a transforming somatic genetic alteration in human NSCLC. Independent comprehensive tumor-genome profiling confirms ALK rearrangements as recurrent somatic driver events in lung adenocarcinoma. The recurrent EML4-ALK fusion is a transforming oncogenic driver in NSCLC. TCGA-integrated human tumor sequencing identified recurrent somatic MET exon 14 skipping as a likely driver alteration in lung adenocarcinoma. Diverse somatic MET exon 14 splice-site alterations occur in human lung cancers and define a recurrent genomic driver class. Authoritative germline ALK susceptibility is established for neuroblastic tumors, not NSCLC. A large NSCLC germline sequencing study found that inherited susceptibility explains only a minority of cases and did not establish ALK as an NSCLC predisposition gene.
EGFR — MET	Translational model validity	B wins	gpt-5.6-luna	80%	MET has the stronger translational-model dossier because it includes explicit pharmacologic efficacy in six MET-amplified NSCLC patient-derived xenograft models and additional patient-derived EGFR-mutant/MET-amplified models, with activity linked to MET-dependent signaling (C2-EV_PERTURBATION_0002, C2-EV_PERTURBATION_0003). Its clinical relevance is also supported in treatment-naive MET exon 14–skipping NSCLC by responses to capmatinib (C2-EV_SAFETY_0002; C2-EV_CLINICAL_0001). EGFR has relevant genetic mouse-model claims and references to PDX and other models, but these are largely catalogue-level summaries without comparably detailed, independently documented model results (C1-CAT-ONCO-BIO-117, C1-CAT-ONCO-BIO-117b). The MET evidence is biomarker-restricted rather than universal, and the dossiers do not provide a direct head-to-head comparison in newly diagnosed, chemotherapy-backed models, so the margin is narrow.	Directly comparable, biomarker-matched patient-derived or humanized in vivo efficacy studies for both targets using newly diagnosed NSCLC models and the specified chemotherapy backbone. For EGFR, a primary-study report with detailed results from the cited EGFR genetic mouse, PDX, organoid, or patient-cell models rather than general catalogue summaries.	Pharmacologic MET inhibition suppresses MET-dependent lung cancer models in vitro and in vivo, with sensitivity linked to MET signaling complexes. A subset of EGFR-mutant, MET-amplified NSCLC models becomes dependent on MET and responds to MET inhibition alone. MET exon 14–altered NSCLC shows clinically meaningful dependence on MET signaling, supporting MET as an essential therapeutic driver in this molecular subset. MET exon 14 skipping is clinically validated as a predictive biomarker in metastatic NSCLC through durable responses to selective MET inhibitors. Welche Tiermodelle validieren EGFR? Welche Ergebnisse wurden in den einzelnen Tiermodellen mit EGFR erzielt? MET perturbational dependency in NSCLC is restricted to a molecularly defined subset rather than being general across lung cancer models. MET amplification has weaker and threshold-dependent clinical validation than MET exon 14 skipping in NSCLC.
KRAS — MET	Resistance and escape liability	B wins	gpt-5.6-luna	79%	SECOND (MET) is judged less vulnerable, but only narrowly. KRAS has direct evidence of multiple escape routes: feedback and parallel-pathway reactivation through RTK–SHP2–SOS–RAS and alternative survival pathways, plus acquired KRAS resistance mutations, allele switching, and amplification under KRAS-G12C inhibition (C1-CAT-ONCO-BIO-004, C1-CAT-ONCO-BIO-013, C1-CAT-ONCO-BIO-063). MET also has documented escape liability: baseline RAS–MAPK alterations can cause primary or reduced response to MET therapy, and MET dependence is heterogeneous and biomarker-dependent (C2-EV_GENETICS_0005, C2-EV_SAFETY_0003, C2-EV_CLINICAL_0006). However, the supplied MET dossier provides less direct evidence of acquired resistance to selective MET inhibition than the KRAS dossier provides for KRAS-directed resistance. Combination and biomarker strategies would mitigate risk for both: KRAS risk could be reduced by combinations targeting RTK/SHP2/SOS or downstream/parallel pathways and by allele-specific selection and monitoring; MET risk could be reduced by selecting MET exon-14 skipping or high-level amplification rather than nonspecific MET expression or low-level amplification, with pathway-combination strategies supported in EGFR/MET bypass models (C1-CAT-ONCO-BIO-015, C2-EV_BIOMARKER_0001, C2-EV_TRACTABILITY_0007, C2-EV_MECHANISM_0005).	Directly matched, variant- and biomarker-specific clinical comparisons of acquired resistance and post-progression escape under KRAS-directed versus MET-directed therapy in NSCLC, including the frequency and durability of MET inhibitor resistance mechanisms.	Gibt es kompensatorische Mechanismen, die die Auswirkungen der Hemmung von KRAS abmildern können? Gibt es biochemische Signalwege, die die Hemmung von KRAS umgehen und dadurch zu einer Resistenz führen können? Ist KRAS von Resistenzmutationen betroffen? Co-occurring RAS–MAPK pathway alterations are common in MET exon 14-altered NSCLC and can limit response to MET-directed therapy. MET exon 14 skipping activates MET by removing the juxtamembrane CBL-binding region, impairing receptor ubiquitination and down-regulation. MET exon 14 skipping is a clinically validated predictive biomarker for selective MET inhibition in NSCLC. Co-occurring RAS–MAPK pathway alterations are common in MET exon 14-altered NSCLC and can limit response to MET-directed therapy. MET dependence in NSCLC is biomarker- and alteration-dependent rather than universal across MET-aberrant tumors. MET amplification has weaker and threshold-dependent clinical validation than MET exon 14 skipping in NSCLC.
EGFR — KRAS	Evidence robustness (meta-evidence)	B wins	gpt-5.6-luna	82%	KRAS has the stronger documented meta-evidence structure: a primary human tumor-genomics cohort, an independent prospective MSK-IMPACT cohort, and a large retrospective cohort with separately described sample bases. These provide clearer source-family diversity and replication than EGFR's supplied evidence, whose strongest genetics item asserts replication but does not identify the contributing studies, cohorts, overlap status, or negative-result availability. The margin is narrow because neither dossier provides comprehensive provenance, overlap, publication-bias, or negative-result metadata, so the comparison remains incomplete.	Explicit provenance metadata, source-family assignments, sample-overlap checks, replication details, and systematic negative-result availability for both candidates; especially an independently documented prospective replication and negative-result record for EGFR.	KRAS is a recurrent, activating somatic driver in human lung adenocarcinoma, supporting KRAS inhibition as a genetically defined therapeutic strategy in NSCLC. Independent prospective clinical sequencing confirms KRAS as one of the dominant oncogenic drivers in metastatic human lung adenocarcinoma. KRAS mutations are common in human NSCLC, particularly lung adenocarcinoma, with KRAS G12C representing the largest mutation subtype and a genetically defined inhibitor-sensitive population. Causal activating mutations
KRAS — ALK	Translational model validity	B wins	gpt-5.6-luna	95%	ALK has the stronger translational efficacy evidence. Its dossier includes selective perturbation in EML4-ALK-positive NSCLC cells, evidence in a patient-derived NSCLC model, genetically engineered lung tumor models, and randomized clinical efficacy in previously untreated ALK-positive NSCLC against platinum-pemetrexed chemotherapy. These models and the clinical setting are directly relevant to NSCLC and newly diagnosed disease. KRAS has strong human genetic and mechanistic support, but the supplied evidence does not provide comparably detailed efficacy results in NSCLC PDX, organoid, primary-cell, or genetically engineered lung models; the cited efficacy example is a MiaPaCa-2 xenograft, which is pancreatic rather than NSCLC. ALK dependence is not uniform across variants or models, but these limitations do not offset the substantially more direct and clinically anchored translational efficacy evidence.	Direct KRAS efficacy results in NSCLC patient-derived xenografts or genetically engineered lung models, specifically using newly diagnosed or treatment-naive disease settings. Direct comparative efficacy data for KRAS and ALK interventions in patient-derived or genetically engineered NSCLC models.	Direct ALK knockdown selectively impairs viability and clonogenic growth of EML4-ALK-positive NSCLC cells. The EML4 portion of EML4-ALK promotes cytoplasmic phase-separated condensates that organize downstream signaling, including STAT3 activation, and support lung tumorigenesis. Alveolar epithelial cells can serve as a relevant lung cellular context for EML4–ALK-driven tumorigenesis in vivo. ALK is clinically validated as a predictive therapeutic biomarker in advanced NSCLC by randomized phase 3 evidence showing superior efficacy of ALK inhibition over platinum-pemetrexed chemotherapy. In untreated advanced ALK-positive NSCLC, biomarker-selected patients derived substantially greater benefit from ALK inhibition with crizotinib than from platinum-pemetrexed chemotherapy. ALK rearrangement does not uniformly establish ALK pathway dependence across NSCLC models. ALK fusion variants produce materially different pathway dependency and inhibitor sensitivity. CRISPR-based in vivo perturbation shows that EML4-ALK variants are functionally heterogeneous rather than a uniform dependency context.

Page 1 of 5Next