In plain terms

The claim being assessed: Sequential somatic driver mutations are the primary initiating and sustaining cause of cancer.

Where the evidence stands: Ε — Causal in a specific setting. It genuinely causes the effect in one well-defined situation, but that does not extend to the wider claim.

What this mechanism does: This is the thing that sets the process off.

Strongest evidence for it: The preregistered scope-matched SC3 gate returns Ε: causal validity with explicit context limits. Driver identification reaches Ζ and defined initiation/maintenance dependencies reach Η through endogenous autochthonous genetics, restoration/rechallenge, human genomic natural history and genotype-selected hard clinical outcomes. The set is not re-capped by culture or xenograft limitations. Broad primacy requires a representative comparison of mutation, tissue and cell-state causation across initiation, maintenance and relapse; that burden is not established by genetic relevance alone. Benign driver clones challenge sufficiency, not the strong causal genetic evidence.

What decisive evidence would look like

In plain terms, researchers would change the proposed cause, confirm it changed as intended, and measure a meaningful health outcome while checking other explanations. The result should be able to support or count against the claim.

Relative explanatory fit: Alternative ranked ahead

What this review means for Somatic Mutation Theory of Cancer

Claim being assessed: Sequential somatic driver mutations are the primary initiating and sustaining cause of cancer.. This review asks whether the proposed factor causes the health outcome—not just whether the two appear together. For the wording assessed here, the evidence is rated Causal in a specific setting (HSEF Ε; Moderate confidence). It genuinely causes the effect in one well-defined situation, but that does not extend to the wider claim. That rating applies to this claim and scope; it does not mean the biological process is irrelevant in every setting.

Implications for health and treatment: This is an assessment of a causal explanation, not proof that a particular supplement, diet, test, or treatment improves health. A treatment claim needs studies that test the intervention in people and measure meaningful benefits and harms. These findings can help frame research and informed clinical questions, but they do not identify what is right for one person.

Limits for individual care: Evidence from cells, animals, associations, or a plausible mechanism may not predict what happens in people. The grade is limited to the exact claim, outcomes, and sources reviewed, and it is not a diagnosis or an individual forecast. Do not start, stop, or change care based on this page; discuss personal decisions with a qualified clinician.

Medical information notice: This review provides general health information, not medical advice. It cannot diagnose a condition or replace an individual discussion with a qualified clinician.

This assessment rests on 70 sources and 74 individually checked claims. 74 of those claims were verified against the original paper.

A low grade is not an accusation. It means the causal work has not been done yet, which is the ordinary state of most actively studied mechanisms in biology.

At a glance

Sub-claims assessed separately

ID	Sub-claim	Type	Level	Pivotal card
SC1	Recurrent somatic driver mutations can be distinguished from passengers in specific cancers.	existence	Ζ	Card ↗
SC2	Driver mutations are causal dependencies for initiation or maintenance in defined cancers.	necessity	Η	Card ↗
SC3	Sequential somatic driver mutations are the dominant initiating and sustaining cause across cancer.	primacy	Ε	Card ↗

Visual Evidence Cards · Pivotal Subclaims

Pivotal subclaim illustrations

Dark-theme simulated-micrograph renders and concise causal evidence summaries for the pivotal subclaims that govern this review's grade. Click any card to inspect the master 2400×2400 render.

Master 2400×2400 render ↗

Master 2400×2400 render ↗

Master 2400×2400 render ↗

Figures

Explanatory figures summarizing this review. They carry no evidential weight of their own; where a figure and the review differ, the review governs.

Full assessment

From here the review is written for a technical reader: caps and thresholds, competing models, the independence basis, and the citation ledger.

Sources: 70 primary reports; claims: 74; verified: 74; unverified: 0. Verification is bounded attribution. Access: 8 partial-body and 62 complete abstract/status examinations. Literature cutoff: 2026-10-06; source examination UTC date: 2026-10-06. No complete-paper audit or blinded calibration claimed.

Adversarial review: Generated by Sol 6.1; adversarially reviewed by Claude Opus 5.5 (Anthropic) on 2026-10-06. 2 corrections applied (see ADVERSARIAL REVIEW AND APPROVAL). Status: Approved as revised.

AT A GLANCE

Somatic drivers cause and sustain many genotype-defined cancers; sequential mutation primacy across all cancer behavior remains context-dependent.

Deciding experiment: Across predeclared representative histologies, use lineage-resolved native cancers with selective driver correction/restoration and factorial niche/chromatin perturbations, preserving viability and measuring both early initiation and established maintenance/relapse.

Belief versus evidence: The preregistered scope-matched SC3 gate returns Ε: causal validity with explicit context limits. Driver identification reaches Ζ and defined initiation/maintenance dependencies reach Η through endogenous autochthonous genetics, restoration/rechallenge, human genomic natural history and genotype-selected hard clinical outcomes. The set is not re-capped by culture or xenograft limitations. Broad primacy requires a representative comparison of mutation, tissue and cell-state causation across initiation, maintenance and relapse; that burden is not established by genetic relevance alone. Benign driver clones challenge sufficiency, not the strong causal genetic evidence.

Best-supported rival: Integrated mutation–tissue model. Use factorial mutation correction/restoration and defined niche/state perturbation in the same native lineage, with lineage-traced malignant endpoints, maintained viability and mutation-sensitive monitoring. An integrated model must make prespecified interaction predictions; flexibility alone cannot win.

COMPETING MECHANISTIC MODELS

Rank 1: Integrated mutation–tissue model (Ζ). Genetic lesions, lineage, niche and regulatory state jointly determine specified cancer behavior. Endogenous APC/NOTUM, context-dependent autochthonous editing, immune-dependent oncogene regression, PM/IL1 promotion and fibroblast depletion show functional interactions across assay classes. Ζ applies to these specified interactions; an unrestricted flexible umbrella is not a tested universal model or proof that tissue changes precede all mutations.

Rank 2: Somatic mutation primacy (Ε). Sequential somatic drivers dominate cancer initiation and maintenance across cancers. Strong endogenous genetic dependencies and genotype-directed survival/progression benefits support many defined contexts. Normal driver-bearing tissues, state plasticity, epigenetic initiation and stage-specific dependencies leave broad dominance incompletely tested.

Rank 3: Nongenetic malignant states (Ε). Chromatin and differentiation memory can initiate or sustain malignant/tolerant states in specified contexts. Transient fly Polycomb depletion, interrupted reprogramming and reversible drug tolerance demonstrate functional state plasticity. Developmental perturbation and culture are not evidence that ordinary human cancer generally lacks genetic drivers; tolerance is not original cancer initiation.

Rank 4: Chromosomal instability first (Ε). Chromosome instability can initiate or facilitate escape in defined models. Transient Mad2-induced tumors and Kras-independent relapse support a bounded instability route. Chromosome changes are genetic; forced checkpoint perturbation and sparse context coverage do not establish universal temporal primacy.

Rank 5: Tissue organization field theory (Γ). Tissue disorganization is the primary initiating and sustaining cause across cancers. Compartment recombination, stromal normalization and integrin/context reversal support contextual causal leads. Transplantation, limited mutation exclusion and heterogeneous stromal effects do not establish a universal endogenous tissue-first dependency; strong genetic dependencies remain.

The tournament does not alter the nominal level. Compatible pathways and flexible alternatives are not automatically proved by explaining observations.

FULL REVIEW

Registered claim and scope

Sequential somatic driver mutations are the primary initiating and sustaining cause of cancer.

Use the index nominal strong claim. Grade mutation presence, causal dependency and cross-cancer primacy separately. Specific mutation-dependent cancers are reported on their own terms; the broad primacy gate is SC3. Normal mutant clones and defined epigenetic/tissue exceptions must be reconciled without recapping a convergent genetic/therapeutic set.

Selection: public-interest rank 25 from the editorial ranking compiled 2026-09-23; categorical prior verdicts were visible. Private fresh review; earlier documents are preserved.

Evidence and causal interpretation

The classic two-event retinoblastoma inference and colorectal stage series established a testable genetic framework. Modern pan-cancer and pediatric genomes identify selected recurrent alterations, including structural and copy-number events. This makes a passenger-only account untenable, but genomic frequency alone is not functional causality or obligatory temporal order (PMID5279523, PMID2841597, PMID24132290, PMID29056346, PMID29489755, PMID31645765).

Normal skin, esophagus, lung, endometrium and blood carry expanding driver-bearing clones without inevitable cancer. The original esophagus body distinguishes estimated mutant-cell coverage from independent variants and contrasts NOTCH1 versus TP53 tumor enrichment. Longitudinal blood clone growth still predicts malignant risk. These results show insufficiency and tissue-specific fitness; they do not refute cooperative driver causality (PMID25999502, PMID30337457, PMID30602793, PMID31996850, PMID32350471, PMID35650444, PMID41062696).

Native genetic evidence is strong. Endogenous Kras generates lung/pancreatic precursor disease; Kras/p53 cooperation yields invasive/metastatic pancreatic cancer. Endogenous p53 restoration regresses native lymphomas/sarcomas. Multiplex lung editing directly tests causal suppressor effects with inert-cutting and Cas9 controls: effects can switch sign by oncogenic context, and clone number is distinct from growth. The assembled case is not culture-only (PMID11751630, PMID14706336, PMID15894267, PMID17251932, PMID37828026).

APC restoration is particularly discriminating. Selected original Results show endogenous protein recovery after reversible shRNA suppression, rapid differentiation, lineage-marked normal crypt behavior and re-growth on re-suppression. Kras/p53-altered tumors regress, with small group sizes and residual microadenomas explicitly retained. This is restoration of endogenous function, not correction of the underlying mutant alleles or elimination of every altered cell (PMID26091037).

Genotype-selected clinical consequence is substantial. IRIS improves cytogenetic and progression outcomes; its ten-year survival is follow-up of the same heavily crossed-over cohort. EGFR adjuvant ADAURA and BRAF melanoma trials demonstrate survival benefit; FLAURA compares two active targeted strategies with a secondary survival report. ALK and RET trials improve progression, while KRAS G12C benefit is more modest. These clinical and endogenous genetic lines complement each other rather than sharing all failure modes (PMID12637609, PMID28273028, PMID37272535, PMID21639808, PMID25399551, PMID29151359, PMID31751012, PMID28586279, PMID37870973, PMID36764316).

Genetic lesions can act through tissue context. APC-mutant NOTUM changes neighbor competition; normal mutant neighbors can purge nascent esophageal tumors. PM2.5 promotes pre-existing EGFR/KRAS substrates through macrophage/IL1beta state changes without increased short-term mutation burden. Fibroblast depletion can worsen survival despite lower tumor weight, with partial immune rescue and embolic complications. Context causality can be downstream or cooperative rather than an exclusive replacement for genes (PMID34079124, PMID34646013, PMID37020004, PMID24856586).

Human organoids reconstruct sequential driver-dependent niche escape, but transplantation alone does not close spontaneous human causality. Matano organoids fail macrometastatic liver colonization under some conditions; niche restoration can substitute for a missing alteration in another orthotopic model. Spontaneous MMR-deficient organoid acquisition still uses culture selection. These specify routes and limits, not an invariant sequence for every cancer (PMID25924068, PMID25706875, PMID28270604, PMID39487295).

Transient native Polycomb depletion initiates fly neoplasia that persists after protein restoration and expands/metastasizes through allografts. Original Methods test pooled tumor/control genomes and threshold variant calls; no recurrent drivers supports this bounded epigenetic route without proving zero DNA change or human prevalence. Interrupted mammalian reprogramming and historical embryonic normalization support plasticity, with weaker mutation-exclusion access. IDH-driven epigenetic insulation is a mutation-mediated route, not an independent refutation (PMID38658752, PMID24529372, PMID1059147, PMID26700815).

Maintenance and relapse can diverge. MYC-off tissues can remain dormant, permanently differentiate or recur depending on context and immunity. Original Kras Results distinguish withdrawal/rechallenge of native pancreatic disease from doxycycline-independent relapse with reactivated Kras* in transplanted cultured lines. Reversible chromatin tolerance, later APOBEC mutation, T790M paths and RET solvent-front resistance coexist. Presence of an original driver after histological transition does not prove retained dependence (PMID15475948, PMID12098700, PMID21035406, PMID23226501, PMID20371346, PMID28607484, PMID37407818, PMID26828195, PMID21430269, PMID31988000).

Consensus supported focused primary discovery and checked indexed APC restoration and partial Polycomb state details. It did not retrieve four of six requested primary papers; original PMC Results supplied normal-esophagus, pollution and tolerance details. Its broad timeline includes secondary and substituted references and is not used as a primary bibliography. This review provides its own source-linked timeline and mechanisms-specific gaps, recording the one browser retry and zero API calls.

HYPOTHESIS STATE EVALUATION FRAMEWORK (HSEF v5.3)

Subclaim matrix and aggregation

Preregistered aggregation: scope-matched. The preregistered scope-matched SC3 gate returns Ε: causal validity with explicit context limits. Driver identification reaches Ζ and defined initiation/maintenance dependencies reach Η through endogenous autochthonous genetics, restoration/rechallenge, human genomic natural history and genotype-selected hard clinical outcomes. The set is not re-capped by culture or xenograft limitations. Broad primacy requires a representative comparison of mutation, tissue and cell-state causation across initiation, maintenance and relapse; that burden is not established by genetic relevance alone. Benign driver clones challenge sufficiency, not the strong causal genetic evidence. Sensitivities: weakest-link Ε; primary-subclaim Ε; scope-matched Ε.

Decisive-perturbation inventory and causal tally

[15] Endogenous mouse esophageal carcinogenesis and competitive-clone induction/inhibition show mutant neighboring epithelium can eliminate nascent tumors. Competition is a causal tissue modifier with retained mutant neoplasms, not proof that genetic lesions are dispensable across cancer.

[19] Oncogenic RAS induces arrest in primary cells; p53/p16 inactivation or E1A permits bypass in defined species and cell systems. Forced expression and cell culture establish a cooperative checkpoint, not endogenous pan-cancer necessity or a universal mutation sequence.

[20] Barcoded autochthonous lung editing across KRAS G12C/G12D, BRAF and EGFR contexts shows causal suppressor effects can reverse sign. Original Results include inert-cutting and Cas9-negative controls and separate clone number from growth. Four lung contexts do not represent all histologies; mutation frequency alone predicts some functional effects poorly.

[21] Apc-mutant cells secrete NOTUM; genetic or pharmacological inhibition reduces wild-type competition loss, mutant fixation and adenomas. This is genetic causality mediated through the niche, not a competing mutation-free account; prevention of adenomas is not universal invasive-cancer maintenance.

[22] Restoring endogenous p53 regresses autochthonous lymphomas and sarcomas with different apoptosis/senescence responses and no detected normal-tissue injury. Defined p53-dependent models establish maintenance causality; they do not show every tumor is dependent or replace initiation with maintenance. Original body unavailable here.

[23] Endogenous Kras G12D and Trp53 R172H cooperate to generate invasive, metastatic pancreatic cancer with genomic instability. Prespecified engineered alleles show joint sufficiency in this tissue, not a universal sequence or mutation-only explanation of tissue context.

[24] Physiological mutant Kras expression induces the PanIN spectrum, with low-frequency spontaneous invasive/metastatic progression. Initiation is genetically imposed and pancreatic-lineage-specific; additional evolution and microenvironmental conditions remain.

[25] Conditional endogenous oncogenic Kras permits timed autochthonous lung tumor induction and progression analysis. Cre induction supplies the initiating alteration; no mutation-loss maintenance test or broad histological generalization follows from this report alone.

[26] Inducible pancreatic Kras withdrawal regresses primary and metastatic tumors; reactivation rapidly restores disease. Original Results show TetO-Kras with endogenous mutant p53 and induced pancreatitis. Off-doxy relapse occurs after culture/transplantation, not in the autochthonous mice; do not conflate those arms or call Kras endogenous expression.

[27] Conditional oncogenic Kras from its endogenous promoter causes fully penetrant myeloproliferation but not transplantable AML. A disease precursor dependency is not AML sufficiency, supporting genetic cooperation and lineage boundaries.

[28] Reversible pancreas-specific Kras expression induces and maintains lesions, with withdrawal causing regression and changes in fibroinflammatory stroma. Transgenic activation and tissue damage condition initiation; stromal mediation can be downstream of a causal mutation rather than an exclusive alternative.

[29] Conditional MYC expression produces hematologic malignancies that regress after transgene inactivation, with differentiation and restored hematopoiesis. Forced oncogene expression without endogenous mutation correction narrows interpretation to the induced dependency.

[30] MYC-off liver tumors differentiate and become dormant; reactivation restores malignancy with the same comparative-genomic signature. Reversal changes driver activity rather than eliminating every mutation; latent tumor identity and tissue-specific responses constrain permanent cure claims.

[31] Inducible pancreatic MYC tumors regress after inactivation but residual cells rapidly recur on re-expression. Marker-positive residual cells are not proof of a universal stem hierarchy; transgene-driven maintenance does not establish natural mutation primacy.

[32] Brief MYC inactivation produces sustained osteogenic tumor regression; reactivation induces apoptosis instead of recurrence. Tissue-specific epigenetic memory modifies response to the same driver; induced expression and inferred epigenetic mechanism limit universal conclusions.

[33] CD4 and thrombospondin functional contrasts show immune remodelling is required for sustained regression after MYC/BCR-ABL inactivation. Oncogene withdrawal and immune context cooperate; immune necessity does not negate genetic dependence. Linked erratum lacks a supplied PMID and details remain unexamined.

[34] Sequential APC/TP53/KRAS/SMAD4 editing produces niche-independent human organoids and invasive xenograft tumors. Culture selection and xenotransplantation alone do not establish endogenous human cancer dependency; engineered order is not proof of obligatory order.

[35] Isogenic five-pathway mutant organoids form tumors but fail liver macrometastatic colonization, unlike chromosome-unstable adenoma-derived organoids. Driver edits support causal sufficiency in the tested niche; invasion/metastasis requires additional lesions or context and is not fully reproduced.

[36] Orthotopic engineered organoids show mutation accumulation supports metastasis; specific niche signals restore growth when one oncogenic alteration is absent. Transplantation and niche rescue are bounded model evidence, not proof of mutation dispensability in all spontaneous cancers.

[37] MMR-deficient organoids acquire spontaneous pathway mutations under sequential niche-factor withdrawal and form xenograft tumors. Repair-defective background and designed selection are not ordinary in vivo initiation; spontaneous acquisition does not remove culture/transplant limits.

[39] Endogenous APC protein restoration reverses autochthonous colorectal adenomas/carcinomas, including Kras/p53-altered settings, with lineage-marked normal crypt behavior. Original Results show reversible shRNA loss rather than precise mutation repair, small Kras/p53 groups, residual microadenomas and repeat regrowth on APC re-silencing; this is powerful defined dependency, not permanent eradication of all mutant cells.

[40] Transient endogenous Polycomb depletion induces persistent native fly neoplasia after protein recovery; serial allografts grow and metastasize. Original Results/Methods test four controls and 12 tumors with pooled whole-genome data, no recurrent drivers and variant/allele-fraction thresholds. This does not prove zero sequence changes or ordinary human mutation-free prevalence; transplantation is not its only initiation evidence.

[41] Matrix stiffness, integrin/Rho/ERK manipulations and contractility reduction alter malignant architecture in an integrated signaling circuit. In vitro phenotype reversion is not pan-cancer initiation or survival causality and does not remove the transformed genotype.

[42] Genetic depletion of proliferating alphaSMA fibroblasts accelerates pancreatic cancer and worsens survival; anti-CTLA4 partially rescues the altered state. Original Results replicate genetic backgrounds and retain poorer survival despite lower tumor weight, with pulmonary emboli and immune changes; stroma can restrain rather than uniformly promote cancer.

[43] Rat mammary stroma age/parity changes tumor incidence and permits normal ducts containing a tumor-origin RAS marker. Transplantation and host developmental/reproductive differences do not establish a universal tissue-first origin; persistent RAS marker supports genotype–phenotype separation.

[44] Brief compression induces normal-like malignant breast structures; nitric-oxide inhibition blocks the architectural response. Three-dimensional morphology is not verified spontaneous cancer elimination or survival; altered context does not erase genomic lesions.

[45] Beta1-integrin inhibitory antibody/Fab reverses malignant breast architecture and reduces nude-mouse tumor formation; stimulatory antibody fails. Reversion is reversible on antibody removal; transplant/culture evidence supports contextual phenotype control but not tissue primacy across endogenous cancers.

[46] PM2.5 promotes EGFR/KRAS-mutant lung lesions; macrophage/IL1beta experiments connect inflammatory promotion to altered epithelial state. Original Results show pre-existing engineered drivers, short-term PM-versus-PBS genome comparisons and concomitant antibody suppression. This is nonmutagenic promotion in a mutant substrate, not proof that mutations are unnecessary.

[47] Rat recombination contrasts place NMU-associated mammary tumor initiation with exposed stroma, independent of epithelial exposure; Ha-ras status does not track all tumors. Different compartment exposure protocols, transplantation and one-gene mutation testing do not establish absence of other drivers or pan-cancer tissue-first necessity.

[48] Transient Mad2 overexpression induces chromosome instability and multiple tumor types; sustained Mad2 is unnecessary for maintenance. Chromosome gains/losses are genetic alterations; forced checkpoint expression is not an endogenous mutation-free origin or universal CIN-first model.

[49] Mad2-mediated chromosome instability increases Kras-independent relapse after initial regression on oncogene withdrawal. Engineered chromosome instability and varied relapse pathways support a context, not all cancers or absence of genetic evolution.

[50] Malignant teratocarcinoma cells contribute to functional normal tissues and germline after blastocyst incorporation. Modal chromosome number and strain markers are not modern genome-wide driver exclusion; rare embryonic reprogramming cannot estimate ordinary adult cancer prevalence.

[51] Interrupted in vivo reprogramming induces dysplastic tumors, and tumor-derived iPSCs can form nonneoplastic kidney cells. Forced reprogramming and developmental reset demonstrate context-specific epigenetic plasticity; reversibility is not sensitive genome-wide proof of zero mutations or broad human predominance.

[52] IDH-mutant glioma insulation is altered; demethylation partly restores regulation and CTCF-site disruption activates PDGFRA in wild-type-IDH spheres. Mutation-to-epigenetic causation is an integrated mechanism, not an independent refutation of mutations. Gliomasphere proliferation is not complete in vivo cancer dependency.

[53] Rare melanoma transcriptional states predict resistance and drug-induced reprogramming stabilizes it. Resistance-state plasticity does not establish nongenetic original initiation; no clear genetic cause is not exhaustive exclusion. Linked correction details are qualified.

[54] Single-clone cancer cultures yield reversible persisters; KDM5A depletion reduces tolerance without detectable parental proliferation loss. Original Results distinguish nine-doubling persister reversal from roughly 90-doubling expanded-persistor reversal and CD133/quiescence controls. These cells retain cancer-driving mutations; clinical relapse necessity is not directly tested.

[55] Therapy induces APOBEC3A in persisters; deletion reduces mutations/structural changes and delays resistance, with concordant patient signatures. Genetic evolution can arise from a nongenetic persister state; patient signatures and cell models do not prove every resistance lineage follows this route.

[56] T790M resistance emerges through either pre-existing clones or evolution from tolerant cells; apoptosis-targeted treatment restores sensitivity in a subset. Patient-derived culture convergence supports distinct paths, not a single universal sequence or permanent dependence on the original inhibitor target.

[58] Serial patient RET resistance variants converge at solvent-front G810, with xenograft and enzyme/cell assays confirming loss of inhibition. Small selected resistance series, shared pathway inhibitors and xenografts do not quantify pan-cancer mutation primacy or exclude state-based escape.

[59] 1106-person IRIS randomization improves cytogenetic response and freedom from blast/accelerated progression with BCR-ABL-directed imatinib. Drug effects are pathway perturbation rather than correction of the chromosome fusion; crossover and additional kinase targets limit isolated genetic mediation.

[61] IMADGIST randomizes 136 high-risk resected GIST patients after three years of imatinib; continuation improves disease-free survival. No survival difference is demonstrated; treatment-selected GIST and KIT/PDGFRA pathway targeting do not generalize to all tumors or establish mutation order.

[62] 556-person FLAURA compares EGFR inhibitors and improves PFS with osimertinib in defined sensitizing mutations. Both arms inhibit EGFR, so comparative efficacy also reflects drug properties; interim overall survival did not meet its statistical boundary.

[63] Planned FLAURA secondary survival analysis favors osimertinib, median 38.6 versus 31.8 months. Same FLAURA patients, not a separate replication; more effective genotype-matched drugs support dependency without proving unselected pan-cancer dominance.

[64] 682-person ADAURA adjuvant randomization improves overall survival after EGFR-mutant lung-cancer resection, with or without prior chemotherapy. Defined stages and mutations plus surgery constrain scope; therapeutic consequence does not directly establish spontaneous initiation or universal sufficiency.

[65] 303-person ALEX improves PFS and CNS progression with alectinib versus crizotinib in ALK-positive lung cancer. Both arms target ALK; pharmacology and CNS penetration differ, and initial report does not establish overall-survival benefit or universal ALK dependence.

[66] LIBRETTO-431 compares selective RET inhibition with chemotherapy with/without pembrolizumab and improves PFS in fusion-positive lung cancer. A structural rearrangement is genetic driver evidence; selected advanced disease and immature/crossover survival do not establish all-cancer primacy.

[67] 675-person BRIM3 improves survival and PFS with vemurafenib versus dacarbazine in BRAF V600E melanoma. Clinical selection and eventual resistance constrain durability; target inhibition does not correct every mutation or remove tissue/immune mediation.

[68] 704-person COMBI-v improves overall survival with BRAF/MEK combination versus vemurafenib in BRAF V600 melanoma. A downstream combination strengthens consequence and resistance control, not mutation-only origin; distinct from BRIM3 despite related drug pathway.

[69] 345-person CodeBreaK200 improves PFS with sotorasib versus docetaxel in previously treated KRAS G12C lung cancer. The benefit is modest and this abstract does not establish overall-survival benefit; mutation-specific dependence varies by disease and treatment stage.

[70] 291-person LIBRETTO-531 improves PFS with selective RET inhibition versus multikinase inhibitors in RET-mutant medullary thyroid cancer. RET alterations may be inherited or somatic; this cohort cannot be counted as a purely somatic initiating-mutation experiment. Both arms target RET-pathway activity.

Tally: 50 discriminating contrasts. Null and negative contrasts are discriminators, not positive threshold crossings. Reports and replications are not independent causal lines. Shared cohorts, experimental platforms, transgenic lineages and repeated outcomes do not constitute fully independent causal lines.

Cap audit, precedence and Β/Γ/Δ boundary

The cancer rubric explicitly accepts endogenous genetic manipulation, matched restoration and functional clonal dependency. Organoid or transplant initiation alone remains Γ for spontaneous broad inference, but it is not the whole evidence set. Endogenous Kras/p53, reversible endogenous APC protein, native clone competition and genotype-selected clinical consequences cover different failure modes. SC2 earns Η in its genotype-defined scope. SC3 earns Ε from real context-specific causal validity with broader dominance unresolved; it is not capped at Γ because some contributing assays are cultured. Epigenetic mark correlations alone are not credited as causal, while transient endogenous Polycomb depletion with persistent native fly neoplasia is distinguished from transplantation alone.

Gamma hard-rule audit: Culture-only and xenograft-only legs cannot establish spontaneous human dependency by themselves. Forced MYC/Mad2 expression without endogenous matched loss does not become Strong solely from tumor formation. The assembled causal set contains endogenous genetic and clinical lines, so no Γ hard rule caps the complete set.

SC1 field-valid definition: “ - Endogenous genetic manipulation in autochthonous models - Patient-derived system perturbation with matched rescue AND in vivo confirmation - Clonal tracking demonstrating persistence of dependency through selection pressure” Positive full-scope result: yes. Recurrent alterations are distinguished from passengers by selection, clonal timing and precise functional perturbations in specific histologies. Independent genome platforms and autochthonous editing reduce sequencing-error and correlation-only explanations. Genomic annotation alone remains insufficient, and fitness in normal tissue is not necessarily cancer-driving fitness. Ζ is supported across assay classes; most claimed variant classes have not individually received therapeutic/endogenous confirmation, so Η is not assigned to every driver-identification inference.

SC2 field-valid definition: “ - Endogenous genetic manipulation in autochthonous models - Patient-derived system perturbation with matched rescue AND in vivo confirmation - Clonal tracking demonstrating persistence of dependency through selection pressure” Positive full-scope result: yes. Endogenous Kras combinations, endogenous p53 restoration, reversible APC loss/restoration and lineage-marked tumor differentiation establish causal dependencies in defined cancers. Multiple genotype-selected clinical trials yield progression or survival consequences across EGFR, BRAF, RET, ALK and BCR-ABL settings. These complementary lines fail differently and satisfy Η in the prespecified genotype-defined scope. The subclaim is initiation OR maintenance, not a requirement that each report establish both. Residual states, relapse and incomplete target response prevent Θ/Ω for most cancer behavior or universal necessity/sufficiency.

SC3 field-valid definition: “ - Endogenous genetic manipulation in autochthonous models - Patient-derived system perturbation with matched rescue AND in vivo confirmation - Clonal tracking demonstrating persistence of dependency through selection pressure” Positive full-scope result: yes. Genetic alterations are primary drivers in many tested cancers, with causal initiation and ongoing dependency, rather than mere passengers. Their phenotypic output depends on lineage, niche, immunity and chromatin; tissue mediation may be downstream of the genetic lesion and is not automatically an exclusive rival. Normal mutant tissues show insufficiency, fly/reprogramming models demonstrate bounded alternative initiation, and relapse can change dependence. Ε credits the real causal model while restricting the broad dominant-sequential initiating-and-sustaining claim. This is not a finding that common cancers generally lack driver mutations.

Upward threshold ladder

SC1 Δ: “Requires at least one meaningful causal perturbation showing the proposed mechanism contributes to initiation, maintenance, or relapse.” Satisfied: yes.

SC1 Ε: “Requires context-specific causal validity with explicit limits. Cancer stem cells often belong here when a stem-like hierarchy exists in some tumor contexts but fails as a universal model.” Satisfied: yes.

SC1 Ζ: “Requires: - Independent causal confirmation across models - Reduced viability of major alternatives - Relevance beyond one assay class” Satisfied: yes.

SC1 Η: “Requires: - Strong endogenous evidence - Therapeutic or mechanistic consequence - Reproducibility across disease settings within the claimed scope” Satisfied: no. Prospectively validate predicted driver-versus-passenger status across prespecified mutation classes with matched endogenous loss/restoration, patient allele-resolved consequence and reproducibility across histologies.

SC2 Δ: “Requires at least one meaningful causal perturbation showing the proposed mechanism contributes to initiation, maintenance, or relapse.” Satisfied: yes.

SC2 Ε: “Requires context-specific causal validity with explicit limits. Cancer stem cells often belong here when a stem-like hierarchy exists in some tumor contexts but fails as a universal model.” Satisfied: yes.

SC2 Ζ: “Requires: - Independent causal confirmation across models - Reduced viability of major alternatives - Relevance beyond one assay class” Satisfied: yes.

SC2 Η: “Requires: - Strong endogenous evidence - Therapeutic or mechanistic consequence - Reproducibility across disease settings within the claimed scope” Satisfied: yes.

SC2 Θ: “Requires that the mechanism explains most of the claimed cancer behavior with only minor exceptions.” Satisfied: no. Demonstrate that these dependencies explain most initiation, maintenance and relapse behavior in the registered genotype-defined settings, with only minor exceptions and durable rescue after verified selective engagement.

SC3 Δ: “Requires at least one meaningful causal perturbation showing the proposed mechanism contributes to initiation, maintenance, or relapse.” Satisfied: yes.

SC3 Ε: “Requires context-specific causal validity with explicit limits. Cancer stem cells often belong here when a stem-like hierarchy exists in some tumor contexts but fails as a universal model.” Satisfied: yes.

SC3 Ζ: “Requires: - Independent causal confirmation across models - Reduced viability of major alternatives - Relevance beyond one assay class” Satisfied: no. In representative predeclared histologies, compare mutation loss/correction and matched restoration with tissue/state interventions while controlling mutation burden and exposure; establish dominant mediation of both initiation and sustained cancer behavior.

Upper ladder: Endogenous genetic natural experiments, APC loss/restoration/rechallenge, oncogene withdrawal and independent genotype-selected hard clinical outcomes were explicitly considered and carry SC2 to Η. No randomized trial of spontaneous mutation acquisition is required. These are not re-capped as a set. They do not establish mutation dominance over tissue/state causes across representative histologies at every cancer stage; this is the missing broad SC3 gate, not an absence of genetic causality.

Independence: Seventy primary reports are not 70 independent causal confirmations. IRIS and FLAURA repeated outcomes are grouped with their original randomizations; inducible pancreatic Kras reports share a platform. Public genome cohorts and organoid/transgene teams can overlap. Autochthonous endogenous genetics, allele/pathway restoration, clonal tissue competition and independent clinical randomizations cover distinct errors, supporting upper levels for defined dependencies. The broad primacy comparison remains underdetermined despite this convergence.

Human driver landscape and normal tissue contrasts [9, 8, 12, 14]: Separate recurrence, normal clone fitness and longitudinal risk; sequencing errors and inferred selection remain distinct from functional dependence.

Endogenous native genetic initiation and restoration [25, 23, 22, 39]: Precise locus and temporal restoration reduce passenger/reverse-causation explanations; different tissues and suppression strategies constrain shared transgene artifacts.

Genotype-selected clinical progression and survival [59, 64, 67, 66]: Randomization and hard outcomes reduce treatment-selection confounding and connect causal pathway activity to human consequence; drug off-targets and selection remain.

Tissue/state interaction perturbations [15, 21, 46, 42]: Competitor, niche and immune manipulations reduce a cell-autonomous-only explanation while retaining genetic substrates.

Alternative state/relapse mechanisms [40, 54, 49, 55]: Separate mutation recurrence, reversible tolerance, instability and acquired genetic escape with distinct assays and contexts.

Falsification and pitfalls

Driver presence is not sufficient for malignancy. Multiple normal-tissue genomes and protective competitive clones contradict an automatic mutant-cell-equals-cancer rule, which is stronger than multistep genetic causality. Additional downgrade: False.

Initiation and maintenance dependencies can differ. Transient Polycomb or Mad2 perturbation, MYC tissue-specific memory and culture/transplant Kras escape contrast with native APC/p53 restoration. Additional downgrade: False.

Context is not uniformly tumor-promoting. Fibroblast depletion can worsen survival, and normal mutant neighbors eliminate nascent tumors. Additional downgrade: False.

Mutation defined only as a point substitution. Pediatric structural drivers, ALK/RET fusions, chromosome instability and biallelic loss.

Benign mutant clones treated as refutation of multistep genetic causality. Normal skin/esophagus/endometrium and blood.

Epigenetic mediation called independent mutation-free cause. IDH, APC/NOTUM, mutated lung substrate and therapy-induced APOBEC.

Reversible state taken as proof of zero DNA lesions. Historical teratocarcinoma and interrupted reprogramming.

Cultured doxycycline-independent relapse (Kras* reactivated) merged with native model. Original PMID23226501 distinguishes arms.

Provider retrieval gap treated as scientific absence. Four of six targeted Consensus papers were not matched; original sources cover several of them.

Adversarial steelman: above, then below

The strongest Ζ case combines widespread tumor driver selection, endogenous genetic initiation/maintenance, matched APC/p53 restoration, resistance mutations and large clinical survival benefits. This substantially reduces passenger-only and universal tissue-first explanations. It should reach Η for defined dependency and must not be pushed back to Γ. For broad sequential dominance at all stages, however, the selected causal contexts do not provide a representative genetic-versus-state comparison; real alternatives and stage-specific escape remain consequential rather than minor exceptions.

The strongest Δ case argues that broad exceptions leave only a single contribution experiment. That understates reproducible context-specific causal validity across pancreas, lung, colorectal, lymphoid and genotype-selected clinical disease. Explicitly bounded causality reaches Ε for the nominal broad claim; it is not merely a plausible cultured mechanism.

Reject Δ because many context-specific causal dependencies are established, and reject Ζ for the broad nominal primacy gate because representative cross-stage dominance is not demonstrated. Preserve SC1 Ζ and SC2 Η without importing them into SC3 or treating the remaining gap as refutation.

Final score, causal direction and M/S/D

The preregistered scope-matched SC3 gate returns Ε: causal validity with explicit context limits. Driver identification reaches Ζ and defined initiation/maintenance dependencies reach Η through endogenous autochthonous genetics, restoration/rechallenge, human genomic natural history and genotype-selected hard clinical outcomes. The set is not re-capped by culture or xenograft limitations. Broad primacy requires a representative comparison of mutation, tissue and cell-state causation across initiation, maintenance and relapse; that burden is not established by genetic relevance alone. Benign driver clones challenge sufficiency, not the strong causal genetic evidence.

In established genotype-defined cancers, engaging drivers initiates native tumors, removing/restoring causal driver function blocks or substantially delays malignant growth, and temporal on/off activity controls progression. Primary-driver status is supported within these tested settings, including APC, p53/Kras and genotype-dependent clinical cancers. It is not a declaration of universal primacy for every histology, initiation route or relapse state. The broad Ε score and the role answer different questions.

Momentum: Mechanistic only. Recent multiplex autochthonous editing, transient Polycomb initiation, APOBEC perturbation and organoid evolution refine specified causal routes; this assembled corpus does not identify an ongoing representative pan-cancer mutation-versus-tissue trial crossing SC3.

Stability: Stable. No verified pending representative comparison is claimed to change the broad primacy grade imminently.

Divergence: Contested; no single implied-level arithmetic. These dimensions do not change level or confidence.

PROPOSED DECISIVE EXPERIMENTS

Across predeclared representative histologies, use lineage-resolved native cancers with selective driver correction/restoration and factorial niche/chromatin perturbations, preserving viability and measuring both early initiation and established maintenance/relapse.

Readout: Native malignant incidence, longitudinal tumor burden, lineage persistence, metastasis, survival and recurrence with allele-sensitive genome/state monitoring. Controls: Matched passenger editing, inactive perturbation, normal mutant clones, unchanged baseline viability, tissue lineage and restoration of the specific driver versus matched state rescue. Supporting: Driver loss/correction suppresses initiation and sustained malignancy across contexts, matched restoration rescues it, and residual tissue/state effects account for only bounded exceptions. Falsifying: Verified driver correction leaves sustained malignancy in representative native contexts, or genotype-matched initiation depends dominantly on a separable tissue/state perturbation despite the same genetic state.

RESEARCH TIMELINE

KNOWLEDGE GAPS

Distinguish selected normal-clone fitness from malignant-driver function with matched normal/tumor lineage perturbations.

Compare mutation and tissue/state causation factorially across representative predeclared histologies; avoid responder-defined scope.

Test initiation and maintenance separately, retaining residual cells and relapse rather than equating regression with eradication.

Determine mutation-independent human cancer frequency with sensitive whole-genome/structural profiling and native functional tests; rare engineered examples cannot estimate prevalence.

Validate precise driver correction/restoration alongside pathway drugs to separate genotype dependency, off-target efficacy and incomplete engagement.

Track therapy-tolerant regulatory states into acquired genetic and nongenetic relapse, with longitudinal allele/state resolution.

Resolve niche-mediated genetic causality and context-specific stromal restraint instead of assuming competing models are mutually exclusive.

METHODS AND CONSENSUS CONTRIBUTION

Consensus was used through the signed-in browser for discovery, counterevidence, research-gap and timeline suggestions, and attempted checks of specific indexed full-text passages. Returned Methods/Results passages were reconciled with original-paper passages where available; failed retrievals, source substitutions and unverified aspects are recorded in the case reconciliation. Successful passage checks are distinguished from provider summaries. Provider screening counts, citation counts and meters were not used as primary counts or HSEF scores. This source log records access depth; it does not claim complete-paper or supplement review. The monthly API allowance was conserved: zero Consensus API calls in this batch.

REVIEWER NOTES

70 primary reports examined, eight with selected original-body sections and 62 at complete abstract/status depth. No complete-paper/supplement audit. Correction metadata and unresolved blank notices are retained. Two distinct Consensus Pro queries plus one retry of a lost connection; no Deep/API. Ranking categorical prior was visible, and historical tau/neuro letters had been visible elsewhere in the batch; no blinded grading or fresh calibration claimed.

Search leads, reviews, protocols and correction notices do not pad the primary count. Downloaded XML does not earn examination credit. Linked notices remain attached; unexamined correction details limit inference. This is a coordinator source/claim consistency assessment, not independent blinded calibration.

ADVERSARIAL REVIEW AND APPROVAL

Authoring model: Sol 6.1. Adversarial reviewer: Claude Opus 5.5 (Anthropic). Date: 2026-10-06.

Scope of checks. Reference PMIDs verified 70/70 against PubMed records (titles, first authors, journals, years; no retractions; one recorded erratum retained in the card for [5]). Load-bearing claims spot-checked: 35 (25 against PubMed abstracts, 10 against the local full-text/ originals for PMID23226501, PMID26091037, PMID24856586, PMID20371346 and PMID38658752). Scoring logic re-derived from SKILL.md Step 3, references/hsef-framework.md (ladder definitions and the v5.2 independence test) and references/evidence-rubrics/02-cancer-biology.md (cap precedence, Γ hard rules, Δ/Ε/Ζ/Η thresholds). Timeline dates, gap statements, cap audit, count consistency and the collective-first-author reference bug were audited separately.

Rejected findings (checked, no change).

1. Scope-matched aggregation. Re-derived against SKILL.md Step 3 — "overall equals the highest subclaim whose scope matches the central claim's scope; narrower subclaims reported separately". The central claim's scope is cancer broadly; only SC3 ("cancer broadly, with tissue and histology set before outcome") matches it, so the overall level is SC3's Ε, with SC1 (Ζ, prespecified histologies/mutation classes) and SC2 (Η, genotype-defined models and patients) reported at their own narrower scopes. The alternatives are reported and arithmetically correct (weakest-link Ε, primary-subclaim Ε). No instance of the "highest subclaim regardless of scope" error. The index prescribes scope-matched for this hypothesis and anticipates SC1–SC2 reaching Η.
2. SC1 at Ζ independence basis. The recorded basis names five lines with distinct systematic errors. SC1 is carried by at least two that could not both fail for one reason: human genomic selection inference (recurrence statistics, normal-tissue clone fitness, longitudinal clone tracking) and autochthonous endogenous perturbation (multiplexed in vivo tumour-suppressor editing with inert-cutting and Cas9-negative controls, matched restoration). Sources pooled within a line are correctly not counted as separate lines, and the review states this explicitly. Ζ, not Η, is the right stopping point because most claimed variant classes have not individually received endogenous or therapeutic confirmation.
3. SC2 at Η and "target usefulness" versus maintenance dependency. Η is carried primarily by intervention-sensitive endogenous experiments — p53 restoration regressing autochthonous lymphomas/sarcomas, Apc re-expression driving sustained regression with regrowth on re-silencing, inducible Kras withdrawal and rechallenge — not by the trials. The genotype-selected randomisations are credited as a separate line for human consequence, and every trial card already records the limitation that both arms perturb the same pathway and that drug inhibition is not allele correction. No overclaim found.
4. Relapse as an unresolved contradiction blocking Η. Drug-tolerant persisters, APOBEC-driven evolution, T790M and solvent-front resistance are escape from an established dependency within the scope, not a contradiction of it; the review uses them to block Θ/Ω rather than Η. Defensible as written.
5. Reference style. Entries 1, 29 and 50 append "et al." to complete one- or two-author lists. Builder formatting, not a citation-integrity error; left unchanged for batch consistency. No collective-first-author (", Lyketsos CG …") instances exist in this review.
6. Citation years. Item [48] (PMID17189715) is printed 2007 and PubMed-dated 2006; within the ±1 print/e-pub convention.

Residual limitations. Access depth is unchanged: 8 partial-body examinations and 62 abstract/status examinations, no complete-paper or supplement audit, so body-level claims outside the five checked full texts rest on abstracts. The competing model ranked 1 (integrated mutation–tissue, Ζ) does not carry its own named independence basis; it reuses the review-level record. Verification here is bounded attribution by a single adversarial reviewer, not blinded calibration.

Level: unchanged at Ε / Moderate / primary driver; subclaims unchanged at SC1 Ζ, SC2 Η, SC3 Ε. No score-card level fields were altered.

Approval: Following adversarial review and the corrections listed above, the revised document is approved. — Claude Opus 5.5 (Anthropic), 2026-10-06.

REFERENCES

1. Knudson AG et al. Mutation and cancer: statistical study of retinoblastoma. Proceedings of the National Academy of Sciences of the United States of America. 1971. PMID 5279523. DOI: 10.1073/pnas.68.4.820. Access: abstract_metadata. Study: PMID5279523.

2. Vogelstein B, Fearon ER, Hamilton SR et al. Genetic alterations during colorectal-tumor development. The New England journal of medicine. 1988. PMID 2841597. DOI: 10.1056/NEJM198809013190901. Access: abstract_metadata. Study: PMID2841597.

3. Martincorena I, Roshan A, Gerstung M et al. Tumor evolution. High burden and pervasive positive selection of somatic mutations in normal human skin. Science (New York, N.Y.). 2015. PMID 25999502. DOI: 10.1126/science.aaa6806. Access: abstract_metadata. Study: PMID25999502.

4. Lee-Six H, Olafsson S, Ellis P et al. The landscape of somatic mutation in normal colorectal epithelial cells. Nature. 2019. PMID 31645730. DOI: 10.1038/s41586-019-1672-7. Access: abstract_metadata. Study: PMID31645730.

5. Martincorena I, Raine KM, Gerstung M et al. Universal Patterns of Selection in Cancer and Somatic Tissues. Cell. 2017. PMID 29056346. DOI: 10.1016/j.cell.2017.09.042. Access: abstract_metadata. Study: PMID29056346.

6. Yoshida K, Gowers KHC, Lee-Six H et al. Tobacco smoking and somatic mutations in human bronchial epithelium. Nature. 2020. PMID 31996850. DOI: 10.1038/s41586-020-1961-1. Access: abstract_metadata. Study: PMID31996850.

7. Priestley P, Baber J, Lolkema MP et al. Pan-cancer whole-genome analyses of metastatic solid tumours. Nature. 2019. PMID 31645765. DOI: 10.1038/s41586-019-1689-y. Access: abstract_metadata. Study: PMID31645765.

8. Ma X, Liu Y, Liu Y et al. Pan-cancer genome and transcriptome analyses of 1,699 paediatric leukaemias and solid tumours. Nature. 2018. PMID 29489755. DOI: 10.1038/nature25795. Access: abstract_metadata. Study: PMID29489755.

9. Kandoth C, McLellan MD, Vandin F et al. Mutational landscape and significance across 12 major cancer types. Nature. 2013. PMID 24132290. DOI: 10.1038/nature12634. Access: abstract_metadata. Study: PMID24132290.

10. Lawson ARJ, Abascal F, Nicola PA et al. Somatic mutation and selection at population scale. Nature. 2025. PMID 41062696. DOI: 10.1038/s41586-025-09584-w. Access: abstract_metadata. Study: PMID41062696.

11. Zhu B, Tapinos A, Koka H et al. Genomes and epigenomes of matched normal and tumor breast tissue reveal diverse evolutionary trajectories and tumor-host interactions. American journal of human genetics. 2024. PMID 39492056. DOI: 10.1016/j.ajhg.2024.10.005. Access: abstract_metadata. Study: PMID39492056.

12. Martincorena I, Fowler JC, Wabik A et al. Somatic mutant clones colonize the human esophagus with age. Science (New York, N.Y.). 2018. PMID 30337457. DOI: 10.1126/science.aau3879. Access: full_text_partial. Study: PMID30337457.

13. Yokoyama A, Kakiuchi N, Yoshizato T et al. Age-related remodelling of oesophageal epithelia by mutated cancer drivers. Nature. 2019. PMID 30602793. DOI: 10.1038/s41586-018-0811-x. Access: abstract_metadata. Study: PMID30602793.

14. Fabre MA, de Almeida JG, Fiorillo E et al. The longitudinal dynamics and natural history of clonal haematopoiesis. Nature. 2022. PMID 35650444. DOI: 10.1038/s41586-022-04785-z. Access: abstract_metadata. Study: PMID35650444.

15. Colom B, Herms A, Hall MWJ et al. Mutant clones in normal epithelium outcompete and eliminate emerging tumours. Nature. 2021. PMID 34646013. DOI: 10.1038/s41586-021-03965-7. Access: abstract_metadata. Study: PMID34646013.

16. Yamaguchi M, Nakaoka H, Suda K et al. Spatiotemporal dynamics of clonal selection and diversification in normal endometrial epithelium. Nature communications. 2022. PMID 35177608. DOI: 10.1038/s41467-022-28568-2. Access: abstract_metadata. Study: PMID35177608.

17. Li R, Di L, Li J et al. A body map of somatic mutagenesis in morphologically normal human tissues. Nature. 2021. PMID 34433965. DOI: 10.1038/s41586-021-03836-1. Access: abstract_metadata. Study: PMID34433965.

18. Moore L, Leongamornlert D, Coorens THH et al. The mutational landscape of normal human endometrial epithelium. Nature. 2020. PMID 32350471. DOI: 10.1038/s41586-020-2214-z. Access: abstract_metadata. Study: PMID32350471.

19. Serrano M, Lin AW, McCurrach ME et al. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell. 1997. PMID 9054499. DOI: 10.1016/s0092-8674(00)81902-9. Access: abstract_metadata. Study: PMID9054499.

20. Blair LM, Juan JM, Sebastian L et al. Oncogenic context shapes the fitness landscape of tumor suppression. Nature communications. 2023. PMID 37828026. DOI: 10.1038/s41467-023-42156-y. Access: full_text_partial. Study: PMID37828026.

21. Flanagan DJ, Pentinmikko N, Luopajärvi K et al. NOTUM from Apc-mutant cells biases clonal competition to initiate cancer. Nature. 2021. PMID 34079124. DOI: 10.1038/s41586-021-03525-z. Access: abstract_metadata. Study: PMID34079124.

22. Ventura A, Kirsch DG, McLaughlin ME et al. Restoration of p53 function leads to tumour regression in vivo. Nature. 2007. PMID 17251932. DOI: 10.1038/nature05541. Access: abstract_metadata. Study: PMID17251932.

23. Hingorani SR, Wang L, Multani AS et al. Trp53R172H and KrasG12D cooperate to promote chromosomal instability and widely metastatic pancreatic ductal adenocarcinoma in mice. Cancer cell. 2005. PMID 15894267. DOI: 10.1016/j.ccr.2005.04.023. Access: abstract_metadata. Study: PMID15894267.

24. Hingorani SR, Petricoin EF, Maitra A et al. Preinvasive and invasive ductal pancreatic cancer and its early detection in the mouse. Cancer cell. 2003. PMID 14706336. DOI: 10.1016/s1535-6108(03)00309-x. Access: abstract_metadata. Study: PMID14706336.

25. Jackson EL, Willis N, Mercer K et al. Analysis of lung tumor initiation and progression using conditional expression of oncogenic K-ras. Genes & development. 2001. PMID 11751630. DOI: 10.1101/gad.943001. Access: abstract_metadata. Study: PMID11751630.

26. Collins MA, Brisset JC, Zhang Y et al. Metastatic pancreatic cancer is dependent on oncogenic Kras in mice. PloS one. 2012. PMID 23226501. DOI: 10.1371/journal.pone.0049707. Access: full_text_partial. Study: inducible-pancreatic-Kras-platform.

27. Chan IT, Kutok JL, Williams IR et al. Conditional expression of oncogenic K-ras from its endogenous promoter induces a myeloproliferative disease. The Journal of clinical investigation. 2004. PMID 14966562. DOI: 10.1172/JCI20476. Access: abstract_metadata. Study: PMID14966562.

28. Collins MA, Bednar F, Zhang Y et al. Oncogenic Kras is required for both the initiation and maintenance of pancreatic cancer in mice. The Journal of clinical investigation. 2012. PMID 22232209. DOI: 10.1172/JCI59227. Access: abstract_metadata. Study: inducible-pancreatic-Kras-platform.

29. Felsher DW, Bishop JM et al. Reversible tumorigenesis by MYC in hematopoietic lineages. Molecular cell. 1999. PMID 10488335. DOI: 10.1016/s1097-2765(00)80367-6. Access: abstract_metadata. Study: PMID10488335.

30. Shachaf CM, Kopelman AM, Arvanitis C et al. MYC inactivation uncovers pluripotent differentiation and tumour dormancy in hepatocellular cancer. Nature. 2004. PMID 15475948. DOI: 10.1038/nature03043. Access: abstract_metadata. Study: PMID15475948.

31. Lin WC, Rajbhandari N, Liu C et al. Dormant cancer cells contribute to residual disease in a model of reversible pancreatic cancer. Cancer research. 2013. PMID 23467612. DOI: 10.1158/0008-5472.CAN-12-2067. Access: abstract_metadata. Study: PMID23467612.

32. Jain M, Arvanitis C, Chu K et al. Sustained loss of a neoplastic phenotype by brief inactivation of MYC. Science (New York, N.Y.). 2002. PMID 12098700. DOI: 10.1126/science.1071489. Access: abstract_metadata. Study: PMID12098700.

33. Rakhra K, Bachireddy P, Zabuawala T et al. CD4(+) T cells contribute to the remodeling of the microenvironment required for sustained tumor regression upon oncogene inactivation. Cancer cell. 2010. PMID 21035406. DOI: 10.1016/j.ccr.2010.10.002. Access: abstract_metadata. Study: PMID21035406.

34. Drost J, van Jaarsveld RH, Ponsioen B et al. Sequential cancer mutations in cultured human intestinal stem cells. Nature. 2015. PMID 25924068. DOI: 10.1038/nature14415. Access: abstract_metadata. Study: PMID25924068.

35. Matano M, Date S, Shimokawa M et al. Modeling colorectal cancer using CRISPR-Cas9-mediated engineering of human intestinal organoids. Nature medicine. 2015. PMID 25706875. DOI: 10.1038/nm.3802. Access: abstract_metadata. Study: PMID25706875.

36. Fumagalli A, Drost J, Suijkerbuijk SJ et al. Genetic dissection of colorectal cancer progression by orthotopic transplantation of engineered cancer organoids. Proceedings of the National Academy of Sciences of the United States of America. 2017. PMID 28270604. DOI: 10.1073/pnas.1701219114. Access: abstract_metadata. Study: PMID28270604.

37. Mizutani T, Boretto M, Lim S et al. Recapitulating the adenoma-carcinoma sequence by selection of four spontaneous oncogenic mutations in mismatch-repair-deficient human colon organoids. Nature cancer. 2024. PMID 39487295. DOI: 10.1038/s43018-024-00841-x. Access: abstract_metadata. Study: PMID39487295.

38. Sell T, Klotz C, Fischer MM et al. Oncogenic signaling is coupled to colorectal cancer cell differentiation state. The Journal of cell biology. 2023. PMID 37017636. DOI: 10.1083/jcb.202204001. Access: abstract_metadata. Study: PMID37017636.

39. Dow LE, O'Rourke KP, Simon J et al. Apc Restoration Promotes Cellular Differentiation and Reestablishes Crypt Homeostasis in Colorectal Cancer. Cell. 2015. PMID 26091037. DOI: 10.1016/j.cell.2015.05.033. Access: full_text_partial. Study: PMID26091037.

40. Parreno V, Loubiere V, Schuettengruber B et al. Transient loss of Polycomb components induces an epigenetic cancer fate. Nature. 2024. PMID 38658752. DOI: 10.1038/s41586-024-07328-w. Access: full_text_partial. Study: PMID38658752.

41. Paszek MJ, Zahir N, Johnson KR et al. Tensional homeostasis and the malignant phenotype. Cancer cell. 2005. PMID 16169468. DOI: 10.1016/j.ccr.2005.08.010. Access: abstract_metadata. Study: PMID16169468.

42. Özdemir BC, Pentcheva-Hoang T, Carstens JL et al. Depletion of carcinoma-associated fibroblasts and fibrosis induces immunosuppression and accelerates pancreas cancer with reduced survival. Cancer cell. 2014. PMID 24856586. DOI: 10.1016/j.ccr.2014.04.005. Access: full_text_partial. Study: PMID24856586.

43. Maffini MV, Calabro JM, Soto AM et al. Stromal regulation of neoplastic development: age-dependent normalization of neoplastic mammary cells by mammary stroma. The American journal of pathology. 2005. PMID 16251424. DOI: 10.1016/S0002-9440(10)61227-8. Access: abstract_metadata. Study: PMID16251424.

44. Ricca BL, Venugopalan G, Furuta S et al. Transient external force induces phenotypic reversion of malignant epithelial structures via nitric oxide signaling. eLife. 2018. PMID 29560858. DOI: 10.7554/eLife.26161. Access: abstract_metadata. Study: PMID29560858.

45. Weaver VM, Petersen OW, Wang F et al. Reversion of the malignant phenotype of human breast cells in three-dimensional culture and in vivo by integrin blocking antibodies. The Journal of cell biology. 1997. PMID 9105051. DOI: 10.1083/jcb.137.1.231. Access: abstract_metadata. Study: PMID9105051.

46. Hill W, Lim EL, Weeden CE et al. Lung adenocarcinoma promotion by air pollutants. Nature. 2023. PMID 37020004. DOI: 10.1038/s41586-023-05874-3. Access: full_text_partial. Study: PMID37020004.

47. Maffini MV, Soto AM, Calabro JM et al. The stroma as a crucial target in rat mammary gland carcinogenesis. Journal of cell science. 2004. PMID 14996910. DOI: 10.1242/jcs.01000. Access: abstract_metadata. Study: PMID14996910.

48. Sotillo R, Hernando E, Díaz-Rodríguez E et al. Mad2 overexpression promotes aneuploidy and tumorigenesis in mice. Cancer cell. 2007. PMID 17189715. DOI: 10.1016/j.ccr.2006.10.019. Access: abstract_metadata. Study: PMID17189715.

49. Sotillo R, Schvartzman JM, Socci ND et al. Mad2-induced chromosome instability leads to lung tumour relapse after oncogene withdrawal. Nature. 2010. PMID 20173739. DOI: 10.1038/nature08803. Access: abstract_metadata. Study: PMID20173739.

50. Mintz B, Illmensee K et al. Normal genetically mosaic mice produced from malignant teratocarcinoma cells. Proceedings of the National Academy of Sciences of the United States of America. 1975. PMID 1059147. DOI: 10.1073/pnas.72.9.3585. Access: abstract_metadata. Study: PMID1059147.

51. Ohnishi K, Semi K, Yamamoto T et al. Premature termination of reprogramming in vivo leads to cancer development through altered epigenetic regulation. Cell. 2014. PMID 24529372. DOI: 10.1016/j.cell.2014.01.005. Access: abstract_metadata. Study: PMID24529372.

52. Flavahan WA, Drier Y, Liau BB et al. Insulator dysfunction and oncogene activation in IDH mutant gliomas. Nature. 2016. PMID 26700815. DOI: 10.1038/nature16490. Access: abstract_metadata. Study: PMID26700815.

53. Shaffer SM, Dunagin MC, Torborg SR et al. Rare cell variability and drug-induced reprogramming as a mode of cancer drug resistance. Nature. 2017. PMID 28607484. DOI: 10.1038/nature22794. Access: abstract_metadata. Study: PMID28607484.

54. Sharma SV, Lee DY, Li B et al. A chromatin-mediated reversible drug-tolerant state in cancer cell subpopulations. Cell. 2010. PMID 20371346. DOI: 10.1016/j.cell.2010.02.027. Access: full_text_partial. Study: PMID20371346.

55. Isozaki H, Sakhtemani R, Abbasi A et al. Therapy-induced APOBEC3A drives evolution of persistent cancer cells. Nature. 2023. PMID 37407818. DOI: 10.1038/s41586-023-06303-1. Access: abstract_metadata. Study: PMID37407818.

56. Hata AN, Niederst MJ, Archibald HL et al. Tumor cells can follow distinct evolutionary paths to become resistant to epidermal growth factor receptor inhibition. Nature medicine. 2016. PMID 26828195. DOI: 10.1038/nm.4040. Access: abstract_metadata. Study: PMID26828195.

57. Sequist LV, Waltman BA, Dias-Santagata D et al. Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors. Science translational medicine. 2011. PMID 21430269. DOI: 10.1126/scitranslmed.3002003. Access: abstract_metadata. Study: PMID21430269.

58. Solomon BJ, Tan L, Lin JJ et al. RET Solvent Front Mutations Mediate Acquired Resistance to Selective RET Inhibition in RET-Driven Malignancies. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer. 2020. PMID 31988000. DOI: 10.1016/j.jtho.2020.01.006. Access: abstract_metadata. Study: PMID31988000.

59. O'Brien SG, Guilhot F, Larson RA et al. Imatinib compared with interferon and low-dose cytarabine for newly diagnosed chronic-phase chronic myeloid leukemia. The New England journal of medicine. 2003. PMID 12637609. DOI: 10.1056/NEJMoa022457. Access: abstract_metadata. Study: IRIS-randomization.

60. Hochhaus A, Larson RA, Guilhot F et al. Long-Term Outcomes of Imatinib Treatment for Chronic Myeloid Leukemia. The New England journal of medicine. 2017. PMID 28273028. DOI: 10.1056/NEJMoa1609324. Access: abstract_metadata. Study: IRIS-randomization.

61. Blay JY, Schiffler C, Bouché O et al. A randomized study of 6 versus 3 years of adjuvant imatinib in patients with localized GIST at high risk of relapse. Annals of oncology : official journal of the European Society for Medical Oncology. 2024. PMID 39241959. DOI: 10.1016/j.annonc.2024.08.2343. Access: abstract_metadata. Study: PMID39241959.

62. Soria JC, Ohe Y, Vansteenkiste J et al. Osimertinib in Untreated EGFR-Mutated Advanced Non-Small-Cell Lung Cancer. The New England journal of medicine. 2018. PMID 29151359. DOI: 10.1056/NEJMoa1713137. Access: abstract_metadata. Study: FLAURA-randomization.

63. Ramalingam SS, Vansteenkiste J, Planchard D et al. Overall Survival with Osimertinib in Untreated, EGFR-Mutated Advanced NSCLC. The New England journal of medicine. 2020. PMID 31751012. DOI: 10.1056/NEJMoa1913662. Access: abstract_metadata. Study: FLAURA-randomization.

64. Tsuboi M, Herbst RS, John T et al. Overall Survival with Osimertinib in Resected EGFR-Mutated NSCLC. The New England journal of medicine. 2023. PMID 37272535. DOI: 10.1056/NEJMoa2304594. Access: abstract_metadata. Study: PMID37272535.

65. Peters S, Camidge DR, Shaw AT et al. Alectinib versus Crizotinib in Untreated ALK-Positive Non-Small-Cell Lung Cancer. The New England journal of medicine. 2017. PMID 28586279. DOI: 10.1056/NEJMoa1704795. Access: abstract_metadata. Study: PMID28586279.

66. Zhou C, Solomon B, Loong HH et al. First-Line Selpercatinib or Chemotherapy and Pembrolizumab in RET Fusion-Positive NSCLC. The New England journal of medicine. 2023. PMID 37870973. DOI: 10.1056/NEJMoa2309457. Access: abstract_metadata. Study: PMID37870973.

67. Chapman PB, Hauschild A, Robert C et al. Improved survival with vemurafenib in melanoma with BRAF V600E mutation. The New England journal of medicine. 2011. PMID 21639808. DOI: 10.1056/NEJMoa1103782. Access: abstract_metadata. Study: PMID21639808.

68. Robert C, Karaszewska B, Schachter J et al. Improved overall survival in melanoma with combined dabrafenib and trametinib. The New England journal of medicine. 2015. PMID 25399551. DOI: 10.1056/NEJMoa1412690. Access: abstract_metadata. Study: PMID25399551.

69. de Langen AJ, Johnson ML, Mazieres J et al. Sotorasib versus docetaxel for previously treated non-small-cell lung cancer with KRASG12C mutation: a randomised, open-label, phase 3 trial. Lancet (London, England). 2023. PMID 36764316. DOI: 10.1016/S0140-6736(23)00221-0. Access: abstract_metadata. Study: PMID36764316.

70. Hadoux J, Elisei R, Brose MS et al. Phase 3 Trial of Selpercatinib in Advanced RET-Mutant Medullary Thyroid Cancer. The New England journal of medicine. 2023. PMID 37870969. DOI: 10.1056/NEJMoa2309719. Access: abstract_metadata. Study: PMID37870969.

Source-specific evidence cards

[1] Mutation and cancer: statistical study of retinoblastoma.

Shows: Original observations on 48 retinoblastoma cases and age/multiplicity statistics support a two-event model distinguishing inherited and sporadic disease.

Does not show: Historical statistical inference does not directly identify the mutations or establish maintenance dependence; the inherited first event is not somatic.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[2] Genetic alterations during colorectal-tumor development.

Shows: Across 172 colorectal neoplasms, RAS mutation and chromosome losses accumulate with pathological progression.

Does not show: Cross-sectional stages do not directly trace a single lineage or prove a universal obligatory order or ongoing dependency.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[3] Tumor evolution. High burden and pervasive positive selection of somatic mutations in normal human skin.

Shows: Deep sequencing of 234 normal skin biopsies from four people finds strong selection of cancer-associated genes while epidermal function is maintained.

Does not show: Driver presence and competitive fitness are not equivalent to malignant sufficiency; benign clones do not refute later cooperative genetic causality.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[4] The landscape of somatic mutation in normal colorectal epithelial cells.

Shows: Whole genomes from normal crypts of 42 individuals reveal mutational processes and probable drivers in around 1 percent of middle-aged crypts.

Does not show: Normal mutant crypts are not inevitable precursors; acquisition alone does not establish cancer or maintenance.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[5] Universal Patterns of Selection in Cancer and Somatic Tissues.

Shows: Selection analysis of 7664 tumors across 29 types distinguishes positively selected coding variants and context-dependent driver burdens.

Does not show: Statistical selection is not matched functional validation of every variant; coding substitution counts omit other genetic classes. Linked erratum metadata is retained.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. ErratumIn: PMID29906452. Subclaims: SC1, SC3.

[6] Tobacco smoking and somatic mutations in human bronchial epithelium.

Shows: Whole genomes from 632 normal bronchial colonies link smoking to mutation burden and drivers; lower-burden cells are more frequent after cessation.

Does not show: Sixteen subjects and normal-cell expansion do not directly estimate cancer prevention mediation; carcinogenic promotion can accompany mutagenesis.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[7] Pan-cancer whole-genome analyses of metastatic solid tumours.

Shows: 2520 metastatic tumor-normal genome pairs show predominantly clonal drivers, frequent genome duplication and biallelic suppressor inactivation.

Does not show: Genomic prevalence and proposed actionability are not experimental necessity or successful treatment of each lesion.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[8] Pan-cancer genome and transcriptome analyses of 1,699 paediatric leukaemias and solid tumours.

Shows: 1699 pediatric cancers reveal driver genes distinct from adult disease and a majority of alteration events as copy-number or structural changes.

Does not show: Restricting mutation to point substitutions would misrepresent this genetic evidence; six histotypes do not establish pan-cancer primacy by perturbation.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[9] Mutational landscape and significance across 12 major cancer types.

Shows: TCGA analysis of 3281 tumors across 12 types identifies significantly mutated genes, pathways and inferred clonal timing.

Does not show: Significance and survival associations cannot establish every alteration as functional or define an obligatory cancer sequence.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[10] Somatic mutation and selection at population scale.

Shows: NanoSeq profiles 1042 oral and 371 blood samples, detecting many selected clones with high sequencing specificity.

Does not show: Population-scale clonal fitness is not malignant sufficiency or progression necessity; these normal tissues are not independent cancer interventions.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[11] Genomes and epigenomes of matched normal and tumor breast tissue reveal diverse evolutionary trajectories and tumor-host interactions.

Shows: Matched normal-adjacent and breast-tumor genome/epigenome profiles from 43 people reveal shared and divergent evolutionary histories.

Does not show: Adjacent tissue and inferred phylogenies cannot alone prove nongenetic initiation or distinguish all tumor-host directions.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[12] Somatic mutant clones colonize the human esophagus with age.

Shows: Normal esophagus has abundant positively selected NOTCH1 and TP53 clones; NOTCH1 alterations are more prevalent in normal tissue than cancer.

Does not show: Original sections distinguish approximate allele-fraction-derived cell coverage from independent mutations, and contrast TP53 enrichment in tumors. Normal fitness does not negate genotype-specific cancer dependencies.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 0:Detection of mutations in normal esophagus, 2:Colonization of the epithelium by NOTCH1-mutant clones. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[13] Age-related remodelling of oesophageal epithelia by mutated cancer drivers.

Shows: 682 normal esophageal samples show age/lifestyle-associated expansion and overrepresentation of some driver genes relative to cancer.

Does not show: Cross-sectional remodelling is not direct malignant-lineage tracking; gene-specific fitness differs between normal and cancer tissue.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[14] The longitudinal dynamics and natural history of clonal haematopoiesis.

Shows: Longitudinal tracking of 697 hematopoietic clones in 385 people links genotype-specific growth and malignant progression risk.

Does not show: Risk enrichment and growth are not deterministic malignancy; age and competitive context condition the same mutation.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[15] Mutant clones in normal epithelium outcompete and eliminate emerging tumours.

Shows: Endogenous mouse esophageal carcinogenesis and competitive-clone induction/inhibition show mutant neighboring epithelium can eliminate nascent tumors.

Does not show: Competition is a causal tissue modifier with retained mutant neoplasms, not proof that genetic lesions are dispensable across cancer.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[16] Spatiotemporal dynamics of clonal selection and diversification in normal endometrial epithelium.

Shows: 1311 normal endometrial glands reveal rhizome-linked clonal expansion and early copy-neutral loss of heterozygosity persisting in normal tissue.

Does not show: Clonal architecture and tolerated alterations are not cancer sufficiency; mutations include structural/allelic events beyond point substitutions.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[17] A body map of somatic mutagenesis in morphologically normal human tissues.

Shows: 1737 normal biopsies across nine organs from five donors show tissue-dependent mutational accumulation and clonal architecture.

Does not show: Spatial restriction and morphology do not exclude occult precursor states or establish a genetic-versus-tissue causal hierarchy.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[18] The mutational landscape of normal human endometrial epithelium.

Shows: Normal endometrial genomes contain driver-bearing clones acquired early and expanded with age, with lower burdens than cancers.

Does not show: Many normal driver-bearing glands never become malignant; driver presence alone is insufficient but can precede cooperative cancer evolution.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[19] Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a.

Shows: Oncogenic RAS induces arrest in primary cells; p53/p16 inactivation or E1A permits bypass in defined species and cell systems.

Does not show: Forced expression and cell culture establish a cooperative checkpoint, not endogenous pan-cancer necessity or a universal mutation sequence.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[20] Oncogenic context shapes the fitness landscape of tumor suppression.

Shows: Barcoded autochthonous lung editing across KRAS G12C/G12D, BRAF and EGFR contexts shows causal suppressor effects can reverse sign.

Does not show: Original Results include inert-cutting and Cas9-negative controls and separate clone number from growth. Four lung contexts do not represent all histologies; mutation frequency alone predicts some functional effects poorly.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Results:paragraph0, 1:Results:paragraph3, 1:Results:paragraph4, 1:Results:paragraph18, 1:Results:paragraph19, 1:Results:paragraph20, 1:Results:paragraph21, 1:Results:paragraph24, 1:Results:paragraph28. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2, SC3.

[21] NOTUM from Apc-mutant cells biases clonal competition to initiate cancer.

Shows: Apc-mutant cells secrete NOTUM; genetic or pharmacological inhibition reduces wild-type competition loss, mutant fixation and adenomas.

Does not show: This is genetic causality mediated through the niche, not a competing mutation-free account; prevention of adenomas is not universal invasive-cancer maintenance.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[22] Restoration of p53 function leads to tumour regression in vivo.

Shows: Restoring endogenous p53 regresses autochthonous lymphomas and sarcomas with different apoptosis/senescence responses and no detected normal-tissue injury.

Does not show: Defined p53-dependent models establish maintenance causality; they do not show every tumor is dependent or replace initiation with maintenance. Original body unavailable here.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[23] Trp53R172H and KrasG12D cooperate to promote chromosomal instability and widely metastatic pancreatic ductal adenocarcinoma in mice.

Shows: Endogenous Kras G12D and Trp53 R172H cooperate to generate invasive, metastatic pancreatic cancer with genomic instability.

Does not show: Prespecified engineered alleles show joint sufficiency in this tissue, not a universal sequence or mutation-only explanation of tissue context.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[24] Preinvasive and invasive ductal pancreatic cancer and its early detection in the mouse.

Shows: Physiological mutant Kras expression induces the PanIN spectrum, with low-frequency spontaneous invasive/metastatic progression.

Does not show: Initiation is genetically imposed and pancreatic-lineage-specific; additional evolution and microenvironmental conditions remain.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[25] Analysis of lung tumor initiation and progression using conditional expression of oncogenic K-ras.

Shows: Conditional endogenous oncogenic Kras permits timed autochthonous lung tumor induction and progression analysis.

Does not show: Cre induction supplies the initiating alteration; no mutation-loss maintenance test or broad histological generalization follows from this report alone.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[26] Metastatic pancreatic cancer is dependent on oncogenic Kras in mice.

Shows: Inducible pancreatic Kras withdrawal regresses primary and metastatic tumors; reactivation rapidly restores disease.

Does not show: Original Results show TetO-Kras with endogenous mutant p53 and induced pancreatitis. Off-doxy relapse occurs after culture/transplantation, not in the autochthonous mice; do not conflate those arms or call Kras endogenous expression.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Results:paragraph0, 1:Results:paragraph19, 1:Results:paragraph20, 1:Results:paragraph22, 1:Results:paragraph24, 1:Results:paragraph26, 1:Results:paragraph29. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[27] Conditional expression of oncogenic K-ras from its endogenous promoter induces a myeloproliferative disease.

Shows: Conditional oncogenic Kras from its endogenous promoter causes fully penetrant myeloproliferation but not transplantable AML.

Does not show: A disease precursor dependency is not AML sufficiency, supporting genetic cooperation and lineage boundaries.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[28] Oncogenic Kras is required for both the initiation and maintenance of pancreatic cancer in mice.

Shows: Reversible pancreas-specific Kras expression induces and maintains lesions, with withdrawal causing regression and changes in fibroinflammatory stroma.

Does not show: Transgenic activation and tissue damage condition initiation; stromal mediation can be downstream of a causal mutation rather than an exclusive alternative.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[29] Reversible tumorigenesis by MYC in hematopoietic lineages.

Shows: Conditional MYC expression produces hematologic malignancies that regress after transgene inactivation, with differentiation and restored hematopoiesis.

Does not show: Forced oncogene expression without endogenous mutation correction narrows interpretation to the induced dependency.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[30] MYC inactivation uncovers pluripotent differentiation and tumour dormancy in hepatocellular cancer.

Shows: MYC-off liver tumors differentiate and become dormant; reactivation restores malignancy with the same comparative-genomic signature.

Does not show: Reversal changes driver activity rather than eliminating every mutation; latent tumor identity and tissue-specific responses constrain permanent cure claims.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[31] Dormant cancer cells contribute to residual disease in a model of reversible pancreatic cancer.

Shows: Inducible pancreatic MYC tumors regress after inactivation but residual cells rapidly recur on re-expression.

Does not show: Marker-positive residual cells are not proof of a universal stem hierarchy; transgene-driven maintenance does not establish natural mutation primacy.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[32] Sustained loss of a neoplastic phenotype by brief inactivation of MYC.

Shows: Brief MYC inactivation produces sustained osteogenic tumor regression; reactivation induces apoptosis instead of recurrence.

Does not show: Tissue-specific epigenetic memory modifies response to the same driver; induced expression and inferred epigenetic mechanism limit universal conclusions.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[33] CD4(+) T cells contribute to the remodeling of the microenvironment required for sustained tumor regression upon oncogene inactivation.

Shows: CD4 and thrombospondin functional contrasts show immune remodelling is required for sustained regression after MYC/BCR-ABL inactivation.

Does not show: Oncogene withdrawal and immune context cooperate; immune necessity does not negate genetic dependence. Linked erratum lacks a supplied PMID and details remain unexamined.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. ErratumIn: PMIDnot supplied. Subclaims: SC2, SC3.

[34] Sequential cancer mutations in cultured human intestinal stem cells.

Shows: Sequential APC/TP53/KRAS/SMAD4 editing produces niche-independent human organoids and invasive xenograft tumors.

Does not show: Culture selection and xenotransplantation alone do not establish endogenous human cancer dependency; engineered order is not proof of obligatory order.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2, SC3.

[35] Modeling colorectal cancer using CRISPR-Cas9-mediated engineering of human intestinal organoids.

Shows: Isogenic five-pathway mutant organoids form tumors but fail liver macrometastatic colonization, unlike chromosome-unstable adenoma-derived organoids.

Does not show: Driver edits support causal sufficiency in the tested niche; invasion/metastasis requires additional lesions or context and is not fully reproduced.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2, SC3.

[36] Genetic dissection of colorectal cancer progression by orthotopic transplantation of engineered cancer organoids.

Shows: Orthotopic engineered organoids show mutation accumulation supports metastasis; specific niche signals restore growth when one oncogenic alteration is absent.

Does not show: Transplantation and niche rescue are bounded model evidence, not proof of mutation dispensability in all spontaneous cancers.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[37] Recapitulating the adenoma-carcinoma sequence by selection of four spontaneous oncogenic mutations in mismatch-repair-deficient human colon organoids.

Shows: MMR-deficient organoids acquire spontaneous pathway mutations under sequential niche-factor withdrawal and form xenograft tumors.

Does not show: Repair-defective background and designed selection are not ordinary in vivo initiation; spontaneous acquisition does not remove culture/transplant limits.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2, SC3.

[38] Oncogenic signaling is coupled to colorectal cancer cell differentiation state.

Shows: Isogenic and patient-derived colorectal organoids link driver combinations to differentiation-dependent signaling and transcription.

Does not show: Single-cell maps and state associations do not isolate organismal necessity; differentiation coupling can coexist with genetic causality.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[39] Apc Restoration Promotes Cellular Differentiation and Reestablishes Crypt Homeostasis in Colorectal Cancer.

Shows: Endogenous APC protein restoration reverses autochthonous colorectal adenomas/carcinomas, including Kras/p53-altered settings, with lineage-marked normal crypt behavior.

Does not show: Original Results show reversible shRNA loss rather than precise mutation repair, small Kras/p53 groups, residual microadenomas and repeat regrowth on APC re-silencing; this is powerful defined dependency, not permanent eradication of all mutant cells.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Results:paragraph7, 1:Results:paragraph8, 1:Results:paragraph9, 1:Results:paragraph10, 1:Results:paragraph13, 1:Results:paragraph14, 1:Results:paragraph15, 1:Results:paragraph16. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[40] Transient loss of Polycomb components induces an epigenetic cancer fate.

Shows: Transient endogenous Polycomb depletion induces persistent native fly neoplasia after protein recovery; serial allografts grow and metastasize.

Does not show: Original Results/Methods test four controls and 12 tumors with pooled whole-genome data, no recurrent drivers and variant/allele-fraction thresholds. This does not prove zero sequence changes or ordinary human mutation-free prevalence; transplantation is not its only initiation evidence.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Epigenetic perturbations initiate tumours:paragraph0, 1:Epigenetic perturbations initiate tumours:paragraph2, 1:Epigenetic perturbations initiate tumours:paragraph3, 1:Epigenetic perturbations initiate tumours:paragraph4, 1:Epigenetic perturbations initiate tumours:paragraph5, 6:EICs are autonomous immortal tumours:paragraph0, 6:EICs are autonomous immortal tumours:paragraph1, 8:Methods:paragraph19, 8:Methods:paragraph29, 8:Methods:paragraph30. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[41] Tensional homeostasis and the malignant phenotype.

Shows: Matrix stiffness, integrin/Rho/ERK manipulations and contractility reduction alter malignant architecture in an integrated signaling circuit.

Does not show: In vitro phenotype reversion is not pan-cancer initiation or survival causality and does not remove the transformed genotype.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[42] Depletion of carcinoma-associated fibroblasts and fibrosis induces immunosuppression and accelerates pancreas cancer with reduced survival.

Shows: Genetic depletion of proliferating alphaSMA fibroblasts accelerates pancreatic cancer and worsens survival; anti-CTLA4 partially rescues the altered state.

Does not show: Original Results replicate genetic backgrounds and retain poorer survival despite lower tumor weight, with pulmonary emboli and immune changes; stroma can restrain rather than uniformly promote cancer.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Results:paragraph0, 1:Results:paragraph1, 1:Results:paragraph6, 1:Results:paragraph7, 1:Results:paragraph8. No complete-paper or supplement audit claimed. ErratumIn: PMIDnot supplied. Subclaims: SC2, SC3.

[43] Stromal regulation of neoplastic development: age-dependent normalization of neoplastic mammary cells by mammary stroma.

Shows: Rat mammary stroma age/parity changes tumor incidence and permits normal ducts containing a tumor-origin RAS marker.

Does not show: Transplantation and host developmental/reproductive differences do not establish a universal tissue-first origin; persistent RAS marker supports genotype–phenotype separation.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[44] Transient external force induces phenotypic reversion of malignant epithelial structures via nitric oxide signaling.

Shows: Brief compression induces normal-like malignant breast structures; nitric-oxide inhibition blocks the architectural response.

Does not show: Three-dimensional morphology is not verified spontaneous cancer elimination or survival; altered context does not erase genomic lesions.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[45] Reversion of the malignant phenotype of human breast cells in three-dimensional culture and in vivo by integrin blocking antibodies.

Shows: Beta1-integrin inhibitory antibody/Fab reverses malignant breast architecture and reduces nude-mouse tumor formation; stimulatory antibody fails.

Does not show: Reversion is reversible on antibody removal; transplant/culture evidence supports contextual phenotype control but not tissue primacy across endogenous cancers.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[46] Lung adenocarcinoma promotion by air pollutants.

Shows: PM2.5 promotes EGFR/KRAS-mutant lung lesions; macrophage/IL1beta experiments connect inflammatory promotion to altered epithelial state.

Does not show: Original Results show pre-existing engineered drivers, short-term PM-versus-PBS genome comparisons and concomitant antibody suppression. This is nonmutagenic promotion in a mutant substrate, not proof that mutations are unnecessary.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:PM-mediated promotion of lung cancer:paragraph0, 1:PM-mediated promotion of lung cancer:paragraph1, 3:PM induces IL-1β production from macrophages. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[47] The stroma as a crucial target in rat mammary gland carcinogenesis.

Shows: Rat recombination contrasts place NMU-associated mammary tumor initiation with exposed stroma, independent of epithelial exposure; Ha-ras status does not track all tumors.

Does not show: Different compartment exposure protocols, transplantation and one-gene mutation testing do not establish absence of other drivers or pan-cancer tissue-first necessity.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[48] Mad2 overexpression promotes aneuploidy and tumorigenesis in mice.

Shows: Transient Mad2 overexpression induces chromosome instability and multiple tumor types; sustained Mad2 is unnecessary for maintenance.

Does not show: Chromosome gains/losses are genetic alterations; forced checkpoint expression is not an endogenous mutation-free origin or universal CIN-first model.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[49] Mad2-induced chromosome instability leads to lung tumour relapse after oncogene withdrawal.

Shows: Mad2-mediated chromosome instability increases Kras-independent relapse after initial regression on oncogene withdrawal.

Does not show: Engineered chromosome instability and varied relapse pathways support a context, not all cancers or absence of genetic evolution.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[50] Normal genetically mosaic mice produced from malignant teratocarcinoma cells.

Shows: Malignant teratocarcinoma cells contribute to functional normal tissues and germline after blastocyst incorporation.

Does not show: Modal chromosome number and strain markers are not modern genome-wide driver exclusion; rare embryonic reprogramming cannot estimate ordinary adult cancer prevalence.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[51] Premature termination of reprogramming in vivo leads to cancer development through altered epigenetic regulation.

Shows: Interrupted in vivo reprogramming induces dysplastic tumors, and tumor-derived iPSCs can form nonneoplastic kidney cells.

Does not show: Forced reprogramming and developmental reset demonstrate context-specific epigenetic plasticity; reversibility is not sensitive genome-wide proof of zero mutations or broad human predominance.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[52] Insulator dysfunction and oncogene activation in IDH mutant gliomas.

Shows: IDH-mutant glioma insulation is altered; demethylation partly restores regulation and CTCF-site disruption activates PDGFRA in wild-type-IDH spheres.

Does not show: Mutation-to-epigenetic causation is an integrated mechanism, not an independent refutation of mutations. Gliomasphere proliferation is not complete in vivo cancer dependency.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2, SC3.

[53] Rare cell variability and drug-induced reprogramming as a mode of cancer drug resistance.

Shows: Rare melanoma transcriptional states predict resistance and drug-induced reprogramming stabilizes it.

Does not show: Resistance-state plasticity does not establish nongenetic original initiation; no clear genetic cause is not exhaustive exclusion. Linked correction details are qualified.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. ErratumIn: PMID29466340. Subclaims: SC2, SC3.

[54] A chromatin-mediated reversible drug-tolerant state in cancer cell subpopulations.

Shows: Single-clone cancer cultures yield reversible persisters; KDM5A depletion reduces tolerance without detectable parental proliferation loss.

Does not show: Original Results distinguish nine-doubling persister reversal from roughly 90-doubling expanded-persistor reversal and CD133/quiescence controls. These cells retain cancer-driving mutations; clinical relapse necessity is not directly tested.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:RESULTS:paragraph3, 1:RESULTS:paragraph4, 1:RESULTS:paragraph6, 1:RESULTS:paragraph7. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[55] Therapy-induced APOBEC3A drives evolution of persistent cancer cells.

Shows: Therapy induces APOBEC3A in persisters; deletion reduces mutations/structural changes and delays resistance, with concordant patient signatures.

Does not show: Genetic evolution can arise from a nongenetic persister state; patient signatures and cell models do not prove every resistance lineage follows this route.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[56] Tumor cells can follow distinct evolutionary paths to become resistant to epidermal growth factor receptor inhibition.

Shows: T790M resistance emerges through either pre-existing clones or evolution from tolerant cells; apoptosis-targeted treatment restores sensitivity in a subset.

Does not show: Patient-derived culture convergence supports distinct paths, not a single universal sequence or permanent dependence on the original inhibitor target.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[57] Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors.

Shows: Biopsies from 37 resistant EGFR-mutant lung cancers retain original drivers but show additional genetic or histological changes, including small-cell transition.

Does not show: Retained driver presence does not prove retained dependency; serial samples are selected resistant cases and not randomized mechanism perturbations.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2, SC3.

[58] RET Solvent Front Mutations Mediate Acquired Resistance to Selective RET Inhibition in RET-Driven Malignancies.

Shows: Serial patient RET resistance variants converge at solvent-front G810, with xenograft and enzyme/cell assays confirming loss of inhibition.

Does not show: Small selected resistance series, shared pathway inhibitors and xenografts do not quantify pan-cancer mutation primacy or exclude state-based escape.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2, SC3.

[59] Imatinib compared with interferon and low-dose cytarabine for newly diagnosed chronic-phase chronic myeloid leukemia.

Shows: 1106-person IRIS randomization improves cytogenetic response and freedom from blast/accelerated progression with BCR-ABL-directed imatinib.

Does not show: Drug effects are pathway perturbation rather than correction of the chromosome fusion; crossover and additional kinase targets limit isolated genetic mediation.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[60] Long-Term Outcomes of Imatinib Treatment for Chronic Myeloid Leukemia.

Shows: IRIS long-term follow-up reports sustained imatinib efficacy and 83.3 percent estimated ten-year survival.

Does not show: Same original randomized cohort; heavy crossover means this survival estimate is not an independent randomized survival comparison or cure of every stem cell.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[61] A randomized study of 6 versus 3 years of adjuvant imatinib in patients with localized GIST at high risk of relapse.

Shows: IMADGIST randomizes 136 high-risk resected GIST patients after three years of imatinib; continuation improves disease-free survival.

Does not show: No survival difference is demonstrated; treatment-selected GIST and KIT/PDGFRA pathway targeting do not generalize to all tumors or establish mutation order.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[62] Osimertinib in Untreated EGFR-Mutated Advanced Non-Small-Cell Lung Cancer.

Shows: 556-person FLAURA compares EGFR inhibitors and improves PFS with osimertinib in defined sensitizing mutations.

Does not show: Both arms inhibit EGFR, so comparative efficacy also reflects drug properties; interim overall survival did not meet its statistical boundary.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[63] Overall Survival with Osimertinib in Untreated, EGFR-Mutated Advanced NSCLC.

Shows: Planned FLAURA secondary survival analysis favors osimertinib, median 38.6 versus 31.8 months.

Does not show: Same FLAURA patients, not a separate replication; more effective genotype-matched drugs support dependency without proving unselected pan-cancer dominance.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[64] Overall Survival with Osimertinib in Resected EGFR-Mutated NSCLC.

Shows: 682-person ADAURA adjuvant randomization improves overall survival after EGFR-mutant lung-cancer resection, with or without prior chemotherapy.

Does not show: Defined stages and mutations plus surgery constrain scope; therapeutic consequence does not directly establish spontaneous initiation or universal sufficiency.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[65] Alectinib versus Crizotinib in Untreated ALK-Positive Non-Small-Cell Lung Cancer.

Shows: 303-person ALEX improves PFS and CNS progression with alectinib versus crizotinib in ALK-positive lung cancer.

Does not show: Both arms target ALK; pharmacology and CNS penetration differ, and initial report does not establish overall-survival benefit or universal ALK dependence.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[66] First-Line Selpercatinib or Chemotherapy and Pembrolizumab in RET Fusion-Positive NSCLC.

Shows: LIBRETTO-431 compares selective RET inhibition with chemotherapy with/without pembrolizumab and improves PFS in fusion-positive lung cancer.

Does not show: A structural rearrangement is genetic driver evidence; selected advanced disease and immature/crossover survival do not establish all-cancer primacy.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[67] Improved survival with vemurafenib in melanoma with BRAF V600E mutation.

Shows: 675-person BRIM3 improves survival and PFS with vemurafenib versus dacarbazine in BRAF V600E melanoma.

Does not show: Clinical selection and eventual resistance constrain durability; target inhibition does not correct every mutation or remove tissue/immune mediation.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[68] Improved overall survival in melanoma with combined dabrafenib and trametinib.

Shows: 704-person COMBI-v improves overall survival with BRAF/MEK combination versus vemurafenib in BRAF V600 melanoma.

Does not show: A downstream combination strengthens consequence and resistance control, not mutation-only origin; distinct from BRIM3 despite related drug pathway.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[69] Sotorasib versus docetaxel for previously treated non-small-cell lung cancer with KRASG12C mutation: a randomised, open-label, phase 3 trial.

Shows: 345-person CodeBreaK200 improves PFS with sotorasib versus docetaxel in previously treated KRAS G12C lung cancer.

Does not show: The benefit is modest and this abstract does not establish overall-survival benefit; mutation-specific dependence varies by disease and treatment stage.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[70] Phase 3 Trial of Selpercatinib in Advanced RET-Mutant Medullary Thyroid Cancer.

Shows: 291-person LIBRETTO-531 improves PFS with selective RET inhibition versus multikinase inhibitors in RET-mutant medullary thyroid cancer.

Does not show: RET alterations may be inherited or somatic; this cohort cannot be counted as a purely somatic initiating-mutation experiment. Both arms target RET-pathway activity.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

Documentary record

Every grade on this site is reproducible from its record. The files below are the posted edition exactly as assessed.

• SHA256.txt
• codex-correction-provenance.json
• consensus-addendum.md
• consensus-validation.json
• edition-approval-audit.json
• somatic-mutation-primacy.review.md
• somatic-mutation-primacy.scorecard.json
• validation-report.json
• validation-round2-2026-10-14.json
• validation.txt

Sources: 70 · load-bearing claims: 74 · verified: 74 · unverified: 0 · contradicted: 0 · schema 5.3