In plain terms

The claim being assessed: Tumor-microenvironment factors are primary drivers of malignant maintenance and progression, dominating cell-autonomous genetic and epigenetic drivers.

Where the evidence stands: Γ — Plausible mechanism. There is a believable biological story and partial support, but the causal step has not been demonstrated.

What this mechanism does: This makes the process worse or keeps it going, but something else starts it.

Strongest evidence for it: Preregistered primary-subclaim SC2 returns Γ. Weakest-link and scope-matched sensitivities also return Γ; SC1 influence reaches Η independently. A narrow macrophage-dependent acquired-resistance state can cross Ε contribution/conditional dominance, but it does not replace the registered broad maintenance/progression claim.

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

A tumor contains cancer cells, immune cells, supporting tissue and blood vessels. This review asks how those surroundings change cancer behavior and whether they generally dominate the genetic programs inside cancer cells.

Implications for health and treatment: Preregistered primary-subclaim SC2 returns Γ. Weakest-link and scope-matched sensitivities also return Γ; SC1 influence reaches Η independently. A narrow macrophage-dependent acquired-resistance state can cross Ε contribution/conditional dominance, but it does not replace the registered broad maintenance/progression claim.

Limits for individual care: 64 primary reports with 9 selected partial-body examinations; density, heterogeneous controls and incomplete access set the confidence, not correctness. The grade applies to the exact registered claim; stronger findings for narrower subclaims remain separate. No complete-paper audit or blinded calibration is claimed. This educational review does not establish an individual treatment recommendation.

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.

Source checking used complete abstracts/status records and selected original partial-body passages. No complete-paper audit or independent scientific peer review is claimed. Detailed access counts and limitations appear in the review.

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

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: 64 primary reports; claims: 67; verified: 67; unverified: 0. Verification is bounded attribution. Access: 9 partial-body and 55 complete abstract/status examinations. Literature cutoff: 2026-10-08; source examination UTC date: 2026-10-08. No complete-paper audit or blinded calibration claimed. Adversarial review: Generated by Sol 6.1; adversarially reviewed by Claude Opus 5.5 (Anthropic) on 2026-10-14. 8 corrections applied (see ADVERSARIAL REVIEW AND APPROVAL). Status: Approved as revised.

AT A GLANCE

Tumor microenvironments causally change growth, immune escape, dormancy and treatment outcomes, but general dominance over tumor genetics is not established.

Deciding experiment: In several endogenous genotype-defined tumors at matched stages, preregister a factorial intrinsic-driver withdrawal and lineage-specific host perturbation, restoring each separately; track clones and host state through treatment-free relapse and orthogonal immune/metabolic bypass.

Belief versus evidence: Preregistered primary-subclaim SC2 returns Γ. Weakest-link and scope-matched sensitivities also return Γ; SC1 influence reaches Η independently. A narrow macrophage-dependent acquired-resistance state can cross Ε contribution/conditional dominance, but it does not replace the registered broad maintenance/progression claim.

Best-supported rival: Dual-function context-dependent stroma. Prospectively compare driver suppression and defined host-compartment perturbation factorially in the same genotype/stage, with equal engagement, restoration of each layer and clone-resolved relapse. This separates joint dependence from true causal priority.

COMPETING MECHANISTIC MODELS

Rank 1: Dual-function context-dependent stroma (Ζ). Immune, stromal and matrix compartments interact with intrinsic programs, with effects changing sign by cell subset, stage and treatment. Opposite FAP/alphaSMA effects, collagen/receptor rescue and genotype/host withdrawal: PMID35348629,33667385,21035406,27199435. Not a fully parameterized predictive theory; related CAF platforms reduce independence and signs require prespecified context.

Rank 2: Immune-compartment causality (Η). Defined immune effectors causally control malignant escape and durable treatment outcomes. Endogenous equilibrium/immunoediting, host restoration and independent checkpoint trials: PMID18026089,22318521,21035406,26028407,27718847,39282897. Some cancers/regimens escape or do not benefit; immune control does not establish all-TME or all-cancer primacy.

Rank 3: Cell-autonomous drivers with permissive niche (Ε). Intrinsic drivers remain required while host conditions supply support or determine eradication/escape. KRAS/MYC/p53 withdrawal/restoration and intrinsic immune escape: PMID22232209,17251932,21035406,27433843. Permissive understates active immune selection and niche-determined resistance; intrinsic lesions can be bypassed by alternative immune routes, PMID32467343.

Rank 4: Stroma as host containment (Ε). Selected matrix/myofibroblast and immune programs restrain growth or maintain dormancy. AlphaSMA/collagen depletion and vascular dormancy: PMID24856586,33667385,23728425. Other CAF compartments promote growth and Hh effects differ by protocol; host containment is not all-stroma universal suppression.

Rank 5: Microenvironment primacy (Γ). Host factors generally dominate intrinsic malignant maintenance/progression. Breast humanization and reversible host-dependent glioma resistance: PMID15051869,27199435. Occult donor clones, endogenous KRAS dependence and combined intrinsic/immune requirements block broad dominance.

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

Tumor-microenvironment factors are primary drivers of malignant maintenance and progression, dominating cell-autonomous genetic and epigenetic drivers.

Use the index primacy distinction, retaining maintenance/progression scope. Influence is tested separately and cannot establish dominance. Immune and fibroblast/matrix compartments can differ in sign by stage; primacy needs explicit paired cell-autonomous comparison.

Selection: public-interest rank 35 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 registered claim separates a settled contribution question from a much harder hierarchy question. Host immune, matrix and fibroblast interventions change malignant growth, escape, dormancy and clinical outcomes. That rejects a universally reactive microenvironment, but does not establish that the host generally dominates intrinsic drivers. The review scores these burdens separately rather than equating intervention success with causal priority.

Early stromal humanization is a serious primacy steelman, with important controls. Olumi uses initiated nontumorigenic epithelium, not genetically verified normal cells (PMID10519415). In Kuperwasser, normal fibroblast addition prevents aberrant outgrowth. Yet growth-factor-modified stroma produces malignancy in only one of ten donor preparations, 16/18 grafts from that preparation; another region of the same breast fails to reproduce malignancy (PMID15051869). DNA fingerprinting verifies donor origin, not absence of occult precancerous clones. These are extrinsic sufficiency/normalization tests in xenografts, not universal endogenous initiation without intrinsic predisposition.

Matrix and fibroblast effects have opposite signs. LOX-linked crosslinking promotes mammary progression, while alphaSMA myofibroblast and collagen depletion accelerates pancreatic disease (PMID19931152,24856586,33667385). Rhim deletes epithelial SHH and reduces stroma; median survival falls from 6.17 to 3.61 months and gross metastasis rises from 15.3 to 43.4 percent (PMID24856585). This is not stromal-only genetic manipulation. Chronic Smo blockade reproduces an adverse phenotype, whereas timing differs from earlier acute treatment; later Hh/CAF work can instead find growth inhibition (PMID33495315). Sign must be bounded by genotype, compartment and schedule.

The strongest stromal controls show that less matrix is not automatically beneficial. Approximately 80 percent alphaSMA depletion shortens survival despite smaller tumor weight, with hypoxia/immune change and more pulmonary emboli (PMID24856586). AntiCTLA4 extends median survival to 65 versus control 47 days without restoring the depleted myofibroblasts. Dual-recombinase alphaSMA Col1 deletion is more compartment precise: collagen and stiffness fall, survival worsens, but vascular integrity remains and Fsp1 deletion does not reproduce the effect (PMID33667385). CXCR2/CCR2 blockade restores survival toward controls through immune bypass. These establish causal restraint without making all collagen protective.

FAP versus alphaSMA depletion makes the compartment distinction explicit: the former improves survival and the latter worsens it (PMID35348629). IL6 is expressed by both populations. Conditional alphaSMA or global IL6 deletion leaves untreated progression and survival similar, yet improves gemcitabine response. The tested dual-checkpoint/gemcitabine combination uses global IL6 loss. Allocation was not randomized, whereas histology was blinded. Reused alphaSMA expression data are disclosed. A therapy-response mechanism can be important without being a baseline rate-limiting driver.

Host state can dominate a restricted resistance phenotype. CSF1R blockade initially regresses glioma; 56 percent rebound, despite continued target engagement, and recurrent cells regain sensitivity in naive hosts (PMID27199435). IGF1R plus continued CSF1R inhibition raises recurrent-tumor median survival from 13 to 63 days, while IGF1R monotherapy does not. Tumor-specific receptor knockdown also tests the adaptive dependency. Passaging and transplantation may reset state, so identical stable genomes cannot be assumed. This is strong conditional environmental control of resistance, not a demonstration covering untreated maintenance across cancers.

GOT2 illustrates why culture rescue must be tested in vivo. Cytosolic, but not mitochondrial, NADH oxidation rescues GOT2 loss in culture; the mitochondrial construct has a functional positive control (PMID35815941). CAF-conditioned pyruvate also rescues culture, but tumor growth remains insensitive to GOT2 loss in the examined xenograft and endogenous pancreatic contexts. Confirmed MCT1 loss or transport/LDH inhibition fails to restore in vivo dependence. Therefore CAF pyruvate is not uniquely established as the in vivo mediator. A separate GOT2 immune/PPARdelta function reinforces context and mechanism distinctions (PMID35894778).

Intrinsic driver withdrawal is a direct opposing contrast, but does not prove immunity-independent genetic dominance. Conditional KRAS withdrawal regresses confirmed pancreatic tumors within two weeks, with selected animals healthy 23 weeks later (PMID22232209). The transgenic driver, p53 background, cerulein injury and small MRI cohort limit scope. In MYC models, transgene suppression and initial apoptosis/arrest occur in immune-deficient hosts, yet durable clearance fails much more often; CD4 restoration recovers control and TSP-deficient immune effectors fail (PMID21035406). BCRABL provides a second tested program. Intrinsic arrest and host-dependent eradication are different endpoints and joint dependencies.

Clinical interventions give independent high-level evidence of host causal contribution. Bevacizumab added to IFL improves colorectal survival; ipilimumab, nivolumab and pembrolizumab improve selected cancer outcomes (PMID15175435,20525992,26028407,27718847). CheckMate067 has durable ten-year outcomes, but the 2015 and long-term reports are one randomized cohort, not two independent trials (PMID26027431,39282897). PDL1-high/EGFR-ALK exclusions in KEYNOTE024 and regimen-specific toxicity delimit generalization. CheckMate914, HALO301 and the glioblastoma bevacizumab trials supply null survival/DFS contrasts despite some response/PFS changes (PMID36774933,32706635,24552317,24552318). None ranks host effects directly against extinguishing tumor drivers.

Tumor genetics can organize immune escape, yet alternatives can bypass that escape. PTEN and JAK/B2M perturbations link intrinsic lesions to host response (PMID26645196,27903500,27433843). Engineered JAK/B2M resistance can be overcome through innate or alternative immune routes (PMID32467343), so immutable genetic dominance is also unsupported. Human multiregion and spatial studies map selection/neighborhoods but remain observational for layer priority (PMID30894752,41533516). The August 2026 Nature expression of concern on exosome-integrin PMID26524530 identifies potentially overlapping images; that report is historical only, with no positive causal burden assigned.

The primary-subclaim score is Γ for general microenvironment dominance, while scoped causal influence reaches Η. Dual-function context-dependent stroma and active immune causality best explain the heterogeneous interventions. Further progress requires paired tumor/host factorial perturbation, restoration and clone-resolved durability rather than deriving hierarchy from a favorable single intervention.

HYPOTHESIS STATE EVALUATION FRAMEWORK (HSEF v5.3)

Subclaim matrix and aggregation

Preregistered aggregation: primary-subclaim. Preregistered primary-subclaim SC2 returns Γ. Weakest-link and scope-matched sensitivities also return Γ; SC1 influence reaches Η independently. A narrow macrophage-dependent acquired-resistance state can cross Ε contribution/conditional dominance, but it does not replace the registered broad maintenance/progression claim. Sensitivities: weakest-link Γ; primary-subclaim Γ; scope-matched Γ.

Decisive-perturbation inventory and causal tally

[1] Carcinoma-associated versus normal fibroblasts promote tumor growth from initiated, nontumorigenic prostate epithelium. Initiated epithelium and xenograft context; not conversion of genetically verified uninitiated cells or general extrinsic dominance.

[2] Growth-factor-modified stroma produces malignant outgrowth in 16/18 grafts from one donor; adding normal fibroblasts prevents abnormal outgrowth. Nine of ten donors remain nonmalignant and a different region of the responding breast fails to reproduce malignancy; DNA identity is not exclusion of occult premalignant clones.

[3] CAFs from human breast cancers promote growth through SDF1/CXCL12-associated angiogenic recruitment and epithelial CXCR4 signaling. Coimplantation with established cancer cells and coupled angiogenesis/proliferation do not establish endogenous universal dominance.

[4] LOX-linked matrix crosslinking promotes integrin signaling and invasion; blocking crosslinking reduces fibrosis, tumor incidence and progression. Oncogene-initiated context; changing mechanics/signaling together is not proof that the matrix overrides all tumor genotypes.

[5] Fibroblast caveolin1 influences matrix remodeling, stiffness and invasion with Rho/p190 signaling contrasts. Global host knockout/development and human prognosis association constrain compartment specificity and primacy.

[6] A wound/fibrotic environment promotes squamous carcinogenesis with TGFbeta/LOX/integrin pathway contrasts. Chemical initiation and chronic injury accompany matrix changes; mutation-independent wording does not prove identical normal epithelial genomes/clones.

[7] Tumor gain-of-function p53 educates fibroblasts into a perlecan-rich niche supporting invasion and therapy resistance. Tumor genotype organizes the host phenotype; dominance of one CAF state over another is not dominance of host over tumor drivers.

[9] IL1/LIF/JAKSTAT promotes inflammatory CAFs while TGFbeta suppresses IL1R1 and favors myofibroblastic states. Tumor ligand output and CAF plasticity are coupled; state selection is not broad malignant maintenance primacy.

[10] FAP-positive CAF depletion improves survival whereas alphaSMA-positive depletion worsens it; alphaSMA-specific IL6 loss improves gemcitabine response. Baseline IL6 deletion leaves progression/survival similar; global IL6 knockout, not solely conditional alphaSMA deletion, supports the tested triple therapy. No allocation randomization, histology blinded; reused alphaSMA data are not independent replication.

[11] SHH loss reduces stroma but median survival falls from 6.17 to 3.61 months and gross metastasis rises from 15.3 to 43.4 percent; VEGFR2 blockade selectively rescues the stroma-low setting. SHH deletion is epithelial, not stromal-only; chronic versus acute blockade can have different signs. A suppressive stromal contribution does not prove opposite universal genetic dominance.

[12] Approximately 80 percent myofibroblast depletion accelerates aggressive disease and shortens survival despite lower tumor weight; antiCTLA4 increases median survival to 65 versus control 47 days. Pulmonary emboli contribute to mortality; sphere/marker enrichment is not CSC proof. Rescue bypasses rather than restores depleted fibroblasts; linked erratum contents unexamined.

[13] Hedgehog inhibition accelerates pancreatic disease whereas pathway activation expands stroma and reduces epithelial proliferation. Stage, genotype and protocol restrict sign; all CAFs and all Hh contexts cannot be called tumor suppressive.

[14] AlphaSMA-specific Col1 deletion reduces fibrillar collagen/stiffness and accelerates PDAC; CXCR2/CCR2 blockade restores survival toward control. Fsp1-directed deletion does not similarly reduce total collagen or survival; immune bypass does not restore matrix. Shows compartment-specific restraint rather than generic collagen protection.

[15] SHH overexpression promotes growth and Smo inhibition impairs it while shifting myCAF/iCAF and immune composition. Contrasts earlier restraining-Hh settings; different stage/protocols prevent one universal directional story.

[16] FAP-cell and CXCL12 contrasts identify immune exclusion; AMD3100 plus antiPDL1 rapidly removes tumor cells in a defined pancreatic model. Combined immune/niche dependence in a driver-bearing tumor does not establish host primacy over all intrinsic programs.

[17] TGFbeta-rich stroma associates with immune exclusion; combined TGFbeta/PDL1 blockade restores penetration and regression in mouse tumors. Human biomarker associations are not randomized compartment comparisons; combination efficacy cannot isolate a universal dominant layer.

[18] TGFbeta blockade prevents metastasis and enables checkpoint control of established lesions in low-mutational-burden driver-mutant CRC. Specific immune-evasion dependence in a defined tumor genotype; does not erase the required epithelial driver program.

[19] CSF1R inhibition regresses glioma and prolongs survival through macrophage-state alteration. TAMs are protected from depletion by glioma signals; functional repolarization, not simply fewer macrophages, explains the tested response.

[20] Continued CSF1R blockade produces rebound in 56 percent; rebound cells regain sensitivity in naive hosts, and IGF1R plus CSF1R inhibition extends recurrent-tumor median survival from 13 to 63 days. Passaging/transplantation can reset state; tumor-specific IGF1R knockdown also restores response. Strong scoped environmental resistance control does not establish general malignant primacy.

[22] CSF1 deficiency reduces macrophages, angiogenic switching and total pancreatic neuroendocrine tumor burden. Established tumors retain growth with compensating neutrophils; developmental effects and redundant myeloid routes limit universal necessity.

[23] Macrophage recruitment promotes an angiogenic switch; blocked infiltration delays progression and local restoration restores vascular effects. A defined angiogenic stage and transgenic platform, not all epithelial initiation or maintenance.

[24] ADAM12-marked slow mesenchymal cells acquire perivascular tumor-supporting states; depletion improves vascular/immune organization and reduces growth. State and lineage restriction, tumor-induced remodeling and different cell populations constrain a unifying stromal hierarchy.

[25] Lymphocyte/IFNgamma competence suppresses tumor development and selects tumor immunogenicity. Host deficiency and carcinogen initiation are broad changes; immune editing is not immunity-only control of every malignant behavior.

[26] Adaptive immune disruption releases occult carcinogen-induced tumors from prolonged equilibrium. Transformed cells preexist; immune containment and eventual escape do not specify universal host primacy.

[27] Immune selection favors preexisting antigen-negative sarcoma clones with defined mutant antigen recognition. Related immunoediting platform; genotype and immunity interact rather than forming separable universal hierarchies.

[28] Type I IFN dependence in host hematopoietic cells differs from IFNgamma tumor-target effects during tumor control. Conditional cell targets, not all immune pathways; linked erratum contents unexamined.

[31] VEGFR1-positive progenitor perturbation and reconstitution alter premetastatic niche formation and metastasis in defined models. Transplant/conditioned-medium and recruitment context; does not prove all spontaneous metastasis requires this niche.

[33] Stable endothelium/TSP1 sustains dormancy whereas sprouting niches and TGFbeta/periostin promote outgrowth. Defined disseminated-cell contexts; unchanged genome cannot be presumed solely from similar cell lineage, and all malignant maintenance is broader.

[34] Tumor-derived tenascinC supports early metastatic colonization before host stromal supply becomes sufficient. The niche effector initially originates from tumor cells; intrinsic/extrinsic boundary is mechanistically coupled.

[35] Restoring p53 regresses established lymphomas and sarcomas through different apoptosis/senescence responses. Tumor-type-dependent genetic restoration can remodel host effectors; not proof that every cancer is intrinsically dominant.

[36] p53-dependent TSP1 induction contributes to regression after MYC inactivation. TSP1 alone does not suppress MYC-on growth; joint oncogene/host mechanism rather than either universal primacy.

[37] p16/Rb/p53 senescence machinery influences durable regression after MYC withdrawal. Cell-state and host effects cooperate; restricted transgenic programs do not define all sporadic cancers.

[38] MYC suppression regresses genetically complex lymphomas, with recurrent tumors acquiring replacement alterations. Driver substitution after selection and transgene withdrawal differ from uniform endogenous human dependence.

[39] Turning off oncogenic KRAS regresses MRI-confirmed invasive tumors within two weeks; selected mice remain tumor-free 23 weeks after withdrawal. Transgenic expression, p53 background and cerulein injury restrict inference; small MRI cohort and simultaneous stromal remodeling do not establish opposite universal genetic primacy.

[40] MYC-off apoptosis/arrest occurs across host immune states, but CD4 loss increases recurrence; CD4 transfer restores durable control and TSP-deficient effectors fail. Initial and durable responses differ; immune bypass and intrinsic withdrawal jointly determine control. PubMed Comment label accompanies genuine original experiments; linked erratum unexamined.

[41] MYC regulates CD47/PDL1 and enforced ligand expression impairs immune remodeling after MYC suppression. Tumor genotype drives immune output; linked erratum 27124464 metadata verified, contents unexamined.

[42] PanKRAS inhibition suppresses diverse preclinical tumors and remodels tumor/microenvironment state; YAP-associated escape is interrogated. Drug specificity and feedback matter; preclinical results are not patient efficacy or universal intrinsic dominance.

[43] GOT2 loss impairs culture but not tested in vivo growth; CAF pyruvate or cytosolic NADH oxidation rescues culture, whereas mitochondrial oxidation does not. Confirmed MCT1 loss/inhibition fails to resensitize tumors in vivo, so CAF pyruvate is not established as the unique in vivo explanation; residual/alternative redox routes remain.

[44] GOT2 supports PPARdelta-linked immune suppression beyond its proliferative metabolic role. Different tumor/immune conditions and enzyme functions prevent treating context papers as a direct universal contradiction.

[45] Adding bevacizumab to IFL in 813 randomized metastatic colorectal patients increases median survival from 15.6 to 20.3 months, HR 0.66. Antiangiogenic/additive clinical benefit is not a matched comparison with extinguishing tumor drivers; delivery and vascular functions are coupled.

[46] Ipilimumab improves median survival versus gp100 from 6.4 to 10.1 months in the defined advanced melanoma trial. Selected HLA/treatment population and immune toxicity; survival benefit does not prove broad extrinsic dominance. Linked erratum contents unexamined.

[47] In 272 previously treated squamous NSCLC patients, nivolumab increases median survival from 6.0 to 9.2 months versus docetaxel, HR 0.59. Specific disease/treatment comparison; PDL1 nonprediction in this study does not mean genotype or tumor state is irrelevant.

[48] In 305 PDL1-high NSCLC patients, pembrolizumab improves PFS from 6.0 to 10.3 months, HR 0.50, versus chemotherapy. Enrollment excludes sensitizing EGFR and ALK alterations and requires PDL1 at least 50 percent; cannot generalize to all genotypes.

[49] CheckMate067 randomizes 945 melanoma patients; combination/nivolumab/ipilimumab median PFS is 11.5/6.9/2.9 months. Initial report emphasizes PFS; later long-term survival is the same trial, not independent replication. Linked erratum 31442371 contents unexamined.

[50] Ten-year CheckMate067 follow-up reports median OS of 71.9/36.9/19.9 months for combination/nivolumab/ipilimumab. Same NCT01844505 cohort as the 2015 report; online 2024 and journal issue 2025, not a new independent trial.

[51] CheckMate914 part A in 816 randomized high-risk resected renal patients finds no significant DFS gain, HR 0.92, CI 0.71–1.19, p=0.53. Specific adjuvant regimen and 32 versus 2 percent toxicity discontinuation; null is not proof immunity never controls RCC.

[52] RTOG0825 randomizes 637 of 978 registered glioblastoma patients; bevacizumab yields OS 15.7 versus 16.1 months and worse selected cognitive/QoL outcomes. PFS 10.7 versus 7.3 months does not meet preset improvement; vascular imaging/response changes are not survival proof.

[53] AVAglio in 921 patients improves PFS 10.6 versus 6.2 months, HR 0.64, with no significant OS benefit, p=0.10. QoL instruments/protocol differ from RTOG0825; do not collapse different QoL findings or infer absent vascular biology.

[54] HALO301 has 494 randomized and 492 analysis patients; hyaluronan targeting yields OS 11.2 versus 11.5 months, HR 1.00, and equal PFS 7.1 months despite greater response. Hyaluronan-high selected disease, toxicity and drug regimen; response is not survival and null does not negate all stromal contributions.

[55] IMmotion150 phase 2 randomizes 305 RCC patients; combination versus sunitinib ITT PFS HR 1.00, CI 0.69–1.45, with immune/angiogenic signatures. TMB is not associated with PFS in this study; biomarker signatures do not prove layer primacy. Linked correction 30291359 contents unexamined.

[56] PTEN loss alters cytokine/autophagy programs, T-cell access and therapy response; PI3Kbeta intervention improves immune control. Intrinsic genetic and host pathways are causally coupled, not separable universal dominant compartments.

[57] JAK1/2 alterations in rare high-mutational-burden nonresponders impair interferon-linked immune response. One of 23 melanoma and one of 16 colorectal patients; rare cases and cell assays do not establish population-wide primacy.

[58] Four acquired-resistance cases reveal JAK1/2 lesions in two and B2M loss in one with altered antigen/IFN response. Small selected series; antigen presentation and IFN mechanisms differ and do not cover all immune escape.

[59] JAK or B2M disruption produces checkpoint resistance; TLR9/PD1 or CD4/NK/IL2 routes restore selected responses. Genetically imposed immune escape can be bypassed, opposing immutable genetic dominance; engineered models do not establish universal immune primacy.

[60] Short VEGF-pathway inhibition/preconditioning can accelerate metastatic growth while suppressing primary tumors. Pretreatment, intravenous/resection context and short schedule cannot be equated with prolonged human therapy effects.

[61] VEGF-targeting reduces primary growth yet increases local invasion/metastasis in selected settings. Stage, hypoxia and vessel mechanism can change sign; no claim that all antiangiogenic clinical treatment increases metastasis.

[63] ITGB5-positive CAFs alter stromal stiffness and pancreatic ductal adenocarcinoma progression in experimental models with clinical association. Prognostic/registration statements in the abstract are not independently verified clinical validation; binding/signaling and matrix state vary together.

[64] Stromal FGF1 maintains tumor MYC under RAF/MEK blockade; adding FAK inhibition enables regression in defined pancreatic models. Joint pathway treatment reinforces tumor/host feedback; a 2024 trial statement does not establish current recruitment or clinical efficacy.

Tally: 58 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

Endogenous genetic/immune perturbations and clinical survival effects cross contribution and reach Η for SC1 when assembled across independent failure modes. Xenograft limits bind broad SC2 dominance, not the whole causal TME literature. Quail resistance/retransplantation supports a restricted environmental state dependence, while endogenous KRAS and MYC/CD4 tests show why it cannot substitute for general primacy. The assembled SC1 convergence is not re-capped by one culture study.

Gamma hard-rule audit: Coimplantation sufficiency, marker signatures and shared-cell transplantation do not establish endogenous general primacy. Outcome after changing one compartment does not compare the counterfactual effect of extinguishing an intrinsic driver; survival benefit alone is not causal hierarchy.

SC2 cap: “Mechanistically plausible cell culture or transplantation-linked models without endogenous in vivo dependency cap at Γ.” Γ: scoped host-dependent resistance and tissue-humanization models support primacy as a mechanistic possibility. Paired evidence also shows continued intrinsic dependencies and joint immune/oncogene requirements. The registered general dominance claim lacks convergent endogenous factorial comparisons; influence cannot substitute for primacy.

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. Η: endogenous compartment-specific loss, immune restoration, relapse/resistance tests and independent clinical survival interventions establish causal TME influence within explicitly defined compartments/settings. They reject the claim that all TME change is reactive; they do not make every stromal factor rate-limiting.

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: no. Γ: scoped host-dependent resistance and tissue-humanization models support primacy as a mechanistic possibility. Paired evidence also shows continued intrinsic dependencies and joint immune/oncogene requirements. The registered general dominance claim lacks convergent endogenous factorial comparisons; influence cannot substitute for primacy.

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: yes.

SC1 Θ: “Requires that the mechanism explains most of the claimed cancer behavior with only minor exceptions.” Satisfied: no. A prespecified causal test matching this subclaim, with mechanistic engagement, meaningful phenotype change and an appropriate negative control.

SC2 Δ: “Requires at least one meaningful causal perturbation showing the proposed mechanism contributes to initiation, maintenance, or relapse.” Satisfied: no. A prespecified causal test matching this subclaim, with mechanistic engagement, meaningful phenotype change and an appropriate negative control.

Upper ladder: Endogenous tumor/host genetics, oncogene withdrawal, clonal immune selection and independent survival interventions are live upper-ladder routes and carry SC1 to Η. They do not establish the relative dominance counterfactual for SC2; no RCT-only barrier is imposed. Θ would require most registered behaviors explained with minor exceptions, and Ω necessity/sufficiency across context.

Independence: Independent genetic, immune and clinical assay classes support contribution while excluding different errors. CAF/collagen papers share lineage and some reused data; MYC and immunoediting papers share platforms. CheckMate067 contributes one trial line despite two reports. Reports, outcomes and citations are not counted as independent causal confirmations of dominance.

Autochthonous stromal lineage/pathway perturbations [11, 14, 10]: Test endogenous stromal contribution with distinct epithelial ligand versus stromal matrix/cell edits; related collagen/alphaSMA laboratory platforms explicitly remain partly dependent.

Host immune loss/restoration with oncogene withdrawal [26, 40]: Exclude solely reactive immunity and distinguish immediate intrinsic response from durable immune-dependent control; related MYC/immunoediting studies are not all independent.

Independent randomized human survival interventions [45, 46, 47, 48, 49, 50]: Exclude complete reliance on transplanted mice or prognostic signatures; different disease/drug populations, with CheckMate067 follow-up one causal trial line.

Environment-sensitive metabolic/resistance controls [20, 43]: Expose culture and naive-host context dependence; drug targets, metabolism, species and investigators differ, but neither alone proves universal primacy.

Falsification and pitfalls

Removing nominally promoting stroma worsens the disease endpoint. AlphaSMA/collagen depletion shortens survival and immune bypass rescues, PMID24856586,33667385. Additional downgrade: False.

Tumor-cell driver remains actionable despite a host contribution. Conditional KRAS withdrawal and MYC/immune factorial contrasts, PMID22232209,21035406. Additional downgrade: False.

A proposed in vitro environmental rescue does not uniquely explain in vivo independence. Confirmed MCT1 disruption fails to resensitize GOT2-deficient tumors, PMID35815941. Additional downgrade: False.

Influence substituted for primacy. Checkpoint/vascular survival benefit is a causal contribution, not comparative dominance.

SHH ligand deletion called stromal-only. Rhim deletes SHH in pancreatic epithelial lineage; tumor-host signaling is altered jointly.

All fibroblasts treated as one directional compartment. FAP and alphaSMA depletion have opposite effects.

Reduced tumor weight equated with better survival. AlphaSMA depletion reduces weight yet increases aggressive disease/embolism mortality.

One long-term trial report counted as independent replication. CheckMate067 2015 and ten-year report share NCT01844505.

Provider failure to retrieve an original equated with original unavailability. Consensus misses GOT2, Quail, KRAS and MYC papers that were independently retrieved and examined.

Expression of concern ignored. Hoshino 2015 remains historical but carries no positive causal burden after the August 2026 notice.

Adversarial steelman: above, then below

Steelman Δ for broad primacy: naive-host retransplantation resensitizes recurrent glioma and normal fibroblast restoration changes breast outgrowth despite ostensibly similar epithelial cells. These are unusually strong conditional extrinsic contrasts. They do not establish broad endogenous driver-independent maintenance, and transplantation/state reset plus occult donor clones constrain the generalization.

Steelman Β: most human spatial/prognostic associations cannot determine dominance. Γ remains justified by mechanistic rescue and host-changing contrasts; the broader evidence also clearly establishes Η influence, so neither association-only TME nor universally reactive stroma is defensible.

Retain Γ for general dominance, Η for compartment-specific causal influence and explicit narrower conditional successes. No additional letter downgrade is imposed for contradictory direction; those results delimit the claim.

Final score, causal direction and M/S/D

Preregistered primary-subclaim SC2 returns Γ. Weakest-link and scope-matched sensitivities also return Γ; SC1 influence reaches Η independently. A narrow macrophage-dependent acquired-resistance state can cross Ε contribution/conditional dominance, but it does not replace the registered broad maintenance/progression claim.

Canonical step 5 describes the broad nominal TME mechanism: it changes maintenance/progression and stabilizes selected tumor or escape states without a demonstrated general dominant initiating hierarchy. Defined immune/matrix states also restrain tumors, and scoped resistance can be rate-limiting; the single broad role does not overwrite those directions.

Momentum: Mechanistic only. 2024 innate bypass, 2025 panKRAS and CAF targeting and 2026 spatial state work test specific mechanisms, but none of them addresses the SC2 boundary check, which requires a genotype- and stage-matched factorial comparison of intrinsic-driver withdrawal against lineage-specific host perturbation. No currently recruiting trial or expected pending result is asserted.

Stability: Stable. No named pending result is asserted to be likely to move the registered score imminently.

Divergence: Contested. Required artifact: the two framings carry different implied levels, so no single implied level and no arithmetic gap can be stated. Review and press framing of the microenvironment as the controlling layer of malignant behavior implies a level well above the assessed Γ for dominance, while the endogenous-genetics and randomized-trial literature cited here supports causal influence (Η for SC1) without a dominance comparison. No manufactured field consensus is asserted. These dimensions do not change level or confidence.

PROPOSED DECISIVE EXPERIMENTS

In several endogenous genotype-defined tumors at matched stages, preregister a factorial intrinsic-driver withdrawal and lineage-specific host perturbation, restoring each separately; track clones and host state through treatment-free relapse and orthogonal immune/metabolic bypass.

Readout: Engagement in tumor and host, absolute tumor burden, metastasis, survival, residual clones and recurrence after restoration. Controls: Equal initial burden, littermate/lineage controls, tumor-only versus host-only editing, depletion efficiency, pharmacokinetics, injury/toxicity and clone-matched reimplantation controls. Supporting: Host perturbation repeatedly controls maintenance despite retained intrinsic drivers, with host restoration restoring growth and no driver-only dependency across the registered contexts. Falsifying: Driver withdrawal remains necessary, host effect is conditional on driver/immune state, or sign reverses across prespecified compartments.

RESEARCH TIMELINE

KNOWLEDGE GAPS

Run genotype/stage-matched factorial host and tumor-driver perturbations rather than ranking layers from separate positive experiments.

Separate immediate arrest from durable eradication, residual clone selection and treatment-free relapse.

Resolve why CAF/SHH effects reverse with subset, stage and intervention schedule using independently replicated protocols.

Test immune/metabolite bypass after endpoint-confirmed transport disruption; CAF pyruvate culture rescue is not the unique in vivo explanation.

Exclude occult premalignant donor clones in normal-tissue humanization with multi-region single-cell genotype/lineage tracking.

Distinguish restored host mechanism from bypass rescue, and retain embolic/toxicity mechanisms in survival analysis.

Validate spatial/biomarker predictors prospectively with intervention and absolute compartment measurements.

Count cohorts/platforms rather than repeated reports as independent evidence and disclose reused expression datasets.

Finish unavailable original bodies/supplements and correction contents; audit questioned raw images before any public figure reuse.

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

64 primary reports examined, nine selected partial original-body examinations, 55 complete abstract/status examinations. PMID21035406 carries a Comment indexing label but contains original conditional-model experiments independently read. Expression-of-concern paper PMID26524530 is historical only and excluded from positive load-bearing evidence. Provider tool limit/missing originals are disclosed; no complete-paper audit or blinded calibration.

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: 2026-10-14.

Scope of checks. All 64 REFERENCES PMIDs were verified against PubMed metadata for title, first author, journal, year and DOI: 64/64 resolve to the stated record, 0 failures. Fourteen load-bearing claims were spot-checked against the primary abstracts: the Rhim versus Ozdemir assignment, which is the easiest confusion in this literature and is correct here (PMID 24856585 is the epithelial Shh deletion that reduces stroma and worsens disease; PMID 24856586 is the alphaSMA myofibroblast depletion that shortens survival and is rescued by anti-CTLA4); the VEGFR-blockade rescue restricted to the stroma-low setting (PMID 24856585); the human-stroma breast reconstruction in which genetically modified stroma produced lesions from ostensibly normal epithelium (PMID 15051869); bevacizumab plus IFL in 813 patients, 15.6 to 20.3 months, hazard ratio 0.66 (PMID 15175435); ipilimumab 6.4 to 10.1 months (PMID 20525992); nivolumab in 272 squamous NSCLC patients, 6.0 to 9.2 months, hazard ratio 0.59 (PMID 26028407); pembrolizumab in 305 PD-L1-high patients, progression-free survival 6.0 to 10.3 months, hazard ratio 0.50, with EGFR and ALK exclusions (PMID 27718847); CheckMate 067 with 945 patients and progression-free survival 11.5, 6.9 and 2.9 months (PMID 26027431) and its ten-year overall survival of 71.9, 36.9 and 19.9 months in the same NCT01844505 cohort (PMID 39282897); CheckMate 914 part A with 816 randomized patients, disease-free survival hazard ratio 0.92, 95% CI 0.71-1.19, p = 0.53 (PMID 36774933); HALO301 with 494 randomized and 492 analysed patients, overall survival 11.2 versus 11.5 months, hazard ratio 1.00 (PMID 32706635); the IMmotion150 progression-free survival hazard ratio of 1.00, 95% CI 0.69-1.45 (PMID 29867230); and the acquired-resistance series of four cases with JAK1/2 lesions in two and B2M loss in one (PMID 27433843). All matched the review. Scoring logic was re-derived independently of the recorded letter: cap order and precedence, the preregistered primary-subclaim aggregation and its two reported alternatives, the five Beta/Gamma/Delta boundary clauses, the SC1 threshold ladder to Eta and its recorded independence basis, and the Momentum, Stability and Divergence artifacts. Both the validator of record (version 2.3) and the newer 2.4 build were re-run after editing.

Findings.

Rejected findings (checked, no change).

1. The Rhim and Ozdemir PMIDs. These two 2014 Cancer Cell papers have adjacent accession numbers and opposite framings, and swapping them would invert the review's central stromal-restraint argument. Checked both ways against PubMed; the assignment is correct throughout the prose, the inventory, the timeline and the cards.
2. Aggregation. The preregistered rule is primary-subclaim on SC2 and returns Gamma, and the review reports that weakest-link and scope-matched also return Gamma. This is the aggregation misread flagged for this round, so it was checked specifically: SC1 is an influence claim whose scope does not match the registered dominance claim, so it cannot set the headline under scope-matched either. Correct as written.
3. SC1 at Eta. The independence basis is recorded and names four lines with the distinct error each excludes, and it correctly treats the two CheckMate 067 reports as one trial line and the shared CAF and immunoediting platforms as partly dependent. No change.
4. Role of mechanism "reinforcing / stabilizing factor" - the step 5 fall-through is derived in the prose and is defensible for a layer that changes maintenance without a demonstrated initiating hierarchy.
5. Stability "Stable" - correct where no pending result is named.
6. The v5.claim_form warning that the entry form is a phenomenon. The registered claim is a full causal proposition with a comparative consequence, so this is a rejected-finding convention for this round.
7. The v5.lay_summary FAIL under validator 2.4. This package is a declared v5.3 review and the lay summary became mandatory only in v5.4. The same FAIL is produced by running 2.4 against the pre-edit original, and the package's own recorded report is from validator 2.3, which passes. Not a defect of this edition and not a regression from these edits.
8. Revision regression check. Compared against the earlier edition of this review: the revision added only the dated Consensus provenance-reconciliation block, which asserts no score change, and altered nothing else. No error was introduced by the revision. The block sits between the framework heading and the subclaim matrix, which is structurally untidy but changes no required content, so it was left in place.

9. The expression of concern on PMID 26524530 is disclosed, dated, and the source is explicitly excluded from positive load-bearing evidence. Checked and retained.

10. Whether the confidence band should move to High once access depth is removed as a determinant. Re-tested against the framework text: the review's own independence section records that the collagen and alphaSMA papers share lineage and reused expression data, that the MYC and immunoediting papers share platforms, and that the two CheckMate 067 reports are one trial line; methodological heterogeneity and partly shared platforms are the Moderate text, not the High text. The claim actually graded, comparative dominance, has no replication at all. The band holds at Moderate. The superseded wording is preserved in the dated audit snapshots in this folder by design.

11. The coordinator-flagged revision artifact in which the spliced Consensus block contradicts preserved reviewer notes asserting that no Consensus work was done. Not present here: this review carries an affirmative "METHODS AND CONSENSUS CONTRIBUTION" section describing signed-in browser use, and its reviewer notes disclose provider limits rather than denying provider use. The two blocks agree.
12. The "Best-supported rival" renderer defect reported elsewhere in this round. Checked against the score card: models_evaluated[0] here is M4, is_nominal: false, rank_in_tournament: 1, and the nominal sits at rank 5. The prose names the rival correctly. No change.

Residual limitations. Access is nine partial-body and 55 complete abstract or status examinations. Body-level figures that the abstracts do not reproduce - the 6.17 to 3.61 month median survival and 15.3 to 43.4 percent metastasis change in PMID 24856585, the approximately 80 percent depletion and 65 versus 47 day anti-CTLA4 survival in PMID 24856586, the 56 percent rebound and 13 to 63 day extension in PMID 27199435, and the one-of-ten-donors and 16 of 18 graft counts in PMID 15051869 - were checked for internal consistency across their multiple appearances in this document but could not be confirmed against a complete paper in this pass. The four unsupplied erratum identifiers remain unresolved for the same reason. No independent blinded calibration or scientific peer review is claimed by this adversarial pass.

Approval: Following adversarial review and the corrections listed above, the revised document is approved. - Claude Opus 5.5 (Anthropic), 2026-10-14.

REFERENCES

1. Olumi AF, Grossfeld GD, Hayward SW et al. Carcinoma-associated fibroblasts direct tumor progression of initiated human prostatic epithelium. Cancer research. 1999. PMID 10519415. DOI: none indexed. Access: abstract_metadata. Study: PMID10519415.

2. Kuperwasser C, Chavarria T, Wu M et al. Reconstruction of functionally normal and malignant human breast tissues in mice. Proceedings of the National Academy of Sciences of the United States of America. 2004. PMID 15051869. DOI: 10.1073/pnas.0401064101. Access: full_text_partial. Study: PMID15051869.

3. Orimo A, Gupta PB, Sgroi DC et al. Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion. Cell. 2005. PMID 15882617. DOI: 10.1016/j.cell.2005.02.034. Access: abstract_metadata. Study: PMID15882617.

4. Levental KR, Yu H, Kass L et al. Matrix crosslinking forces tumor progression by enhancing integrin signaling. Cell. 2009. PMID 19931152. DOI: 10.1016/j.cell.2009.10.027. Access: abstract_metadata. Study: PMID19931152.

5. Goetz JG, Minguet S, Navarro-Lérida I et al. Biomechanical remodeling of the microenvironment by stromal caveolin-1 favors tumor invasion and metastasis. Cell. 2011. PMID 21729786. DOI: 10.1016/j.cell.2011.05.040. Access: abstract_metadata. Study: PMID21729786.

6. Mittapalli VR, Madl J, Löffek S et al. Injury-Driven Stiffening of the Dermis Expedites Skin Carcinoma Progression. Cancer research. 2016. PMID 26676755. DOI: 10.1158/0008-5472.CAN-15-1348. Access: abstract_metadata. Study: PMID26676755.

7. Vennin C, Mélénec P, Rouet R et al. CAF hierarchy driven by pancreatic cancer cell p53-status creates a pro-metastatic and chemoresistant environment via perlecan. Nature communications. 2019. PMID 31406163. DOI: 10.1038/s41467-019-10968-6. Access: abstract_metadata. Study: PMID31406163.

8. Öhlund D, Handly-Santana A, Biffi G et al. Distinct populations of inflammatory fibroblasts and myofibroblasts in pancreatic cancer. The Journal of experimental medicine. 2017. PMID 28232471. DOI: 10.1084/jem.20162024. Access: abstract_metadata. Study: PMID28232471.

9. Biffi G, Oni TE, Spielman B et al. IL1-Induced JAK/STAT Signaling Is Antagonized by TGFβ to Shape CAF Heterogeneity in Pancreatic Ductal Adenocarcinoma. Cancer discovery. 2019. PMID 30366930. DOI: 10.1158/2159-8290.CD-18-0710. Access: abstract_metadata. Study: PMID30366930.

10. McAndrews KM, Chen Y, Darpolor JK et al. Identification of Functional Heterogeneity of Carcinoma-Associated Fibroblasts with Distinct IL6-Mediated Therapy Resistance in Pancreatic Cancer. Cancer discovery. 2022. PMID 35348629. DOI: 10.1158/2159-8290.CD-20-1484. Access: full_text_partial. Study: PMID35348629.

11. Rhim AD, Oberstein PE, Thomas DH et al. Stromal elements act to restrain, rather than support, pancreatic ductal adenocarcinoma. Cancer cell. 2014. PMID 24856585. DOI: 10.1016/j.ccr.2014.04.021. Access: full_text_partial. Study: PMID24856585.

12. 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.

13. Lee JJ, Perera RM, Wang H et al. Stromal response to Hedgehog signaling restrains pancreatic cancer progression. Proceedings of the National Academy of Sciences of the United States of America. 2014. PMID 25024225. DOI: 10.1073/pnas.1411679111. Access: abstract_metadata. Study: PMID25024225.

14. Chen Y, Kim J, Yang S et al. Type I collagen deletion in αSMA+ myofibroblasts augments immune suppression and accelerates progression of pancreatic cancer. Cancer cell. 2021. PMID 33667385. DOI: 10.1016/j.ccell.2021.02.007. Access: full_text_partial. Study: PMID33667385.

15. Steele NG, Biffi G, Kemp SB et al. Inhibition of Hedgehog Signaling Alters Fibroblast Composition in Pancreatic Cancer. Clinical cancer research : an official journal of the American Association for Cancer Research. 2021. PMID 33495315. DOI: 10.1158/1078-0432.CCR-20-3715. Access: abstract_metadata. Study: PMID33495315.

16. Feig C, Jones JO, Kraman M et al. Targeting CXCL12 from FAP-expressing carcinoma-associated fibroblasts synergizes with anti-PD-L1 immunotherapy in pancreatic cancer. Proceedings of the National Academy of Sciences of the United States of America. 2013. PMID 24277834. DOI: 10.1073/pnas.1320318110. Access: abstract_metadata. Study: PMID24277834.

17. Mariathasan S, Turley SJ, Nickles D et al. TGFβ attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells. Nature. 2018. PMID 29443960. DOI: 10.1038/nature25501. Access: abstract_metadata. Study: PMID29443960.

18. Tauriello DVF, Palomo-Ponce S, Stork D et al. TGFβ drives immune evasion in genetically reconstituted colon cancer metastasis. Nature. 2018. PMID 29443964. DOI: 10.1038/nature25492. Access: abstract_metadata. Study: PMID29443964.

19. Pyonteck SM, Akkari L, Schuhmacher AJ et al. CSF-1R inhibition alters macrophage polarization and blocks glioma progression. Nature medicine. 2013. PMID 24056773. DOI: 10.1038/nm.3337. Access: abstract_metadata. Study: PMID24056773.

20. Quail DF, Bowman RL, Akkari L et al. The tumor microenvironment underlies acquired resistance to CSF-1R inhibition in gliomas. Science (New York, N.Y.). 2016. PMID 27199435. DOI: 10.1126/science.aad3018. Access: full_text_partial. Study: PMID27199435.

21. Bowman RL, Klemm F, Akkari L et al. Macrophage Ontogeny Underlies Differences in Tumor-Specific Education in Brain Malignancies. Cell reports. 2016. PMID 27840052. DOI: 10.1016/j.celrep.2016.10.052. Access: abstract_metadata. Study: PMID27840052.

22. Pyonteck SM, Gadea BB, Wang HW et al. Deficiency of the macrophage growth factor CSF-1 disrupts pancreatic neuroendocrine tumor development. Oncogene. 2012. PMID 21822305. DOI: 10.1038/onc.2011.337. Access: abstract_metadata. Study: PMID21822305.

23. Lin EY, Li JF, Gnatovskiy L et al. Macrophages regulate the angiogenic switch in a mouse model of breast cancer. Cancer research. 2006. PMID 17114237. DOI: 10.1158/0008-5472.CAN-06-1278. Access: abstract_metadata. Study: PMID17114237.

24. Di Carlo SE, Raffenne J, Varet H et al. Depletion of slow-cycling PDGFRα+ADAM12+ mesenchymal cells promotes antitumor immunity by restricting macrophage efferocytosis. Nature immunology. 2023. PMID 37798557. DOI: 10.1038/s41590-023-01642-7. Access: abstract_metadata. Study: PMID37798557.

25. Shankaran V, Ikeda H, Bruce AT et al. IFNgamma and lymphocytes prevent primary tumour development and shape tumour immunogenicity. Nature. 2001. PMID 11323675. DOI: 10.1038/35074122. Access: abstract_metadata. Study: PMID11323675.

26. Koebel CM, Vermi W, Swann JB et al. Adaptive immunity maintains occult cancer in an equilibrium state. Nature. 2007. PMID 18026089. DOI: 10.1038/nature06309. Access: abstract_metadata. Study: PMID18026089.

27. Matsushita H, Vesely MD, Koboldt DC et al. Cancer exome analysis reveals a T-cell-dependent mechanism of cancer immunoediting. Nature. 2012. PMID 22318521. DOI: 10.1038/nature10755. Access: abstract_metadata. Study: PMID22318521.

28. Dunn GP, Bruce AT, Sheehan KC et al. A critical function for type I interferons in cancer immunoediting. Nature immunology. 2005. PMID 15951814. DOI: 10.1038/ni1213. Access: abstract_metadata. Study: PMID15951814.

29. Houlahan KE, Khan A, Greenwald NF et al. Germline-mediated immunoediting sculpts breast cancer subtypes and metastatic proclivity. Science (New York, N.Y.). 2024. PMID 38815010. DOI: 10.1126/science.adh8697. Access: abstract_metadata. Study: PMID38815010.

30. Rosenthal R, Cadieux EL, Salgado R et al. Neoantigen-directed immune escape in lung cancer evolution. Nature. 2019. PMID 30894752. DOI: 10.1038/s41586-019-1032-7. Access: abstract_metadata. Study: PMID30894752.

31. Kaplan RN, Riba RD, Zacharoulis S et al. VEGFR1-positive haematopoietic bone marrow progenitors initiate the pre-metastatic niche. Nature. 2005. PMID 16341007. DOI: 10.1038/nature04186. Access: abstract_metadata. Study: PMID16341007.

32. Hoshino A, Costa-Silva B, Shen TL et al. Tumour exosome integrins determine organotropic metastasis. Nature. 2015. PMID 26524530. DOI: 10.1038/nature15756. Access: abstract_metadata. Study: PMID26524530.

33. Ghajar CM, Peinado H, Mori H et al. The perivascular niche regulates breast tumour dormancy. Nature cell biology. 2013. PMID 23728425. DOI: 10.1038/ncb2767. Access: abstract_metadata. Study: PMID23728425.

34. Oskarsson T, Acharyya S, Zhang XH et al. Breast cancer cells produce tenascin C as a metastatic niche component to colonize the lungs. Nature medicine. 2011. PMID 21706029. DOI: 10.1038/nm.2379. Access: abstract_metadata. Study: PMID21706029.

35. 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.

36. Giuriato S, Ryeom S, Fan AC et al. Sustained regression of tumors upon MYC inactivation requires p53 or thrombospondin-1 to reverse the angiogenic switch. Proceedings of the National Academy of Sciences of the United States of America. 2006. PMID 17056717. DOI: 10.1073/pnas.0608017103. Access: abstract_metadata. Study: PMID17056717.

37. Wu CH, van Riggelen J, Yetil A et al. Cellular senescence is an important mechanism of tumor regression upon c-Myc inactivation. Proceedings of the National Academy of Sciences of the United States of America. 2007. PMID 17664422. DOI: 10.1073/pnas.0701953104. Access: abstract_metadata. Study: PMID17664422.

38. Karlsson A, Giuriato S, Tang F et al. Genomically complex lymphomas undergo sustained tumor regression upon MYC inactivation unless they acquire novel chromosomal translocations. Blood. 2003. PMID 12517816. DOI: 10.1182/blood-2002-10-3091. Access: abstract_metadata. Study: PMID12517816.

39. 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: full_text_partial. Study: PMID22232209.

40. 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: full_text_partial. Study: PMID21035406.

41. Casey SC, Tong L, Li Y et al. MYC regulates the antitumor immune response through CD47 and PD-L1. Science (New York, N.Y.). 2016. PMID 26966191. DOI: 10.1126/science.aac9935. Access: abstract_metadata. Study: PMID26966191.

42. McAndrews KM, Paradiso F, Stalnecker CA et al. An allele-agnostic mutant-KRAS inhibitor suppresses tumor maintenance signals and reprograms tumor immunity in pancreatic cancer. Science translational medicine. 2025. PMID 40901924. DOI: 10.1126/scitranslmed.adt5511. Access: abstract_metadata. Study: PMID40901924.

43. Kerk SA, Lin L, Myers AL et al. Metabolic requirement for GOT2 in pancreatic cancer depends on environmental context. eLife. 2022. PMID 35815941. DOI: 10.7554/eLife.73245. Access: full_text_partial. Study: PMID35815941.

44. Abrego J, Sanford-Crane H, Oon C et al. A Cancer Cell-Intrinsic GOT2-PPARδ Axis Suppresses Antitumor Immunity. Cancer discovery. 2022. PMID 35894778. DOI: 10.1158/2159-8290.CD-22-0661. Access: abstract_metadata. Study: PMID35894778.

45. Hurwitz H, Fehrenbacher L, Novotny W et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. The New England journal of medicine. 2004. PMID 15175435. DOI: 10.1056/NEJMoa032691. Access: abstract_metadata. Study: PMID15175435.

46. Hodi FS, O'Day SJ, McDermott DF et al. Improved survival with ipilimumab in patients with metastatic melanoma. The New England journal of medicine. 2010. PMID 20525992. DOI: 10.1056/NEJMoa1003466. Access: abstract_metadata. Study: PMID20525992.

47. Brahmer J, Reckamp KL, Baas P et al. Nivolumab versus Docetaxel in Advanced Squamous-Cell Non-Small-Cell Lung Cancer. The New England journal of medicine. 2015. PMID 26028407. DOI: 10.1056/NEJMoa1504627. Access: abstract_metadata. Study: PMID26028407.

48. Reck M, Rodríguez-Abreu D, Robinson AG et al. Pembrolizumab versus Chemotherapy for PD-L1-Positive Non-Small-Cell Lung Cancer. The New England journal of medicine. 2016. PMID 27718847. DOI: 10.1056/NEJMoa1606774. Access: abstract_metadata. Study: PMID27718847.

49. Larkin J, Chiarion-Sileni V, Gonzalez R et al. Combined Nivolumab and Ipilimumab or Monotherapy in Untreated Melanoma. The New England journal of medicine. 2015. PMID 26027431. DOI: 10.1056/NEJMoa1504030. Access: abstract_metadata. Study: NCT01844505-CheckMate067.

50. Wolchok JD, Chiarion-Sileni V, Rutkowski P et al. Final, 10-Year Outcomes with Nivolumab plus Ipilimumab in Advanced Melanoma. The New England journal of medicine. 2025. PMID 39282897. DOI: 10.1056/NEJMoa2407417. Access: abstract_metadata. Study: NCT01844505-CheckMate067.

51. Motzer RJ, Russo P, Grünwald V et al. Adjuvant nivolumab plus ipilimumab versus placebo for localised renal cell carcinoma after nephrectomy (CheckMate 914): a double-blind, randomised, phase 3 trial. Lancet (London, England). 2023. PMID 36774933. DOI: 10.1016/S0140-6736(22)02574-0. Access: abstract_metadata. Study: PMID36774933.

52. Gilbert MR, Dignam JJ, Armstrong TS et al. A randomized trial of bevacizumab for newly diagnosed glioblastoma. The New England journal of medicine. 2014. PMID 24552317. DOI: 10.1056/NEJMoa1308573. Access: abstract_metadata. Study: PMID24552317.

53. Chinot OL, Wick W, Mason W et al. Bevacizumab plus radiotherapy-temozolomide for newly diagnosed glioblastoma. The New England journal of medicine. 2014. PMID 24552318. DOI: 10.1056/NEJMoa1308345. Access: abstract_metadata. Study: PMID24552318.

54. Van Cutsem E, Tempero MA, Sigal D et al. Randomized Phase III Trial of Pegvorhyaluronidase Alfa With Nab-Paclitaxel Plus Gemcitabine for Patients With Hyaluronan-High Metastatic Pancreatic Adenocarcinoma. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2020. PMID 32706635. DOI: 10.1200/JCO.20.00590. Access: abstract_metadata. Study: PMID32706635.

55. McDermott DF, Huseni MA, Atkins MB et al. Clinical activity and molecular correlates of response to atezolizumab alone or in combination with bevacizumab versus sunitinib in renal cell carcinoma. Nature medicine. 2018. PMID 29867230. DOI: 10.1038/s41591-018-0053-3. Access: abstract_metadata. Study: PMID29867230.

56. Peng W, Chen JQ, Liu C et al. Loss of PTEN Promotes Resistance to T Cell-Mediated Immunotherapy. Cancer discovery. 2016. PMID 26645196. DOI: 10.1158/2159-8290.CD-15-0283. Access: abstract_metadata. Study: PMID26645196.

57. Shin DS, Zaretsky JM, Escuin-Ordinas H et al. Primary Resistance to PD-1 Blockade Mediated by JAK1/2 Mutations. Cancer discovery. 2017. PMID 27903500. DOI: 10.1158/2159-8290.CD-16-1223. Access: abstract_metadata. Study: PMID27903500.

58. Zaretsky JM, Garcia-Diaz A, Shin DS et al. Mutations Associated with Acquired Resistance to PD-1 Blockade in Melanoma. The New England journal of medicine. 2016. PMID 27433843. DOI: 10.1056/NEJMoa1604958. Access: abstract_metadata. Study: PMID27433843.

59. Torrejon DY, Abril-Rodriguez G, Champhekar AS et al. Overcoming Genetically Based Resistance Mechanisms to PD-1 Blockade. Cancer discovery. 2020. PMID 32467343. DOI: 10.1158/2159-8290.CD-19-1409. Access: abstract_metadata. Study: PMID32467343.

60. Ebos JM, Lee CR, Cruz-Munoz W et al. Accelerated metastasis after short-term treatment with a potent inhibitor of tumor angiogenesis. Cancer cell. 2009. PMID 19249681. DOI: 10.1016/j.ccr.2009.01.021. Access: abstract_metadata. Study: PMID19249681.

61. Pàez-Ribes M, Allen E, Hudock J et al. Antiangiogenic therapy elicits malignant progression of tumors to increased local invasion and distant metastasis. Cancer cell. 2009. PMID 19249680. DOI: 10.1016/j.ccr.2009.01.027. Access: abstract_metadata. Study: PMID19249680.

62. Lyubetskaya A, Rabe B, Kavran A et al. In situ multi-modal characterization of pancreatic cancer reveals tumor cell identity as a defining factor of the surrounding microenvironment. Cell reports. 2026. PMID 41533516. DOI: 10.1016/j.celrep.2025.116827. Access: abstract_metadata. Study: PMID41533516.

63. Yang A, Gu C, Liu Y et al. ITGB5-mediated biomechanical regulation in pancreatic ductal adenocarcinoma stroma impacts tumor progression and prognosis. Journal of translational medicine. 2025. PMID 41121153. DOI: 10.1186/s12967-025-07119-5. Access: abstract_metadata. Study: PMID41121153.

64. Liu X, Baer JM, Stone ML et al. Stromal reprogramming overcomes resistance to RAS-MAPK inhibition to improve pancreas cancer responses to cytotoxic and immune therapy. Science translational medicine. 2024. PMID 39441902. DOI: 10.1126/scitranslmed.ado2402. Access: abstract_metadata. Study: PMID39441902.

Source-specific evidence cards

[1] Carcinoma-associated fibroblasts direct tumor progression of initiated human prostatic epithelium.

Shows: Carcinoma-associated versus normal fibroblasts promote tumor growth from initiated, nontumorigenic prostate epithelium.

Does not show: Initiated epithelium and xenograft context; not conversion of genetically verified uninitiated cells or general extrinsic dominance.

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

[2] Reconstruction of functionally normal and malignant human breast tissues in mice.

Shows: Growth-factor-modified stroma produces malignant outgrowth in 16/18 grafts from one donor; adding normal fibroblasts prevents abnormal outgrowth.

Does not show: Nine of ten donors remain nonmalignant and a different region of the responding breast fails to reproduce malignancy; DNA identity is not exclusion of occult premalignant clones.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 0:Publisher HTML paragraphs:paragraph16, 0:Publisher HTML paragraphs:paragraph17, 0:Publisher HTML paragraphs:paragraph39, 0:Publisher HTML paragraphs:paragraph40, 0:Publisher HTML paragraphs:paragraph42, 0:Publisher HTML paragraphs:paragraph43, 0:Publisher HTML paragraphs:paragraph44, 0:Publisher HTML paragraphs:paragraph45, 0:Publisher HTML paragraphs:paragraph46, 0:Publisher HTML paragraphs:paragraph47, 0:Publisher HTML paragraphs:paragraph48, 0:Publisher HTML paragraphs:paragraph53. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

[3] Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion.

Shows: CAFs from human breast cancers promote growth through SDF1/CXCL12-associated angiogenic recruitment and epithelial CXCR4 signaling.

Does not show: Coimplantation with established cancer cells and coupled angiogenesis/proliferation do not establish endogenous universal dominance.

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

[4] Matrix crosslinking forces tumor progression by enhancing integrin signaling.

Shows: LOX-linked matrix crosslinking promotes integrin signaling and invasion; blocking crosslinking reduces fibrosis, tumor incidence and progression.

Does not show: Oncogene-initiated context; changing mechanics/signaling together is not proof that the matrix overrides all tumor genotypes.

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

[5] Biomechanical remodeling of the microenvironment by stromal caveolin-1 favors tumor invasion and metastasis.

Shows: Fibroblast caveolin1 influences matrix remodeling, stiffness and invasion with Rho/p190 signaling contrasts.

Does not show: Global host knockout/development and human prognosis association constrain compartment specificity and primacy.

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

[6] Injury-Driven Stiffening of the Dermis Expedites Skin Carcinoma Progression.

Shows: A wound/fibrotic environment promotes squamous carcinogenesis with TGFbeta/LOX/integrin pathway contrasts.

Does not show: Chemical initiation and chronic injury accompany matrix changes; mutation-independent wording does not prove identical normal epithelial genomes/clones.

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

[7] CAF hierarchy driven by pancreatic cancer cell p53-status creates a pro-metastatic and chemoresistant environment via perlecan.

Shows: Tumor gain-of-function p53 educates fibroblasts into a perlecan-rich niche supporting invasion and therapy resistance.

Does not show: Tumor genotype organizes the host phenotype; dominance of one CAF state over another is not dominance of host over tumor drivers.

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

[8] Distinct populations of inflammatory fibroblasts and myofibroblasts in pancreatic cancer.

Shows: Spatially distinct inflammatory and myofibroblastic CAF states emerge in pancreatic models and tissue.

Does not show: Phenotype/state mapping alone is not a survival perturbation or causal hierarchy.

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

[9] IL1-Induced JAK/STAT Signaling Is Antagonized by TGFβ to Shape CAF Heterogeneity in Pancreatic Ductal Adenocarcinoma.

Shows: IL1/LIF/JAKSTAT promotes inflammatory CAFs while TGFbeta suppresses IL1R1 and favors myofibroblastic states.

Does not show: Tumor ligand output and CAF plasticity are coupled; state selection is not broad malignant maintenance primacy.

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

[10] Identification of Functional Heterogeneity of Carcinoma-Associated Fibroblasts with Distinct IL6-Mediated Therapy Resistance in Pancreatic Cancer.

Shows: FAP-positive CAF depletion improves survival whereas alphaSMA-positive depletion worsens it; alphaSMA-specific IL6 loss improves gemcitabine response.

Does not show: Baseline IL6 deletion leaves progression/survival similar; global IL6 knockout, not solely conditional alphaSMA deletion, supports the tested triple therapy. No allocation randomization, histology blinded; reused alphaSMA data are not independent replication.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 0:Publisher HTML paragraphs:paragraph24, 0:Publisher HTML paragraphs:paragraph25, 0:Publisher HTML paragraphs:paragraph28, 0:Publisher HTML paragraphs:paragraph35, 0:Publisher HTML paragraphs:paragraph36, 0:Publisher HTML paragraphs:paragraph39, 0:Publisher HTML paragraphs:paragraph40, 0:Publisher HTML paragraphs:paragraph41, 0:Publisher HTML paragraphs:paragraph42, 0:Publisher HTML paragraphs:paragraph43, 0:Publisher HTML paragraphs:paragraph44, 0:Publisher HTML paragraphs:paragraph56, 0:Publisher HTML paragraphs:paragraph67, 0:Publisher HTML paragraphs:paragraph80, 0:Publisher HTML paragraphs:paragraph38, 0:Publisher HTML paragraphs:paragraph45, 0:Publisher HTML paragraphs:paragraph46, 0:Publisher HTML paragraphs:paragraph50, 0:Publisher HTML paragraphs:paragraph57, 0:Publisher HTML paragraphs:paragraph47, 0:Publisher HTML paragraphs:paragraph48, 0:Publisher HTML paragraphs:paragraph49. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

[11] Stromal elements act to restrain, rather than support, pancreatic ductal adenocarcinoma.

Shows: SHH loss reduces stroma but median survival falls from 6.17 to 3.61 months and gross metastasis rises from 15.3 to 43.4 percent; VEGFR2 blockade selectively rescues the stroma-low setting.

Does not show: SHH deletion is epithelial, not stromal-only; chronic versus acute blockade can have different signs. A suppressive stromal contribution does not prove opposite universal genetic dominance.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:RESULTS:paragraph0, 1:RESULTS:paragraph1, 1:RESULTS:paragraph2, 1:RESULTS:paragraph4, 1:RESULTS:paragraph6, 1:RESULTS:paragraph7, 1:RESULTS:paragraph11, 1:RESULTS:paragraph13. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

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

Shows: Approximately 80 percent myofibroblast depletion accelerates aggressive disease and shortens survival despite lower tumor weight; antiCTLA4 increases median survival to 65 versus control 47 days.

Does not show: Pulmonary emboli contribute to mortality; sphere/marker enrichment is not CSC proof. Rescue bypasses rather than restores depleted fibroblasts; linked erratum contents unexamined.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Results:paragraph0, 1:Results:paragraph1, 1:Results:paragraph5, 1:Results:paragraph6, 1:Results:paragraph8. No complete-paper or supplement audit claimed. Erratum in: PMID not supplied. Subclaims: SC1, SC2.

[13] Stromal response to Hedgehog signaling restrains pancreatic cancer progression.

Shows: Hedgehog inhibition accelerates pancreatic disease whereas pathway activation expands stroma and reduces epithelial proliferation.

Does not show: Stage, genotype and protocol restrict sign; all CAFs and all Hh contexts cannot be called tumor suppressive.

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

[14] Type I collagen deletion in αSMA+ myofibroblasts augments immune suppression and accelerates progression of pancreatic cancer.

Shows: AlphaSMA-specific Col1 deletion reduces fibrillar collagen/stiffness and accelerates PDAC; CXCR2/CCR2 blockade restores survival toward control.

Does not show: Fsp1-directed deletion does not similarly reduce total collagen or survival; immune bypass does not restore matrix. Shows compartment-specific restraint rather than generic collagen protection.

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, 1:RESULTS:paragraph11, 1:RESULTS:paragraph20. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

[15] Inhibition of Hedgehog Signaling Alters Fibroblast Composition in Pancreatic Cancer.

Shows: SHH overexpression promotes growth and Smo inhibition impairs it while shifting myCAF/iCAF and immune composition.

Does not show: Contrasts earlier restraining-Hh settings; different stage/protocols prevent one universal directional story.

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

[16] Targeting CXCL12 from FAP-expressing carcinoma-associated fibroblasts synergizes with anti-PD-L1 immunotherapy in pancreatic cancer.

Shows: FAP-cell and CXCL12 contrasts identify immune exclusion; AMD3100 plus antiPDL1 rapidly removes tumor cells in a defined pancreatic model.

Does not show: Combined immune/niche dependence in a driver-bearing tumor does not establish host primacy over all intrinsic programs.

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

[17] TGFβ attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells.

Shows: TGFbeta-rich stroma associates with immune exclusion; combined TGFbeta/PDL1 blockade restores penetration and regression in mouse tumors.

Does not show: Human biomarker associations are not randomized compartment comparisons; combination efficacy cannot isolate a universal dominant layer.

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

[18] TGFβ drives immune evasion in genetically reconstituted colon cancer metastasis.

Shows: TGFbeta blockade prevents metastasis and enables checkpoint control of established lesions in low-mutational-burden driver-mutant CRC.

Does not show: Specific immune-evasion dependence in a defined tumor genotype; does not erase the required epithelial driver program.

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

[19] CSF-1R inhibition alters macrophage polarization and blocks glioma progression.

Shows: CSF1R inhibition regresses glioma and prolongs survival through macrophage-state alteration.

Does not show: TAMs are protected from depletion by glioma signals; functional repolarization, not simply fewer macrophages, explains the tested response.

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

[20] The tumor microenvironment underlies acquired resistance to CSF-1R inhibition in gliomas.

Shows: Continued CSF1R blockade produces rebound in 56 percent; rebound cells regain sensitivity in naive hosts, and IGF1R plus CSF1R inhibition extends recurrent-tumor median survival from 13 to 63 days.

Does not show: Passaging/transplantation can reset state; tumor-specific IGF1R knockdown also restores response. Strong scoped environmental resistance control does not establish general malignant primacy.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 0:A subset of GBMs develop resistance to CSF-1R inhibition in long-term preclinical trials, 2:Resistance to CSF-1R inhibition is mediated by the microenvironment, 6:Combination of CSF-1R and IGF-1R inhibition improves outcome. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

[21] Macrophage Ontogeny Underlies Differences in Tumor-Specific Education in Brain Malignancies.

Shows: Nonirradiative lineage tracing distinguishes brain-resident and recruited macrophage origins with different programs in glioma.

Does not show: Ontogeny and epigenetic association do not alone establish growth dependency or primacy.

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

[22] Deficiency of the macrophage growth factor CSF-1 disrupts pancreatic neuroendocrine tumor development.

Shows: CSF1 deficiency reduces macrophages, angiogenic switching and total pancreatic neuroendocrine tumor burden.

Does not show: Established tumors retain growth with compensating neutrophils; developmental effects and redundant myeloid routes limit universal necessity.

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

[23] Macrophages regulate the angiogenic switch in a mouse model of breast cancer.

Shows: Macrophage recruitment promotes an angiogenic switch; blocked infiltration delays progression and local restoration restores vascular effects.

Does not show: A defined angiogenic stage and transgenic platform, not all epithelial initiation or maintenance.

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

[24] Depletion of slow-cycling PDGFRα+ADAM12+ mesenchymal cells promotes antitumor immunity by restricting macrophage efferocytosis.

Shows: ADAM12-marked slow mesenchymal cells acquire perivascular tumor-supporting states; depletion improves vascular/immune organization and reduces growth.

Does not show: State and lineage restriction, tumor-induced remodeling and different cell populations constrain a unifying stromal hierarchy.

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

[25] IFNgamma and lymphocytes prevent primary tumour development and shape tumour immunogenicity.

Shows: Lymphocyte/IFNgamma competence suppresses tumor development and selects tumor immunogenicity.

Does not show: Host deficiency and carcinogen initiation are broad changes; immune editing is not immunity-only control of every malignant behavior.

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

[26] Adaptive immunity maintains occult cancer in an equilibrium state.

Shows: Adaptive immune disruption releases occult carcinogen-induced tumors from prolonged equilibrium.

Does not show: Transformed cells preexist; immune containment and eventual escape do not specify universal host primacy.

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

[27] Cancer exome analysis reveals a T-cell-dependent mechanism of cancer immunoediting.

Shows: Immune selection favors preexisting antigen-negative sarcoma clones with defined mutant antigen recognition.

Does not show: Related immunoediting platform; genotype and immunity interact rather than forming separable universal hierarchies.

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

[28] A critical function for type I interferons in cancer immunoediting.

Shows: Type I IFN dependence in host hematopoietic cells differs from IFNgamma tumor-target effects during tumor control.

Does not show: Conditional cell targets, not all immune pathways; linked erratum contents unexamined.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. Erratum in: PMID not supplied. Subclaims: SC1.

[29] Germline-mediated immunoediting sculpts breast cancer subtypes and metastatic proclivity.

Shows: Germline epitope repertoire associates with recurrent tumor alterations across 5870 lesions.

Does not show: Computational selection and population associations are not randomized immune perturbation; earlier preprint is the same study.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. UpdateOf: PMID36993286. Subclaims: SC1, SC2.

[30] Neoantigen-directed immune escape in lung cancer evolution.

Shows: 258 regions from 88 lung tumors reveal antigen-presentation loss and neoantigen expression/loss trajectories compatible with immune selection.

Does not show: Observational clonal tracking infers selection but does not experimentally compare intrinsic and host primacy.

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

[31] VEGFR1-positive haematopoietic bone marrow progenitors initiate the pre-metastatic niche.

Shows: VEGFR1-positive progenitor perturbation and reconstitution alter premetastatic niche formation and metastasis in defined models.

Does not show: Transplant/conditioned-medium and recruitment context; does not prove all spontaneous metastasis requires this niche.

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

[32] Tumour exosome integrins determine organotropic metastasis.

Shows: The original report proposed organotropic exosomal integrins directing uptake and metastatic niche formation.

Does not show: Nature expression of concern published 5 August 2026 identifies potentially overlapping images; excluded from positive load-bearing causal evidence. No raw-image audit or retraction claimed.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. ExpressionOfConcernIn: PMID42557336. Subclaims: SC1.

[33] The perivascular niche regulates breast tumour dormancy.

Shows: Stable endothelium/TSP1 sustains dormancy whereas sprouting niches and TGFbeta/periostin promote outgrowth.

Does not show: Defined disseminated-cell contexts; unchanged genome cannot be presumed solely from similar cell lineage, and all malignant maintenance is broader.

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

[34] Breast cancer cells produce tenascin C as a metastatic niche component to colonize the lungs.

Shows: Tumor-derived tenascinC supports early metastatic colonization before host stromal supply becomes sufficient.

Does not show: The niche effector initially originates from tumor cells; intrinsic/extrinsic boundary is mechanistically coupled.

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

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

Shows: Restoring p53 regresses established lymphomas and sarcomas through different apoptosis/senescence responses.

Does not show: Tumor-type-dependent genetic restoration can remodel host effectors; not proof that every cancer is intrinsically dominant.

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

[36] Sustained regression of tumors upon MYC inactivation requires p53 or thrombospondin-1 to reverse the angiogenic switch.

Shows: p53-dependent TSP1 induction contributes to regression after MYC inactivation.

Does not show: TSP1 alone does not suppress MYC-on growth; joint oncogene/host mechanism rather than either universal primacy.

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

[37] Cellular senescence is an important mechanism of tumor regression upon c-Myc inactivation.

Shows: p16/Rb/p53 senescence machinery influences durable regression after MYC withdrawal.

Does not show: Cell-state and host effects cooperate; restricted transgenic programs do not define all sporadic cancers.

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

[38] Genomically complex lymphomas undergo sustained tumor regression upon MYC inactivation unless they acquire novel chromosomal translocations.

Shows: MYC suppression regresses genetically complex lymphomas, with recurrent tumors acquiring replacement alterations.

Does not show: Driver substitution after selection and transgene withdrawal differ from uniform endogenous human dependence.

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

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

Shows: Turning off oncogenic KRAS regresses MRI-confirmed invasive tumors within two weeks; selected mice remain tumor-free 23 weeks after withdrawal.

Does not show: Transgenic expression, p53 background and cerulein injury restrict inference; small MRI cohort and simultaneous stromal remodeling do not establish opposite universal genetic primacy.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 0:Publisher HTML paragraphs:paragraph71, 0:Publisher HTML paragraphs:paragraph73, 0:Publisher HTML paragraphs:paragraph74, 0:Publisher HTML paragraphs:paragraph75, 0:Publisher HTML paragraphs:paragraph76, 0:Publisher HTML paragraphs:paragraph80, 0:Publisher HTML paragraphs:paragraph83, 0:Publisher HTML paragraphs:paragraph84, 0:Publisher HTML paragraphs:paragraph85. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

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

Shows: MYC-off apoptosis/arrest occurs across host immune states, but CD4 loss increases recurrence; CD4 transfer restores durable control and TSP-deficient effectors fail.

Does not show: Initial and durable responses differ; immune bypass and intrinsic withdrawal jointly determine control. PubMed Comment label accompanies genuine original experiments; linked erratum unexamined.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Results:paragraph0, 1:Results:paragraph1, 1:Results:paragraph2, 1:Results:paragraph4, 1:Results:paragraph5, 1:Results:paragraph10, 1:Results:paragraph14, 1:Results:paragraph15, 1:Results:paragraph16, 1:Results:paragraph17. No complete-paper or supplement audit claimed. Erratum in: PMID not supplied. Subclaims: SC1, SC2.

[41] MYC regulates the antitumor immune response through CD47 and PD-L1.

Shows: MYC regulates CD47/PDL1 and enforced ligand expression impairs immune remodeling after MYC suppression.

Does not show: Tumor genotype drives immune output; linked erratum 27124464 metadata verified, contents unexamined.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. Erratum in: PMID27124464. Subclaims: SC1, SC2.

[42] An allele-agnostic mutant-KRAS inhibitor suppresses tumor maintenance signals and reprograms tumor immunity in pancreatic cancer.

Shows: PanKRAS inhibition suppresses diverse preclinical tumors and remodels tumor/microenvironment state; YAP-associated escape is interrogated.

Does not show: Drug specificity and feedback matter; preclinical results are not patient efficacy or universal intrinsic dominance.

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

[43] Metabolic requirement for GOT2 in pancreatic cancer depends on environmental context.

Shows: GOT2 loss impairs culture but not tested in vivo growth; CAF pyruvate or cytosolic NADH oxidation rescues culture, whereas mitochondrial oxidation does not.

Does not show: Confirmed MCT1 loss/inhibition fails to resensitize tumors in vivo, so CAF pyruvate is not established as the unique in vivo explanation; residual/alternative redox routes remain.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Results:paragraph14, 1:Results:paragraph17, 1:Results:paragraph23, 1:Results:paragraph27, 1:Results:paragraph34, 1:Results:paragraph36, 1:Results:paragraph40, 1:Results:paragraph55. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

[44] A Cancer Cell-Intrinsic GOT2-PPARδ Axis Suppresses Antitumor Immunity.

Shows: GOT2 supports PPARdelta-linked immune suppression beyond its proliferative metabolic role.

Does not show: Different tumor/immune conditions and enzyme functions prevent treating context papers as a direct universal contradiction.

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

[45] Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer.

Shows: Adding bevacizumab to IFL in 813 randomized metastatic colorectal patients increases median survival from 15.6 to 20.3 months, HR 0.66.

Does not show: Antiangiogenic/additive clinical benefit is not a matched comparison with extinguishing tumor drivers; delivery and vascular functions are coupled.

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

[46] Improved survival with ipilimumab in patients with metastatic melanoma.

Shows: Ipilimumab improves median survival versus gp100 from 6.4 to 10.1 months in the defined advanced melanoma trial.

Does not show: Selected HLA/treatment population and immune toxicity; survival benefit does not prove broad extrinsic dominance. Linked erratum contents unexamined.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. Erratum in: PMID not supplied. Subclaims: SC1.

[47] Nivolumab versus Docetaxel in Advanced Squamous-Cell Non-Small-Cell Lung Cancer.

Shows: In 272 previously treated squamous NSCLC patients, nivolumab increases median survival from 6.0 to 9.2 months versus docetaxel, HR 0.59.

Does not show: Specific disease/treatment comparison; PDL1 nonprediction in this study does not mean genotype or tumor state is irrelevant.

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

[48] Pembrolizumab versus Chemotherapy for PD-L1-Positive Non-Small-Cell Lung Cancer.

Shows: In 305 PDL1-high NSCLC patients, pembrolizumab improves PFS from 6.0 to 10.3 months, HR 0.50, versus chemotherapy.

Does not show: Enrollment excludes sensitizing EGFR and ALK alterations and requires PDL1 at least 50 percent; cannot generalize to all genotypes.

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

[49] Combined Nivolumab and Ipilimumab or Monotherapy in Untreated Melanoma.

Shows: CheckMate067 randomizes 945 melanoma patients; combination/nivolumab/ipilimumab median PFS is 11.5/6.9/2.9 months.

Does not show: Initial report emphasizes PFS; later long-term survival is the same trial, not independent replication. Linked erratum 31442371 contents unexamined.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. Erratum in: PMID31442371. Subclaims: SC1.

[50] Final, 10-Year Outcomes with Nivolumab plus Ipilimumab in Advanced Melanoma.

Shows: Ten-year CheckMate067 follow-up reports median OS of 71.9/36.9/19.9 months for combination/nivolumab/ipilimumab.

Does not show: Same NCT01844505 cohort as the 2015 report; online 2024 and journal issue 2025, not a new independent trial.

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

[51] Adjuvant nivolumab plus ipilimumab versus placebo for localised renal cell carcinoma after nephrectomy (CheckMate 914): a double-blind, randomised, phase 3 trial.

Shows: CheckMate914 part A in 816 randomized high-risk resected renal patients finds no significant DFS gain, HR 0.92, CI 0.71–1.19, p=0.53.

Does not show: Specific adjuvant regimen and 32 versus 2 percent toxicity discontinuation; null is not proof immunity never controls RCC.

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

[52] A randomized trial of bevacizumab for newly diagnosed glioblastoma.

Shows: RTOG0825 randomizes 637 of 978 registered glioblastoma patients; bevacizumab yields OS 15.7 versus 16.1 months and worse selected cognitive/QoL outcomes.

Does not show: PFS 10.7 versus 7.3 months does not meet preset improvement; vascular imaging/response changes are not survival proof.

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

[53] Bevacizumab plus radiotherapy-temozolomide for newly diagnosed glioblastoma.

Shows: AVAglio in 921 patients improves PFS 10.6 versus 6.2 months, HR 0.64, with no significant OS benefit, p=0.10.

Does not show: QoL instruments/protocol differ from RTOG0825; do not collapse different QoL findings or infer absent vascular biology.

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

[54] Randomized Phase III Trial of Pegvorhyaluronidase Alfa With Nab-Paclitaxel Plus Gemcitabine for Patients With Hyaluronan-High Metastatic Pancreatic Adenocarcinoma.

Shows: HALO301 has 494 randomized and 492 analysis patients; hyaluronan targeting yields OS 11.2 versus 11.5 months, HR 1.00, and equal PFS 7.1 months despite greater response.

Does not show: Hyaluronan-high selected disease, toxicity and drug regimen; response is not survival and null does not negate all stromal contributions.

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

[55] Clinical activity and molecular correlates of response to atezolizumab alone or in combination with bevacizumab versus sunitinib in renal cell carcinoma.

Shows: IMmotion150 phase 2 randomizes 305 RCC patients; combination versus sunitinib ITT PFS HR 1.00, CI 0.69–1.45, with immune/angiogenic signatures.

Does not show: TMB is not associated with PFS in this study; biomarker signatures do not prove layer primacy. Linked correction 30291359 contents unexamined.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. Erratum in: PMID30291359. Subclaims: SC1.

[56] Loss of PTEN Promotes Resistance to T Cell-Mediated Immunotherapy.

Shows: PTEN loss alters cytokine/autophagy programs, T-cell access and therapy response; PI3Kbeta intervention improves immune control.

Does not show: Intrinsic genetic and host pathways are causally coupled, not separable universal dominant compartments.

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

[57] Primary Resistance to PD-1 Blockade Mediated by JAK1/2 Mutations.

Shows: JAK1/2 alterations in rare high-mutational-burden nonresponders impair interferon-linked immune response.

Does not show: One of 23 melanoma and one of 16 colorectal patients; rare cases and cell assays do not establish population-wide primacy.

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

[58] Mutations Associated with Acquired Resistance to PD-1 Blockade in Melanoma.

Shows: Four acquired-resistance cases reveal JAK1/2 lesions in two and B2M loss in one with altered antigen/IFN response.

Does not show: Small selected series; antigen presentation and IFN mechanisms differ and do not cover all immune escape.

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

[59] Overcoming Genetically Based Resistance Mechanisms to PD-1 Blockade.

Shows: JAK or B2M disruption produces checkpoint resistance; TLR9/PD1 or CD4/NK/IL2 routes restore selected responses.

Does not show: Genetically imposed immune escape can be bypassed, opposing immutable genetic dominance; engineered models do not establish universal immune primacy.

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

[60] Accelerated metastasis after short-term treatment with a potent inhibitor of tumor angiogenesis.

Shows: Short VEGF-pathway inhibition/preconditioning can accelerate metastatic growth while suppressing primary tumors.

Does not show: Pretreatment, intravenous/resection context and short schedule cannot be equated with prolonged human therapy effects.

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

[61] Antiangiogenic therapy elicits malignant progression of tumors to increased local invasion and distant metastasis.

Shows: VEGF-targeting reduces primary growth yet increases local invasion/metastasis in selected settings.

Does not show: Stage, hypoxia and vessel mechanism can change sign; no claim that all antiangiogenic clinical treatment increases metastasis.

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

[62] In situ multi-modal characterization of pancreatic cancer reveals tumor cell identity as a defining factor of the surrounding microenvironment.

Shows: Tumor identity and malignant programs associate with spatial immune/stromal neighborhoods in untreated pancreatic ductal adenocarcinoma from 39 patients.

Does not show: Before-cutoff 2026 report; observational spatial association does not identify directional causal dominance.

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

[63] ITGB5-mediated biomechanical regulation in pancreatic ductal adenocarcinoma stroma impacts tumor progression and prognosis.

Shows: ITGB5-positive CAFs alter stromal stiffness and pancreatic ductal adenocarcinoma progression in experimental models with clinical association.

Does not show: Prognostic/registration statements in the abstract are not independently verified clinical validation; binding/signaling and matrix state vary together.

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

[64] Stromal reprogramming overcomes resistance to RAS-MAPK inhibition to improve pancreas cancer responses to cytotoxic and immune therapy.

Shows: Stromal FGF1 maintains tumor MYC under RAF/MEK blockade; adding FAK inhibition enables regression in defined pancreatic models.

Does not show: Joint pathway treatment reinforces tumor/host feedback; a 2024 trial statement does not establish current recruitment or clinical efficacy.

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

Documentary record

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

• ADVERSARIAL_REVIEW.md
• SHA256.txt
• adversarial-findings.json
• codex-correction-provenance.json
• consensus-addendum.md
• consensus-validation.json
• edition-approval-audit.json
• reader-summary.json
• tumor-microenvironment-primacy.review.md
• tumor-microenvironment-primacy.scorecard.json
• validation-report.json
• validation-round2-2026-10-14.json
• validation.txt

Sources: 64 · load-bearing claims: 67 · verified: 67 · unverified: 0 · contradicted: 0 · schema 5.3