# --------------------------------------------------------------------------- # SEARCH HEADLINE: rules as in gulf-of-tonkin/claims.yaml (1-7). # It may only state a claim that is established or refuted at HIGH confidence with # anchor_checked: primary. The headline claim (cvc-no-registry-surge) is established at # MODERATE confidence (the 2026 US report was read as an abstract and a press release, the # English registry not at all), so the status stays DRAFT until a reviewer and the owner move it. # --------------------------------------------------------------------------- headline: "Do COVID Vaccines Cause 'Turbo Cancer'? Not Shown; Two Disputed Studies Report More Cancer" headline_claim: cvc-no-registry-surge headline_status: draft # draft until the headline claim reaches high confidence with a primary check search_summary: > Do COVID vaccines cause "turbo cancer"? Not shown: no controlled study measured it, and the US and Japanese registries show no surge. Two cohort studies report more cancer after vaccination; both are disputed and one carries an editor's notice of concern. url_slug: do-covid-vaccines-cause-turbo-cancer # --------------------------------------------------------------------------- # ASSESSMENT (SCHEMA N23): the "Where it stands" box the site shows first. # Drawn from short_answer below; no percentages (N23). # --------------------------------------------------------------------------- assessment: question: "Do COVID-19 vaccines cause cancer, including the rapidly progressing cancers called 'turbo cancer'?" answer: leans_no headline: "Leans no (moderate confidence). No study has measured 'turbo cancer'; registries show no surge; the two cohorts that report more cancer are disputed and their authors stop short of cause." text: > The term "turbo cancer" means cancers that appear or progress unusually fast after COVID-19 vaccination. We found no controlled study that defines such cancers in advance and compares them between vaccinated and unvaccinated people (cvc-rapid-progression-gap), so the claim has never been tested directly. What can be checked is indirect. National cancer series for the United States and Japan show a drop in diagnoses in 2020, when screening was disrupted, and a return toward earlier levels by 2021 to 2022; overall age-standardised cancer death rates kept falling (cvc-no-registry-surge). Those are population totals with no vaccination status, so they cannot rule out a rise in a small group, and the same totals cannot show cause in either direction. Two cohort studies compared vaccinated with unvaccinated people. A Seoul study of adults (a Letter in Biomarker Research, September 2025) reports a hazard ratio of 1.27 for any cancer within one year, and an Italian study of Pescara province residents aged 11 and over reports 1.23 for hospital admission with cancer. The Seoul authors write that their findings do not establish causal relationships; the Italian authors call theirs preliminary because healthy-vaccinee bias and unmeasured confounders could not be quantified. Two published commentaries point to different start dates for the two groups and a low-incidence comparison group in the Seoul study; in the Italian study the authors' own sensitivity analysis shows the one-dose association no longer significant, and the three-dose hazard ratio below 1, when a 365-day minimum lag is used; and the journal has carried a notice of concern on the Seoul letter since October 2025, with no outcome yet (cvc-cancer-incidence-higher-after-vaccination). A Japanese paper that reported excess cancer deaths after the third dose was retracted in June 2024. Case reports describe individual patients whose cancer appeared or progressed after vaccination, but they have no comparison group and their compilers say they cannot show how often this happens. Laboratory work shows a spike-protein effect on p53 in two engineered cell lines, which its authors call preliminary. A mouse case report that was widely shared as proof carries an addendum in which its authors reject the "turbo cancer" reading. Statistics cannot prove a general claim false, so "vaccines cause cancer" and "vaccines cause turbo cancer" are filed as proposed, not refuted. What health bodies and politicians say, in both directions, is shown separately as adoption, not evidence: NCI, EMA, the American Cancer Society and ASCO say there is no evidence of a link; a Florida official, two members of a CDC advisory workgroup and a Senate subcommittee chair have said the opposite or called for more study. The FDA label says the vaccine has not been evaluated for carcinogenicity, which is neither a finding of harm nor of safety. key_points: - "No study has measured 'turbo cancer' by vaccination status; US and Japanese registries show a 2020 dip and then no surge, but cannot see a small subgroup." - "Two cohorts report more cancer after vaccination (Seoul 1.27, Pescara 1.23); both disputed, their authors stop short of cause, and the Seoul letter has a concern notice." - "Adoption is not evidence: a retracted 2024 Japan paper and a mouse case report whose authors rejected the turbo-cancer reading circulated widely; NCI, EMA, ASCO and a Senate hearing took opposite sides." basis: [cvc-rapid-progression-gap, cvc-no-registry-surge, cvc-cancer-incidence-higher-after-vaccination, cvc-korea-cohort-bias-concerns, cvc-italy-cohort-lag, cvc-case-reports-no-denominator, cvc-gibo-paper-retracted, cvc-mechanism-spike-p53] rated_claim: cvc-turbo-cancer-vaccine-caused lean: toward: "no" short: "leans no" strength: "moderate" because: > Registries show no surge in all-cancer rates after the rollout, no controlled study has found rapid progression after vaccination, the two cohorts that report more cancer have documented design problems and in one the one-dose association is no longer significant at a 12-month lag (breast and bladder stay raised), and the mechanism work is in cell lines. caveats: > Population totals cannot see a rise in a small group. No study has tested the claim as its proponents state it. The two cohorts cannot be set aside as noise, and the Seoul letter's concern notice is unresolved. Absence of a finding is a state of the search, not proof of absence. would_settle: > Not yet read, and listed as sub-questions of this question (claims cvc-sq-...): the Seoul journal's outcome on its notice of concern; the English national registry for 2021 to 2022; the 2026 US report in full; the Kruger interview and the "Died Suddenly" film by a human viewer; the randomised trials' supplementary adverse-event tables; the French cohort's cancer-specific result; the residual-DNA papers behind the FDA and EMA statements; and the cdc.gov and UK official pages. What would settle the question itself: a matched cohort that defines rapid progression in advance and compares vaccinated and unvaccinated people over several years (dt-rapid-progression-cohort), a registry-linked analysis with individual vaccination status (dt-registry-linked-incidence), and a replication of the Seoul analysis with a common start date and a lag period (dt-lag-analysis). # --------------------------------------------------------------------------- # SHORT ANSWER (owner decision 2026-10-02, log L-02): every dig answers the question # and names the exceptions where they exist. Each line points to the claim that carries # its weight. The answer is only as strong as those claims. # --------------------------------------------------------------------------- short_answer: question: "Do COVID-19 vaccines cause 'turbo cancer'?" answer: "Probably not, to the best of our knowledge: no study has shown it, and the national totals show no surge." summary: > "Turbo cancer" is a social-media term, first found in a post of 4 August 2022, for cancers said to appear or grow unusually fast after COVID-19 vaccination. No controlled study has defined such cancers and compared them by vaccination status. National cancer registries in the United States and Japan show a 2020 dip in diagnoses and no surge after the vaccine rollout. Two cohort studies, in Seoul and in Pescara province, report modestly higher cancer rates among vaccinated people; the Seoul authors say their findings do not establish cause, the Pescara authors call theirs preliminary, and both are disputed. These figures are studies; what health agencies and politicians say is recorded separately, as adoption. summary_figures: - figure: "US all-cancer incidence: 461.3 per 100,000 (2017-2019 and 2021), 425.6 in 2020" detail: > Annual Report to the Nation 2025: age-standardised incidence, registries covering 97 percent of the US population; 2020 was about 8 percent lower than the other years and 2021 "returned to prepandemic levels" per the authors. Mortality 2018 to 2022: 146.0 per 100,000, falling 1.5 per year on average. source: src-arn-2025 checked: primary - figure: "Japan all-cancer age-standardised mortality: 278.9 (2019) to 266.7 (2024) per 100,000" detail: > National Cancer Center Japan, both sexes, falling every year in that span. National incidence: 282.1 (2019), 263.6 (2020), 275.7 (2021), 271.9 (2022), 272.5 (2023). source: [src-ncc-mortality-2024, src-ncc-incidence-2023] checked: primary - figure: "Seoul cohort, hazard ratio 1.27 (1.21 to 1.33) for any cancer in one year" detail: > Kim et al., 2,975,035 matched adults, Seoul, one-year follow-up from index dates that differ by group (1 January 2022 for the unvaccinated). Authors: findings do not establish causal relationships. Disputed by two published critiques; a notice of concern has been on the article since 22 October 2025. source: [src-kim-2025-page, src-kim-2025-supp, src-locquet-2026, src-roccetti-2026] checked: primary - figure: "Pescara cohort, hazard ratio 1.23 (1.11 to 1.37) for cancer hospital admission" detail: > Acuti Martellucci et al., 296,015 residents aged 11 and over, 180-day minimum lag; at a 365-day lag the one-dose association was not significant and the three-dose hazard ratio was 0.90 (0.83 to 0.98). The authors call the findings "inevitably preliminary". source: src-acuti-martellucci-2025 checked: primary - figure: "Not summed" detail: > The two cohorts use different populations, outcomes (new cancer diagnosis claims versus hospital admission) and designs; their hazard ratios are not pooled here. source: [src-kim-2025-supp, src-acuti-martellucci-2025] checked: primary basis: > No controlled study measures rapid progression by vaccination status (cvc-rapid-progression-gap, a gap and not a finding). Registry totals show no surge but are ecological (cvc-no-registry-surge). The two cohorts that report more cancer are observational, carry design problems and are disputed (cvc-cancer-incidence-higher-after-vaccination, cvc-korea-cohort-bias-concerns, cvc-italy-cohort-lag). A 2024 Japanese paper was retracted (cvc-gibo-paper-retracted) and a widely shared mouse case report carries an addendum in which its authors reject the "turbo cancer" reading (cvc-eens-proved-turbo-cancer). except: - claim: cvc-cancer-incidence-higher-after-vaccination text: > What remains open, 1: two cohorts report more cancer among vaccinated people. In Seoul, any cancer in one year: hazard ratio 1.27 (1.21 to 1.33); in Pescara province, hospital admission with cancer, hazard ratio 1.23 (1.11 to 1.37) with at least one dose. We cannot say these are bias, and we cannot say they are not. scale: > Two studies in two populations. The Seoul cohort followed adults for one year, and its authors and critics agree that solid tumours rarely arise that fast. In Pescara, at a 12-month lag the one-dose association is not significant overall and the three-dose hazard ratio is below 1, but breast and bladder cancer admissions stay raised with at least one dose. scale_sources: [src-kim-2025-page, src-kim-2025-supp, src-acuti-martellucci-2025, src-locquet-2026] what_it_is: - "Two published observational comparisons of vaccinated with unvaccinated people (Seoul: adults 20 and over; Pescara: residents 11 and over), each reporting a hazard ratio above 1 for cancer." - "Reported with the authors' own cautions: the Seoul authors write that their findings do not establish causal relationships; the Pescara authors call theirs inevitably preliminary." what_it_is_not: - "Not evidence of 'turbo cancer': neither study defines or measures rapid progression." - "Not an independent replication: different populations, outcomes and designs." - "Not settled: critics describe design problems, and the Seoul letter carries an unresolved notice of concern." - claim: cvc-case-reports-no-denominator text: > What remains open, 2: individual patients. A scoping review lists 66 reports of 333 patients in 27 countries whose cancer appeared, recurred or progressed after vaccination or infection. These are real sequences in real patients. scale: > 66 article-level reports, mostly single patients; no comparison group and no count of how many vaccinated people were not affected. The compilers say they cannot estimate risk or incidence. scale_sources: [src-kuperwasser-2026, src-goldman-2021] what_it_is: - "A collection of case reports and small series, compiled by two authors who also hold the view that a signal may exist." what_it_is_not: - "Not an incidence or relative risk, and not a test of whether the order of events is more than coincidence." - "Not limited to vaccination: the review also covers cancers after infection." - claim: cvc-mechanism-spike-p53 text: > What remains open, 3: a laboratory result. In two engineered human cancer cell lines, spike protein expressed from a plasmid reduced p53 activity; the authors call it preliminary. A route from vaccination to cancer has not been shown in animals or people. scale: > Two cell lines, one laboratory, spike produced from a plasmid and not delivered by a vaccine. Other proposed routes (IgG4, interferon, spread to bone marrow, residual DNA) are hypotheses, reviews or serology without tumour data. scale_sources: [src-zhang-2024, src-eldeiry-hypothesis-2026, src-irrgang-2023, src-isidoro-2025] what_it_is: - "A cell-line experiment that found spike lowered p53 reporter activity and blunted drug-induced responses." what_it_is_not: - "Not a test of the vaccine, not an animal or human tumour result, and not an oncogenesis assay." - claim: cvc-gibo-paper-retracted text: > What remains open, 4: a retracted paper that is still cited. A Japanese study reporting excess cancer deaths after the third dose was retracted by its journal on 26 June 2024; its authors disagree. It cannot be counted as evidence, but it is still used as support. scale: > One paper; national all-cancer age-standardised mortality in Japan fell every year from 2019 to 2024. scale_sources: [src-gibo-retraction-2024, src-ncc-mortality-2024] what_it_is: - "An ecological time-series with no individual vaccination status, retracted over the authors' objection." what_it_is_not: - "Not proof the opposite is true: a retraction removes a paper, it does not test the claim." - claim: cvc-fda-label-not-evaluated text: > What remains open, 5: the label. The FDA prescribing information for COMIRNATY says the vaccine "has not been evaluated" for carcinogenicity and genotoxicity. scale: > One sentence in section 13.1 of the label, common to many products whose non-clinical studies were not run for those endpoints; we did not check this for other products. scale_sources: [src-comirnaty-pi] what_it_is: - "A statement that a test was not run." what_it_is_not: - "Not a finding that the vaccine causes cancer, and not a finding that it does not." - claim: cvc-rapid-progression-gap text: > A limit, not an exception: no controlled study we found defines rapid progression in advance and compares it between vaccinated and unvaccinated people. Until one is done, "no evidence of turbo cancer" means "not measured", and registry totals cannot see a small subgroup. scale: > Our search found survival comparisons in patients starting immunotherapy and no recurrence or progression comparison outside them. Search record in the claim. scale_sources: [src-natmed-2026, src-grippin-2025] also_true: - claim: cvc-ici-patients-no-worse-survival text: > The reverse direction has also been looked at. In four observational cohorts of patients starting immunotherapy for cancer, those vaccinated against COVID-19 around the start of treatment had no worse survival, and most had better. The largest independent cohort (95,015 French patients) attributes the early difference to transient or healthy-vaccinee effects. scale: > Hazard ratios for death 0.51 (Nature 2025, 884 patients), 0.615 (281), 0.86 (7,218 matched) and 0.90 at one to three months, no significant association after 12 months (95,015). One cohort reports more immune-related adverse events (hazard ratio 1.22). scale_sources: [src-grippin-2025, src-natmed-2026, src-ejc-2026, src-jitc-2026] # --------------------------------------------------------------------------- # SOURCE NOTICES (rule PS4): shown first on the page, above every other section. # --------------------------------------------------------------------------- source_notices: - source: src-gibo-2024 notice: > RETRACTED. The Cureus paper by Gibo et al. (2024) reporting excess age-adjusted cancer mortality in Japan after the third mRNA dose was retracted on 26 June 2024 after an expression of concern on 12 June 2024. The notice says the correlation between mortality rates and vaccination status cannot be proven with the data presented; the authors disagree. It is cited here only to record what it claimed and what happened to it. - source: src-kim-2025 notice: > NOTICE OF CONCERN. The Biomarker Research letter by Kim et al. (September 2025) has carried an editor's notice since 22 October 2025 that concerns have been raised and that editorial action will follow once they are investigated. It has not been retracted or corrected as of 5 October 2026. - source: src-eens-2023 notice: > CORRECTED BY ADDENDUM. The mouse case report by Eens et al. (Frontiers in Oncology, 2023) is not retracted. Its authors published an addendum on 20 October 2023 disassociating themselves from the term "turbo cancer" and stating what a case report does not show. subject: covid-vaccine-cancer title: "Do COVID-19 Vaccines Cause Cancer? The 'turbo cancer' claim (2020 to 2026)" description: > The question is resolvable in principle (a matched cohort with a pre-defined outcome would answer it) and has not been answered. This dig keeps five things apart: (1) the direct tests, two cohort studies, filed as `synthetic` evidence with their disputes; (2) the population registry series, which are ecological and cannot see individual exposure; (3) case reports and laboratory work, which are filed with what they cannot show; (4) a retracted paper and a mouse case report, whose authors reject the "turbo cancer" reading, that circulated as proof; and (5) the institutional record, filed as adoption in both directions, never as evidence. Scope was set by the owner to the popular claim of rapidly progressing cancers after vaccination ("turbo cancer"), with a timeline. # --------------------------------------------------------------------------- # VERIFICATION LEVELS used below (see log L-04): # anchor_checked: primary the full document was fetched and read directly (shell copy, sha256 in sources/MANIFEST.yaml) # anchor_checked: secondary read only as an abstract, through a summarizing fetch tool, or as quoted by others # anchor_checked: no not read # Evidential weight is capped at 4 unless anchor_checked is primary (SCHEMA rule 11, proposed). # Population statistics are filed as `synthetic` (CONTRIBUTING: statistics and computed results). # --------------------------------------------------------------------------- divergence_note: > The reach of this claim and the evidence for it come apart. Online, the phrase "turbo cancer" was reported by one monitoring vendor (not audited) as about 1.6 million English-language posts and 22 billion impressions from November 2022 to November 2024. From 2023 to 2026 the idea was linked, by a CBS-quoted physician, to a Florida surgeon general's January 2024 recommendation to providers (his documents themselves do not use the term), and the debate reached two slides at a CDC advisory meeting (September 2025), a Senate subcommittee hearing and a few review papers; the same period produced official statements that there is "no evidence" of a link. None of that is evidence either way. The evidence is thinner than either side's reach: no controlled study has measured rapid progression by vaccination status; two cohorts report a modest excess and are disputed; registry totals show no surge; case reports and cell-line work cannot bear the weight put on them. The open edge is the unmeasured part: a rise in a small subgroup would not show in all-cancer registry rates, and the study that would see it has not been done. Strata records the gap and gives the crowd, on every side, no evidential weight. claims: # ===================================================================== # A. THE CORE QUESTION # ===================================================================== - id: cvc-turbo-cancer-vaccine-caused state: proposed statement_kind: judgment statement: > COVID-19 vaccination causes unusually rapid or aggressive cancers (called "turbo cancer") in vaccinated people. evidence_class: interpretive confidence: provisional evidential_weight: 1 # of 5 anchor_checked: secondary # the Kruger interview was not watched; case reports and testimony read adoption_weight: 3 # of 5, shown separately; never combined with evidential_weight adoption_note: > No survey of belief was found. Indirect indicators only: one monitoring vendor (not audited) counted about 1.6 million English-language posts and 22 billion impressions containing the term from November 2022 to November 2024; the claim was raised at a Senate subcommittee hearing (3 June 2026). Adoption is information, not evidence. would_change_if: > a matched cohort that defines rapid progression in advance finds it clearly more common after vaccination, or a registry-linked analysis with individual vaccination status finds excess early-stage-to-advanced progression in vaccinated people anchor: type: case-reports-and-testimony description: > Rests on case reports and clinical impressions. Etana Hecht's post of 4 Aug 2022 summarises an interview with Dr Ute Kruger, a pathologist, and says "Dr. Kruger initially thought that these turbo cancers, as she calls them, were due to delayed doctor appointments from Covid lockdowns, but that period is long over" (summary of an interview not watched here); Hecht calls the testimony "by nature anecdotal". Angus Dalgleish's written Senate testimony (3 Jun 2026) says he has "no doubt" the vaccine "likely played a significant role" in "these unexpected cancers"; the sentence follows his account of relapses and aggressive presentations in his own practice and a passage on cancer diagnoses disclosed by vaccinated public figures, and he adds that he raises this "not to imply certainty regarding any individual case". The testimony has no data tables. Case reports are in cvc-case-reports-no-denominator. No source read defines the term or measures it. accessible: true sources: [src-hecht-turbo-2022, src-psi-dalgleish, src-kuperwasser-2026, src-goldman-2021] positions: - label: "Yes" holder: "Ute Kruger (as summarised by Hecht), William Makis, Angus Dalgleish, others who use the term" ground: > Clinicians report tumours that are larger, multiple, in younger patients or recurring after vaccination, from their own practice. - label: "No / not shown" holder: "ASCO (Gralow), NCI, EMA, the authors of the Eens report" ground: > No clinical evidence proving it; cancer "does not develop suddenly" (ASCO written testimony); the Eens authors say "nor do we recognize it as a legitimate medical term". basis: > Filed as `proposed` rather than `refuted`. Rule 9 lets a refutation rest only on material or primary-text evidence, and what bears on this claim (registry series and cohorts) is statistics, which CONTRIBUTING.md classes as `synthetic`. The owner decided (vaccines-autism log L-19) that statistics cannot prove a claim false but can show that no evidence supports it. The claim is not retired by a test: no test has been run on it as stated (cvc-rapid-progression-gap). Registry totals (cvc-no-registry-surge) are consistent with no population-wide surge and cannot see a small subgroup. - id: cvc-vaccines-cause-cancer state: proposed statement_kind: judgment statement: > COVID-19 vaccination raises the risk of cancer in vaccinated people (in general, not only in the rapid form). evidence_class: interpretive confidence: provisional evidential_weight: 2 # of 5; two observational cohorts with disputed designs, one mechanism experiment in cell lines anchor_checked: primary # Kim 2025 (letter and additional files) and Acuti Martellucci 2025 read in full adoption_weight: 3 # of 5 adoption_note: > Cited as support in three review or opinion papers (Kuperwasser and El-Deiry 2026, Gentilini 2026, Kakeya 2025); Senate subcommittee hearing title (3 Jun 2026) assumes "Plausible Mechanisms of COVID-19 Injections Causing Cancer". Uses, not replications. would_change_if: > a cohort with a common start date and a lag period replicates a clear excess, or a mechanism is shown in vaccinated animals or people flagged_sources: [src-kim-2025] # expression of concern; shown beside this claim (rule PS3) anchor: type: observational-cohorts-and-lab-work description: > Two cohorts report hazard ratios above 1 (cvc-korea-cohort-reports-excess, cvc-italy-cohort-lag), both disputed (cvc-korea-cohort-bias-concerns); one cell-line experiment (cvc-mechanism-spike-p53); mechanism reviews that name no tumour data (cvc-mechanism-other-hypotheses). In the Seoul cohort the overall hazard ratio is highest for adenoviral-vector vaccines (1.472) and lowest for mRNA only (1.199), the reverse of what an mRNA-specific mechanism would suggest, as our observation, not the authors' claim. accessible: true sources: [src-kim-2025, src-kim-2025-supp, src-acuti-martellucci-2025, src-zhang-2024] positions: - label: "Yes (what exists)" holder: "Kim et al. (Seoul); Acuti Martellucci et al. (Pescara); Kuperwasser and El-Deiry" ground: > Hazard ratios of 1.27 and 1.23 for cancer among vaccinated adults, and a compiled pattern of case reports. - label: "No / not shown" holder: "NCI; EMA; ASCO; Locquet et al.; Roccetti; Kutsuna and Suzuki (on Japanese cancer mortality)" ground: > No evidence of a causal link; one-year solid-tumour carcinogenesis is not biologically plausible (Locquet; Kutsuna and Suzuki, who write that a measurable rise in cancer deaths within the year of a dose is biologically implausible); design bias in the cohorts. basis: > Filed as `proposed` rather than `refuted` for the Rule 9 reason in cvc-turbo-cancer-vaccine-caused. It is not a sign the claim is settled: the question has two disputed positive studies and one disputed set of negatives (registry totals). - id: cvc-cancer-incidence-higher-after-vaccination state: contested statement_kind: judgment statement: > COVID-19-vaccinated people are diagnosed with cancer more often than unvaccinated people (hazard ratio 1.27 in a Seoul cohort and 1.23 for cancer hospital admission in a Pescara cohort), and this reflects the vaccination and not how the groups were defined and followed. evidence_class: synthetic confidence: low evidential_weight: 2 # of 5 anchor_checked: primary # both papers, the Seoul additional files and the critiques read in full adoption_weight: 2 # of 5 adoption_note: > Cited as support by three papers; criticised by two published commentaries and by FactCheck.org and Science Feedback (secondary, not opened). Adoption is information, not evidence. would_change_if: > a cohort with one common start date, a 6 to 12 month lag and a pre-specified outcome finds a hazard ratio clearly above 1 (the claim would strengthen), or the same design finds it at 1 (it would weaken); or the Seoul journal's investigation ends in a correction or retraction flagged_sources: [src-kim-2025] disputed_by: - {who: "Locquet M and colleagues, Frontiers in Public Health commentary (2026)", kind: work, position: "Different index dates in the two groups create calendar-time bias; surveillance bias with national screening; 30 uncorrected subgroup tests; one-year solid-tumour carcinogenesis does not seem biologically plausible; low-quality evidence by GRADE.", source_read: true, via: src-locquet-2026} - {who: "Roccetti M, AIMS Public Health (2026)", kind: work, position: "The Seoul unvaccinated group has far fewer people aged 65 and over than national data and a cancer rate 45 percent below the national rate, which would inflate hazard ratios.", source_read: true, via: src-roccetti-2026} - {who: "Kakeya H and colleagues, JMA Journal reply (2026)", kind: work, position: "Reply to a critique of their Japanese opinion paper; cites the Seoul letter as 'a recent peer-reviewed paper from Korea' reporting an association and writes that they 'cannot rule out' vaccination worsening pre-existing cancers; says their paper did not assume causality. Not an analysis of the Seoul cohort.", source_read: true, via: src-kakeya-reply-2026} - {who: "National Cancer Institute and European Medicines Agency", kind: institution, position: "No evidence that COVID-19 vaccines cause cancer.", source_read: true, via: src-nci-2023} anchor: type: live-dispute description: > Seoul: Kim et al., Biomarker Research 2025;13:114 (a Letter), Korean National Health Insurance claims, Seoul residents aged 20 and over, 595,007 unvaccinated and 2,380,028 vaccinated after 1:4 propensity matching; one-year follow-up; any cancer hazard ratio 1.27 (95% CI 1.21 to 1.33); the authors write "our findings do not establish causal relationships" (Additional File 2). Pescara: Acuti Martellucci et al., EXCLI J 2025;24:690, 296,015 residents aged 11 and over, outcome first hospital admission with cancer; hazard ratio 1.23 (1.11 to 1.37) for at least one dose and 1.09 (1.02 to 1.16) for three or more; the authors call the findings "inevitably preliminary". Their critics and the cohort limits are in cvc-korea-cohort-bias-concerns and cvc-cohort-limits. accessible: true sources: [src-kim-2025, src-kim-2025-supp, src-acuti-martellucci-2025, src-locquet-2026, src-roccetti-2026, src-kakeya-reply-2026] positions: - label: "Higher after vaccination" holder: "Kim et al. (Seoul, findings do not establish causal relationships); Acuti Martellucci et al. (Pescara, findings inevitably preliminary)" ground: "Hazard ratios above 1 for cancer in vaccinated compared with unvaccinated people." - label: "Not shown to be an effect" holder: "Locquet et al.; Roccetti; NCI; EMA" ground: "Different start dates, a low-incidence control group, surveillance bias and no lag period; no mechanism that acts within a year." basis: > Contested because both the data and their meaning are in dispute and the authors of both studies stop short of cause. The incidence difference itself is reported in the papers; what is contested is whether it is more than design. - id: cvc-no-registry-surge state: established statement_kind: judgment statement: > In the national cancer series we read for the United States and Japan, age-standardised all-cancer incidence fell in 2020 and returned toward earlier levels by 2021 to 2022, and age-standardised all-cancer mortality kept falling to 2023 or 2024, so these series show no surge in cancer after the 2021 vaccine rollout beyond long-running trends at particular sites. evidence_class: synthetic confidence: moderate evidential_weight: 4 # of 5; the series were read directly, but they are ecological anchor_checked: primary # US 2025 report (full text) and the two Japanese national files read directly adoption_weight: 3 # of 5 adoption_note: > Stated in the Senate record by Sen. Blumenthal ("not supported") and in NCI, EMA and ASCO statements. Adoption is information, not evidence. would_change_if: > a later year of the US or Japanese series, or the English registry, shows an age-standardised all-cancer rise that cannot be traced to a long-running site trend; or an analysis with individual vaccination status finds a rise in a subgroup that totals hide anchor: type: registry-series-read-directly description: > US: Sherman et al., Annual Report to the Nation 2025 (Cancer 2025, read in full): incidence 2017-2019 and 2021 aggregate 461.3 per 100,000, 2020 425.6; joinpoint all sites 2012-2021 annual percent change 0.1 (95% CI -0.1 to 0.3); mortality 2018-2022 146.0, average annual change -1.5; joinpoint 2016-2019 -2.1 and 2019-2022 -1.3, "slowing" but not reversing; the 2026 report (abstract and CDC release only) reports mortality 2019-2023 falling 1.6 per year in men and 1.1 in women and incidence rising slightly 2018 to 2022, "driven by increases in breast cancer among women and prostate cancer among men" (wording as recorded from the CDC release, which was not opened by the independent reviewer; it is not in the abstract). Japan: National Cancer Center files read directly: age-standardised all-cancer mortality, both sexes, 278.9 (2019), 276.8 (2020), 275.0, 273.9, 268.4, 266.7 (2024); registry incidence 282.1 (2019), 263.6 (2020), 275.7, 271.9, 272.5 (2023). Not flat at every site: Japanese female breast cancer mortality rose from 20.6 to 21.9 in 2022 and fell to 21.3 in 2023; leukaemia rose from 6.7 to 7.1 in 2022 and stayed there; pancreas has risen steadily since 2010. We did not test breakpoints at the vaccine rollout ourselves. accessible: true sources: [src-arn-2025, src-arn-2026-abstract, src-cdc-arn-2026-release, src-seer-arn-2026-page, src-ncc-mortality-2024, src-ncc-incidence-2023] positions: - label: "No surge" holder: "The registry series above" ground: "Rates dipped in 2020 and then followed earlier trends; death rates kept falling." - label: "Surge hidden by totals" holder: "Not asserted by any source read; stated here as a limit" ground: "Totals include every age and cancer type; a rise in a small subgroup would not move them." basis: > Our judgment, from the series themselves. Moderate, not high: the 2026 US report was read only as an abstract and a press release, the English registry (403 to our proxy) and Danish data were not read, and the series stop at 2023 (US mortality) or 2024 (Japan). Ecological series have no vaccination status, so they cannot test causation in either direction and cannot see a small group. The female incidence rise and the prostate rise pre-date the vaccines in the US report's joinpoint tables (segments starting 2003 and 2014). The 2020 dip complicates any reading of counts after it (cvc-pandemic-diagnosis-disruption). - id: cvc-turbo-cancer-entity state: proposed statement_kind: judgment statement: > "Turbo cancer" is a recognised medical term for a defined kind of cancer. evidence_class: interpretive confidence: provisional evidential_weight: 1 # of 5 anchor_checked: primary # Szebeni 2025 (Oct), the Eens addendum and the ASCO testimony read adoption_weight: 3 # of 5 adoption_note: "See cvc-turbo-cancer-vaccine-caused. Adoption is information, not evidence." would_change_if: > a published case definition for the term is adopted by a cancer society or tested in a cohort, or an oncology guideline uses it anchor: type: usage-in-the-literature description: > The phrase is in a small number of review papers: our Europe PMC phrase searches returned 5 or 6 hits (the two research streams counted differently; log L-09), none an epidemiological study. Szebeni (Pharmaceutics 2025, Oct) writes that the close timing of blood-cell cancers and vaccination led to the concept of 'turbo cancer', "a highly debated assertion on the acceleration of malignancy and rise in rapidly progressing new cancers following vaccinations", and that "This concept is not accepted by mainstream medicine"; the Eens authors write in their addendum that they "unequivocally disassociate ourselves from this term" and "nor do we recognize it as a legitimate medical term"; ASCO's written testimony says the appearance of late-stage aggressive tumours within weeks or months of an injection is "biologically incompatible" with decades of research on cancer causes. accessible: true sources: [src-szebeni-2025-oct, src-szebeni-2025-apr, src-eens-addendum-2023, src-psi-gralow] flagged_sources: [src-eens-2023] basis: > Filed as `proposed`. No source read gives the term a case definition (see cvc-rapid-progression-gap). Its use in journals records adoption of the word, not a finding. - id: cvc-rapid-progression-gap state: searched_gap statement_kind: search_result statement: > No controlled study we found defines rapid cancer progression, or aggressive cancer at diagnosis, in advance and compares it, or cancer recurrence, between vaccinated and unvaccinated people. The nearest are survival comparisons in patients starting immunotherapy. absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 2 # of 5, for the existence of the gap anchor_checked: secondary # the dig's own search records adoption_weight: 3 # of 5 adoption_note: > Kuperwasser and El-Deiry say there are "no published population studies in the US with mortality or cancer incidence follow-up beyond 42 days" (quoted, not verified by us). That is their claim, recorded as adoption. would_change_if: "a study of rapid progression or recurrence by vaccination status is published" next_step: > Match vaccinated and unvaccinated adults newly diagnosed with the same cancer type and stage in a registry linked to vaccination records, define "rapid progression" before looking (stage shift, time to recurrence, survival by stage), and report it with a 6 to 12 month lag and a common start date (dt-rapid-progression-cohort). anchor: type: search-record description: > Searched PubMed and Europe PMC, 5 Oct 2026: "COVID-19 vaccination cancer risk cohort"; "turbo cancer COVID vaccine"; TITLE:"cancer" AND TITLE:"vaccinated" AND TITLE:"unvaccinated"; vaccination terms with Clalit or Israel; with Vaccine Safety Datalink or Kaiser; with Netherlands, Denmark, Sweden, Norway or Finland; "excess mortality" AND cancer AND COVID; and the citation lists of the Seoul, Pescara and Japanese papers. Hits were screened by title and abstract. None is a study whose primary outcome is rapid progression or recurrence by vaccination status outside immunotherapy cohorts (cvc-ici-patients-no-worse-survival). The one Vaccine Safety Datalink hit was a mortality study with no cancer outcome. A title and abstract screen can miss a study; the absence is a state of our search. accessible: true basis: > A gap is not evidence of harm and not evidence of safety. It is the central open item: the claim as proponents state it has not been tested, so the totals and the cohorts stand in for it. # ===================================================================== # B. THE EVIDENCE TESTS # ===================================================================== - id: cvc-korea-cohort-reports-excess state: established statement_kind: reported statement: > Kim et al. (Biomarker Research, 26 Sep 2025, a Letter) report that among 2,975,035 matched Seoul adults aged 20 and over in the Korean national health insurance database, vaccinated people had more first cancer diagnoses in one year of follow-up than unvaccinated people: overall hazard ratio 1.27 (95% CI 1.21 to 1.33), 42.62 against 33.43 per 10,000 at one year. Site hazard ratios were 1.351 (thyroid), 1.335 (gastric), 1.283 (colorectal), 1.533 (lung), 1.197 (breast) and 1.687 (prostate). evidence_class: primary_text confidence: high evidential_weight: 4 # of 5, for what the paper reports; the paper is not a test of cause (see cvc-korea-cohort-bias-concerns) anchor_checked: primary # letter, PMC XML and Additional Files 1 and 2 read adoption_weight: 3 # of 5 adoption_note: > Described as showing "27% higher overall" risk in Kuperwasser and El-Deiry and in Gentilini et al.; cited as support in a JMA Journal letter (uses, not replications); criticised in two commentaries and by FactCheck.org and Science Feedback (secondary). would_change_if: > the journal corrects or retracts the letter, or the Additional Files differ from the PMC copy read flagged_sources: [src-kim-2025] # expression of concern (rule PS3) anchor: type: the-document-itself description: > Kim HJ, Kim MH, Choi MG, Chun EM, Biomark Res 2025;13:114 (received 11 Jul 2025, accepted 21 Aug 2025), PubMed type Letter. Not "national": the cohort is Seoul residents. 8,407,849 screened (679,479 unvaccinated, 7,728,370 vaccinated); after exclusions and 1:4 greedy propensity matching, 595,007 unvaccinated and 2,380,028 vaccinated; index date 1 Jan 2022 for the unvaccinated and the day after finishing the primary series for the vaccinated; one-year follow-up; matching on age, sex, income proxy, Charlson index and prior SARS-CoV-2 infection (77.22 percent in both groups). The overall figure is in the additional files; the letter's abstract gives site hazard ratios. Cumulative incidence per 10,000 at 1 month 2.91 against 1.63, at 1 year 42.62 against 33.43. By vaccine type, overall hazard ratio: adenoviral only (the paper's "cDNA vaccine" group) 1.472, heterologous 1.339, mRNA only 1.199. Booster against no booster (n=1,067,688): overall 1.01 (0.95 to 1.06), pancreatic 2.25 (1.44 to 3.50), leukaemia 0.56 (0.35 to 0.92). Authors' limits (Additional File 2): "our findings do not establish causal relationships"; one-year follow-up is short for solid tumours and "the possibility of reverse causation or surveillance bias cannot be excluded". accessible: true sources: [src-kim-2025, src-kim-2025-supp, src-kim-2025-page] basis: > This claim records what the paper reports. Whether the difference is an effect is cvc-cancer-incidence-higher-after-vaccination (contested). Our observation, not the authors': the gap between the two groups is already present at one month, and the two groups have different index-date definitions. - id: cvc-korea-cohort-bias-concerns state: contested statement_kind: judgment statement: > The excess in the Seoul cohort is explained at least in part by how the two groups were defined and followed: different start dates, an unvaccinated control group with a cancer rate well below national figures, screening and surveillance differences, and many uncorrected subgroup tests. evidence_class: interpretive confidence: low evidential_weight: 2 # of 5 anchor_checked: primary # the cohort's additional files, Locquet and Roccetti read adoption_weight: 2 # of 5 adoption_note: "Two published critiques; also criticised by FactCheck.org and Science Feedback (secondary, not opened)." would_change_if: > the authors publish a reply or a re-analysis with a common start date, a lag period and delayed-entry modelling and the excess remains (the claim weakens) or disappears (it strengthens) flagged_sources: [src-kim-2025] disputed_by: - {who: "Kim HJ and colleagues (authors of the Seoul letter)", kind: group, position: "Single-payer access reduces surveillance differences; they acknowledge reverse causation and surveillance bias cannot be excluded and that the findings do not establish cause. No reply to the commentaries was found.", source_read: true, via: src-kim-2025-supp} anchor: type: live-dispute description: > Locquet et al., Front Public Health 2026;14:1772406: different index dates (vaccination date against 1 Jan 2022) create "calendar time" bias without time-dependent or delayed-entry modelling; matching targets the effect in the vaccinated; 77.22 percent prior infection with no sensitivity analysis excluding it; surveillance bias for screened cancers (thyroid, breast, gastric, colorectal); 30 subgroup analyses uncorrected for multiplicity; "One-year solid tumor carcinogenesis does not seem biologically plausible"; asks for a lag of 6 to 12 months; rates the evidence low quality by GRADE. One author reports personal fees from pharmaceutical and diagnostic firms. Roccetti, AIMS Public Health 2026 (single author, benchmark audit): in the cohort 12.2 percent were aged 65 and over against 18 percent nationally, and cancer incidence in the unvaccinated aged 65 and over was 85.2 against 155.2 per 10,000 nationally (45.1 percent lower); argues a deflated control baseline inflates the hazard ratios. Our caveat: Roccetti compares a Seoul-only matched cohort with national figures and uses figures taken from the letter. accessible: true sources: [src-locquet-2026, src-roccetti-2026, src-kim-2025-supp] positions: - label: "Design explains much of it" holder: "Locquet et al.; Roccetti" ground: "Asymmetric start dates, a low-baseline control group, surveillance and multiplicity." - label: "Authors' position" holder: "Kim et al." ground: "Report the ratios; say they do not establish cause; argue access to care is similar." basis: > Contested, not established: the critics' arguments are specific and checkable, but no re-analysis has been published, and Roccetti's benchmark is a single author's comparison with national data. An unresolved notice of concern is on the article (cvc-sq-korea-concern-outcome). - id: cvc-italy-cohort-lag state: established statement_kind: reported statement: > Acuti Martellucci et al. (EXCLI Journal, July 2025) report that in 296,015 residents of Pescara province aged 11 and over, first hospital admission with cancer was more frequent after at least one dose (hazard ratio 1.23, 1.11 to 1.37) and three or more doses (1.09, 1.02 to 1.16) than with none, and that with a 365-day minimum lag the one-dose association was not significant overall and the three-dose hazard ratio was 0.90 (0.83 to 0.98); the paper adds that at 365 days "breast and bladder cancer hospitalizations maintained their positive association with ≥1 dose". evidence_class: primary_text confidence: high evidential_weight: 4 # of 5, for what the paper reports anchor_checked: primary # the paper (PMC) and the exchange of letters read adoption_weight: 2 # of 5 adoption_note: "Cited in the scoping review and in Gentilini et al. as an association. Four letters and replies were published." would_change_if: "the authors or a critic publish a corrected analysis that changes the figures" anchor: type: the-document-itself description: > Acuti Martellucci C et al., EXCLI J 2025;24:690-707, DOI 10.17179/excli2025-8400. Follow-up to 31 Dec 2023; the unvaccinated were followed from 27 Jun 2021 (one-dose comparison) or 26 Dec 2021 (three-dose comparison) and the vaccinated from 180 days after the first (or third) dose, a 180-day minimum lag between exposure and outcome (90 and 365 days as sensitivity runs). Unvaccinated 49,265, at least one dose 246,750, three or more 183,999. Outcome: first admission with a cancer diagnosis in the regional discharge database (skin cancer excluded; people with a prior admission for the same cancer in the previous ten years excluded), not registry incidence; 3,134 admissions. Cox models adjusted for age, sex, prior infection, diabetes, hypertension, cardiovascular disease, COPD, kidney disease and prior cancer admission; no smoking, screening or healthcare-use data. Crude proportions 0.85 percent (unvaccinated) and 1.15 percent (at least one dose). Without a recorded prior infection the hazard ratios were 1.31 (1.16 to 1.47) and 1.11 (1.03 to 1.20). Breast, bladder and colorectal were raised at one dose; lung, ovary and thyroid were not. By sex, the raised risk was seen only in men with at least one dose (1.31, 1.12 to 1.52). All-cause death hazard ratio 0.42 (0.39 to 0.44) for at least one dose, which the authors ascribe to healthy-vaccinee bias. The authors write: "Given that it was not possible to quantify the potential impact of the healthy vaccinee bias and unmeasured confounders, these findings are inevitably preliminary." Letters: Malatesta et al. argue the mortality result reflects immortal-time bias; Manzoli et al. reply that age-adjusted models align time zero. Their main argument is about mortality; Malatesta et al. also note that the excess of tumours "could also be due to factors unrelated to actual increases". accessible: true sources: [src-acuti-martellucci-2025, src-malatesta-2025, src-manzoli-reply-2025] basis: > Reported, with the lag result alongside: an association that weakens or loses significance when the early months and then a full year are excluded is the pattern the critics of cohort designs predict for bias, and the pattern an effect that takes years to arise would also show for a one-year lag. We do not adjudicate; the authors' own words are that the results are preliminary. The critics of the Italian paper are not mainly on the reassurance side: their main argument is that its mortality benefit is an artefact, though one of them also says the cancer excess may be unrelated to a real increase. - id: cvc-cohort-limits state: established statement_kind: judgment statement: > Both cohorts compare people who chose to be vaccinated with people who did not, take their outcomes from claims or hospital records and not from cancer registries, and have no data on smoking or screening; the Seoul authors state that the results do not establish cause and the Pescara authors call theirs preliminary. evidence_class: primary_text confidence: high evidential_weight: 4 # of 5 anchor_checked: primary adoption_weight: 2 # of 5 adoption_note: "Not measured." would_change_if: > a cohort with randomised or registry-verified exposure and outcome, and data on screening, is published flagged_sources: [src-kim-2025] anchor: type: the-documents-themselves description: > Seoul: outcome is the first primary diagnosis of cancer in insurance claims; Additional File 2 says "the possibility of reverse causation or surveillance bias cannot be excluded". Pescara: outcome is hospital admission with a cancer diagnosis, which the authors say "represent only a proxy of the total new cases of cancer"; smoking unknown. Pescara's all-cause death hazard ratio of 0.42 for at least one dose, which no one reads as a protective effect of that size, is the authors' own evidence that bias between vaccinated and unvaccinated groups in this setting is large (our reading; the authors say so for mortality). Neither study can say whether cancers were found earlier because vaccinated people use more healthcare, or found later because they did not. accessible: true sources: [src-kim-2025-supp, src-acuti-martellucci-2025] basis: > Our judgment from the authors' stated limits. It is a limit on both positive cohorts, not a conclusion that they are wrong. Bias could also run the other way: the Italian authors say healthy-vaccinee bias can cause underestimation of harm for lifestyle-linked cancers. - id: cvc-case-reports-no-denominator state: established statement_kind: judgment statement: > Published case reports and small series of cancers appearing, recurring or progressing after COVID-19 vaccination describe individual patients; they have no comparison group and no denominator, so they cannot show how often this happens or whether it is more than expected. evidence_class: primary_text confidence: high evidential_weight: 4 # of 5 anchor_checked: primary # the review and Goldman 2021 read in full; other case reports as abstracts adoption_weight: 3 # of 5 adoption_note: > The review was cited in testimony to a Senate subcommittee ("nearly 70 publications ... more than 300 reported cancer cases from 27 countries") and in two slides at a CDC advisory meeting. Adoption is information, not evidence. would_change_if: > a registry-linked analysis of the same cancer types shows the reported patterns more often in vaccinated than unvaccinated people anchor: type: the-document-itself description: > Kuperwasser C, El-Deiry WS, Oncotarget 2026 (3 Jan 2026): 69 publications (January 2020 to October 2025 per the abstract), 66 article-level reports of 333 patients in 27 countries; reports of lymphomas, leukaemias, breast, lung, melanoma, sarcoma, pancreas, glioma, Kaposi and Merkel; recurring themes of "unusually rapid progression, recurrence, or reactivation". The review also covers cancers after infection. Its authors' limits: most reports are single-patient descriptions or small series, "highly susceptible to publication bias and selective reporting", and the review is "not designed to estimate cancer risk or incidence, nor to draw causal inferences". The authors report searching PubMed, Scopus, Web of Science, Google Scholar and React19 (their abstract calls it a systematic search) and, after PubMed keyword searches returned little, an expanded search with general web queries; the abstract gives the search period as January 2020 to October 2025 and the results text as January 2020 to April 2025. The Editor-in-Chief, El-Deiry, is also an author, was not involved in the peer review, and emailed the manuscript to the NCI Director on 12 Dec 2025 (editorial note in the paper). accessible: true sources: [src-kuperwasser-2026, src-goldman-2021, src-vaccines-2025-cutaneous, src-ijo-2022, src-jeadv-2022, src-gentilini-2026] basis: > Counts in the sources differ in how they split the 69 items (one reading: 50 case reports, 5 series, 3 reviews, 8 population items; the paper's text: 55 of 69, 81 percent, single-patient reports or small series; log L-09). The conclusion does not depend on which. Mistiming and confounding are possible in any single report; so is a real effect; a report cannot say which. - id: cvc-goldman-aitl-case state: established statement_kind: reported statement: > Goldman et al. (Frontiers in Medicine, 2021) report one 66-year-old man with angioimmunoblastic T-cell lymphoma whose whole-body total lesion glycolysis was 5.3 times higher on a second PET scan 22 days after the first, 8 days after a booster dose, with new lesion sites; the authors call the progression "highly unexpected" and say extrapolating to other patients is premature. evidence_class: primary_text confidence: high evidential_weight: 3 # of 5, for what the report documents anchor_checked: primary adoption_weight: 2 # of 5 adoption_note: "Not measured. Included in the scoping review's list." would_change_if: "not applicable: a report of one patient; a cohort would test the pattern" anchor: type: the-document-itself description: > Goldman S et al., Front Med 2021;8:798095. Biopsy after a PET/CT on 8 Sept 2021 showing stage IV-pattern disease; two vaccine doses 5 to 6 months earlier; booster 14 days after the first PET; steroids started right after; antibody titres unchanged. The authors write that it is "reasonable to postulate" the booster was the trigger, and also that extrapolation of the findings to other patients is premature; no comparison patients. Our observation, not the authors': the same 22 days include an untreated aggressive lymphoma, a flu-like illness at presentation, and a short interval between scans. accessible: true sources: [src-goldman-2021] basis: > An example of what a case report contains: a documented sequence and a plausible reading, without a way to separate the vaccine from the disease's own course. - id: cvc-eens-mouse-case-report state: established statement_kind: reported statement: > Eens et al. (Frontiers in Oncology, 1 May 2023) report one BALB/c mouse, one of a group of 14 given high-dose intravenous BNT162b2 in a myocarditis model, that was found dead two days after the booster with a B-cell lymphoblastic lymphoma; the authors say "a demonstration of direct causality remains difficult" and, in an addendum, that over 70 other BALB/c mice they injected did not develop haematologic malignancy. evidence_class: primary_text confidence: high evidential_weight: 3 # of 5, for what the report documents anchor_checked: primary # paper (PubMed record and text) and addendum read adoption_weight: 3 # of 5 adoption_note: "Widely circulated; see cvc-eens-proved-turbo-cancer. Adoption is information, not evidence." would_change_if: "the journal retracts the paper or the authors publish further animal data" flagged_sources: [src-eens-2023] # corrected by addendum (rule PS3) anchor: type: the-document-itself description: > Eens S et al., Front Oncol 2023;13:1158124 (PMID 37197431). Single mouse; CD19-, TdT- and c-MYC-positive lymphoid neoplasm; vaccination intravenous (not intramuscular) and high dose; found dead two days after the booster at 14 weeks. Addendum (doi 10.3389/fonc.2023.1267904, 20 Oct 2023): "we have injected over 70 BALB/c mice ... Excluding the published case report, none of these other vaccinated animals developed hematologic malignancies of any type"; spontaneous lymphoma is known in the strain; "case reports ... do not test correlation between variables, nor are they used to demonstrate causality". FactCheck.org (2023-08-31, secondary) says 1 of 14 high-dose mice died of lymphoma. accessible: true sources: [src-eens-2023, src-eens-addendum-2023, src-factcheck-2023] basis: > Not retracted. The paper stays in the record with its addendum. - id: cvc-eens-proved-turbo-cancer state: refuted statement_kind: reported refutes_target: > Social-media posts of 2023 presenting Eens et al. as proving that Pfizer mRNA induced "turbo cancer" (post text "SHOCKING: New Study proves Pfizer mRNA induced turbo cancer", as quoted by FactCheck.org, 31 Aug 2023) statement: > The Eens et al. 2023 mouse case report proved that Pfizer's mRNA vaccine induced "turbo cancer". evidence_class: primary_text refutation_class: narrative_drift confidence: high evidential_weight: 5 # of 5, for the refutation anchor_checked: primary # the authors' own addendum read adoption_weight: 3 # of 5 adoption_note: "Spread widely in 2023 (FactCheck.org). Adoption is information, not evidence." would_change_if: > the authors retract their addendum, or the paper is shown to say what the posts claimed flagged_sources: [src-eens-2023] anchor: type: the-document-itself description: > The paper is a case report of one mouse in which the authors write that "a demonstration of direct causality remains difficult". Their addendum of 20 Oct 2023 says: "we wish to unequivocally disassociate ourselves from this term. In our case report, there is not a single reference to a condition called 'turbo cancer', nor do we recognize it as a legitimate medical term"; the vaccine was given intravenously at high dose in a myocarditis-model replication. accessible: true sources: [src-eens-2023, src-eens-addendum-2023, src-factcheck-2023] basis: > Refuted on primary text: the source of the claim says it does not claim this. It is the post that drifted from the paper (`narrative_drift`), not the paper that is wrong. A refutation of the posts does not test whether vaccines cause cancer. - id: cvc-ici-patients-no-worse-survival state: proposed statement_kind: judgment statement: > In four observational cohorts of patients starting immune checkpoint inhibitors, those vaccinated against COVID-19 around the start of treatment had no worse survival than unvaccinated patients (hazard ratios 0.51 to 0.90 for death or survival); one cohort also reports more immune-related adverse events (hazard ratio 1.22), and the largest independent cohort found no significant association after 12 months. evidence_class: synthetic confidence: low evidential_weight: 2 # of 5 anchor_checked: secondary # Grippin read in full; the other three read as abstracts adoption_weight: 2 # of 5 adoption_note: "The Nature paper is widely noticed; Kuperwasser and El-Deiry criticise its interpretation." would_change_if: > a prospective study or a cohort with a negative-control vaccine finds worse survival in vaccinated patients, or confirms a durable benefit anchor: type: observational-cohorts description: > Grippin et al., Nature 2025;647:488-497: MD Anderson, 180 vaccinated within 100 days of starting immunotherapy and 704 not (the non-small-cell lung cohort; the paper also reports melanoma cohorts); adjusted for 39 covariates; median survival 20.6 against 37.3 months as printed, adjusted hazard ratio 0.51 (0.37 to 0.71); no benefit around chemotherapy alone or with influenza or pneumonia vaccines; mice and healthy volunteers show a rise in type I interferon. Single-centre cohort, retrospective. Nature Medicine 2026 (abstract): French national claims, 95,015 adults; pre-treatment mRNA vaccination hazard ratio for death 0.90 (0.87 to 0.94) at 1 to 3 months, 0.96 (0.93 to 1.00) at 6 months, "no significant association after 12 months", with "similar early patterns with non-mRNA COVID-19 and other adult vaccines" suggesting "modest, transient associations or healthy-vaccinee effects". European Journal of Cancer 2026 (abstract): 281 patients, 47 vaccinated, median 51.0 against 21.0 months, hazard ratio 0.615 (0.407 to 0.928). Journal for ImmunoTherapy of Cancer 2026 (abstract): 3,609 matched pairs, mortality hazard ratio 0.86 (0.80 to 0.93), immune-related adverse events 1.22 (1.17 to 1.28), severe 1.11, ICU admission 0.43. accessible: true sources: [src-grippin-2025, src-natmed-2026, src-ejc-2026, src-jitc-2026] basis: > Filed as `proposed`: observational, mostly one-institution or abstract-level, and subject to the healthy-vaccinee effects the French cohort names. It bears on the question only in one direction (no sign of harm in this group) and covers only patients on immunotherapy; it is not a test of recurrence or progression in other patients. - id: cvc-pandemic-diagnosis-disruption state: established statement_kind: reported statement: > In 2020 new cancer diagnoses fell sharply in many countries because of disrupted screening and care, and then returned toward expected levels; this complicates any reading of cancer counts from 2020 onward. evidence_class: synthetic confidence: high evidential_weight: 4 # of 5 anchor_checked: primary # the US report and the Japanese files read directly; the other series as abstracts adoption_weight: 3 # of 5 adoption_note: "Not disputed in any source read." would_change_if: "a later series shows the 2020 deficit was not recovered or was larger than recorded" anchor: type: registry-series-read-directly description: > US (Annual Report to the Nation 2025, read in full): 2020 incidence 425.6 against 461.3 for the other years, about 8 percent lower; earlier reports cited 2020 diagnoses "9%-10% lower than expected overall", with screenable sites (prostate rate ratio 0.83 to 0.86, thyroid about 0.81, lung 0.86 to 0.87) falling most; the authors write that "additional data years are needed to correctly interpret this decline and assess whether cases went undiagnosed or underreported, because the rates in 2021 were not greater than expected". Japan: registry incidence 282.1 (2019) to 263.6 (2020), 275.7 in 2021, still below 2019 in 2022 and 2023; Abstract-level: US Bayesian model 6.7 percent fewer cases than expected in 2020 to 2022 and "substantial numbers of unaccounted-for cases remained"; seven-country benchmarking, 16 percent deficit April to December 2020; England prostate down 31 percent in 2020 and back to expected by 2022; Nordic first-wave declines of 7.9 to 21.7 percent; Hungary 12.8 percent lower in men. accessible: true sources: [src-arn-2025, src-ncc-incidence-2023, src-seer-cross-2025, src-icbp-2026, src-opensafely-prostate-2024, src-breast-england-2023, src-nordic-2025, src-hungary-2026] basis: > It bears on the claim two ways: a later catch-up could produce later-stage diagnoses for reasons unrelated to vaccination, and a rise in late-stage cases is what both sides would predict. A "rise after 2021" cannot be read from counts without this. - id: cvc-military-nhl-counts state: established statement_kind: reported statement: > A US military surveillance report counted incident non-Hodgkin lymphoma in active-duty service members as 40, 34, 30, 29, 30, 38 and 43 cases in 2017 to 2023 (rates 3.0, 2.5, 2.2, 2.1, 2.2, 2.8 and 3.3), mentions no vaccination and attributes nothing to it; a 2026 review cites it as a "~50% increase" (2.2 in 2021 to 3.3 in 2023), while 2023 is 10 percent above 2017. evidence_class: synthetic confidence: high evidential_weight: 3 # of 5 anchor_checked: primary adoption_weight: 2 # of 5 adoption_note: "Cited in the scoping review as a population-level signal." would_change_if: "the report is revised or a later report attributes the change to a cause" anchor: type: the-document-itself description: > Non-Hodgkin lymphoma incidence in active component U.S. service members, MSMR 2025 (PMID 40019944), Defense Medical Surveillance System, 621 incident cases 2017-2023; "Specified and unspecified NHL" counts and rates as above; mature T/NK counts 9, 13, 9, 8, 17, 14, 14. The report notes a "modest increase". The units are inconsistent inside the document (table footnote per 10,000 person-years, text per 100,000). Counts are in the tens. The 10 percent figure is Strata's own arithmetic from 3.0 and 3.3. accessible: true sources: [src-msmr-nhl-2025, src-kuperwasser-2026] basis: > A surveillance report, not a peer-reviewed study. Whether 2022 and 2023 counts are higher than chance is not tested in it. The comparison base matters: against 2021 the rise is large; against 2017 it is small. - id: cvc-gibo-paper-retracted state: established statement_kind: reported statement: > A 2024 Cureus paper by Gibo et al. reporting excess age-adjusted cancer mortality in Japan after the third mRNA dose (2.1 percent above an extrapolated line in 2022) was retracted on 26 June 2024 after an expression of concern on 12 June, with the editors writing that the correlation between mortality rates and vaccination status cannot be proven with the data presented; the authors disagree, and the paper is still cited as support. evidence_class: primary_text confidence: high evidential_weight: 5 # of 5, for the retraction as a fact anchor_checked: primary # retraction notice (Europe PMC copy) and the paper read; the expression of concern as a PubMed record only adoption_weight: 2 # of 5 adoption_note: > Still cited as support in Gentilini et al. (2026, which acknowledges the retraction) and in a JMA Journal opinion paper (Kakeya et al. 2025), which Kutsuna and Suzuki say cites a book reflecting the same arguments as the retracted paper. Adoption is information, not evidence. would_change_if: "the journal reverses the retraction" flagged_sources: [src-gibo-2024] # retracted; shown beside this claim (rule PS3) anchor: type: retraction-notice-and-the-paper description: > Retraction notice, Cureus 2024 (doi 10.7759/cureus.r143), read in the Europe PMC copy: "The Editors-in-Chief have retracted this article. Upon post-publication review, it has been determined that the correlation between mortality rates and vaccination status cannot be proven with the data presented in this article. As this invalidates the conclusions of the article, the decision has been made to retract. The authors disagree with this retraction." The paper (doi 10.7759/cureus.57860): descriptive ecological time-series of Japanese vital statistics, logistic extrapolation of 2010-2019 (10 points) to predict 2020-2022, no individual vaccination status (vaccination coverage appears only as a monthly series), 20 cancer types tested with no multiplicity correction described; excess for all cancers -0.4 (2020), 1.1 (2021) and 2.1 (2022), 7,162 excess deaths in 2022 (95% CI 4,786 to 9,522); observed all-cancer age-adjusted rates 275.5, 275.8 and 274.6 per 100,000, so the observed rate did not rise in 2022. Its own limits: "has not been clinically validated"; further analytical statistics by vaccination status are needed. The paper also notes that 2020 screening fell. Cureus pages returned 403 here; the expression-of-concern text was not read. accessible: true sources: [src-gibo-2024, src-gibo-retraction-2024, src-gibo-eoc-2024, src-ncc-mortality-2024] basis: > A retraction removes a paper from the citable record; it does not show the opposite conclusion. Another retraction touches the wider claim: a 2024 Cureus review (Mead et al.) that listed cancer among serious adverse events was retracted on 26 Feb 2024 for unreliable conclusions and misrepresented references (read in c2 notes; not in this dig's manifest, log L-08). # ===================================================================== # C. MECHANISM AND LABEL # ===================================================================== - id: cvc-mechanism-spike-p53 state: established statement_kind: reported statement: > Zhang and El-Deiry (Oncotarget, May 2024) report that in two human cell lines lacking p53 (U2OS p53-knockout and HCT116 p53-null) given wild-type p53, spike protein expressed from a plasmid reduced MDM2-p53 co-precipitation and p53 reporter activity and blunted the rise of p21, DR5 and MDM2 after cisplatin or nutlin; the authors did not detect spike bound to p53, saw no cell-cycle arrest, and call the observations preliminary. evidence_class: primary_text confidence: moderate evidential_weight: 3 # of 5, for what the experiment reports anchor_checked: primary adoption_weight: 3 # of 5 adoption_note: > Presented as a mechanism in Senate testimony (El-Deiry, 3 Jun 2026) and in a 2026 hypothesis paper by the same author; El-Deiry says his work has been attacked on PubPeer and that "we corrected minor errors where appropriate" (not verified). Adoption is information, not evidence. would_change_if: > the effect is replicated in vaccinated animals or in tumour tissue from vaccinated patients (it would strengthen), or fails to replicate in an independent laboratory (it would weaken) anchor: type: the-document-itself description: > Zhang S, El-Deiry WS, Oncotarget 2024 (PMID 38709242; received 16 Apr, accepted 30 Apr 2024). Plasmid pcDNA3.1-SARS2-spike transfected into cancer cell lines with co-transfected wild-type p53; effects on co-immunoprecipitation, a PG13-luciferase reporter and downstream proteins; cisplatin-treated spike cells had higher viability. The paper frames its finding as a possible mechanism for SARS-CoV-2 infection and chemotherapy sensitivity, not for vaccination. Authors: "did not detect SARS-CoV-2 spike bound with p53", "preliminary observations ... warrant further studies". Competing interest in the paper: the corresponding author is cofounder of Oncoceutics, p53-Therapeutics and SMURF-Therapeutics. A 2026 hypothesis paper by El-Deiry (PMID 41498241) adds an anecdote of a recurrence after vaccination relayed from a third party. accessible: true sources: [src-zhang-2024, src-eldeiry-hypothesis-2026, src-psi-el-deiry] basis: > Records the observation. What it does not show: no vaccine-delivered spike, no animal or human tumour data, no transformation or oncogenesis assay, and the effect is in cancer cell lines under overexpression. Whether a vaccine can cause cancer by this route is cvc-mechanism-other-hypotheses (proposed). - id: cvc-mechanism-igg4-tumour-link state: proposed statement_kind: judgment statement: > The shift toward IgG4 antibodies after repeated mRNA doses weakens the body's antibody defence against tumours. evidence_class: interpretive confidence: provisional evidential_weight: 1 # of 5 anchor_checked: secondary # the serology papers searched in full text for tumour terms; the tumour link taken from other literature not opened adoption_weight: 2 # of 5 adoption_note: "Quoted in the retracted Japan paper, in Kakeya et al. and in Szebeni's review, and in the ACIP workgroup slides. Adoption is information, not evidence." would_change_if: > a study shows spike-specific IgG4 impairing anti-tumour antibody effects in vaccinated people, or higher cancer incidence in people with higher IgG4 after vaccination anchor: type: serology-without-tumour-data description: > Irrgang et al., Sci Immunol 2023: the IgG4 share of spike-specific IgG rose from 0.04 percent after dose 2 to 19.27 percent after dose 3; not seen after adenoviral vaccines; reduced phagocytosis and complement activity. Our text search of this paper and of the 2024 Immunity and Ageing paper for "cancer", "tumor", "tumour" and "malignan" found no hits. Cell Reports 2025: spike IgG4 is "only inhibitory when directly competing with functional IgG subclasses" and cytotoxicity in plasma "was not depleted by adding a cocktail of Spike-specific IgG4 monoclonal antibodies". The link from IgG4 to tumours is borrowed from other literature (one mouse study cited by Gibo, not opened). accessible: true sources: [src-irrgang-2023, src-cellrep-igg4-2025, src-immunageing-2024] basis: > The IgG4 rise itself is a reported serology finding; the claim is the step from it to cancer, which no source read tests. - id: cvc-mechanism-interferon-suppression state: proposed statement_kind: judgment statement: > mRNA vaccination suppresses type I interferon signalling enough to impair cancer surveillance. evidence_class: interpretive confidence: provisional evidential_weight: 1 # of 5 anchor_checked: secondary adoption_weight: 2 # of 5 adoption_note: "Quoted in the retracted Japan paper and by Isidoro. Adoption is information, not evidence." would_change_if: "measurements in vaccinated people show a lasting fall in interferon responses linked to tumour outcomes" anchor: type: narrative-review description: > Seneff et al., Food and Chemical Toxicology 2022;164:113008: a hypothesis and narrative review using VAERS counts and the literature, no tumour data. A 2023 letter (Barriere et al., FCT 178:113897) lists "major misunderstandings of the literature" and says VAERS was misused; a corrigendum to that letter was published 25 Sep 2026 (not read). Grippin et al. (Nature 2025) report the opposite direction, "a substantial increase in type I interferon" after COVID mRNA vaccines in mice and humans. accessible: true sources: [src-seneff-2022, src-barriere-2023, src-grippin-2025] basis: > A hypothesis paper; the one direct measurement we read points the other way. - id: cvc-mechanism-other-hypotheses state: proposed statement_kind: judgment statement: > COVID-19 vaccination causes cancer through one of several other proposed routes: spread of vaccine mRNA or spike to bone marrow with lymphopenia, microRNA in exosomes, spike binding oestrogen receptor alpha or BRCA and p53 proteins, integration of residual vaccine DNA, or promotion of dormant disease in a "multi-hit" model. evidence_class: interpretive confidence: provisional evidential_weight: 1 # of 5 anchor_checked: secondary # reviews read for what they say; underlying papers not opened adoption_weight: 3 # of 5 adoption_note: "Raised in Senate testimony and the Florida correspondence (residual DNA). Adoption is information, not evidence." would_change_if: > integration of vaccine DNA into human cells, or tumour formation by it, is shown in vaccinated people or animals with validated assays anchor: type: reviews-and-hypotheses description: > Narrative reviews, in silico docking, in vitro work and case reports. Isidoro, Cancers 2025: "it appears extremely unlikely that SARS-CoV-2 and anti-COVID-19 mRNA vaccines elicit genotoxic events and cause neo-cancerogenesis in a short time", proposing non-genotoxic effects and writing "While a causal link cannot be established at this stage". Szebeni, Pharmaceutics 2025 (Oct): DNA integration evidence "is heavily argued in the cancer field". Valdes Angues and Perea Bustos, Cureus 2023, is a review; its conclusion urges caution for patients with cancer. EMA: no evidence linking residual plasmid DNA to side effects, and high-DNA claims "are based on results from unvalidated tests used by third parties". FDA (Marks): residual DNA fragments reaching the nucleus and chromosomes is "quite implausible". Underlying residual-DNA papers were not opened (cvc-sq-dna-contamination). accessible: true sources: [src-szebeni-2025-oct, src-isidoro-2025, src-valdes-angues-2023, src-ema-key-facts, src-fda-marks-2023] basis: > None was tested in vaccinated tumour-bearing people. Regulators and some of the reviews' own authors rate the genotoxic routes unlikely; that is expert opinion, not a measurement. - id: cvc-fda-label-not-evaluated state: established statement_kind: reported statement: > The FDA prescribing information for COMIRNATY (revised August 2026), section 13.1, says: "COMIRNATY has not been evaluated for the potential to cause carcinogenicity, genotoxicity, or impairment of male fertility." evidence_class: primary_text confidence: high evidential_weight: 5 # of 5, for what the label says anchor_checked: primary adoption_weight: 2 # of 5 adoption_note: "Cited by people on the claim's side as an admission. Adoption is information, not evidence." would_change_if: "the label is revised to report such an evaluation" anchor: type: the-document-itself description: > Section 13.1 of the prescribing information, read in the FDA copy (https://www.fda.gov/media/151707/download, formula 2026-2027, revised 8/2026); excerpt saved in sources/. The sentence says "not evaluated". accessible: true sources: [src-comirnaty-pi] basis: > What it is: a statement that a non-clinical test was not run. What it is not: a finding that the vaccine causes cancer, and not a finding that it does not. We did not check whether other products' labels carry the same sentence. # ===================================================================== # D. ADOPTION: WHAT INSTITUTIONS AND PEOPLE SAID # ===================================================================== - id: cvc-health-bodies-no-evidence state: established statement_kind: reported statement: > The US National Cancer Institute (page updated 10 Oct 2023), the European Medicines Agency (page dated 8 Apr 2026), the American Cancer Society (2024 version of its page) and ASCO (written testimony, 3 Jun 2026) each state that there is no evidence, information or clinical evidence that COVID-19 vaccines cause cancer. evidence_class: primary_text confidence: high evidential_weight: 4 # of 5, for what the statements say anchor_checked: primary adoption_weight: 4 # of 5 adoption_note: > The NCI sentence was read into the Senate record by Sen. Blumenthal on 3 Jun 2026. Statements are adoption: they say the evidence at a date and confer no weight on either side. would_change_if: "an agency page is revised or withdrawn" anchor: type: the-documents-themselves description: > NCI: "There is no evidence that COVID-19 vaccines cause cancer, lead to recurrence, or lead to disease progression. Furthermore, COVID-19 vaccines do not change your DNA". EMA: "There is no evidence that COVID-19 vaccines may cause side effects, such as cancer, in the long term" and "No side effects linked to gene mutations, such as cancer, have been observed after vaccination with mRNA vaccines". ACS (Last Revised 9 Sep 2024, Wayback): "There is no information that suggests that COVID-19 vaccines cause cancer." ASCO (Gralow): "Currently, there is no clinical evidence proving that mRNA COVID-19 vaccines cause cancer". The live ACS page (revised 24 Sep 2025) no longer has that section; we do not know why (cvc-sq-acs-wording-change). We found no WHO, current CDC or UK statement on the question (cvc-sq-who-cdc-uk-statements). accessible: true sources: [src-nci-2023, src-ema-key-facts, src-acs-2024, src-acs-2025, src-psi-gralow, src-psi-blumenthal] positions: - label: "No evidence of a link" holder: "NCI, EMA, ACS (2024), ASCO" ground: "None found in their review of the literature at the dates given." - label: "A signal needs study" holder: "Kuperwasser and El-Deiry (ACIP slides, Senate testimony)" ground: "Case reports and mechanism papers; they call it an early phase of potential safety-signal detection." basis: > "No evidence" is a statement about the state of the evidence at the page's date, not a proof of absence, and these bodies did not run a new analysis. The statements are filed as adoption. - id: cvc-fda-marks-letter state: established statement_kind: reported statement: > On 14 December 2023 the FDA's CBER director Peter Marks replied to Florida Surgeon General Ladapo that "with over a billion doses of the mRNA vaccines administered, no safety concerns related to residual DNA have been identified" and that "there is nothing to indicate harm to the genome, such as increased rates of cancers". evidence_class: primary_text confidence: high evidential_weight: 4 # of 5, for what the letter says anchor_checked: primary adoption_weight: 3 # of 5 adoption_note: "A regulator's position statement. Adoption is information, not evidence." would_change_if: "the letter is withdrawn or the surveillance data it cites are shown not to support it" anchor: type: the-document-itself description: > Marks to Ladapo, 14 Dec 2023 (read in the FDA PDF; text saved in sources/): residual DNA fragments reaching the nucleus and chromosomes is "quite implausible"; the 2007 plasmid DNA vaccine guidance "was developed for DNA vaccines themselves, not for DNA as a contaminant in other vaccines, and is not applicable to the mRNA COVID-19 vaccines". Points to global surveillance data and animal studies (FDA documents 151733 and 155931, not opened). accessible: true sources: [src-fda-marks-2023, src-fda-plasmid-guidance-2007] basis: > What it is: an agency letter answering a state official, stating the FDA's position and manufacturing facts. What it is not: a peer-reviewed study or a new analysis of cancer rates; "no increased rates" is surveillance, which cannot see a small effect. - id: cvc-florida-surgeon-general state: established statement_kind: reported statement: > On 6 December 2023 Florida Surgeon General Ladapo wrote to the FDA that "DNA integration could theoretically impact a human's oncogenes", quoting the 2007 guidance, and on 3 January 2024 his department said that if DNA integration risks "have not been assessed", mRNA vaccines "are not appropriate for use in human beings"; neither document reports a cancer case or uses the phrase "turbo cancer". evidence_class: primary_text confidence: high evidential_weight: 4 # of 5, for what the documents say anchor_checked: primary adoption_weight: 3 # of 5 adoption_note: "A state official's statement. CBS quoted an outside physician linking it to the 'turbo cancer' idea; that is CBS's and the physician's characterisation, not Ladapo's wording." would_change_if: "the department publishes data on cancer in vaccinated people" anchor: type: the-documents-themselves description: > Letter of 6 Dec 2023 to FDA Commissioner Califf (Wayback copy; floridahealth.gov returned 403): "The presence of SV40 promoter/enhancer DNA may also pose a unique and heightened risk of DNA integration into host cells"; three questions about integration and mutagenesis. Release of 3 Jan 2024 (Wayback copy): "DNA integration poses a unique and elevated risk to human health and to the integrity of the human genome"; providers should "prioritize patient access to non-mRNA COVID-19 vaccines and treatment". The word "cancerous" appears only in the quoted 2007 guidance. accessible: true sources: [src-fl-ladapo-letter-2023, src-fl-halt-release-2024, src-cbs-ladapo-2024] basis: > What it is: questions and a recommendation to providers, framed as a risk not yet assessed. What it is not: a ban or order, a finding that cancer has occurred, or new data. - id: cvc-acip-slides state: established statement_kind: reported statement: > At the CDC's ACIP meeting of 18-19 September 2025, two workgroup participants, Wafik El-Deiry and Charlotte Kuperwasser, presented slides saying "Cancers have been reported in mRNA vaccinated individuals in temporal association to immunization" and listing "Genomic integration in tissues or tumors in vaccinated patients" among gaps in knowledge; the slides do not use the phrase "turbo cancer". evidence_class: primary_text confidence: moderate evidential_weight: 4 # of 5, for what the slides say anchor_checked: primary adoption_weight: 3 # of 5 adoption_note: "FactCheck.org (7 Nov 2025, secondary) says experts judged the DNA-limit and IgG4 claims flawed, and that it was the first time such claims were given legitimacy at an ACIP meeting." would_change_if: "the slides are shown to differ from the Wayback copy read" anchor: type: the-document-itself description: > Workgroup slides dated 19 Sep 2025, read in a Wayback copy (cdc.gov blocked here), PDF created 18 Sep 2025: lists case reports (38 case reports and a study of 96 pancreatic cancer outcomes against IgG4) and states residual DNA "Exceeds limits by ~36-153-fold" for the Pfizer vaccine. The slides were not stored in this dig's sources folder (rights not checked), so quotations are marked for manual check. accessible: true sources: [src-acip-slides-2025, src-factcheck-acip-2025] basis: > What it is: a presentation at a public advisory meeting by two workgroup participants (neither a voting member, per FactCheck.org). What it is not: an ACIP vote, a CDC finding or a causal conclusion; the slides frame the cancer items as reported in temporal association. - id: cvc-senate-hearing state: established statement_kind: reported statement: > On 3 June 2026 the Senate Permanent Subcommittee on Investigations, chaired by Sen. Ron Johnson, held a hearing titled "Plausible Mechanisms of COVID-19 Injections Causing Cancer and Attacks on Scientific Publications and Research"; the written statements read contain no new cancer data and none uses the phrase "turbo cancer". evidence_class: primary_text confidence: moderate evidential_weight: 4 # of 5, for what was said anchor_checked: primary # written statements only; oral testimony not read adoption_weight: 3 # of 5 adoption_note: "Majority-selected witnesses; a hearing makes no scientific finding." would_change_if: "the oral testimony or a committee report contains data not in the written statements" anchor: type: the-documents-themselves description: > Witnesses: Dalgleish, El-Deiry, Hazan, Mostert, Malhotra, Gralow (ASCO), Felder. Johnson's opening is about media and pharma influence and "attacks on scientific publications"; Dalgleish (written): "I have no doubt in my mind that the mRNA vaccine likely played a significant role in the development of these unexpected cancers", from clinical observations without data tables; El-Deiry: the literature review and p53 experiments, and a PubPeer complaint; Gralow: no clinical evidence proving a link; Blumenthal: "the idea that a rollout of COVID-19 vaccines has precipitated a surge in cancer is simply not supported". Johnson's release of 5 Jun 2026 says the hearing "exposed the corruption" and cites "over 39,000 deaths" associated with the injection (source not given). Oral testimony and Q&A not read (no transcript found). accessible: true sources: [src-psi-hearing-page, src-psi-johnson-opening, src-psi-blumenthal, src-psi-dalgleish, src-psi-el-deiry, src-psi-gralow, src-psi-other-statements, src-johnson-release-2026] basis: > What it is: testimony before a subcommittee. What it is not: an agency action, a regulatory finding or a report with data. The 39,000 figure is asserted and not sourced in what we read. - id: cvc-term-first-use state: established statement_kind: reported statement: > The earliest document we read that uses "turbo cancer" about COVID-19 vaccines is Etana Hecht's Substack post "Turbo-Cancer" of 4 August 2022, a summary of an interview with the Swedish pathologist Ute Kruger; two earlier write-ups of the same interview (27 and 28 July 2022) say "aggressive and unusual cancers" and not "turbo". evidence_class: primary_text confidence: moderate evidential_weight: 3 # of 5 anchor_checked: primary # via Wayback copies adoption_weight: 2 # of 5 adoption_note: "Republished by Global Research on 8 Aug 2022. Reach and republication are adoption." would_change_if: "an earlier document using the phrase about COVID-19 vaccines is found (cvc-sq-term-origin-earlier)" anchor: type: the-documents-themselves description: > Hecht (Wayback copy of etana.substack.com/p/turbo-cancer): "Dr. Kruger initially thought that these turbo cancers, as she calls them"; she writes that she "hate[s] the title" and that Kruger's testimony is "by nature anecdotal". Global Research republication, metadata 8 Aug 2022. RAIR Foundation (27 Jul 2022) and myMedicalFreedom (28 Jul 2022) text-searched: no "turbo". Later amplification: DailyClout Report 61 (10 Mar 2023) titled with "Turbo Cancers". Secondary: Gorski (SBM, 19 Dec 2022) says he could not find who coined the term and cites an ar15.com comment of 30 Nov 2020 that we could not open. Whether Kruger herself used the term in the interview video is not verified (video not watched). accessible: true sources: [src-hecht-turbo-2022, src-globalresearch-2022, src-rair-kruger-2022, src-mymedfreedom-kruger-2022, src-dailyclout-61, src-sbm-turbo-2022] basis: > "Earliest we read" is a statement about our search, not a finding of who coined the term. It is the origin of the word, not evidence for the claim. # ===================================================================== # E. SEARCHED GAPS: SOURCES NOT READ OR NOT FOUND # None carries evidential weight for or against the question. # ===================================================================== - id: cvc-sq-term-origin-earlier state: searched_gap statement_kind: search_result statement: > Sub-question: was "turbo cancer" used about COVID-19 vaccines before 4 August 2022, as in an ar15.com comment of 30 November 2020 that Gorski cites? absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 1 # of 5, for the existence of the gap anchor_checked: "no" adoption_weight: 1 # of 5 adoption_note: "Not measured." would_change_if: "the comment is opened and says what Gorski reports, or an earlier use is found" next_step: > Someone with access to ar15.com (403 to us, also via Wayback) can open the November 2020 thread and confirm the comment's date and wording. anchor: type: search-record description: > Gorski (secondary) gives the date and a quotation; we could not open the page. We found nothing earlier in a primary document. accessible: false sources: [src-sbm-turbo-2022] - id: cvc-sq-kruger-interview state: searched_gap statement_kind: search_result statement: > Sub-question: did Ute Kruger herself say "turbo cancers" in the Doctors for Covid Ethics interview of July 2022, and what did she report? absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 1 # of 5 anchor_checked: "no" adoption_weight: 1 # of 5 adoption_note: "Not measured." would_change_if: "the interview or her 18 Sep 2022 symposium talk (on archive.org) shows she used or did not use the term" next_step: > Watch the interview and the symposium talk (archive.org item dr.-ute-kruger-covid-19-vaccination-observations-of-a-pathologist-2nd-medical-sy) and log what she said, with times. anchor: type: search-record description: "Only Hecht's summary and two earlier write-ups were read; the video was not watched." accessible: true sources: [src-hecht-turbo-2022, src-rair-kruger-2022] - id: cvc-sq-died-suddenly-term state: searched_gap statement_kind: search_result statement: > Sub-question: does the 2022 film "Died Suddenly" use the phrase "turbo cancer", and what does it say about cancer? absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 1 # of 5 anchor_checked: secondary adoption_weight: 1 # of 5 adoption_note: "Not measured." would_change_if: "a human viewing finds the phrase or a different account of the cancer segments" next_step: > A person watches 22:00 to 27:00 of the 68-minute video and transcribes the cancer segments and speakers. Our automated transcript (base.en, error-prone) has no "turbo" and has cancer statements at about 22:08 and 25:34 (a military-doctor segment citing a "300% increase"). anchor: type: search-record description: > Automated speech-to-text of a downloaded copy; the film is a venue for a general vaccine-cancer claim via DMED data, not a documented origin of the term. Lead Stories (secondary) reviewed the film for sudden death. accessible: true sources: [src-died-suddenly-asr, src-died-suddenly-film, src-leadstories-2022] - id: cvc-sq-other-cohorts state: searched_gap statement_kind: search_result statement: > No cohort of cancer after COVID-19 vaccination from Kaiser or the Vaccine Safety Datalink, Israel, the Netherlands, the Nordic countries or the UK was found, beyond the Seoul and Pescara studies. absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 1 # of 5 anchor_checked: secondary adoption_weight: 1 # of 5 adoption_note: "Not measured." would_change_if: "such a cohort is found or published" next_step: > Query the Vaccine Safety Datalink, Clalit, Danish and Swedish registers for cancer incidence by vaccination status with a lag period, and publish the analysis plan first. anchor: type: search-record description: > PubMed and Europe PMC queries listed in cvc-rapid-progression-gap, 5 Oct 2026. The only Vaccine Safety Datalink hit was a mortality study in which cancer is an exclusion criterion. Absence from our searches is not proof none exists. accessible: true sources: [src-vsd-mortality-2024] - id: cvc-sq-rct-malignancy-data state: searched_gap statement_kind: search_result statement: > The main texts of the two Pfizer-BioNTech randomised trial reports report no cancer outcome, and their supplementary adverse-event tables were not opened, so malignancy counts by arm are unread. absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 1 # of 5 anchor_checked: secondary adoption_weight: 1 # of 5 adoption_note: "Not measured." would_change_if: "the supplementary tables or the FDA review show malignancy counts by arm" next_step: > Read the supplementary adverse-event tables and the FDA clinical review for malignancy counts by arm; note the follow-up length (about 6 months) and when the placebo group was offered vaccine, which we did not verify. anchor: type: search-record description: "Text search of the open full text of Polack 2020 and Thomas 2021 for malignan, cancer, neoplasm and tumor: no cancer outcome in the main text." accessible: true sources: [src-polack-2020, src-thomas-2021] - id: cvc-sq-semenzato-cancer-mortality state: searched_gap statement_kind: search_result statement: > Does the French cohort of 28.7 million adults aged 18 to 59 (Semenzato et al., JAMA Network Open, Dec 2025) report lower cancer mortality in vaccinated people, as Sen. Blumenthal said? absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 1 # of 5 anchor_checked: "no" adoption_weight: 1 # of 5 adoption_note: "Blumenthal described it in his opening statement; unverified by us." would_change_if: "the paper's cause-specific results are read" next_step: > Read the full paper and its supplement for a cancer-specific mortality result (hazard ratio for all-cause death 0.75 over 4 years is in the abstract; the abstract says "regardless of the cause"). anchor: type: search-record description: "Abstract read; no copy retained, so the manifest entry has no sha256." accessible: true sources: [src-semenzato-2025, src-psi-blumenthal] - id: cvc-sq-who-cdc-uk-statements state: searched_gap statement_kind: search_result statement: > We found no WHO statement, no current CDC page and no UK (NHS, UKHSA, MHRA) statement that addresses whether COVID-19 vaccines cause cancer. absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 1 # of 5 anchor_checked: secondary adoption_weight: 1 # of 5 adoption_note: "Not measured." would_change_if: "such a statement is found" next_step: > Read current cdc.gov vaccine safety pages (403 to us), UKHSA and MHRA pages, Health Canada and the TGA. Absence of a statement is not an endorsement either way. anchor: type: search-record description: > WHO vaccine Q&A and the 49th GACVS report (March 2026) text-searched for cancer, turbo, malignan, residual DNA and SV40: no hits on causation. Archived CDC pages of 2023 contain no cancer statement. accessible: true sources: [src-who-qa, src-who-gacvs49, src-cdc-archive-myths, src-cdc-archive-safety, src-cdc-archive-adverse] - id: cvc-sq-england-registry state: searched_gap statement_kind: search_result statement: > England's national cancer registry figures for 2021 and 2022 (NHS England) were not read. absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 1 # of 5 anchor_checked: "no" adoption_weight: 1 # of 5 adoption_note: "Not measured." would_change_if: "the figures show a post-2021 all-cancer rise" next_step: "Open the NHS England cancer registration statistics for 2021 and 2022 (403 to our proxy) and the Danish NORDCAN series." anchor: type: search-record description: "digital.nhs.uk returned 403; the ONS 2023 deaths bulletin has no all-cancer age-standardised rate. England prostate and breast series were read as abstracts." accessible: false sources: [src-opensafely-prostate-2024, src-breast-england-2023] - id: cvc-sq-dna-contamination state: searched_gap statement_kind: search_result statement: > The residual-DNA papers behind the FDA and EMA statements and behind the 36 to 153-fold claim were not opened; this dig scopes the DNA-contamination question out. absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 1 # of 5 anchor_checked: "no" adoption_weight: 2 # of 5 adoption_note: "Raised by Florida and the ACIP slides; see those claims." would_change_if: "the underlying papers contradict how the agencies and reviews describe them" next_step: > Read the papers EMA cites (Kaiser et al., Vaccine 2025, PMID 40120438; Achs et al., npj Vaccines 2025) and the residual-DNA papers behind the slide claim (Kammerer et al.), as a separate dig. anchor: type: search-record description: "Statements read; underlying measurements not opened." accessible: true sources: [src-ema-key-facts, src-fda-marks-2023, src-acip-slides-2025] - id: cvc-sq-japan-mortality-rerun state: searched_gap statement_kind: search_result statement: > We did not re-run the retracted Japan paper's regression against current national data; this is scoped out. absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 1 # of 5 anchor_checked: secondary adoption_weight: 1 # of 5 adoption_note: "Not measured." would_change_if: "a re-run shows excess or none against a pre-specified method" next_step: "Fit the 2010-2019 trend to the National Cancer Center series and test 2020-2024, with prediction intervals and a correction for the 2020 screening drop." anchor: type: search-record description: "Only the paper's reported figures and the national series were compared." accessible: true sources: [src-ncc-mortality-2024] flagged_sources: [src-gibo-2024] - id: cvc-sq-korea-concern-outcome state: searched_gap statement_kind: search_result statement: > What editorial action will Biomarker Research take on the Seoul letter after its notice of concern of 22 October 2025? absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 1 # of 5 anchor_checked: secondary adoption_weight: 1 # of 5 adoption_note: "Not measured." would_change_if: "the journal corrects, retracts or clears the letter" next_step: "Re-read the article page and Crossref for an update; ask the editors." flagged_sources: [src-kim-2025] anchor: type: search-record description: "Article page and PMC change history read 5 Oct 2026: notice present, no outcome; Crossref shows no update-to relation." accessible: true sources: [src-kim-2025-page, src-kim-2025] - id: cvc-sq-acs-wording-change state: searched_gap statement_kind: search_result statement: > Why does the American Cancer Society page, revised 24 September 2025, no longer carry the 2024 statement that nothing suggests COVID-19 vaccines cause cancer? absence_anchor: true evidence_class: primary_text confidence: provisional evidential_weight: 1 # of 5 anchor_checked: secondary adoption_weight: 1 # of 5 adoption_note: "Not measured." would_change_if: "ACS says the removal was editorial or a change of position" next_step: "Ask ACS; compare the Wayback versions between September 2024 and September 2025." anchor: type: search-record description: "Text difference between the 2024 Wayback version and the live page noted; the ACS mRNA vaccines page (revised 29 Aug 2025) makes no causation statement." accessible: true sources: [src-acs-2024, src-acs-2025]