Preamble and Research Purpose

Critical Methodological Constraints

These constraints apply to every domain without exception.

• All-cause mortality is the only acceptable primary endpoint for any claim that a diagnostic intervention “saves lives.” Disease-specific mortality relies on cause-of-death coding, which is frequently unvalidated by autopsy, subject to attribution bias, and inadmissible as a hard endpoint.
• The appropriate comparator for contrast-enhanced MRI is not “no imaging.” It is non-contrast MRI of equivalent technical quality and optimised sequences, or the best available alternative diagnostic modality.
• Any study claiming GBCA administration changes the clinical diagnosis must demonstrate the finding was genuinely undetectable on non-contrast sequences — not merely less conspicuous. Studies not meeting this standard should be identified and set aside.
• A change in image appearance is not a change in diagnosis. A change in diagnosis is not a change in management. A change in management is not a change in outcome. Each step requires independent evidence.
• The harm side of any cost-benefit calculation must be assessed on the same evidential standard as the benefit side.
• The conclusion that macrocyclic GBCAs are safer than linear agents rests on evidence of lower tissue retention as measured by radiological signal change — a surrogate endpoint. Do not treat macrocyclic-is-safer as established unless supported by clinical outcome data beyond retention metrics.
• Any study assessing adverse reactions to GBCAs must be evaluated for follow-up duration and method. Studies capturing only in-scanner or immediate post-procedure reactions are inadequate. Adverse reactions can manifest hours, days, and weeks after the patient has left the imaging department. A study without several weeks of active surveillance cannot be considered a complete assessment of the adverse reaction profile.
• Any observational study in a GBCA-exposed population must be evaluated for survivorship bias. Studies relying on clinic attendance or voluntary enrolment will systematically underrepresent patients who have died, become too ill to participate, or disengaged from healthcare. Report whether all-cause mortality and dropout rates with reasons are recorded.
• For any condition with a known or suspected latency or prodromal phase — including neurodegenerative diseases, autoimmune conditions, fibrotic disease, and chronic inflammatory conditions — a study of inadequate duration is not merely underpowered. It is operating on the wrong timescale and cannot detect an association even if one exists. Such studies must not be cited as evidence of absence. For each, compare follow-up duration to the estimated latency period.
• When a putative confounder lies on the hypothesised causal pathway between exposure and outcome, correcting for it is a methodological error. If GBCA exposure leads to psychiatric symptoms which leads to psychotropic prescribing, adjusting for psychotropic use removes part of the mechanism under study. Identify any such studies and treat their conclusions with caution.
• For every study cited, report the funding source and declared author conflicts of interest. Industry-funded pharmaceutical safety studies have a documented pattern of producing more favourable conclusions than independently funded studies. Where funding is not declared, note the absence.
• Evaluate whether there is evidence of publication bias in the GBCA literature. Are there indications — through registered trial registries, conference abstracts, or meta-analytic funnel plot asymmetry — that negative studies or studies showing harm have been selectively unpublished? The shape of the published evidence is as important as its content.
• Statistical power and the Rule of Three: Any safety study must be evaluated for whether its sample size was adequate to detect adverse effects at clinically meaningful frequencies. The Rule of Three states that to exclude adverse effects occurring in 1% of subjects with 95% confidence, a study requires approximately 300 subjects per group; for 0.1%, approximately 3,000. A study with n=10–20 per group cannot exclude adverse effects at any standard ADR frequency threshold — including “very common” (≥10%), “common” (≥1%), or “uncommon” (≥0.1%). Such a study is mathematically incapable of detecting harm and cannot be cited as evidence of safety. Where no power calculation is reported, note this omission explicitly as a violation of standard regulatory and statistical guidance (ICH E9, ICH S5).
• The “explains away” pattern: When studies acknowledge positive findings but attribute them to confounders — such as evolving disease process, concurrent medications, surgical effects, inter-examiner variability, or spontaneous incidental occurrence — evaluate whether the explanatory claim was tested or merely asserted. An attributed explanation that has not been empirically tested is itself an assumption. Where a pattern exists of systematically classifying every concerning finding as “non-adverse,” “incidental,” or “not related” by investigators with conflicts of interest, identify this pattern explicitly. One-directional interpretive bias consistently in the direction of safety, exhibited across multiple findings by manufacturer-affiliated authors, is a substantive methodological problem, not a series of individual judgement calls.

Domain 1 — The “Saves Lives” Claim: Origin and Hard Endpoints ★ Priority

• Is there any high-quality RCT, prospective cohort study, or longitudinal data demonstrating that routine GBCA use directly reduces all-cause mortality compared to standardised, high-quality non-contrast MRI protocols?
• Is there a primary source for the “countless lives” claim, or an identifiable point in the literature where it appears without supporting references? Who said it first, in what publication, and what data did they cite? If unreferenced, trace the chain of subsequent citations to establish how the claim propagated.
• What conditions or clinical scenarios are most commonly cited as evidence that GBCAs save lives? List these, then evaluate each individually against the methodological constraints above.

Domain 2 — Claimed Indications: What Do Radiologists and Manufacturers Assert?

• What is the full list of diagnostic indications for which GBCA administration is recommended in current radiology society guidelines (ACR, ESR, RCR, and equivalent bodies)?
• What claims do GBCA manufacturers make in prescribing information and marketing materials?
• For each claimed indication, is the supporting evidence a demonstration that contrast changes the clinical diagnosis against an optimised non-contrast comparator, or a demonstration that contrast produces different image characteristics? These are not the same thing.
• Are there indications where the evidence was established when non-contrast MRI technology was substantially less capable than today, and has never been re-evaluated against modern protocols?

Domain 3 — Does Contrast Change the Diagnosis?

Part A: Contrast vs non-contrast MRI

• Is there robust comparative evidence that GBCA administration changes the clinical diagnosis — not merely image appearance — compared to non-contrast MRI of equivalent quality and sequence optimisation?
• Any study claiming contrast changes the diagnosis must demonstrate, with original non-contrast images reviewed by a blinded radiologist, that the finding was genuinely invisible on non-contrast sequences. Studies not meeting this standard should be identified explicitly.

Part B: Contrast MRI vs alternative modalities

• Where contrast MRI has been compared head-to-head against alternatives — PET-CT, CEUS, targeted biopsy, functional imaging, advanced non-contrast sequences (DWI, SWI, T1 mapping, T2 mapping, arterial spin labelling) — how does it perform?
• Are there clinical areas where contrast MRI demonstrably underperforms available alternatives?

Part C: Conditions where radiology creates false reassurance

• Identify clinical contexts where radiology — with or without contrast — is likely to return a normal result in the presence of genuine pathology, including: metabolic and nutritional deficiencies, mitochondrial dysfunction, toxic illness (including gadolinium toxicity), mould illness/CIRS, infectious conditions (SIBO, fungal overgrowth, cystitis, sepsis, tick-borne illness), dysautonomia, and autoimmune conditions without structural correlates.
• What is the clinical consequence when a normal contrast MRI is treated as authoritative and further investigation not pursued?

Domain 4 — The Outcome Chain: Detection, Treatability, and Benefit

The “saves lives” claim requires all three links to hold. Evaluate each independently:

• (a) Detection: Does GBCA administration detect clinically significant findings undetectable by any other investigation, including optimised non-contrast MRI?
• (b) Treatability: Where such findings are detected, is there an intervention with proven efficacy?
• (c) Outcome: Does intervening improve PROMs, quality of life, or all-cause survival compared to not intervening?

A positive answer to (a) does not imply (b), and (b) does not imply (c). All three must hold. Where studies conflate these steps, identify the conflation explicitly.

Domain 5 — Brain Tumour Detection: Is Contrast Actually Necessary?

• Is there evidence from blinded comparative studies that clinically significant brain tumours are undetectable on non-contrast MRI of adequate quality, and detectable only on contrast-enhanced MRI? Evaluate in both paediatric and adult populations.
• In symptomatic patients — presenting with headache, focal neurology, seizure, or raised intracranial pressure — what is the sensitivity of non-contrast MRI for detecting the mass lesion driving symptoms?
• Is there evidence from the historical transition away from contrast-enhanced CT in paediatric neuroimaging that diagnostic accuracy for brain tumours was compromised?

• Is there comparable evidence in adult oncology comparing contrast MRI, non-contrast MRI, and PET-CT, with specific attention to whether clinically significant cancers were missed without contrast? Where evidence exists only in one age group, explicitly note the evidence gap.

The dismissal pattern in brain tumour contrast research

• In longitudinal brain tumour imaging studies, have researchers attributed changes in imaging appearance to disease progression, surgery, radiation, or chemotherapy rather than to the presence or absence of contrast? Was the explanatory claim supported by independent analysis, or asserted as a default?

Domain 6 — Cardiac MRI, Myocarditis, and Cardiomyopathy

• What role does gadolinium-enhanced cardiac MRI (late gadolinium enhancement, LGE) actually play in diagnosis compared to standard workup: ECG, echocardiography, troponin, cardiac enzymes, inflammatory markers, and aetiological investigation?
• Do non-contrast parametric MRI sequences (T1 mapping, T2 mapping) — now incorporated in the updated Lake Louise Criteria — achieve comparable diagnostic accuracy to LGE-based protocols?
• Does gadolinium cardiac MRI change management or hard outcomes, or does it change image characterisation without changing clinical decisions?
• Is there evidence that paediatric cardiology manages myocarditis and cardiomyopathy effectively without gadolinium cardiac MRI?

Domain 7 — MR Angiography and Non-Contrast Alternatives

• What is the current evidence comparing non-contrast MRA techniques (time-of-flight, phase contrast, arterial spin labelling, quiescent-interval single-shot) to contrast-enhanced MRA across all major vascular territories: intracranial, cervical, thoracic aortic, cardiac, abdominal, renal, mesenteric, and peripheral?
• For each territory, for what specific indications is contrast-enhanced MRA demonstrably superior, and what is the magnitude of the clinical difference?
• Where non-contrast MRA is equivalent or near-equivalent, what are the barriers to wider adoption — and are these clinical, technical, or primarily organisational and financial?
• Has the comparison been made against modern non-contrast techniques, or against older protocols predating current hardware and sequence optimisation?

Domain 8 — False Positives: Incidentalomas, Misidentification, and Misattribution

Incidentalomas

• What is the incidence of incidentalomas detected on contrast-enhanced MRI compared to non-contrast MRI? What proportion lead to further investigation, procedural intervention, or surgery, and what proportion are ultimately benign?

Misidentification

• What is the rate at which contrast-enhanced MRI findings are misclassified — benign findings interpreted as malignant, or one pathology misidentified as another?
• Enhancement patterns (tumour, inflammation, infection, demyelination, vascular malformation) are not diagnostically specific. What is the false positive rate for specific common diagnoses such as MS lesion attribution or tumour versus inflammatory lesion?

Misattribution

• What is the rate at which a contrast-detected finding is real but coincidental to the patient's presenting symptoms, leading to clinical anchoring and failure to investigate the actual cause of illness?
• Evaluate the lumbar spine and back pain surgery literature as a case study in image-driven misattribution and overtreatment.

Domain 9 — False Negatives, Missed Diagnoses, and Medical Labelling

False negatives and clinical consequences

• Is there evidence on conditions that are missed or falsely reassured by contrast MRI, where the consequence is delayed or absent diagnosis of a treatable condition?
• Is there evidence on the extent to which clinicians weight radiological findings over other investigative modalities, and whether that weighting is clinically justified?

The medical labelling pipeline

• Is there evidence on clinical consequences for patients labelled as “difficult,” “heartsink,” “medically unexplained,” or “functional” following normal or non-diagnostic imaging — including rates of subsequent investigation, specialist referral, analgesic prescribing, and mortality from missed diagnosis?
• Is there evidence that a negative label in the clinical record durably influences subsequent clinicians across providers and settings?

Financial and systemic costs to patients

• What is the evidence on direct financial costs to patients who receive incorrect or misleading diagnoses, including follow-up imaging, unnecessary procedures, private investigation, and lost employment?
• What is the evidence on indirect costs: disability benefit applications, medico-legal expenses, and healthcare avoidance?

Domain 10 — Biopsy Rates

• Is there evidence that contrast MRI reduces downstream biopsy rates compared to non-contrast MRI or alternative investigations?
• Is there evidence that contrast MRI increases biopsy rates by detecting additional ambiguous enhancing lesions that cannot be definitively classified without tissue sampling? Enhancement is a physiological signal, not a pathological label.
• What proportion of biopsies triggered by contrast-detected findings return benign or clinically insignificant results?
• What are the complication rates and patient burden of biopsies triggered by contrast findings?

Domain 11 — Vulnerable Populations and Risk Factors Including EDS

Vulnerable populations

• What evidence exists on GBCA safety and diagnostic outcomes specifically in: children and adolescents; pregnant women; women of childbearing age; elderly patients; patients with renal impairment; patients with connective tissue disorders. Where evidence is absent, state this explicitly. Absence of safety data is not the same as evidence of safety.

EDS, hypermobility, and mast cell activation as risk factors

• Is there evidence that EDS, hypermobility syndromes, or MCAS are associated with higher rates of GBCA adverse reactions, including both acute and longer-term symptom development?
• Is EDS or hypermobility documented as a risk factor in any safety guidelines, manufacturer prescribing information, or pharmacovigilance literature?
• Are there biological mechanisms by which connective tissue disorders might plausibly increase GBCA adverse reaction risk, including altered vascular permeability, mast cell reactivity, or impaired xenobiotic clearance?

Domain 12 — GBCA Harm: Beyond Retention ★ Priority

The dominant research framing equates the threshold of radiological visibility with the threshold of biological harm. These are not the same. Research within this framing is structurally unable to detect harm through mechanisms that do not produce T1 signal change. The macrocyclic-is-safer conclusion rests almost entirely on this inadequate metric.

Documented toxicity mechanisms in the published literature

The following mechanisms of gadolinium and GBCA toxicity are documented in the peer-reviewed literature. For each mechanism, identify the primary supporting studies (verifying author details, journal, year, and DOI independently), and evaluate whether any GBCA safety study has tested for the relevant endpoint using an instrument capable of detecting it. A documented mechanism for which no adequate safety endpoint exists represents a specific and named evidence gap. Do not rely on pre-supplied citations for this section — locate and verify each source independently.

• Mitochondrial dysfunction: Published evidence documents Gd3+ inhibition of mitochondrial calcium-activated F1FO-ATPase, mitochondrial membrane potential loss, altered mitochondrial metabolites in persistently symptomatic GBCA-exposed individuals, and altered functional mitochondrial protein levels in neuron-derived extracellular vesicles of gadolinium deposition disease patients. Identify the primary papers supporting each of these claims, verify details, and assess whether any GBCA regulatory safety study included endpoints capable of detecting mitochondrial dysfunction.
• Endoplasmic reticulum stress and unfolded protein response: Published evidence documents Gd3+-triggered UPR in primary rat cortical astrocytes, and elevated ER stress markers (including PERK, DDIT3, ATF6) in rodent brain tissue following GBCA exposure. Identify and verify the primary papers, and assess whether any safety study measured ER stress endpoints.
• Golgi apparatus disruption and glycosylation impairment: Toxicogenomic profiling studies have identified disrupted glycosylation as a primary affected cellular function following gadolinium exposure, with vesicle-mediated Golgi transport and endocytosis identified as key affected pathways. Identify and verify the primary papers.
• Immune dysregulation: Published evidence documents multiple significantly elevated pro-inflammatory cytokines in gadolinium deposition disease patients compared to controls, with specific cytokines (including TNF, IL-6, and MCP-3) linked to pain symptomatology, and dynamic cytokine changes during chelation correlating with gadolinium mobilisation. Identify and verify the primary papers, and assess whether any safety study included immune or cytokine endpoints.
• Endocrine disruption: Published evidence documents GBCA alteration of thyroid hormone receptor-mediated transcription and effects on Purkinje cell dendritic arborisation in thyroid hormone-treated cultures. Identify and verify the primary papers, and assess whether any safety study included thyroid hormone assays or thyroid receptor endpoints beyond standard H&E histopathology.
• Calcium channel interference and ionic mimicry: Gd3+ is a potent blocker of stretch-activated and voltage-gated calcium channels. Its ionic radius (approximately 1.053 Angstrom) closely mimics that of Ca2+ (approximately 1.00 Angstrom), providing a plausible mechanism for displacement at calcium-dependent enzyme and channel sites and disruption of calcium-dependent cellular signalling. Identify published literature supporting this mechanism and assess whether any safety study included calcium-dependent functional endpoints.
• Transmetallation: Published evidence documents displacement of physiologically important metals (Zn2+, Cu2+, Fe3+, Ca2+) from enzyme binding sites by Gd3+, with brain regions showing highest T1 hyperintensity also being areas of highest endogenous iron, copper, and zinc concentration. Insoluble extracellular gadolinium foci have been identified in skin specimens with gadolinium localised in macrophages, lysosomes, and fibrocytes. Identify and verify the primary papers.
• Oxalate-mediated dechelation (critical finding — verify independently): A 2025 paper by Henderson and Wagner (MRI journal, approximately 2025) reported that oxalic acid displaces gadolinium from both linear AND macrocyclic chelates at lysosomal pH, with serum albumin accelerating the process and forming insoluble gadolinium oxalate precipitates. If confirmed, this provides a documented chemical mechanism for macrocyclic agent toxicity previously considered impossible under the stable-chelate model. Locate, verify, and evaluate this paper and its implications for the macrocyclic-is-safer regulatory position.
• In vivo nanoparticle formation: Emerging evidence suggests in vivo generation of gadolinium-containing nanoparticles as a distinct toxicity pathway operating through different mechanisms than free Gd3+ ions. A paper by Domingo and Semelka (Archives of Toxicology, approximately 2025) is reported to address this. Locate, verify, and assess the evidence.
• Oxidative stress: Published evidence documents substantial reactive oxygen species generation in multiple cell types following gadolinium exposure, with associated DNA fragmentation, chromatin condensation, and altered apoptotic pathway markers. Identify and verify the primary papers.
• P5P depletion and downstream endogenous oxalate production: P5P (pyridoxal-5-phosphate, active vitamin B6) is the cofactor for aminotransferase reactions that convert glyoxylate to glycine. If GBCA administration depletes P5P, glyoxylate would be shunted toward endogenous oxalate production, potentially triggering oxalate toxicity as a downstream consequence of GBCA exposure, independent of gadolinium retention. Has this specific mechanism been investigated in the published literature?
• Antioxidant mineral depletion: Is there evidence that GBCA administration depletes minerals critical to antioxidant and xenobiotic defence, including zinc (superoxide dismutase cofactor), selenium (glutathione peroxidase cofactor), manganese (mitochondrial SOD cofactor), and copper? Has depletion of these minerals been measured in GBCA-exposed populations?

Epigenetic and persistent effects

• Is there evidence that GBCA exposure triggers epigenetic changes — DNA methylation alterations, histone modification, or durable changes in gene expression — that persist after gadolinium has been cleared? A hit-and-run mechanism of this type would cause lasting harm without any radiologically detectable trace.

Single dose vs cumulative exposure

• Is the harm profile the same for single-dose and cumulative multi-dose exposure? The Walsh et al. finding that the majority reacted severely to their first dose suggests harm is not primarily a cumulative dose effect for a significant subgroup. Does the published literature adequately distinguish these populations?
• Does the absence of a detectable threshold below which harm does not occur have implications for the safety of any dose?

Documented deaths and serious harm

• Identify all categories of death attributable to GBCA administration in the published literature. Do not limit to NSF or intrathecal accidents — these are known examples only. The search should be open to all documented mechanisms and routes. Assess whether under-reporting is probable given the surveillance infrastructure's inability to capture delayed-onset reactions.
• What does the regulatory restriction or withdrawal of certain GBCA agents by the EMA, FDA, and MHRA imply about the adequacy of original pre-approval safety evaluation?

The macrocyclic-is-safer assumption

• Given the Henderson and Wagner (2025) oxalate dechelation finding, what is the current chemical basis for the macrocyclic-is-safer regulatory position?
• Is the macrocyclic-is-safer claim supported by clinical outcome evidence, or only by lower T1 brain signal retention? Is there evidence from clinical outcome studies that macrocyclic patients have lower rates of chronic symptoms or lower all-cause mortality than linear patients?
• Has a controlled study comparing contrast-naive patients to those who have received one or more GBCA administrations of any type been conducted using clinical hard endpoints? If not, characterise this gap explicitly.

Domain 13 — Adverse Reaction Timing, Surveillance Failures, and Neuropsychiatric Outcomes ★ Priority

The pharmacovigilance infrastructure for GBCAs is designed around the in-scanner acute reaction model. This is structurally inadequate for a drug that can cause multisystemic illness developing hours to weeks after the patient has left the imaging department.

Reaction timing and surveillance adequacy

• What follow-up duration and method is used in published GBCA adverse reaction studies? Are delayed-onset reactions — occurring hours, days, or weeks after the patient has left — captured? If not, characterise the resulting evidence gap.
• What is the current pharmacovigilance infrastructure for capturing delayed GBCA adverse reactions? Is there a mechanism by which a GP or specialist seeing a patient weeks after an MRI would reliably attribute a new symptom complex to the prior contrast administration?
• What is the evidence on the proportion of GBCA adverse reactions captured by current reporting systems versus those never attributed or attributed to other causes?

Neuropsychiatric outcomes and suicidal ideation

• Has any published GBCA safety study assessed neuropsychiatric outcomes, including depression, anxiety, mood disturbance, cognitive impairment, or suicidal ideation? If not, state this explicitly and characterise the significance of the omission.
• What validated neuropsychological tools would be appropriate for investigating post-GBCA brain fog and cognitive dysfunction? Has any published study used these tools in a GBCA-exposed population? If not, note this as a specific methodological gap.
• Is there evidence on disability outcomes in GBCA-exposed patients — including new disability diagnoses and loss of employment — in the period following contrast administration?

Domain 14 — Neurodegenerative Disease, Chronic Conditions, Study Design, and the Latency Problem ★ Priority

A study of insufficient duration for a condition with a long prodromal phase is not merely underpowered — it is asking the wrong question on the wrong timescale. This principle applies to all chronic conditions, not only neurodegenerative disease.

Latency, prodromal phases, and study duration

• For Parkinson's and Alzheimer's disease, what is the estimated prodromal phase duration, and what does this imply for the minimum follow-up required to detect increased incidence in an exposed population?
• What follow-up durations have been used in published GBCA studies for conditions with long latency periods? For each, compare the follow-up duration to the estimated latency. Studies of inadequate duration must not be cited as evidence of absence of association.
• Apply this principle broadly to all chronic conditions in the GBCA literature where pathological processes precede clinical presentation. Identify all studies where follow-up duration is inadequate relative to the condition being investigated.

The psychotropic medication confounding problem

• Identify studies that adjusted for psychotropic medication use and evaluate whether this variable is an independent confounder or an intermediate on the causal pathway.
• What is the evidence on the rate at which patients with post-GBCA symptoms are subsequently prescribed psychotropic medications, anti-epileptics, or drugs used in pain clinic settings including gabapentin?

Brain fog and survivorship bias

• Has any published GBCA study used validated neuropsychological assessment tools to measure brain fog, cognitive dysfunction, processing speed, or working memory? If not, characterise this as a specific remediable gap: a common, disabling, measurable symptom that has never been measured using appropriate instruments.
• For all observational GBCA studies, report whether all-cause mortality is tracked, dropout rates recorded with reasons, and non-completer analysis provided. Is survivorship bias discussed or accounted for in any published GBCA safety study?

Domain 15 — Informed Consent

• What information has been provided to patients before GBCA administration at each stage of the evolving evidence base? Were patients informed about gadolinium retention in brain and bone, and about the possibility of chronic symptoms?
• What do current guidelines and manufacturer prescribing information recommend regarding consent prior to GBCA administration? Is this adequate given current knowledge?
• What are the legal and ethical implications of administering a drug with an incompletely characterised long-term safety profile, without full informed consent, to hundreds of millions of patients? Has this been addressed in the bioethics literature?
• Is there evidence that patients are routinely asked about prior GBCA exposure before further doses? Given evidence of first-dose severe reactions, what does the absence of systematic prior exposure assessment imply?

Domain 16 — Functional Neurological Disorder and GBCA Toxicity

FND is one of the most common diagnostic labels applied to patients presenting with post-GBCA symptoms. Its positive diagnostic criteria do not exclude metabolic, toxic, or nutritional causes of atypical and fluctuating neurological presentations. The FND label frequently terminates further investigation.

• What is the evidence on all-cause mortality in FND patients compared to matched controls?
• Does the FND diagnostic framework adequately exclude gadolinium toxicity, thiamine and B-vitamin deficiencies including P5P depletion, mitochondrial dysfunction, and other metabolic causes of atypical neurological presentation?
• Is there evidence on treatment response to nutritional repletion — particularly thiamine — in patients meeting FND criteria?
• Is there published evidence on FND as a diagnostic outcome specifically in patients presenting with post-GBCA symptoms?
• What are the clinical and ethical implications of applying this label to post-GBCA patients who have not had metabolic and toxic causes systematically excluded?

Domain 17 — Full Cost Accounting and False Economy

• What is the healthcare system cost of downstream investigations triggered by contrast-detected findings that are ultimately benign, non-contributory, or misattributed?
• What is the cost of managing all documented categories of GBCA complication?
• Is there published economic modelling including missed diagnosis costs, patient financial harm, lost productivity, and litigation? If not, characterise what an honest accounting would require.
• What is the trajectory of medico-legal exposure for institutions and clinicians recommending contrast MRI, given emerging evidence and questions of inadequate informed consent?
• What is the cost argument for transitioning to optimised non-contrast protocols when full downstream costs of GBCA use are included?

Domain 18 — Historical Parallels: Medical Claims Overturned by Scrutiny

Identify cases where a widely-held clinical belief was overturned by scrutiny. For each: the original claim, its plausibility mechanism, the accepted evidence, the methodological move that dismantled it, and the parallel to the GBCA “saves lives” claim.

Minimum cases: HRT and cardiovascular protection (WHI, 2002); antiarrhythmic drugs post-MI (CAST, 1989); arthroscopic knee surgery for osteoarthritis — note that the Moseley et al. sham surgery trial (NEJM, 2002) and the FIDELITY trial on arthroscopic partial meniscectomy (Sihvonen et al., NEJM, 2013) are separate studies making related but distinct arguments and should be treated as such; mammography screening and overdiagnosis; high-dose chemotherapy with bone marrow transplant for breast cancer; PSA screening. Identify additional cases where a diagnostic claim was overturned.

Core question: does the “GBCAs save countless lives” claim fit the known pattern of beliefs that propagated on plausible mechanism and surrogate endpoints, without hard endpoint evidence, and persisted through institutional momentum and industry interest?

Domain 19 — FDA-Mandated Preclinical Safety Studies: Are They Adequate? ★ Priority

In December 2017, the FDA issued postmarketing requirements to all four GBCA NDA holders — Bayer AG, Bracco Imaging, GE Healthcare, and Guerbet — to conduct preclinical safety studies on gadolinium retention effects. The resulting publications are the studies currently cited as evidence that GBCAs have been found safe under FDA mandate. This domain asks whether these studies were capable of detecting the harm they were mandated to investigate.

• All four NDA holders co-authored all three studies. The manufacturers whose products were being evaluated for safety co-designed, co-conducted, and co-authored the research that determined whether regulatory action was taken against those products. Was any independent oversight, independent statistical analysis, or blinded endpoint assessment described? What does this structural conflict of interest imply for interpreting the conclusions?
• Apply the Rule of Three (Methodological Constraint 13) to the sample sizes used: Lewis (2024a) used n=10/sex/group for behavioural endpoints; Smieja (2025) used n=12/group total. Calculate the 95% upper confidence bound for undetected adverse effects at these sample sizes. Were power calculations performed? Were sample sizes justified against ICH S5(R3) guidance?
• Cross-reference the study endpoints against the documented toxicity mechanisms identified in Domain 12. For each mechanism, was an endpoint capable of detecting it included? How many of the documented mechanisms were directly assessed?
• The Smieja (2025) study found a difference in learning in the high-dose gadobenate dimeglumine group, which was classified as “non-adverse” by the manufacturer-employed investigators. What is the basis for this classification, and was it made independently? Apply Methodological Constraint 14 (the “explains away” pattern) to evaluate whether a consistent interpretive bias exists across the positive findings in these studies.
• Are these studies currently cited as evidence of GBCA safety by manufacturers, regulators, or clinicians? If so, given the methodological limitations identified above, is that citation valid? What would adequately powered, independently conducted, mechanism-specific studies need to look like?

Synthesis ★ Priority