Repurposed Cancer Drugs: Case Evidence Registry and Outcomes Audit

October 12, 2026

SmartCancer | Evidence Methods

Cancer patients and families increasingly encounter reports about medicines originally developed for other conditions that are being investigated or used experimentally for cancer. These reports range from peer-reviewed case studies to laboratory research, social media testimonials and claims of long-term remission.

The Repurposed Cancer Drugs: Case Evidence Registry and Outcomes Audit establishes a structured framework for cataloguing these reports, tracing them to their sources, assessing the quality of the available documentation and distinguishing reported outcomes from independently verified clinical responses.

Registry status and scope

This page defines the registry's proposed methodology, data fields and audit rules. It does not represent a completed systematic review, a fully populated clinical registry or an independently verified collection of patient outcomes. Individual records and aggregate statistics should be published only after the relevant source checks and audit procedures have been completed.

Methodology version: 1.0 | Last reviewed: October 13, 2026

1. What Is the Repurposed Cancer Drugs Case Evidence Registry?

Drug repurposing involves investigating an existing medicine or compound for a new therapeutic purpose. In oncology, this may involve a medicine approved for a non-cancer indication, an established human medicine used off-label, an investigational compound or a veterinary product proposed for a human cancer indication.

The biological rationale for repurposing a drug does not, by itself, demonstrate that it is safe or effective against a human cancer. Evidence must be evaluated according to the type of research performed, the clinical context, the quality of the outcome documentation and the possibility of alternative explanations.

This registry framework is designed to make those distinctions visible. Rather than treating every reported remission as an equivalent success, it records what was reported, where the information originated, how the outcome was assessed, what other treatments were administered and whether later publications changed the interpretation.

Core questions the registry should answer

What drug or compound was used, and was its formulation and regulatory status documented?
What cancer type, histology, stage and biomarker profile were reported?
Which standard treatments had already been given, and what happened before the repurposed drug was introduced?
Was the drug used alone or alongside chemotherapy, immunotherapy, radiation, surgery, supplements or other interventions?
What response or safety outcome was reported, by whom and using which assessment method?
Was the outcome supported by imaging, pathology, clinical documentation or only a personal account?
Was follow-up available, and did the patient subsequently experience progression, recurrence, another treatment response or a serious adverse event?
Has the original publication been corrected, challenged or retracted?

Guiding principle: The registry records evidence about reported cases. It does not assume that a drug caused an observed improvement merely because exposure preceded that improvement.

2. What Belongs in the Registry?

A useful registry needs explicit inclusion rules. Otherwise, published case reports, repeated social media posts, laboratory experiments and patient testimonials may be mixed together and incorrectly presented as a single body of clinical evidence.

Source category	Proposed handling	Important limitation
Published human case reports	Create or link a patient-level case record when relevant details can be extracted.	Publication does not establish causation or independent verification.
Published case series	Extract individual cases when identifiable without duplication and when adequate information is available.	Patients in later publications may overlap with earlier reports.
Observational studies	Link the study-level record and extract individual cases only where permitted and methodologically appropriate.	Aggregate results should not be counted as additional individual patients.
Clinical trials	Link trial records and published results to the relevant drug evidence page.	Trial-level outcomes are not automatically patient-level case reports.
Adverse-event and safety reports	Record or link traceable human safety reports under a separate safety classification.	A report of harm does not by itself establish drug causality or incidence.
Patient testimonials and social media	Keep traceable reports in a clearly separated, unverified category where appropriate.	Do not treat these reports as confirmed clinical responses.
Cell and animal research	Link to a separate preclinical evidence record.	Preclinical findings are not human patient outcomes.

Minimum inclusion criteria for a patient-level record

A traceable source or clearly identified origin for the report.
A human cancer diagnosis or a sufficiently described oncology context.
An identifiable repurposed drug or compound.
Enough information to determine what is being claimed about exposure, outcome or harm.
A documented classification for missing information, uncertainty and source reliability.

Reports that do not meet these criteria should not silently disappear. Where relevant, they can be retained in a separate screening log with the reason for exclusion or the information needed before review.

A formal systematic review or scoping review should use an appropriate published protocol and transparent study-selection process. PRISMA 2020 provides reporting guidance for systematic reviews; it is not a substitute for doing the search, screening and analysis themselves.^[2]

3. The Case-Level Data Model

Every eligible patient-level report should receive a unique, non-identifying Case ID. Original articles, follow-up publications, safety notices and corrections should receive separate Source IDs linked to that Case ID.

This approach allows the registry to track one reported patient across multiple publications without inflating the apparent number of cases. It also makes later updates auditable.

Data domain	Fields to capture	Audit purpose
Record identity	Case ID, Source ID, first publication date, DOI, PMID, source URL, related publications and update history.	Traceability, version control and duplicate detection.
Cancer diagnosis	Primary site, histology, stage, metastatic sites, date of diagnosis and relevant diagnostic evidence.	Establish the clinical context and avoid grouping unlike cancers.
Tumour biology	Reported molecular alterations, receptor status, actionable biomarkers and assay details when available.	Assess biological context and potential differences between cases.
Prior treatment	Surgery, radiation, systemic anticancer treatments, dates, previous responses and progression history.	Assess competing explanations and treatment history.
Repurposed drug	Generic name, formulation, intended indication, route, reported exposure, dates, source of product information and regulatory status by jurisdiction.	Identify the actual intervention without turning published exposures into dosing recommendations.
Concurrent interventions	Other medicines, supplements, diet changes, surgery, radiation, immunotherapy and supportive treatment.	Identify possible confounding and combination-treatment effects.
Baseline status	Disease burden, imaging date, laboratory results, symptoms, functional status and available biomarkers.	Establish the documented starting point.
Clinical outcome	Reported response, method of assessment, assessment date, response duration, progression, recurrence and survival follow-up where documented.	Separate clinical events from interpretation and ensure adequate follow-up.
Safety outcomes	Suspected adverse events, reported severity, laboratory abnormalities, discontinuation, management and subsequent course.	Capture possible harms alongside claimed benefits.
Source credibility	Publication status, conflicts of interest, funding, corrections, retractions and author-reported limitations.	Make material changes and credibility concerns visible.

Recommended public case-record format

A public summary should expose the most decision-relevant fields without publishing identifiable patient information. The following is a field template, not a real patient record.

Field	Example format
Case ID	RCD-CASE-0001
Drug	Generic drug name; formulation documented or unknown
Cancer	Primary site, histology and stage as reported
Evidence design	Individual case report, case series or other predefined category
Outcome	Author-reported response, independently assessed response, harm or unknown
Verification	V0–V3 according to the proposed framework below
Publication status	Current status checked on a specified date
Source	DOI, PMID or other traceable original source

4. Evidence-Design Classification: E0–E5

The proposed E0–E5 classification labels the type of evidence associated with a record. It is an editorial classification designed for this registry, not a validated universal evidence-rating system. It must not replace a formal risk-of-bias assessment or the grading system used by an applicable clinical guideline.

Label	Evidence category	Typical material	What it can establish
E0	Preclinical evidence	Cell experiments, molecular studies and animal models.	Biological hypotheses and experimental findings in the tested model.
E1	Anecdotal reports	Testimonials, social media narratives and unverified personal accounts.	That an outcome has been reported, not that the outcome is verified or caused by the drug.
E2	Human case reports and case series	Individual published cases and small uncontrolled series.	A documented clinical observation that may justify further investigation.
E3	Observational human evidence	Cohort studies, registries and non-randomized comparisons.	Patterns and associations, subject to confounding and study quality.
E4	Controlled clinical studies	Randomized trials and other appropriately controlled intervention studies.	Comparative treatment effects when design, execution and analysis support that inference.
E5	Evidence synthesis and clinical guidance	Systematic reviews, meta-analyses and evidence-based guidelines.	A synthesis of the available evidence, depending on the underlying studies and the quality of the review.

These labels are not interchangeable with proof of efficacy. For example, an E5 systematic review can still conclude that evidence is insufficient if the included studies are weak, inconsistent or at high risk of bias. An E4 label likewise does not automatically make a trial reliable.

For that reason, every record should have additional fields for risk of bias, source limitations, missing data and applicability to the cancer population being considered. When multiple evidence types are linked to a drug, the registry should preserve their distinct identities rather than assigning one overall score to the entire drug.

5. Outcome Verification: V0–V3

Evidence design and outcome verification answer different questions. A peer-reviewed paper can provide a traceable source without giving the reader access to underlying clinical records. Conversely, a testimonial may contain photographs or test results that have not been independently authenticated.

The registry therefore proposes a second, independent verification label.

Label	Verification status	Required description
V0	Uncorroborated report	An outcome is claimed, but no supporting clinical documentation is available to the audit.
V1	Traceable source	A publication or identifiable primary report exists. The outcome remains classified according to what that source actually documents.
V2	Supporting documentation reviewed	Relevant records, such as imaging reports or clinical documents, have been reviewed under a documented protocol with appropriate permissions.
V3	Independent outcome adjudication	An independent qualified reviewer or panel has assessed the available evidence using predefined criteria, where feasible.

A V1 designation does not mean a patient outcome has been independently confirmed. V2 and V3 should be assigned only when the stated review actually took place, and the public record should identify the scope of the review.

Neither this verification scheme nor the E0–E5 classification is a formal clinical or regulatory standard. Both are proposed operational tools that should be tested for consistency before they are used to generate aggregate research conclusions.

6. Standardized Outcome Definitions

One of the most important tasks is to distinguish what an author or patient reported from what the available clinical evidence supports.

For solid tumours, RECIST 1.1 provides standardized criteria for measuring tumour response in clinical research. Its response categories include complete response, partial response, stable disease and progressive disease.^[3] These criteria should be applied only when the required information is available and the criteria are appropriate to the case.

Outcome field	Meaning	Registry treatment
Complete response (CR)	A response meeting the applicable complete-response criteria.	Record whether this was author-reported, documented under appropriate criteria or independently adjudicated.
Partial response (PR)	A response meeting the applicable partial-response criteria.	Capture the assessment method, dates and supporting evidence.
Stable disease (SD)	Disease that meets neither the applicable response nor progression criteria.	Record assessment timing and the duration of stability when available.
Progressive disease (PD)	Disease meeting the applicable progression criteria.	Capture the date, progression site and assessment method when reported.
Symptom improvement	A reported change in pain, appetite, fatigue or another symptom.	Keep symptom outcomes separate from tumour response.
Biochemical or biomarker change	A change in a laboratory result, tumour marker or other measured variable.	Record the assay, baseline value, follow-up date and clinical context.
Progression-free or overall survival	Time-to-event outcomes assessed using defined endpoints.	Record only when the relevant dates, definitions and follow-up support the calculation.
Not assessable or unknown	Available information does not permit a reliable classification.	Preserve this category rather than inferring a response from incomplete descriptions.

If a case report simply states that a tumour became smaller without sufficient measurements to assess formal response criteria, the registry should preserve that wording as an author-reported observation. It should not automatically convert it into a verified partial response.

Similarly, a normal tumour marker, an improvement in symptoms or an isolated favourable scan does not necessarily establish complete remission. Outcome definitions must be cancer-specific where appropriate, and follow-up must be considered before conclusions about durability are made.

7. How the Outcomes Audit Works

A reproducible audit requires a written procedure. Reviewers should apply the same rules to apparently positive reports, negative reports, adverse events and cases with uncertain outcomes.

Step 1: Establish source identity

Verify the original article, DOI, PMID or other traceable source. Record the publication date, article type, authors, journal and any subsequent corrections or retractions.

Step 2: Detect duplicate patients and publications

Search for overlapping case details across articles, follow-up reports and different websites. Similarity alone does not prove that two reports describe the same patient; uncertain matches should be flagged for review. Link verified duplicates rather than counting them as new independent cases.

Step 3: Confirm the reported cancer context

Extract the primary cancer site, histology, stage, metastatic sites and available biomarkers. Record whether details were supported by pathology, imaging or clinical documentation, as described in the source.

Step 4: Reconstruct the treatment timeline

Record previous treatments, the date the repurposed drug was introduced, reported exposure, treatment interruptions, concurrent interventions and subsequent changes in therapy. The timeline is essential when several interventions could explain an observed outcome.

Step 5: Assess the baseline and outcome

Compare the reported outcome against baseline disease status. Capture imaging or laboratory dates, the assessment method, response duration and the date of last available follow-up. Do not assume that missing follow-up means the patient remained in remission.

Step 6: Audit adverse events

Extract reported harms, relevant laboratory abnormalities, treatment discontinuations and recovery or deterioration after exposure. Where available, record the source's causality assessment and note other plausible causes. Do not assume that an event was caused by the repurposed drug merely because it occurred after exposure.

Step 7: Review conflicts and publication status

Check author disclosures, funding, corrections, expressions of concern and retraction notices. A retraction must remain visible in the record, even if the original article continues to circulate through other websites.

Step 8: Grade the record and document missing information

Assign the evidence-design and outcome-verification labels separately. Log unavailable fields explicitly, document reviewer disagreements and state whether the record was independently checked.

Step 9: Preserve a version history

Record when the case was created and last reviewed, which fields changed, why the change was made and which source supports it. Where a paper is retracted after a case has been published in the registry, update the record without deleting its audit history.

Consistency rule: Reviewers should not use a more favourable interpretation for a preferred drug than they would use for an alternative drug. Audit rules should be predefined, version-controlled and applied consistently.

8. Why Publication Status and Safety Matter

A registry becomes more useful when it demonstrates how new information can change the interpretation of an existing report.

Example A: A retracted fenbendazole case series

The paper titled Fenbendazole as an Anticancer Agent? A Case Series of Self-Administration in Three Patients was published in Case Reports in Oncology in May 2025. The publisher and editor formally retracted it on January 21, 2026, after identifying an undeclared potential conflict of interest involving the first author's provision of services related to the paper's subject. The retraction notice states that the explanation provided by the author was unsatisfactory and that the undeclared activity could have affected interpretation and recommendations.^[4]

An evidence registry should therefore retain the source's bibliographic identity while prominently marking its retracted status and linking the retraction notice. It should not continue presenting the article as an unqualified source of clinical evidence.

Equally, a retraction for an undeclared conflict of interest is not, by itself, proof that every individual clinical observation in the paper was false. The correct treatment is to document the retraction's stated grounds and prevent the paper from being interpreted as an unqualified, currently valid publication.

Primary records: Original article and retraction information | Formal retraction notice

Example B: A suspected drug-related liver injury

A 2021 case report described severe liver injury in a patient with advanced non-small cell lung cancer who had been receiving pembrolizumab and subsequently disclosed self-administration of fenbendazole. The report stated that liver dysfunction resolved after fenbendazole was discontinued and that tumour shrinkage was not observed.^[5]

A separate report published in 2026 described severe hepatocellular liver injury following increased self-administration of fenbendazole in a patient with metastatic colorectal cancer receiving nivolumab and relatlimab. The authors discussed the difficulty of differentiating possible fenbendazole-related injury from immunotherapy-associated hepatitis.^[6]

These case reports are relevant safety observations, not estimates of how frequently liver injury occurs or definitive proof of causality in every circumstance. They illustrate why a registry must record concurrent therapy, timing, laboratory findings, discontinuation and subsequent clinical course.

Primary records: 2021 fenbendazole-associated liver injury report | 2026 liver injury and causality assessment report

These examples illustrate the audit method only. They are not a complete fenbendazole evidence review, and the registry must not infer a drug's overall benefit-risk profile from two selected cases.

9. Data Analysis: What Can and Cannot Be Concluded?

A case registry can be useful for identifying research questions, documenting reported outcomes, finding gaps in follow-up and detecting patterns that warrant further investigation. It cannot automatically establish that a repurposed drug caused a remission or prolonged survival.

Why apparent success rates can mislead

Publication bias: Unusual or favourable outcomes may be more likely to be published than ordinary outcomes or treatment failures.
Selection bias: Online collections may selectively include cases that support a preferred treatment.
Duplicate reports: The same patient may appear in a paper, interview, social media post and several secondary articles.
Concurrent therapies: Surgery, radiation, standard systemic treatment or other interventions may contribute to the outcome.
Differences between cancers: Two tumours with the same primary site may differ substantially in histology, stage, molecular biology and natural history.
Incomplete follow-up: A favourable scan at one time point does not establish durable remission or improved survival.
Missing denominators: A collection of selected reports usually does not reveal how many patients used the drug without benefit or experienced harm.
Confounding and attribution: The sequence of events may be compatible with a drug effect without demonstrating that the drug was responsible.

Rules for publishing aggregate statistics

The registry should not publish an overall success percentage calculated from a curated collection of positive case reports or testimonials. An aggregate estimate should be produced only where the inclusion criteria, denominator, duplicate handling, outcome definition, follow-up rules and missing-data treatment have been specified in advance and are suitable for the question being asked.

Where those requirements cannot be met, descriptive reporting is more appropriate. For example, a report may count the number of traceable published cases with a particular outcome while explicitly stating that the total is not an efficacy rate and that the registry cannot estimate the probability of response among all people exposed to the drug.

Patient-level observations and study-level clinical results should also remain analytically distinct. A randomized trial, a retrospective cohort and a patient testimonial cannot be combined into one numerator simply because all three mention the same drug.

10. Patient Safety, Ethics and Privacy

Evidence about a medicine's potential anticancer activity is only one part of its clinical assessment. Formulation, pharmacokinetics, interactions, organ toxicity, dose exposure, manufacturing quality and suitability for a specific patient may all affect the benefit-risk balance.

Veterinary products, investigational compounds, medicines approved for a different indication and off-label use of an approved human medicine should be identified as distinct categories. Regulatory status must be checked against the relevant jurisdiction and the date of review; it should not be inferred from an internet seller's description.

The National Cancer Institute explains the distinction between complementary medicine used alongside standard care and alternative medicine used instead of standard treatment. It also notes that the safety and effectiveness of many complementary and alternative cancer approaches are uncertain.^[7]

Minimum safety and governance requirements

Never present an unverified story as a treatment recommendation or clinical outcome guarantee.
Do not publish a case's dose history as an instruction for self-treatment.
Record negative outcomes, adverse events and uncertainty alongside reports of apparent benefit.
Use appropriately de-identified information and avoid publishing patient names, dates of birth, medical record numbers or identifying images without a lawful and justified basis.
Set clear rules for consent, data access, retention, correction requests and removal of improperly disclosed information.
Any prospective collection of patient-level health information should undergo appropriate legal, privacy and ethics review before launch.

Patient guidance: This registry is an evidence resource, not individualized medical advice. Do not start, stop or replace prescribed cancer treatment based on a case report or online claim. Discuss investigational or repurposed medicines with the treating oncology team, including possible interactions, toxicity and appropriate clinical trials.

11. Building a Living Repurposed Cancer Drugs Registry

The methodology should be implemented in stages so that the published resource does not imply a degree of completeness or verification that the underlying data do not yet support.

Phase 1: Source-based case collection

Create the initial dataset from traceable published human cases and relevant safety reports. Store the source citation, original publication status, screening decision and reason for inclusion. Keep preclinical studies and aggregate clinical trial results in linked evidence tables rather than counting them as individual patients.

Phase 2: Standardized extraction and review

Apply the common case-record fields, duplicate-detection rules, E0–E5 evidence-design labels and V0–V3 verification labels. Document unclear coding decisions. Use a second reviewer for important outcome fields where feasible and report how disagreements are resolved.

Phase 3: Public case pages and audit dashboard

Publish records with traceable sources and clearly displayed evidence, verification and publication-status labels. A useful dashboard could summarize case records by drug, cancer type, source type, outcome category, follow-up completeness, adverse events and current publication status.

Dashboard totals should be calculated from the maintained dataset, with documented inclusion rules and a visible last-updated date. Until those data and checks exist, display the methodology rather than invented totals or success statistics.

Phase 4: Corrections and ongoing surveillance

Establish a process for monitoring important publications, corrections, retractions, new clinical trials and safety findings. Preserve the previous state of each record in the audit log and explain meaningful changes to the public summary.

Phase 5: Controlled submissions

Consider accepting patient or clinician submissions only after appropriate privacy, consent, data-security and review procedures have been established. Keep self-reported information separate from published evidence and independently reviewed outcomes.

12. Connecting the Registry to the Cancer Knowledge Graph

A case evidence registry becomes more valuable when each record is connected to the wider cancer evidence landscape instead of being treated as an isolated story.

The proposed knowledge graph connects the following entities:

Cancer: Primary site, histology, stage and metastatic context.
Biomarker: Molecular alteration, receptor status, assay and date.
Treatment: Repurposed drug, standard anticancer therapies and concurrent interventions.
Outcome: Response category, progression, survival follow-up and adverse events.
Evidence: Case reports, observational studies, trials, reviews and safety notices.
Source status: Publication, correction, retraction, review date and audit history.
Uncertainty: Missing information, confounding factors, verification level and competing explanations.

This structure supports questions such as: Which cancer types have published human observations for a specific compound? Are outcomes documented for a particular biomarker subgroup? Were concurrent therapies reported? How much follow-up is available? Have the conclusions changed following a retraction or a new safety report?

Such queries are useful for evidence mapping and hypothesis generation. They should not be presented as personalized treatment recommendations or as proof that a drug works for a particular cancer subtype.

Related OneDayMD research collections

Fenbendazole cancer reports archive — a historical collection that can be re-examined using source-level deduplication, outcome verification, safety tracking and publication-status auditing.

The central registry should own the canonical case IDs. Drug-specific and cancer-specific pages should reference those records rather than independently recounting the same patients.

13. Frequently Asked Questions

What is a repurposed cancer drug?

A repurposed cancer drug is a medicine or compound originally developed or used for a different indication that is investigated for a possible role in cancer prevention or treatment. The term describes the proposed new use, not proof of anticancer efficacy.

Does a published case report prove that a drug works?

No. A case report can document a clinically interesting observation, but it usually cannot isolate the effect of one treatment from concurrent therapies, natural disease variation or other explanations. Controlled human research is generally needed to establish comparative treatment effects.

What is the difference between a reported remission and a verified remission?

A reported remission is an outcome described by a patient, author or other source. A verified outcome requires supporting documentation reviewed under a stated process. The degree of verification and the clinical response criteria used should always be disclosed.

Does a larger number of success stories indicate a higher response rate?

Not necessarily. Online reports may duplicate the same patient, omit treatment failures and lack a denominator of all exposed patients. Without systematic case ascertainment, reliable outcome definitions and adequate follow-up, the number of stories cannot establish a response rate.

Why should retracted papers remain visible in the registry?

Retaining a record with a prominent retraction notice preserves the publication history, prevents accidental reuse as current evidence and allows readers to understand why its status changed. It is more transparent than silently removing a source after its conclusions have circulated elsewhere.

Can this registry recommend a repurposed drug for an individual patient?

No. The registry is intended to catalogue evidence and its limitations. Treatment decisions require individualized clinical evaluation, and a collection of cases should not substitute for oncology care, established treatment guidelines or appropriate clinical trials.

Will every case have an independently verified outcome?

No. Some publications provide only limited information, and some records cannot be independently checked. The registry should disclose the available documentation and verification level rather than imply that all records have been externally confirmed.

14. References and Reporting Standards

The following sources support the proposed reporting, outcome-assessment and publication-audit framework. They are reference standards and illustrative examples, not an exhaustive systematic review of repurposed cancer drugs.

CARE Guidelines. Consensus-based guidance intended to improve the accuracy, transparency and usefulness of published case reports.
CARE official website
PRISMA 2020. Reporting guidance for systematic reviews, including transparent reporting of review methods, study selection and results.
PRISMA 2020 official guidance
RECIST 1.1. Standardized response criteria used to measure tumour response in clinical research involving solid tumours.
RECIST 1.1 guidelines
Retraction statement. Makis W, Baghli I, Martinez P. Fenbendazole as an Anticancer Agent? A Case Series of Self-Administration in Three Patients. Case Reports in Oncology. 2025. Retracted January 21, 2026.
Retraction notice, DOI: 10.1159/000549387
Original article record, DOI: 10.1159/000546362
Drug-induced liver injury after fenbendazole self-administration. Published case report involving a patient with advanced non-small cell lung cancer receiving pembrolizumab.
PubMed record, PMID: 34248555
Differentiating fenbendazole-induced liver injury from immunotherapy hepatitis: the importance of structured causality assessment. Published case report, 2026.
PubMed record, PMID: 41608149
National Cancer Institute. Complementary and Alternative Medicine for Patients.
NCI patient guidance