Regulatory & Ethics in Oncology Research

Note: This page is educational and reflects the state of regulation and bioethics in 2025. It does not replace legal or regulatory advice.

TL;DR

Modern oncology research lives at the intersection of safety regulation, research ethics, and data protection law. The key rules to internalize: trials require prospective ethics review and informed consent; drugs and diagnostics require regulatory approval (FDA/EMA/ANVISA); sensitive health data require a legal basis (HIPAA, GDPR, LGPD in Brazil). For technologists, the most overlooked area is algorithmic fairness and accountability when AI is used in clinical or screening contexts. "Move fast and break things" is the wrong default in oncology.

---

1. Two distinct ethical regimes

It is critical to keep two regimes mentally separated:

• Treatment ethics — what you may do for one patient in their best interest, with their consent.
• Research ethics — what you may do to learn something that benefits future patients, with stricter oversight.

Research ethics is stricter precisely because the participant may not benefit personally — yet they accept risk for the common good.Mixing the two is a classic source of ethical breach (e.g., "compassionate use" framed to bypass trial requirements). Sources: [1], [2]

---

2. Foundational documents (read these once)

• Nuremberg Code (1947) — born from Nazi medical atrocities; introduced voluntary consent.
• Declaration of Helsinki (WMA, 1964; latest revision 2024) — the international ethical standard for medical research with humans.
• Belmont Report (1979) — the three principles: respect for persons, beneficence, justice.
• CIOMS International Ethical Guidelines (latest 2016) — operational guidance, especially for low- and middle-income contexts.
• ICH-GCP E6(R3) — Good Clinical Practice; the operational standard adopted by regulators globally.

These are not historical curiosities — they are the source of every modern protocol clause.

---

3. Regulatory pathways for drugs and diagnostics

Drugs

Region	Approval pathway	Key accelerators
USA (FDA)	NDA / BLA	Fast Track, Breakthrough Therapy, Accelerated Approval, Priority Review
Europe (EMA)	Centralised MAA	PRIME, Conditional Approval
Brazil (ANVISA)	Registro de medicamento	Priority review for unmet need; orphan-drug provisions
Japan (PMDA)	NDA + Sakigake	Sakigake (pioneer) designation

Accelerated paths use surrogate endpoints (response rate, PFS) — earlier access in exchange for post-marketing confirmation. The downside: when confirmatory trials fail, withdrawal is politically and clinically painful.

Diagnostics and companion diagnostics

• FDA — IVD regulation (PMA / 510(k)); CDx specifically reviewed alongside their drug.
• Europe — IVDR (in vitro diagnostic regulation) since 2022 — much stricter than the previous IVDD.
• Brazil — ANVISA RDC 36/2015 and IN 50/2019 govern IVDs; updates ongoing.

Software / AI as a Medical Device (SaMD / AIaMD)

• FDA — pre-cert program, Predetermined Change Control Plan (PCCP) for AI/ML.
• EU — MDR + AI Act (2024); high-risk AI in healthcare requires conformity assessment, transparency, human oversight.
• Brazil — ANVISA RDC 657/2022 and RDC 751/2022 cover SaMD; AI-specific guidance evolving.

---

4. Informed consent — beyond the form

A signed document is not consent. Consent is a process: information disclosure, comprehension, voluntariness, decision capacity, authorization. Common failure modes: Sources: [1], [2]

• Therapeutic misconception — the patient believes the trial is treatment optimized for them, not research.
• Coercion — the only access to a promising agent is via the trial; the choice is not free.
• Information overload — 30-page consent forms that no one understands.
• Vulnerability — children, prisoners, people with cognitive impairment, populations with limited education or healthcare access.
• Cross-cultural mismatch — concepts like "randomization" or "placebo" do not translate cleanly across cultures.

A modern view treats consent as a fiduciary relationship — the physician/researcher is obligated to act in the participant's interest, calibrated to that participant's preferred level of decision delegation. Sources: [2]

For Brazilian protocols: the consent form (TCLE — Termo de Consentimento Livre e Esclarecido) is regulated, and templates are reviewed by the local CEP and CONEP.

---

5. Data protection in oncology

Cancer data are sensitive on multiple axes — genomic data alone can re-identify; combined with clinical data they are highly identifying.

Legal frameworks

• HIPAA (US, 1996) — covers covered entities and business associates; safe harbor de-identification.
• GDPR (EU, 2018) — special category data (health, genetic, biometric); explicit consent or other legal basis required.
• LGPD (Brazil, Lei 13.709/2018) — operates on principles parallel to GDPR; the ANPD is the regulator. Health data are dado pessoal sensível.
• Common Rule (US, 45 CFR 46) — federal research protections.

Practical implications

• Pseudonymization ≠ anonymization. Genomic data are inherently re-identifiable; promise re-identification risks honestly.
• Data minimization. Collect only what the protocol needs.
• Access controls and audit logs — not optional for sensitive cohorts.
• Cross-border transfer — standard contractual clauses, adequacy decisions, or explicit consent.
• Right to withdraw — clarify whether withdrawal includes existing data and downstream analyses.
• Secondary use — broad consent vs. dynamic consent vs. specific consent each have trade-offs.

For deep dive: Data governance & LGPD.

---

6. AI fairness and accountability in oncology

Algorithmic decisions in oncology — risk scores, screening triage, treatment recommendation — carry the same ethical weight as a physician's. Specific failure modes:

• Distributional shift — model trained on US/EU patients fails on Brazilian, African, or Asian populations.
• Label bias — historical care disparities baked into "ground truth" labels.
• Spurious shortcuts — model classifies cancer using acquisition artifacts of the imaging device, not biology.
• Calibration failure — apparent accuracy on average, but systematic over/under-confidence in subgroups.
• Lack of explainability — clinicians cannot interrogate the model's reasoning.
• Automation bias — clinicians defer to the model even when they shouldn't.
• Accountability gap — who is responsible when the model is wrong?

The EU AI Act, FDA's PCCP framework, and emerging Brazilian guidance all push toward transparency, monitoring, and human oversight. See ML pitfalls in oncology for technical countermeasures.

---

7. Special populations

Oncology touches groups that need extra protection:

• Children — pediatric oncology trials require additional review and assent (not just parental consent).
• Pregnant individuals — usually excluded; this creates evidence gaps.
• Older adults with cognitive impairment — surrogate decision-making rules vary by jurisdiction.
• Indigenous populations — collective consent and benefit-sharing (especially relevant in Brazil).
• Genetic relatives — implications of returning germline findings extend beyond the consenting individual.
• Severely ill patients — therapeutic misconception risk is highest.

---

8. Conflicts of interest and integrity

Industry ties pervade oncology. Disclosure is necessary but not sufficient. Key questions:

• Is the principal investigator employed or paid by the sponsor?
• Were the analyses pre-specified, or were endpoints adjusted post-hoc?
• Was the comparator deliberately weak (placebo where standard of care exists)?
• Were negative trials published or buried?
• Are KOL (key opinion leader) honoraria distorting practice patterns?

Resources: ICMJE conflict-of-interest disclosure standards; Open Payments (Sunshine Act, US).

---

9. Post-trial access and benefit-sharing

A trial that works ethically must also end ethically. Key considerations:

• Post-trial access — does the participant continue to receive the drug after the trial ends? Lei 14.874/2024 strengthens this in Brazil.
• Pricing and reimbursement — does the population that took the risk get to use the resulting therapy?
• Data and biospecimens — were they used as authorized? Returned or discarded as agreed?
• Publication — including negative trials (registry-based commitments).

These are not optional courtesies; they are part of the social contract that makes research possible.

---

10. A short checklist for technologists building in this space

Before deploying anything that touches a patient, ask:

1. Is this research, treatment, or product? Different rules apply.
2. Who is the regulator? FDA / EMA / ANVISA / domestic equivalent.
3. What is the legal basis for processing this data? (Consent, legitimate interest, public health…)
4. Is the model auditable? (Versioned, logged, monitored for drift.)
5. Where can things fail silently? (Subgroup performance, distribution shift, label noise.)
6. Who is accountable when it does fail? (Have you said this in writing?)
7. Have ethics review and a clinical owner signed off?
8. Have you read the protocol? (Yes, the whole thing.)

---

See also

• Clinical trials (overview)
• Clinical trials 101
• Data governance & LGPD
• ML pitfalls in oncology
• Biomarkers & companion diagnostics

---

References

1. del Carmen MG, Joffe S. Informed consent for medical treatment and research: a review. Oncologist 2005;10:636-641. PMID 16177288. https://doi.org/10.1634/theoncologist.10-8-636
2. Ludewigs S, Narchi J, Kiefer L, Winkler EC. Ethics of the fiduciary relationship between patient and physician: the case of informed consent. J Med Ethics 2024;51:59-66. PMID 36564172. https://doi.org/10.1136/jme-2022-108539
3. World Medical Association. Declaration of Helsinki (revised 2024). https://www.wma.net/policies-post/wma-declaration-of-helsinki/
4. U.S. National Cancer Institute. Clinical trials information. https://www.cancer.gov/about-cancer/treatment/clinical-trials
5. ANVISA — Agência Nacional de Vigilância Sanitária. https://www.gov.br/anvisa/pt-br
6. CONEP / Plataforma Brasil. https://plataformabrasil.saude.gov.br
7. Lei nº 14.874, de 28 de maio de 2024 — pesquisa clínica com seres humanos. https://www.planalto.gov.br/ccivil_03/_ato2023-2026/2024/lei/L14874.htm
8. Lei nº 13.709, de 14 de agosto de 2018 — Lei Geral de Proteção de Dados (LGPD). https://www.planalto.gov.br/ccivil_03/_ato2015-2018/2018/lei/L13709.htm
9. A.C. Camargo Cancer Center. https://accamargo.org.br
10. Fundação do Câncer (Brasil). https://www.cancer.org.br/
11. Ministério da Saúde / BVS. ABC do câncer. https://bvsms.saude.gov.br/bvs/publicacoes/abc_do_cancer.pdf
12. American Cancer Society. Cancer A-Z. https://www.cancer.org/cancer.html
13. Cleveland Clinic. Cancer (overview). https://my.clevelandclinic.org/health/diseases/12194-cancer