Homework

Weekly homework submissions:

  • Week 1 HW: Principles and Practices

    ENDOMETRIOSIS ORGAN-ON-A-CHIP Engineering women’s health through microfluidic systems with ethical precision. A research concept at the intersection of microfluidics, endometriosis biology, and responsible innovation.

  1. Project Vision Concept Overview I propose developing an endometriosis organ-on-a-chip using microfluidics. The system would model key aspects of the disease in a controlled environment:

Subsections of Homework

Week 1 HW: Principles and Practices

ENDOMETRIOSIS ORGAN-ON-A-CHIP

Engineering women’s health through microfluidic systems with ethical precision.

A research concept at the intersection of microfluidics, endometriosis biology, and responsible innovation.

Organ-on-a-Chip Model Organ-on-a-Chip Model

1. Project Vision

Concept Overview

I propose developing an endometriosis organ-on-a-chip using microfluidics. The system would model key aspects of the disease in a controlled environment:

  • Endometrial-like tissue cells (stromal + epithelial)
  • Optional immune components (inflammation relevance)
  • Hormone cycling (estrogen/progesterone)
  • Controlled fluid flow to mimic physiological conditions

Purpose:
To create a more human-relevant research model than traditional cell culture or animal systems.

Why it matters:
Endometriosis remains difficult to study. Patient samples are limited, animal models do not fully replicate human biology, and standard 2D cell culture oversimplifies the disease. This platform aims to improve research quality and drug screening relevance.

2. Governance & Ethical Goals

To ensure that this organ-on-a-chip platform contributes to an ethical biological future, the governance framework must address harm prevention, scientific integrity, fairness, and responsible bioengineering from the beginning — not as an afterthought.

Goal A — Prevent Harm

The primary ethical priority is avoiding misuse or overinterpretation of results.

  • A1: Prevent individuals from using chip results to make personal medical decisions (“this drug will work for me”).
  • A2: Prevent exaggerated or misleading claims that could promote unsafe treatments, unregulated products, or false hope.
  • A3: Ensure communication clearly states that the platform is a research tool, not a diagnostic or treatment-selection device.

Goal B — Scientific Reliability & Integrity

A model that looks sophisticated but lacks reliability can mislead research and waste resources.

  • B1: Ensure repeatability (same input → similar output across chips, days, and operators).
  • B2: Validate the system against known biological markers and reference compounds.
  • B3: Report limitations transparently, including variability and boundary conditions.

Goal C — Equity & Representation

Endometriosis affects diverse populations. The model should not reinforce biological bias.

  • C1: Avoid building a platform based only on one donor profile (one ancestry, age group, or disease subtype).
  • C2: Encourage inclusion of diverse donor-derived materials when possible.
  • C3: Promote accessibility through shared protocols and cost-conscious design, so the tool is not limited to elite institutions.

Goal D — Responsible Bioengineering & Biosafety

Human-derived materials and engineered systems require careful oversight.

  • D1: Follow strict biosafety standards for handling and disposal of human cells and materials.
  • D2: Ensure clear consent frameworks for any donor-derived materials.
  • D3: Prevent misuse through overstated “personalized medicine” claims without strong clinical evidence.

3. Governance Actions

To ensure the organ-on-chip platform develops responsibly, governance must combine regulatory-style standards, communication limits, transparency mechanisms, and equity incentives.

Action 1 — Minimum Validation Standard

Actors: Academic labs, journals, funding agencies; regulators if commercialized.

Purpose:
Currently, organ-on-chip studies vary widely in methods and reporting. Many appear promising but are difficult to reproduce. The proposed change is to require a minimum validation package before strong scientific or commercial claims are made.

Design:
A standardized checklist that includes:

  • Verification of tissue identity (marker expression)
  • Demonstration of at least one core functional response (inflammatory signaling)
  • Testing against 1–2 known reference compounds (positive and negative controls)
  • Reporting variability across chips, batches, and experimental days

Journals and funders would require this package before publication or grant approval.

Assumptions:
Standardization improves trust and reproducibility without overly restricting innovation.

Risks of Failure & “Success”:

  • Failure: Labs treat validation as a box-checking exercise.
  • Success risk: Excessive standardization may slow innovation or disadvantage smaller labs with fewer resources.

Action 2 — No Personalized Treatment Claims (Early Stage)

Actors: Companies, universities, ethics boards, communications teams.

Purpose:
The largest ethical risk is patients interpreting chip results as personalized medical advice. The proposed change is to strictly limit claims in early-stage research.

Design:

  • Mandatory labeling as “research tool only”
  • Prohibition of phrases implying patient-specific treatment prediction
  • Any move toward personalized clinical claims requires formal clinical validation and regulatory review

Assumptions:
Clear communication boundaries reduce misuse and overinterpretation.

Risks of Failure & “Success”:

  • Failure: Marketing language subtly implies personalization despite formal restrictions.
  • Success risk: Conservative messaging may reduce investor enthusiasm or media visibility.

Action 3 — Open Reporting Standards & Audit Trails

Actors: Journals, scientific conferences, funding bodies, research institutions.

Purpose:
Replication often fails due to incomplete method reporting. The proposed change is mandatory transparency in chip design and experimental parameters.

Design:
Require reporting of:

  • Chip schematics or design files
  • Flow rates and media composition
  • Cell sources and passage numbers
  • Quality control procedures (leak tests, bubble management)
  • Raw data for key outputs

This creates an audit trail and enables meaningful replication.

Assumptions:
Transparency improves scientific reliability and reduces hype.

Risks of Failure & “Success”:

  • Failure: Researchers withhold details under intellectual property concerns.
  • Success risk: Poorly understood replications may create inconsistent results and confusion.

Action 4 — Equity Incentives for Diverse Cell Sourcing

Actors: Funding agencies, biobanks, universities, industry partners.

Purpose:
If the model is built from narrow donor populations, it may fail to reflect disease diversity. The change is to incentivize broader biological representation.

Design:

  • Funding preference for projects including diverse donor-derived cells
  • Requirement for subgroup performance reporting when claims are made
  • Shared biobank access to support smaller institutions

Assumptions:
Biological diversity improves generalizability and fairness.

Risks of Failure & “Success”:

  • Failure: Logistical and financial barriers slow progress.
  • Success risk: Expanded donor sourcing increases privacy complexity and consent challenges.

Action 5 — Consent & Privacy Framework for Human-Derived Materials

Actors: Ethics committees (IRBs), hospitals, biobanks, universities.

Purpose:
To protect donor autonomy and prevent misuse of sensitive information.

Design:

  • Clear consent scope (research use, sharing, commercial use)
  • De-identification and restricted data access
  • Defined limits on re-contacting donors

Assumptions:
Transparent consent frameworks build long-term trust and legal compliance.

Risks of Failure & “Success”:

  • Failure: Overly broad consent becomes ethically questionable.
  • Success risk: Overly restrictive policies may limit data utility and slow research.

4. Governance Action Scoring

Scoring scale:
1 = strongest alignment with the goal
2 = moderate alignment
3 = limited alignment
n/a = not directly applicable

Governance actionA: Non-malfeasanceB: ReliabilityC: EquityD: Privacy & biosafety
Action 1: Minimum validation standard2122
Action 2: “No personalized treatment claims” rule1222
Action 3: Open reporting standards + audit trail2122
Action 4: Equity incentive (diverse cell sourcing + fairness checks)2212
Action 5: Consent + privacy framework2221

Interpretation

  • Action 2 scores highest for non-malfeasance because it directly prevents patient harm from misinterpretation.
  • Actions 1 and 3 score highest for reliability, as they strengthen scientific rigor and reproducibility.
  • Action 4 scores highest for equity, ensuring broader biological representation.
  • Action 5 is strongest for privacy and biosafety, protecting donor autonomy and data governance.

5. Prioritization & Strategic Recommendation

Based on the scoring framework, I would prioritize a sequenced combination of governance actions, rather than a single intervention.

Priority 1 — Action 2: No Personalized Treatment Claims

The most immediate ethical risk is patient harm through misinterpretation.
Therefore, the first priority is enforcing strict limits on personalized treatment claims.

This ensures the organ-on-chip platform remains a research tool, not a premature clinical decision-making device. Preventing harm at the communication level is both low-cost and high-impact.

Priority 2 — Actions 1 and 3: Validation + Transparency

Once misuse risk is controlled, scientific integrity becomes the next priority.

  • Action 1 (Minimum validation standard) strengthens methodological rigor.
  • Action 3 (Open reporting and audit trails) strengthens reproducibility and accountability.

Together, they reduce hype, increase comparability across labs, and protect long-term credibility.

Priority 3 — Action 5: Consent & Privacy Framework

If human-derived materials are used, privacy and consent protections must be embedded early.
Trust is foundational in biomedical research, and weak governance at this stage could undermine the entire platform.

Priority 4 — Action 4: Equity Incentives (Scaling Phase)

Equity becomes especially important as the platform expands beyond proof-of-concept.
Diverse cell sourcing and subgroup analysis improve fairness, but also require more resources and coordination. For this reason, equity incentives should scale alongside platform maturity.

Trade-offs Considered

  • Speed vs. safety: Strong governance may slow deployment, but prevents harm and reputational damage.
  • Standardization vs. innovation: Validation standards could constrain experimentation, but improve reliability.
  • Equity vs. cost: Broader donor representation increases complexity and expense, but avoids structural bias.
  • Privacy vs. data richness: Limiting donor data protects autonomy but may reduce model detail.

Key Assumptions & Uncertainties

  • The chip meaningfully captures core endometriosis biology.
  • Validation benchmarks correlate with clinically relevant insights.
  • Clear communication boundaries reduce misuse.
  • Governance mechanisms will be adopted by journals, funders, and institutions.

This strategy prioritizes immediate harm prevention while building a foundation for scientific credibility, donor protection, and equitable scaling.

The following list contains the references used.

References

  1. Bendifallah, S. et al. (2024). Assessing the Utility of Artificial Intelligence in Endometriosis: Promises and Pitfalls. Journal of Nanobiotechnology, 22, Article 146.

    https://jnanobiotechnology.biomedcentral.com/articles/10.1186/s12951-024-02651-w

  2. Shen, Y. et al. (2021). Machine learning identifies important biomarkers for endometriosis using gene expression data. Scientific Reports, 11, 24688.

    https://www.nature.com/articles/s41598-021-04637-2

  3. Clinical Use of AI in Endometriosis: A Scoping Review.

    https://pmc.ncbi.nlm.nih.gov/articles/PMC11062212/

  4. Jovic, D. et al. (2025). Applications of AI and ML in diagnosing gynecological diseases: A cross-sectional review. npj Digital Medicine.

    https://www.nature.com/articles/s41746-025-01597-z

  5. Zhou, H. et al. (2023). Insights into transcriptomic signatures of endometriosis: A comparative analysis. BioSystems, 231, 104893.

    https://doi.org/10.1016/j.biosystems.2023.104893

  6. Turco, M. Y. & Giudice, L. C. (2021). Modeling human endometrial physiology in vitro. In: Human Reproductive and Prenatal Genetics, Academic Press, pp. 971–987.

    https://www.sciencedirect.com/science/article/abs/pii/B9780128178270000576

  7. OpenAI ChatGPT (2025). Language model assistance with writing, editing, and data interpretation. Accessed May 2025.

    https://chat.openai.com

  8. World Health Organization (WHO). Endometriosis – Fact Sheet.

    https://www.who.int/news-room/fact-sheets/detail/endometriosis

  9. NIH – National Library of Medicine. (2023). Endometriosis – StatPearls.

    https://www.ncbi.nlm.nih.gov/books/NBK519540/

  10. Aston University. (2024). Multiple factors delay timely endometriosis diagnosis, study shows.

    https://www.aston.ac.uk/latest-news/multiple-factors-delay-timely-endometriosis-diagnosis-study-shows

  11. Endometriosis Clinic UK. What is Endometriosis?

    https://www.endometriosisclinic.co.uk/copy-of-what-is-endometriosis

  12. McKinnon, B. et al. (2024). The genetics of endometriosis: Advances and future directions. Nature Genetics, 56, 277–287.

    https://www.nature.com/articles/s41588-023-01323-z

  13. Coccia, M. E. et al. (2024). Novel imaging approaches in endometriosis diagnosis. Best Practice & Research Clinical Obstetrics & Gynaecology, 95, 103317.

    https://www.sciencedirect.com/science/article/pii/S1471491424001667

  14. Zondervan, K. T. et al. (2020). Endometriosis. Nature Reviews Disease Primers, 6, Article 9.

    https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7352633/

  15. Lawrence, J. (2023). Painful progress: why endometriosis remains so difficult to diagnose and treat. Nature Spotlight on Women’s Health.

    https://www.nature.com/articles/d43747-023-00021-3