Organoid-AI platforms need integrated governance in drug discovery

Area of Science:

• Drug discovery and development
• Artificial intelligence in medicine
• Organoid technology

Background:

• Organoid-AI platforms are evolving into critical decision systems in drug discovery.
• Current governance practices often validate biological and computational models separately, despite their interdependence.
• This fragmented validation can lead to false assurance and obscure issues like donor imbalance, batch effects, and poor biological transportability.

Purpose of the Study:

• To propose a new framework for governing Organoid-AI platforms in drug discovery.
• To ensure the credibility and reliability of these platforms as preclinical gatekeepers.
• To address the limitations of separate validation for biological and computational models.

Main Methods:

• Developing a platform-level governance approach.
• Implementing a single context of use for validation.
• Establishing linked provenance tracking.
• Conducting transportability testing.
• Defining predefined fallback rules scaled to decision stakes.

Main Results:

• Separate validation of biological and computational models can create false assurance.
• Algorithmic shortcuts can arise from issues like donor imbalance and culture drift.
• Confident model outputs may mask weak biological transportability.
• A proportionate, platform-level governance can enhance credibility.

Conclusions:

• Integrated governance is essential for Organoid-AI platforms to function as reliable decision systems.
• Platform-level validation, considering the context of use and transportability, is crucial.
• This approach aims to increase adoption by building trust and ensuring credibility in preclinical drug discovery.