Scaling AI for Clinical Trials - with Damion Nero of Takeda

Unknown Source May 07, 2025 27 min

artificial-intelligence investment

🎧 Listen to Original

18 Companies

38 Key Quotes

2 Topics

1 Insights

🎯 Summary

Podcast Episode Summary: Scaling AI for Clinical Trials - with Damion Nero of Takeda

This 26-minute episode of the AI and Business Podcast features Damion Nero, Head of Data Science in US Medical at Takeda Pharmaceuticals, discussing the transformative role of AI and digital technologies in modernizing clinical trials and drug development. The core narrative focuses on the shift toward decentralized trial models, the systemic challenges hindering widespread AI adoption, and the strategic path forward for realizing tangible ROI.

---

1. Focus Area: AI in Pharmaceutical Clinical Trials and Drug Development. Key themes include the acceleration of decentralized clinical trials (DCTs), leveraging real-world data (RWD), precision medicine applications, and overcoming operational and regulatory hurdles for scaling AI solutions.

2. Key Technical Insights:
   • Decentralized Trial Acceleration: AI is critical for processing the influx of data from remote patient interactions (telehealth, electronic consent, chatbots for symptom tracking) inherent in decentralized trial models.
   • Computer Vision in Diagnostics: AI, particularly computer vision, is making significant strides in objective tumor diagnosis by rapidly screening and assessing imaging data, overcoming human subjectivity driven by emotional factors.
   • Data Integrity and Hallucination Risk: AI models trained on fractured, inconsistent historical site data risk “exaggerating” effects or generating misleading insights (hallucinations), necessitating strong human curation by data scientists.
3. Market/Investment Angle:
   • Focus on High-Volume Areas: AI impact is currently greatest in areas with large patient populations (e.g., oncology) where data volume can adequately train models, rather than in rare disease trials where data scarcity is a major constraint.
   • ROI Driven by Operational Efficiency: Initial investment justification relies on “low-hanging fruit”—demonstrating clear operational efficiencies and cost savings behind the scenes to secure leadership buy-in for more complex, outcome-driving AI projects.
   • Competitive Catalyst: Significant industry-wide adoption is anticipated once a competitor achieves a major, verifiable success with advanced AI, forcing others to follow suit to satisfy investor demands.
4. Notable Companies/People:
   • Damion Nero (Takeda): Guest expert providing an insider’s view on AI deployment, data challenges, and internal alignment within a major pharmaceutical company.
   • Takeda Pharmaceuticals: The context for the discussion, highlighting current strategies in data science and clinical operations.
   • Metable: Sponsor of the special series on AI and clinical workflow.
5. Regulatory/Policy Discussion:
   • Global Regulatory Stringency: The US FDA is considered relatively “looser” compared to regulatory bodies in the EU and Asia, which impose significantly higher standards, complicating the deployment of novel AI-driven trial materials globally.
   • Internal Alignment Challenge: Achieving consensus among clinical, regulatory, and legal teams regarding the use of AI-generated results is a major internal hurdle, often leading companies to keep AI applications strictly in administrative or background support roles.
   • Black Box Concerns: Reliance on third-party vendors for AI development often results in proprietary “black box” algorithms, which prevents internal teams from fully understanding or explaining the methodology to regulators.
6. Future Implications:
   • Discovery Space Growth: AI will likely see the fastest growth in the drug discovery phase (virtual simulations, candidate generation) as it is the least outward-facing and requires less immediate regulatory approval.
   • Virtualization as Gold Standard: The long-term goal is to move toward fully virtualized trials, minimizing in-person HCP interactions to reduce patient aversion and operational costs.
   • Shift to Proactive Systems: The ultimate trajectory involves moving beyond operational efficiencies to proactive systems that flag risks, draft submissions, and support decisions before issues arise, transforming clinical workflows from reactive to predictive.
7. Target Audience: Pharmaceutical Professionals, Clinical Operations Leaders, Data Science Executives in Healthcare, and Investors focused on HealthTech/Biotech. Professionals needing a realistic, candid assessment of AI scaling challenges in highly regulated environments will find this most valuable.

🏢 Companies Mentioned

So Damien ✅ unknown

The EU ✅ unknown

And I ✅ unknown

Now AI ✅ unknown

Affordable Care Act ✅ unknown

United States ✅ unknown

With AI ✅ unknown

But I ✅ unknown

So I ✅ unknown

Takeda Pharmaceuticals ✅ unknown

US Medical ✅ unknown

Data Science ✅ unknown

Damian Nero ✅ unknown

Emerge AI Research ✅ unknown

Editorial Director ✅ unknown

💬 Key Insights

"Takeda is working toward models that flag risks, draft submissions, and support decisions before challenges arise, shifting clinical workflows from reactive to predictive."

Impact Score: 10

"Beyond that, as we get more and more into decentralized trials and really just virtualizing a lot of that process as well, where really the interactions with HCPs are kept to maybe once during the entire trial, ideally or maybe never. That is really, I think, the gold standard that we're moving towards as well to get, move past any sort of patient aversion, reduce our costs as well..."

Impact Score: 10

"The EU much more organized, but also having a way higher standard when it comes to personal privacy, the risk categories for different kinds of AI systems than the United States, which is a far more piecemeal system."

Impact Score: 10

"We're not really in a position right now where there isn't A, enough trust even internally to do that and B, we could actually explain it well. Because also keep in mind, we're not building a lot of this in-house. A lot of this is being outsourced to third parties. And for many of them, this is a black box type of situation where they'll build it for us. Sure, they won't tell us how they did it. They'll say, it's a proprietary algorithm that we own and then that's it."

Impact Score: 10

"Something that isn't commonly talked about is that the US actually has some of the looser regulations. FDA is pretty stringent about the drug approval process. You go to the EU, you go to Asia, forget it. The standards are much, much higher there and there are a lot of requirements that they have."

Impact Score: 10

"AI tends to have a capacity to, I mean, for lack of a better term, exaggerate. In a sense that it will take a bit of information and turn it into something that is totally not the case."

Impact Score: 10

📊 Topics

#artificialintelligence 57 #investment 1

🧠 Key Takeaways