Transforming Regenerative Medicine Manufacturing with AI
How Autonomy Is Scaling Autologous Cell Therapies for Broader Patient Access
At the Precision Health Summit 2024, Matthias Wagner, Co-Founder and Chief Technology Officer of Cellino, presents a groundbreaking approach to manufacturing regenerative medicines using artificial intelligence (AI). With a distinguished career spanning manufacturing automation, machine vision, and photonic chip manufacturing, Wagner delves into how AI is poised to overcome the critical challenges of scalability and variability in producing autologous cell therapies.
Unlocking the Potential of Autologous Cell Therapies
Autologous cell therapies, which use a patient's own cells for treatment, hold immense promise for transforming healthcare. Diseases such as dry age-related macular degeneration (AMD), Parkinson's disease, and type 1 and type 2 diabetes could see revolutionary treatments through the replacement or supplementation of dysfunctional cells.
Wagner highlights a remarkable case from China, where an autologous induced pluripotent stem cell (iPSC) therapy enables a patient with severe type 1 diabetes to discontinue insulin injections after just 75 days. "Fat cells are harvested from the patient, reprogrammed into iPSCs, and then differentiated into beta islet cells capable of secreting insulin," Wagner explains. This case exemplifies the transformative potential of autologous therapies.
The Scalability Challenge: A Barrier to Accessibility
Despite their promise, autologous therapies face significant hurdles in scalability and cost. "These are extremely expensive to produce. It takes several months to make these cells from the source cells," Wagner notes. The current manufacturing process is labor-intensive, involving manual steps performed by highly trained experts, often at the PhD level. Each patient sample typically requires a dedicated clean room for the entire duration of cell processing, which significantly escalates costs and limits patient access.
This bottleneck has real-world implications. With first-generation cell therapies like CAR-T treatments approved in 2017, only about 20% of eligible patients in the U.S. receive them, and the global percentage is even lower. Wagner emphasizes, "We want to fix that problem for the next wave of cell therapies."
Beyond Automation: The Shift to Autonomy
A common misconception is that traditional automation can solve these challenges. Wagner dispels this notion by highlighting the inherent variability in cell cultures. "In reality, this process is extremely variable. There is a high degree of patient-to-patient variability, cell-to-cell variability," he says. This unpredictability renders one-size-fits-all automated processes ineffective.
He likens the manufacturing process to "driving off-road with unexpected obstacles" rather than running a train on fixed tracks. To navigate this complex landscape, Cellino is moving beyond automation to autonomy. "We combine two elements: a closed-loop control system that can steer the process and AI to have a generalized way to react to the dynamics of the cell culture," Wagner explains.
The AI-Powered Closed-Loop Control System
Cellino's innovative approach involves a closed-loop control system that integrates AI to manage the cell manufacturing process proactively. The system comprises several key components:
1. Daily Label-Free Imaging: The system performs daily imaging of the entire cell culture using transmission light imaging, avoiding the need for labels that are unsuitable for cells intended for patient use.
2. AI-Driven Cellular Mapping: Advanced AI algorithms generate detailed maps highlighting live cells, local cell density, and pluripotency levels. "We use AI to generate a series of maps," Wagner states, emphasizing the technology's ability to analyze and interpret complex cellular data.
3. Precision Laser Cell Removal: Guided by the AI-generated maps, a proprietary laser system precisely removes unwanted cells, thereby enriching the desired cell population. This targeted removal is crucial for maintaining healthy cell densities and ensuring the pluripotency of iPSCs.
4. Automated Liquid Handling and Incubation: Automation in these areas reduces manual labor and minimizes the risk of contamination, enhancing overall efficiency.
Addressing Safety Concerns with Laser Technology
Safety is paramount in any medical manufacturing process, and Wagner addresses potential concerns regarding the use of lasers. He assures that the laser light used passes straight through the cells without absorption. "The laser light, if directed at our cells, would pass straight through them—the cells don't absorb any of the light," he explains. Instead, the energy is absorbed by a specialized coating applied to the surface just beneath the cells. "This gives us very high precision of where the energy is absorbed and turned into mechanical energy of the bubbles," ensuring that only unwanted cells are removed without affecting healthy ones.
Enhancing Quality Control Through Predictive AI
Quality control is a critical aspect of cell therapy manufacturing. Wagner discusses how AI enhances this process. "We are predicting quality within our process," he says. While AI predictions do not replace standard quality control tests, they significantly improve efficiency and yield. "We're hoping that the in-process quality predictions give us a much higher yield through those QC tests," Wagner states. Whereas current industry QC pass rates hover around 10-20%, Cellino aims for 70-80% pass rates, drastically improving the scalability and cost-effectiveness of therapies.
领英推荐
Simulation and Reinforcement Learning: A New Frontier
Cellino has developed a simulation environment to further refine the manufacturing process. By simulating cell cultures and their variability, the company tests the effectiveness of management algorithms using the laser system. "We've trained a multi-agent reinforcement learning model that makes this essentially into a game," Wagner explains. The model learns to manage multiple cell colonies within a confined space, avoiding collisions and overgrowth, which is crucial for maintaining cell health and viability.
This simulation allows for "millions of hours of training in silico before transferring it to in vitro," accelerating development cycles and optimizing cell culture management strategies.-
Scaling Ambitions: From Hundreds to Millions
When questioned about the scalability of Cellino's solution, Wagner expresses ambitious yet achievable goals. "Even starting with being able to do hundreds of patients for Phase 3 type trials is a big win," he acknowledges. However, the vision extends far beyond. "We really would like to be able to produce cells for tens of thousands of patients from centralized facilities and are thinking about how to make a system that could be distributed geographically to serve hundreds of thousands, and ultimately millions of patients," he says.
A significant milestone in this journey is the recent $25 million funding from ARPA-H (Advanced Research Projects Agency for Health). This investment supports the development of a highly scalable infrastructure based on Cellino's optical bioprocessing technology, bringing the company closer to its goal of making autologous cell therapies widely accessible.
Navigating Multidisciplinary Challenges
Integrating multiple disciplines—AI, optics, biology, and automation—is inherently complex. Wagner candidly addresses these challenges: "Making sure that these groups can all communicate and act as one is a constant challenge." The interdisciplinary nature of the work requires robust collaboration and effective communication to ensure each component aligns seamlessly.
Beyond internal hurdles, Wagner highlights the need for industry-wide standardization. "We would really like to see more standardization around iPSCs as starting material," he emphasizes. Establishing agreed-upon quality metrics for iPSCs would make them an interchangeable part in the cell differentiation process, benefiting the entire industry by reducing variability and improving scalability.
Implications for the Healthcare Industry
Cellino's approach offers valuable insights for business leaders and professionals in the healthcare industry:
- Innovative Process Design: Embracing autonomy over automation can address variability in complex manufacturing processes.
- Interdisciplinary Collaboration: Success in cutting-edge fields necessitates integrating diverse expertise, highlighting the importance of fostering a collaborative organizational culture.
- Leveraging AI and Simulation: Utilizing AI and simulation technologies can accelerate development cycles, improve quality control, and enhance scalability.
- Advocating for Standardization: Pushing for industry-wide standards can reduce barriers to scalability and facilitate broader adoption of transformative therapies.
Pioneering the Future of Regenerative Medicine
Matthias Wagner's insights at the Precision Health Summit 2024 illuminate a promising path toward making regenerative medicines more accessible and affordable. By harnessing AI and developing autonomous systems, the critical challenges of variability and scalability in cell therapy manufacturing are being addressed head-on.
As the field progresses, the integration of AI in manufacturing processes is not just advantageous but essential. It represents a paradigm shift in how therapies are developed and scaled. "Our goal is to get these therapies to hundreds of thousands, and ultimately millions of patients," Wagner summarizes.
The journey ahead requires continued collaboration, standardization, and significant investment. However, the potential rewards—a substantial improvement in patient outcomes and a reduction in healthcare costs—make this endeavor of paramount importance.
>>> WATCH THE VIDEO OF THE SESSION HERE: https://1businessworld.com/precision-health-summit-hub/precision-regenerative-medicine-manufacturing-with-ai-matthias-wagner/