Table of Contents
In the rapidly evolving life sciences landscape, artificial intelligence (AI) has continued to gain momentum as a potentially transformative tool to revolutionize everything from drug discovery to patient care. As industry leaders and decision-makers navigate these choppy waters, understanding the current and future use of AI in biopharma will be integral for manufacturers that are eager to increase market share with the potential of getting more efficacious treatments into patients’ hands faster. However, realizing these benefits requires a nuanced understanding of the various AI technologies, their applications, and the challenges the models themselves may present.
History of AI Models in Life Sciences
The first usage of AI in life sciences was in computational biology. The field has experienced rapid acceleration in recent years due to advances in machine learning techniques and increased data availability. Today, the life sciences industry employs a diverse array of AI models, each with its own strengths and complexities.
In the 1950s, foundational concepts like the Turing Test and the first artificial neural networks laid the groundwork for AI. The 1960s and 1970s saw the development of one of the first artificial neural networks, the Adaptive Linear Neuron (ADALINE), as well as a better understanding of single layer neural network limitations. Significant advances in deep learning were introduced in the 1980s with early forms of cognitive computing, model-free reinforced learning, and backpropagation in neural networks. By the 1990s, the first practical applications of AI in drug discovery began to emerge.
The 2000s marked the rise of deep learning models, with significant milestones like the development of convolutional neural networks (CNNs) and recurrent neural networks (RNNs). The 2010 and 2020s saw the introduction of generative adversarial networks (GANs), transformer-based models, and the revolution of natural language processing.
Model Differentiation
The AI ecosystem in life sciences encompasses a diverse portfolio of technologies, each with its own risk-reward profile. At the foundation are traditional models such as decision trees and support vector machines (SVMs). These tools, while less glamorous than their more advanced counterparts, offer robust solutions for structured data analysis and often serve as the backbone for many biotech and pharma applications due to their interpretability and regulatory compliance.
Moving up the value chain, we encounter more sophisticated AI solutions like deep learning models. Neural networks, for example, have the ability to process complex high-dimensional data and are driving breakthroughs in areas such as image analysis and genomic sequencing. The investment in these advanced technologies is significantly higher but hold the promise of yielding more attractive return multiples in the form of accelerated research timelines and novel insights.
At the cutting edge of AI innovation, we find generative models and natural language processing systems. These technologies are opening new frontiers in drug discovery and literature analysis, potentially reshaping the R&D landscape. However, as Christopher Wallace said, increased complexity carries with it increased challenges, requiring careful cost-benefit analysis before deployment.
The complexity of these models varies significantly, reflecting a spectrum of capabilities and challenges. On one end, we have relatively simple algorithms like linear regression, with complexity ratings and at the other extreme, deep learning and generative models. This increasing complexity often correlates with more powerful capabilities but also brings challenges in interpretability and computational demands.
Current Applications in Life Sciences
AI is currently being applied across the entire life sciences value chain, significantly impacting workflow in drug discovery and design, development, operations, commercial activities, and beyond. The appetite for the application within healthcare is clear, $2.8 billion has been invested into the space in YTD with an estimate of $11.1 billion by the end of 2024. Additionally, 42% of the total investment deal volume from life sciences and healthcare is being directed towards the utilization of AI in some fashion, and it’s easy to understand why. R&D executives aim to accelerate the speed to market for treatments addressing unmet patient needs, however, the consistently high and rising costs of R&D are also front of mind. With an average development time over a decade and costs reaching well over a billion dollars per drug for top biopharma companies, the return on investment has decreased significantly, highlighting the need for more efficient processes. By implementing comprehensive digital transformations and utilizing AI in combination with other technological tools, biopharma and device manufacturers have the potential to significantly enhance the efficiency of the current bench-to-bedside process.
Drug Discovery
AI driven methods in drug discovery have the potential to streamline and enhance the identification, refinement, and design of new therapeutic candidates; significantly accelerating the drug discovery process and ultimately reducing costs and increasing the return on investment.
AI systems have the ability to process and analyze a wide array of data types, including genetic information, proteomic profiles, and clinical records. By leveraging these diverse datasets, algorithms can identify potential therapeutic targets that may have been missed or too difficult to discern through traditional research methods. Thus, providing researchers with the ability to uncover disease-associated targets and molecular pathways that play a pivotal role in the development and progression of various medical conditions.
Scientists can utilize reinforcement learning and generative models to aid in the design of innovative medications that can effectively modulate biological processes. By targeting specific molecular pathways or disease-associated targets, these AI-designed drugs may have the ability to deliver more precise and potent therapeutic effects. This targeted approach not only enhances the likelihood of developing successful treatments but also minimizes the risk of adverse side effects, as the medications are engineered to interact with specific biological mechanisms rather than broadly impacting the entire system.
The use of AI in drug optimization is most often in processing data related to the chemical structure, biological activity, and toxicological profiles of potential molecules. Algorithms can fine-tune chemical structure and properties of therapeutic molecules to enhance their effectiveness. This may involve modifying functional groups, adjusting molecular weight, or improving solubility to ensure optimal drug absorption and distribution. Moreover, AI can predict potential side effects by analyzing the interactions between the drug candidate and various biological targets. In addition to efficacy and safety, AI algorithms can analyze pharmacokinetic properties of drug candidates, including factors such as absorption, distribution, metabolism, and excretion. By predicting how the drug will behave in the body, AI can guide the design of molecules with improved bioavailability, longer half-life, and reduced toxicity. This optimization ensures that the drug reaches its intended target at the appropriate concentration and duration, maximizing its therapeutic potential.
As previously mentioned, there are many other areas within drug discovery that AI could be leveraged at this time including virtual screening, QSAR modeling, drug repurposing, combination therapy analysis, dosage and delivery design, nanomedicines, and many others. Currently, there are companies who are using some or parts of these different technologies, however, there are many more that have yet to realize the potential impact.
Clinical Trials
The conventional linear, randomized clinical trial process is resource-intensive, intricate, and heavily regulated. Despite the potential of digital technologies, automation tools, and patient-experience solutions to reduce manual tasks and overall cycle time and cost, their impact on clinical trial productivity has been incremental rather than transformative.
BioPharma companies are progressively employing technology-enabled approaches that harness R&D data to guide decision-making. Consequently, the volume of data generated during clinical trials is growing exponentially, with Phase III trials producing an average of 3.6 million data points in 2021, triple the amount collected a decade ago. This is precisely where implementation of generative AI to process and learn from enormous volumes of structured and unstructured data could be a significant value add for biopharma and CRO companies alike.
One of the key areas where AI can have a significant positive impact is patient recruitment. Clinical trial recruitment is often a time-consuming and labor-intensive process, requiring the manual review of electronic health records and patient interviews. However, with the advancement of natural language processing techniques, AI systems can efficiently process complex electronic health record data, identifying potential trial participants more quickly and accurately. Studies have shown promising results in reducing the time to recruitment and minimizing the workload for clinical trial designers. A specific focused area within patient recruitment that has been challenging for life sciences companies has been how best to manage a patient centric approaches to recruitment and retention, while also adhering to development timelines and cost requirements. Large language model-based applications that facilitate conversational interfaces with patients can provide timely answers to questions about studies at critical junctures, such as during recruitment or consent processes. This can lead to improved engagement and compliance among participants
AI can also play a crucial role in optimizing clinical trial design and cohort selection. By analyzing vast amounts of data from previous trials and real-world evidence, AI algorithms can help identify the most suitable patient subgroups for a given intervention, ensuring a more targeted and efficient trial design. This approach can lead to smaller, more focused trials that require fewer participants, ultimately reducing costs and accelerating the development process.
Moreover, AI can enhance data management and analysis in clinical trials. With the increasing complexity and volume of data generated during trials, traditional data management methods can be cumbersome and prone to errors. AI-powered tools can automate data capture, validation, and analysis, ensuring higher data quality and reducing the risk of human error. Machine learning algorithms can also identify patterns and insights that may be overlooked by human analysts, leading to more accurate and comprehensive trial results.
Another area where AI can make a significant impact is in the translation of clinical trial findings into clinical practice. By leveraging risk stratification techniques, AI can help identify patient subgroups that are most likely to benefit from a particular intervention, enabling more targeted and personalized treatment approaches. This can improve the efficiency and effectiveness of healthcare delivery, ensuring that the right therapies are delivered to the right patients at the right time.
Furthermore, AI can contribute to the development of novel clinical trial designs, such as basket, umbrella, and adaptive enrichment strategies. These innovative designs allow for the testing of targeted therapeutics across multiple patient subpopulations or tumor types, enabling a more efficient and flexible approach to drug development. AI algorithms can support the implementation of these designs by identifying suitable patient subgroups, adapting trial parameters in real-time, and analyzing the complex data generated by these trials.
As AI technologies continue to evolve and mature, their integration into clinical trials is expected to become increasingly prevalent, ultimately leading to faster, more efficient, and more effective drug development for the benefit of patients worldwide.
Supply Chain
The integration of AI into the biopharma supply chain has already made an impact on the industry; however further adoption and openness to innovation will inevitably enhance efficiencies, reduce costs, and ensure the integrity of sensitive medications.
One of the key areas where AI can have a significant positive impact is demand forecasting. AI algorithms can analyze vast amounts of historical data, market trends, and other relevant factors to generate accurate demand predictions. This enables biopharma companies to better align their production and inventory levels with anticipated demand, reducing the risk of stockouts or overstocking. AI-powered demand forecasting can also help companies respond more effectively to sudden changes in market conditions or customer requirements, enhancing their agility and responsiveness.
AI can also streamline inventory management processes, ensuring optimal stock levels and minimizing waste. By continuously monitoring inventory levels and analyzing data from various sources, AI systems can provide real-time insights and recommendations for inventory optimization. This can help biopharma companies reduce carrying costs, improve inventory turnover, and minimize the risk of expired or obsolete products. AI-driven inventory management can also facilitate better collaboration and information sharing among different stakeholders in the supply chain, enabling more efficient and coordinated operations.
In addition, AI can optimize logistics and distribution processes, ensuring the timely and efficient delivery of biopharma products to customers. AI algorithms can analyze transportation networks, weather patterns, and other factors to identify the most optimal routes and delivery schedules. This can help reduce transportation costs, improve delivery times, and enhance customer satisfaction. AI can also facilitate the tracking and monitoring of shipments in real-time, providing greater visibility and control over the supply chain.
Furthermore, AI can play a crucial role in ensuring the integrity and quality of biopharma products throughout the supply chain. By leveraging AI-powered quality control systems, companies can detect and prevent potential quality issues, such as contamination or degradation, in real-time. A tangential and added benefit is the identification and management of counterfeit drugs, which pose a significant threat to patient safety and supply chain integrity. By leveraging machine learning algorithms and computer vision techniques, AI systems can detect and identify counterfeit medications based on their packaging, labeling, or chemical composition. This helps prevent the infiltration of fake drugs into the legitimate supply chain, protecting patients and maintaining the trust in pharmaceutical products.
Hype vs. Reality
AI has shown tremendous promise; however, it's crucial to distinguish between what is real and what is vapor. The exaggeration of AI claims in the market stems from various factors; the oversimplification of complex biological systems, limitations in data quality and quantity, regulatory hurdles that slow implementation, integration challenges with existing workflows, and ethical concerns surrounding data privacy and AI use in healthcare.
These challenges can be observed across the board on a daily basis by the excessive use of AI buzzwords across media, investor pitch materials, and corporate literature. Further complicating the situation is that a number of the models in use today which tout an AI foundation instead leverage predefined rules, various statistical methods, or manual processes to achieve results without the actual support of artificial intelligence or machine learning. The goal of this statement is not to undermine one model type or another, it’s simply to highlight the need for a more measured and realistic approach to AI in the life sciences sector, focusing on validated results and practical applications, as well as increased transparency and clarity to treatment manufacturers, investors, and key decision makers regarding what application differentiation.
Future Developments and Opportunities
As it has since the 1950s, artificial intelligence and its application will continue to advance, offering new opportunities for improved patient care and enhanced therapeutic performance.
The emergence of quantum computing also holds immense potential for the pharmaceutical and biotech industries. Quantum computers can perform complex calculations and simulations that are beyond the capabilities of classical computers. In the context of drug discovery, quantum computing could enable the rapid exploration of vast chemical spaces, identifying novel drug candidates that may have been overlooked by traditional methods.
One promising development is the integration of edge computing with AI-enabled medical devices. Edge computing allows for data processing and analysis to occur directly on the device, reducing the need for constant communication with centralized servers. This decentralized approach can improve the responsiveness and reliability of AI-powered medical devices, enabling real-time decision-making and personalized interventions. For example, an AI-enabled wearable device equipped with edge computing capabilities could continuously monitor a patient's vital signs and provide immediate feedback or alerts without relying on a stable internet connection. AI-powered medical devices are expected to become increasingly sophisticated and autonomous. With advancements in machine learning and sensor technology, medical devices could adapt to individual patient needs and provide personalized treatments. For instance, an AI-enabled insulin pump could continuously monitor a patient's blood glucose levels and automatically adjust insulin delivery based on real-time data and predictive algorithms. Such devices could improve patient outcomes, reduce the burden on healthcare providers, and enhance the overall quality of care.
The View from the Crow’s Nest
With all of this in mind, it still remains important to consider the integration of human intelligence alongside AI to address societal issues and ensure responsible innovation in biomedical engineering and medicine. Collaborative efforts between industry stakeholders, regulators, and academic institutions will be essential in establishing best practices and standards for the development, validation, and deployment of AI technologies in these domains.
As we move forward, the companies that will thrive are those that can effectively harness AI to not only optimize current processes but also to unlock new business models and revenue streams. The convergence of AI and life sciences is not just a technological shift—it's a fundamental reimagining of how we approach health and medicine. For forward-thinking organizations, this represents an unparalleled opportunity to shape the future of the industry and deliver unprecedented value to patients and stakeholders alike.
If you are interested in learning more, get in touch at strategy@spinnakerLS.com.
Spinnaker offers true partnership and comprehensive guidance to help leaders navigate the complexities of the Life Sciences industry and chart a path to success. From early-stage market assessment through commercial execution and ongoing lifecycle management, we deliver tailored solutions to ensure optimized practicable results.
Sources:
Revolutionizing Drug Formulation: Harnessing Artificial Intelligence and Machine Learning for Enhanced Stability, Formulation Optimization, and Accelerated Development, A. Dangeti et al, International Journal of Pharmaceutical Sciences and Medicine, 2023
Harnessing the Power of AI in Pharmacokinetics and Pharmacodynamics: A Comprehensive Review, V. Pawar et al, International Journal of Pharmaceutical Quality Assurance, 2023
Tufts Center for the Study of Drug Development (http://csdd.tufts.edu)
The role of artificial intelligence in hastening time to recruitment in clinical trials. Ismail A. et al, BJR Open. 2023
A Scalable AI Approach for Clinical Trial Cohort Optimization, X. Liu, 2021
Improving Clinical Trial Efficiency: Thinking outside the Box, Mandrekar SJ et al, Am Soc Clin Oncol Educ Book. 2015
Leveraging Meta Data Management: Powering Cognitive Automation in Clinical Trial Processes, 2018
Integrating clinical trial findings into practice through risk stratification: the case of heart failure management, Smith DH et al, Popul Health Manag, 2010
Acquiring supply chain agility through information technology capability: the role of demand forecasting in retail industry, B Bai et al, Kybernetes, 2022
AI - Driven Supply Chain Optimization: Enhancing Inventory Management, Demand Forecasting, and Logistics within ERP Systems, S. Choudhuri et al, International Journal of Science and Research, 2024
Enhancing manufacturing productivity: A review of AI-Driven supply chain management optimization and ERP systems integration, O. Adenekan et al, International Journal of Management & Entrepreneurship Research, 2024
Transforming Next-Gen Therapy Supply Chains Into Patient-Connected Value Chains, L. Pernenkil et al, 2019
Blockchain and AI-based Solutions to Combat Coronavirus (COVID-19)-like Epidemics: A Survey
Manipulating Supply Chain Demand Forecasting With Targeted Poisoning Attacks, J. Chen et al, IEEE Transactions on Industrial Informatics, 2023
AI Isn't the Magic Bullet to Simplify Drug Discovery, N. Narain et al, genengnews.com, 2024
Generative AI in the pharmaceutical industry: Moving from hype to reality, B. Shah et al, 2024
Two-thirds of top 20 pharmas have banned ChatGPT: report, fiercepharma.com
Emerging technology trends in Pharmaceutical industry, G. Patel et al, World Journal of Biology Pharmacy and Health Sciences, 2024
Patient-centric smart health-care systems for handling COVID-19 variants and future pandemics: technological review, research challenges, and future directions, A. Kumar et al, The Internet of Medical Things: Enabling technologies and emerging applications, 2021
Healthcare and the Internet of Medical Things: Applications, Trends, Key Challenges, and Proposed Resolutions, I. Al Khatib et al, Informatics, 2024
The future of healthcare in Singapore. How an integrated use of A.I., Internet-of-Medical things IoMT, Blockchain-based technologies, and Cloud-computing-based Medtech and Digital Health solutions will radically address medical data integrity concerns, J. Ow et al, SSRN Electronic Journal, 2021
Integrating Artificial and Human Intelligence: A Partnership for Responsible Innovation in Biomedical Engineering and Medicine, Dzobo K et al, OMICS, 2020