In 2021, the number of drug regulatory submissions to the U.S. Food and Drug Administration (FDA) based on machine learning increased about tenfold (132 submissions) compared to 2020. This serves as strong evidence that the introduction of artificial intelligence (AI) in drug development is accelerating.
Anthony Costello, CEO of Medidata, noted in an interview with ChosunBiz on Aug. 11 that the introduction of AI-based clinical trials is "not an option but a necessity."
AI is utilized in all phases of clinical trials, including trial design, data management, and trial execution, and it is essential for minimizing errors in the complex new drug development process and lowering the failure rate.
Founded in 1999 and headquartered in New York, Medidata provides cloud-based solutions for clinical trials. The company offers solutions necessary for every stage of clinical trials, from planning and design to completion. It established a branch in Korea in 2014.
Costello has 30 years of experience in clinical trials. He participated in global clinical trials for innovative cancer drugs Herceptin and Avastin at Genentech and co-founded the clinical data electronic collection startup NextTrials (now acquired by PRA Health Sciences) where he served as senior vice president for product development and virtual clinical trials.
He also co-founded Mytrus, a startup in the field of electronic consent, and joined Medidata following the company's acquisition of Mytrus in 2017. He was appointed CEO last year.
The following is a Q&A with him.
– What is the secret to Medidata becoming a powerhouse in clinical trial data?
Medidata has been at the forefront of converting data collection processes in clinical trials to electronic formats for over 25 years. Medidata's electronic data capture (EDC) platform holds the number one global market share in this field.
Thanks to operating the EDC platform for a long time, it has been able to create data collection and analysis platforms necessary for the entire clinical trial process, from new drug development and planning through execution to commercialization.
As a result, it has provided unparalleled insights from over 11 million patients and more than 36,000 clinical trials. More than 70% of the new drugs approved by the U.S. FDA in 2024 utilized Medidata technology.
Since 2019, Medidata has actively provided machine learning-based clinical trial data analysis and prediction services, adopting "AI Everywhere" as its top strategy.
– As a result of utilizing AI, how much has clinical efficiency improved?
Notable AI solutions from Medidata include Simulants, which generates synthetic data similar to actual data while protecting patient privacy, allowing for accurate decision-making in the design phase and identifying potential risks; Intelligent Trials, which predicts effective site selection during patient recruitment and operational phases and accelerates patient enrollment; and Synthetic Control Arm, which provides external control arms generated from past clinical trial data in disease areas where recruiting control groups is challenging. A recent product, "Clinical Data Studio," is AI-based and supports real-time quality control and risk management by integrating and analyzing clinical data collected from various sources.
AI capabilities have been embedded not only in individual products but also across the entire clinical trial process—from planning to outcomes—enabling anomaly detection, outcome prediction, optimized trial design, and patient targeting. As a result, study startup time can be reduced by up to 75% and study build time by as much as 80%.
The intersection of clinical trials and AI is naturally expanding into decentralized clinical trials (DCT). By collecting real-time data from sources such as remote monitoring, electronic consent (eConsent), and electronic clinical outcome assessments (eCOA), clinical trial efficiency is improved.
AI is essential for integrating, visualizing, and managing clinical data from multiple sources in a single environment, while reducing risks and supporting decision-making.
– Globally, clinical trials for cancer drugs are notoriously complicated in various ways.
"Cancer drug clinical trials are truly complex, and the protocols are often changed. The burden on patients participating in clinical trials is also significant. Medidata recently unveiled an AI-based solution, "Protocol Optimization," at the American Society of Clinical Oncology (ASCO) to improve complex clinical trial designs.
This solution offers functions such as AI-based predictive modeling, digital trial protocols, and simulation of trial performance before any patient is enrolled. This is expected to significantly reduce protocol changes and patient enrollment delays.
– What are some specific performance cases utilizing AI-based solutions?
A multinational pharmaceutical company, Eisai, is one of the first clients to adopt the AI-based "Clinical Data Studio." Utilizing this solution, the company has resolved the hidden obstacle of data silos inherent in AI and has perfectly integrated this solution into existing software to maintain quality and integrity across all data sources.
Another pharmaceutical company, Bristol-Myers Squibb (BMS), solved the problem of low registration rates in clinical trials through its predictive modeling solution, "Performance Analytics." As a result, it was able to shorten the new drug development timeline by over six months.
A case study of ovarian cancer clinical trials involving AbbVie gained attention due to its sophisticated analytical results. In the study conducted in collaboration with AbbVie and Medidata focusing on ovarian cancer patients, it revealed that the median progression-free survival (PFS) after treatment with a PARP (Poly ADP-ribose polymerase) inhibitor involved in DNA damage repair in cancer cells was approximately 6.11 months. It also confirmed that the combination therapy involving two drugs showed higher efficacy than monotherapy.
– What is the trend for utilizing AI analytical technologies in the pharmaceutical and biotechnology sectors?
In 2016 to 2017, there was about one submission per year. By 2021, the number of submissions reached 132. The number of AI-based FDA drug regulatory submissions in 2021 saw a nearly tenfold increase compared to 2020.
By 2030, AI is expected to be utilized in 60–70% of all clinical trials. The pharmaceutical industry estimates it could save up to $30 billion (about 41 trillion won) annually through reductions in clinical trial durations via AI.
– The application of AI in clinical trials is rapidly expanding across the Asia-Pacific region, with China actively adopting technology.
About 700 companies in Asia are actively utilizing AI throughout the entire process of new drug development. Notably, there are more than 100 AI-based pharmaceutical companies in China. In 2021, investments in AI technology for new drug development in China exceeded $1.26 billion (1.7434 trillion won), thanks to relaxed regulations encouraging AI-based clinical trials and therapeutic effects.
– It's been a year since you took office as CEO. What philosophy are you running the company on?
My leadership style is entrepreneurial. I often feel frustrated with bureaucracy. I always strive to exceed limits to move things along quickly. The clinical trial field is essentially slow and heavily regulated. Paradoxically, there is great satisfaction that comes from pursuing bold innovation.
When I worked at a startup, constant change was an essential virtue for survival. Toward that end, I fostered a culture that encourages transparent and open-minded thinking and actively derives the best ideas through lively discussions.
Recently, Medidata has attempted considerable changes in its internal strategies. A prime example is transitioning from a product-oriented approach to an experience-oriented approach in its operations and strategies.
Medidata has integrated its 35 product lines into three categories: Patient Experience, Data Experience, and Study Experience.
This transformation ensures that all our innovations start with the experience, aiming to improve the efficiency and accuracy of clinical trials with patient engagement, and optimized study design and execution.
– It seems that workforce restructuring or reduction is inevitable with the introduction of AI.
In fact, hiring is increasing. We are striving to recruit experts not only in computer science and data science but also in biotechnology and life sciences.
Experts in computer science and data science are expanding their work beyond simple programming to design and operate AI systems, understand AI ethics, and interpret complex data.
Experts in biotechnology and life sciences play an important role in effectively applying AI solutions to clinical trials. Ensuring data quality, setting the learning direction for models, and contextualizing various materials ultimately rely on people. The role of people is also vital in ensuring compliance with ethical standards.
To stay ahead in the fierce competition for talent, Medidata offers over 200 training courses and certification programs, as well as global mentoring opportunities to facilitate the growth and mobility of its internal talent and recruit newcomers.
– Can you share examples of collaboration with Korean pharmaceutical companies?
Since establishing its Korean branch, Medidata has collaborated with over 150 corporations, including the top 10 pharmaceutical companies in Korea. For instance, Yuhan Corporation utilized Medidata's patient-centered electronic clinical outcome assessment (eCOA) in a Phase 1 clinical trial targeting allergy medications to enable efficient data collection, while JW Pharmaceutical implemented five clinical solutions, including the Rave Electronic Data Capture (EDC) platform, in a global Phase 3 clinical trial for gout treatments.
– While Korean companies are actively entering new drug development, there have not been many achievements yet.
Korea consistently ranks among the top 5-10 countries globally for the number of clinical trial registrations each year, based on its excellent execution capabilities, quality, and hospital infrastructure. The problem is that the adoption of digital technology in clinical trials is still in its early stages.
Only about 6.4% of multinational clinical trials in Korea incorporate Decentralized Clinical Trial (DCT) elements. The utilization rate of AI is also low; as of 2023, out of 104 AI-integrated pipelines, only 7 cases have applied AI to clinical trials, with most focused on the early stages of new drug development.
As the Asia-Pacific market grows, Korea must strengthen its competitiveness by building a new drug development ecosystem based on AI technology and data-driven innovation and by making various regulations more ecosystem-friendly.
– What advice would you give to Korean pharmaceutical and biotech companies?
The success of clinical trials depend on two key factors: first, compliance with global regulations; second, risk management and data integrity using the latest technologies. For this, it is crucial to adhere to the recommendations of global regulatory agencies such as the FDA from the early stages.
Even if the research is aimed at technology transfer, the global pharmaceutical companies receiving the technology also require data with quality and integrity that meet strict global standards, so preparation for this is essential.