메뉴보기

KOREAN

ABOUT

PRINCIPAL INVESTIGATOR

PEOPLE

RESEARCH

PUBLICATIONS

NEWS

SEARCH

    HOME

    OVERVIEW

    • up
    • down

    Use of Artificial Intelligence for Clinical Drug Development

    • Artificial Intelligence (AI) is a discipline of computer science that studies the ways to mimic and reproduce human intelligence processes such as learning, knowledge representation, decision-making and reasoning by machines. Several AI-based approaches have been applied to drug discovery and development to increase the efficiency while reducing the time and cost, resulting in a mix of success and failure.

      Clinical trials are an important research tool to determine the safety and efficacy of the drug in humans, which s an indispensable knowledge for physicians and patients for the best and optimal care. With the advent of digital health care technology, the fragmented nature of clinical data capture in the previous era is about to change and the way that clinical trials are performed can be also revolutionized. In particular, digital healthcare technology implemented in wearable devices and smart phones has enabled uninterrupted continuous data collection from patients in a real-life setting.



      Furthermore, there has been a growing demand for customized, user-friendly digital healthcare platforms that serve each patient’s specialized needs to identify and locate a clinical trial [s]he may be eligible for. Patient recruitment is one of the most expensive, time-consuming, and inefficient steps in any clinical trials. What makes the matter worse is that eligibility assessment is manually conducted by humans (i.e., physicians or study coordinators) d on eyeballing of large amount of patient records and related information. To overcome this drawback, AI-based deep learning algorithms can help match eligible patients to a specific clinical trial and recruiting them into it. Therefore, AI can turn lengthy, laborious, and complex procedures of patient eligibility assessment into several quick and easy clicks on a machine-learning system.



      Likewise, it is almost impossible for an average patient, who is lacking the domain knowledge, to find a list of potential clinical trials that [s]he might be eligible for without his/her physician’s help. A simpler, but more efficient, way may be to leverage the utility of AI to identify an appropriate list of clinical trials d on patient’s diagnosis, disease stage, severity and conditions, geographical location, and even personal preferences. Chatbot and the voice-activated assistant can play an important role too in this process.

      CCADD has focused on the application of AI technology to clinical drug development, particularly the operational aspects of clinical trials such as eligibility assessment and patient recruit. With the advance of information and communication technology, AI-approach can increase the efficiency of patient recruitment process by providing automated eligibility screening. To analyze heterogeneous patient data, a variety of machine learning algorithms can be used to assess whether a certain patient meets the eligibility criteria, on his/her age, gender, stage of disease, medical history, and clinical conditions. Additionally, dynamic deep neural network can be also used to select clinical features from the electronic medical records (EMR) for eligibility screening. Topic modeling is another helpful tool.

      To enhance EMR usage for clinical trials, the two most important clinical trials' resources - information on the potential pool and clinical trials being conducted (or have been conducted) - should be fully integrated. By the year 2018, CCADD has successfully developed a clinical trial resource integration system named 'AI-based Clinical Trial Resource Information System' or ACTRiS. What makes ACTRiS unique is its active employment of state-of-the-art user interface (UI) & user experience (UX) technologies, and implementation fo AI technologies. This integrated system will provide a sound basis to design clinical trials that are feasible and practical to perform.


      CCADD is currently working on a research project named 'A Dimensionality Reduction Model to increase the efficiency and accuracy of clinical trial feasibility assessment using electronic medical records'. This research aimed to develop a machine learning-basedimensionality reduction model to select the discriminant subset of eligibility features, which 

      adequately returns a sufficient number of eligible patients. This algorithm has the potential to significantly increase the efficiency and performance standard of the traditional approach for patient eligibility screening, contributing to better and more economic conduct of clinical trials. 


      Also, since spring 2019, CCADD has participated in the three-year project named 'Development an AI-model to predict and evaluate drug-basedrug interactions (DDIs)'. In this project, CCADD is responsible for collecting and curating drug-food interaction information (DFI) from publicly available research papers. Now, we are developing NLP models that recognize name entities of drug and food and classify whether a sentence in an abstract of scientific literature contains a valid DFI information or not. Also, we plan to validate a developed system by verifying whether predicted drug-basedrug pairs as having DDI cause some meaningful change of safety and efficacy of victim drug using a Common Data Model (CDM) of SNUH.

    Systems Pharmacology and Pharmacometrics

    Systems pharmacology is an innovative approach to predict in vivo drug effects, in which biological networks rather than a single transduction pathway are viewed as the basis of drug action and disease progression. Likewise, pharmacometrics is the science of developing and applying mathematical and statistical methods to describe, characterize, understand, and predict a drug’s pharmacokinetic (PK) and pharmacodynamic (PD). When combined complementarily, systems pharmacology and pharmacometrics can be an indispensable tool to improve decision making processes in every phase and stage of drug development.  
     
    Pharmacokinetic/Pharmacodynamic (PK/PD) modeling using NONMEM program 
    Top-down approach

    - Describes the time course of individual patient exposure and response to a new treatment
    - PK is the study of how the body processes a drug and PD is the study of how the drug acts on the body
    - Population PK/PD models are used to determine how patient factors such as demographics, disease status and progression, and co-medication, might affect patient exposure to drug and their subsequent response

    Physiologically Based Pharmacokinetic (PBPK) and PK/PD modeling using SimCYP program
    Bottom-up approach

    - Simulation of parent drug and metabolite concentration-time profiles and prediction of volume of distribution based on the lipophilicity, ionization, protein and tissue composition data 
    - Prediction of the extent of metabolism-based drug-drug interactions
    - Simulation for virtual patient populations including North European Caucasians, Japanese, healthy volunteers (for virtual Phase I studies) as well as obese/morbidly obese individuals and patients with renal impairment (moderate or severe) and liver cirrhosis (Child-Pugh A, B or C)
    - A full PBPK model, developmental physiology and the ontogeny of drug elimination pathways allows prediction of PK in neonates, infants and children

    Hyun A Lee, PhD, an alumna of CCADD, has published a paper entitled ‘A physiologically-based pharmacokinetic model adequately predicted the human pharmacokinetic profiles of YH4808, a novel K+-competitive acid blocker’ in the European Journal of Pharmaceutical Sciences. In this paper, a simulation d on the human PBPK model indicated that the pH-dependent solubility of YH4808 could have resulted in the reduced exposure after multiple administration. 



    Policy Analysis for an Integrated Clinical Research System in Korea

    CCADD has done research on studying the strategic policy issues to efficiently and seamlessly integrate standalone clinical research systems in Korea such as drug- or disease-registries (public or private), government's open data, and hospital's electronic medical records (EMRs). This research was sponsored by the Korea National Enterprise for Clinical Trials (KoNECT). In this research, CCADD has first enumerated the available clinical trial resources in Korea followed by a systematic investigation of the current legal, societal, and cultural works that have prevented clinical trial systems being fully integrated. CCADD has also studied the following topics:


    - Current status of integration of EMRs in Korea

    - Technologies available for integration of public health data and EMRs

    - Availability of a metadata model for integrated clinical research systems in other countries

    - Roles and responsibilities of the centralized coordinating center for integrated clinical research systems


    CCADD performed an exhaustive and systematic search for various articles and papers that described the current status and critical obstacles in integrating EMR data to be used as a source of clinical trials, including data quality and healthcare stakeholder engagement. The result of this study was written as a white paper titled "임상시험자원 통합정보시스템 구축을 위한 기획연구: 병원간의 전자의무기록 표준화 통합을 위한 구조적 접근 ". In this paper, we suggest a number of potential solutions to address current challenges. 







    A New Statistical Method in Clinical Trials with Biosimilars

    A biosimilar is a biological product that is similar to a reference biologic in efficacy, safety, and immunogenicity, albeit minor differences deemed clinically not important. The U.S. Food and Drug Administration (FDA) defines biosimilar as “a biological product that is highly similar to and has no clinically meaningful differences from an existing FDA-approved reference product.” With several biologics fast approaching their patents expiration, many biosimilars are currently under development. 

    Generics of small-molecule drugs are chemically synthesized (i.e., made) using the same active ingredients as the reference drug. However, biologics, whether reference or biosimilar, are derived (i.e., grown) from living cells or organisms such as bacteria, animals, and even humans, frequently through the use of recombinant DNA technology. The generic is likely to behave almost the same way as the reference drug as long as bioequivalence is shown. However, even minor modifications in the manufacturing process can introduce unintended changes, often adversarial, to the final biological product. These changes may negatively affect the clinical safety and effectiveness of biosimilars. 

    Because manufacturing processes for biologics are inherently complex and complicated, it is still challenging to fully characterize structural and functional properties of biologics. Furthermore, it is even intricate to define similarity of two biologics d on physicochemical and biological attributes. Instead, clinical trials are playing a critical role to detect any clinically significant differences in efficacy, safety, and immunogenicity between the biosimilar and a reference biologic. Therefore, scientifically and clinically valid statistical methods are to be applied to the design, conduct, analysis, and report of clinical trials with biosimilars to assess their equivalence to and interchangeability with a reference product.

    The equivalence margin is the largest difference that is clinically acceptable between the test (i.e., experimental) drug and the active control (i.e., reference) drug (Lee, 2012). It is important to accurately set the equivalence margin because the consumer risk that mistakes biosimilars for erroneously declaring equivalence when in fact it is not should be minimized in the approval of any biosimilar products. Based on this understanding, CCADD has developed a new statistical methodology to appropriately derive the equivalence margin in hypothesis testing between the biosimilar and a reference biologic. The core principle is similar to the calculation of the non-inferiority margin recommended by the U.S. FDA and two-step fixed-margin approach. The first margin is estimated using the historical data obtained from mostly the placebo-controlled trials of the reference drug. The second margin is then set by narrowing the first margin d on clinical judgement. 




    Pharmacoepigenetics/Pharmacoepigenomics Study

    Epigenetics is defined as “the study of changes in organisms caused by modification of gene rather than alteration of the genetic code itself.” The epigenetic study investigates changes in gene function that is explained without modification in the sequence of nucleic acids. The two most important epigenetic changes are DNA methylation and histone modification, both of which can influence how genes are expressed without changing the DNA sequences.

    Pharmacoepigenetics and pharmacoepigenomics aim to elucidate the mechanism of epigenenetic modifications affecting the of a drug and its response through changes in the pharmacokinetics and pharmacodynamics. The genotype varies between individuals, whereas the epigenetic patterns varies in five comparable states: cell type, inter-individuals difference, aging, exercise, and drug intake.






    CCADD focuses on studying the sensitive pharmacoepigenetic regions for drug intake, which may lead to an identification of optimal epigenetic state to maximize the efficacy of a drug while minimizing the likelihood of developing adverse drug reactions. To this end, machine learning (ML) technology has been used to identify susceptible regions for DNA methylation. For example, CCADD has tried to amalgamate pharmacoepigenetics/pharmacoepigenomics and ML technology to find differently methylated regions in genes that encode drug-metabolizing enzymes.

    Dr. Jeong-An Gim has published a paper ' A Machine Learning-Based Identification of Genes Affecting the Pharmacokinetics of Tacrolimus Using the DMETTM Plus Platform' in International Journal of Molecular Sciences. In this paper, genetic variants to predict tacrolimus exposure was identified using machine learning algorithms. A decision tree coupled with random forest analysis has been found to be an efficient tool for predicting the exposure to tacrolimus d on genotype.



    Writing an Introductory Book on Biologics

    Over the last decade, we have witnessed how technology innovations, such as smartphones and artificial intelligence, can quickly change our lives. The biopharmaceutical industry is not an exception. Since the human genome project was successfully completed, the cost of genetic analysis has rapidly decreased, making genetic analysis more affordable for routine patient care. Furthermore, personalized medicine combined with artificial intelligence technology shed bright light on the possibility of reanalyzing the relationship between patients and diseases.

    More extensive development and use of biological agents or biologics d on advanced medical technologies is one of those new trends. To support this notion, it is estimated that almost 50% of research expense by big pharmaceutical companies has been spent in the development of biologics. Biologics currently occupy approximately a quarter of the total pharmaceutical market in terms sales, which is on the sharp rise. For example, seven out of the top 10 best-selling drugs in 2019 were biologics.

    However, lay people know little about biologics. This is rather odd because Humulin, the first biopharmaceutical developed by Genentech, was approved by FDA more than 35 years ago (1982). Even OKT3, the first monoclonal antibody drug, was first approved in 1986. Their lack of appreciation on biologics may have something do with the fact that the development and production of biologics involves a variety of modern biological disciplines including molecular biology, to which lay people are not easily accessible.  

    Based on this understanding, we decided to write an introductory book on biologics that may help lay people understand the core principles of biologics and their follow-on drugs or biosimilars in association with their development, manufacturing, and regulatory implications. We hope that the readers will be able to understand the dynamic changes that are happening in the course of developing biologics and the marketplace. The book was finally published in November 2019.


    Microtracing/microdosing Study

    The best model for humans is human. Drug development has long been suffering from the lack of translatability and predictability of the preclinical animal experiments to humans. Dr. Richard Klausner, former Director of the US National Cancer Institute, commented in 1998, “the history of cancer research has been a history of curing cancer in the mouse. … and it simply didn’t work in humans.” Twenty years fast forward, drug development scientists are still struggling with the same issue.




    Microtracing/microdosing is an innovative technology that can revolutionize the current paradigm of clinical drug development. Typically, a very small amount of the drug, i.e., ‘microdose’, which is less than 100 micrograms (or 30 nmoles for proteins), is administered to humans. Since this is much smaller than 1/100 of the pharmacologically active dose, microtracing/microdosing technology can be employed at a very early stage of clinical drug development even when there is limited animal toxicology data. Furthermore, in order to trace minute doses, an accelerator mass spectrometer (AMS) is required and the compound should be labeled, typically with 14C. The microtracing/microdosing study allows clinical drug development scientists for generating the intravenous pharmacokinetics, mass balance, metabolite profiling, and absolute bioavailability data much easier, faster, and at a lower cost.  

    Dr. Howard Lee has spearheaded the employment of microtracing/microdosing technology for drug development in Korea since 2012, supported by three government grants. Dr. Lee has successfully d a new research ecosystem for the microtracing/microdosing study, which helped him perform the first microtracing/microdosing study under a full investigational new drug (IND) application in 2014. The result of this study was published at the most reputable journal in the field. Dr. Lee has conducted two additional microtracing/microdosing studies, and has recently won a two-year grant of $1 mil, in which microtracing/microdosing technology will be applied to a new drug development study with biologics.

    CCADD

    Center for Convergence Approaches in Drug Development, Graduate School of Convergence Science and Technology, Seoul National University

    Room C-208, 145 Gwanggyo-ro, Yeongtong-gu, Suwon-si, Gyeonggi-do, 16229, SOUTH KOREA (Gwanggyo)

    Room 406, Building 17, Seoul National University College of Medicine, 103 Daehak-ro, Jongno-gu, Seoul, SOUTH KOREA (Yeon-gun)

    Tel: +82-31-888-9189 (Gwanggyo); +82-2-3668-7381 (Yeon-gun)

    Fax: +82-31-888-9575

    Email: ccadd.snu@gmail.com