Asia‐Inclusive Clinical Research and Development Enabled by Translational Science and Quantitative Clinical Pharmacology: Toward a Culture That Challenges the Status Quo

Abstract

Access lag to innovative therapies in Asian populations continues to present a challenge to global health. Recent progressive changes in the global regulatory landscape, including newer guidelines, are enabling simultaneous global drug development and near‐simultaneous global drug registration. The International Conference on Harmonization (ICH) E17 guideline outlines general principles for the design and analysis of multiregional clinical trials (MRCTs). We posit that translational research and quantitative clinical pharmacology tools are core enablers for Asia‐inclusive global drug development aligned with ICH E17 principles. Assessment of ethnic sensitivity should be initiated early in the development lifecycle to inform the need for, and extent of, Asian phase I ethno‐bridging data. Relevant ethno‐bridging data may be generated as standalone Asian phase I trials, as part of Western First‐In‐Human trials, or under accelerated development settings as a lead‐in phase in an MRCT. Quantitative understanding of human clearance mechanisms and pharmacogenetic factors is vital to forecasting ethnic sensitivity in drug exposure using physiologically‐based pharmacokinetic models. Stratification factors to control heterogeneity in MRCTs can be identified by reverse translational research incorporating pharmacometric disease models and model‐based meta‐analyses. Because epidemiological variations can extend to the molecular level, quantitative systems pharmacology models may be useful in forecasting how molecular variation in therapeutic targets or pathway proteins across populations might impact treatment outcomes. Through prospective evaluation of conservation in drug‐ and disease‐related intrinsic and extrinsic factors, a pooled East Asian region can be implemented in Asia‐inclusive MRCTs to maximize efficiency in substantiating evidence of benefit‐risk for the region at‐large with a Totality of Evidence approach.

Evaluating the impact of race and ethnicity on variability in drug exposure and treatment response is a key component of clinical pharmacology planning in drug development. ¹ , ² , ³ A recent survey of drug approvals by the US Food and Drug Administration (FDA), indicated that 10% of new molecular entities approved between 2014 and 2019 showed a difference in exposure and/or response based on race/ethnicity or pharmacogenetic factors known to vary in frequency across global populations. ⁴ For the cohort of FDA approvals between 2008 and 2013, this percentage was 21%. ⁴ Drug development programs initiated in the Western Hemisphere have traditionally expanded to include Asian populations at later timepoints, applying principles outlined in the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH) E5 (ethnic bridging) guideline. This has led to delays in bringing new therapies to Asian regions. Adopted and in effect in its final form in 2018, the ICH E17 (multiregional clinical trials (MRCTs)) builds upon the principles of ICH E5 and provides general guidelines for the design and analysis of MRCTs, accounting for population sources of heterogeneity. Timely participation in MRCTs is vital to decreasing the delay in regional approval of drugs (i.e., country‐level drug lags), thereby decreasing the associated delay in access to innovative therapies (i.e., access lag). This is evident in the results of retrospective analyses of factors influencing drug lags in Japan and China. ⁵ , ⁶ , ⁷ In a recent assessment of drug lags in development initiation, New Drug Application, and approval between Japan and the United States, the cohort of 2016 to 2020 new drug approvals in Japan was analyzed. ⁶ MRCT‐based development was associated with shorter approval lag by ~ 3 years compared with strategies based on local trials. With the recent regulatory reform in China together with the country joining ICH as a full regulatory member, early participation in clinical development of novel investigational agents is now feasible due to substantial streamlining of regulatory and administrative processes. This further opens up opportunities for decreasing drug lag in China through strategic planning for inclusion of the region in the geographic footprint of pivotal MRCTs leveraging principles of the ICH E17. ⁸ Furthermore, a recent FDA guidance offers valuable points to consider for enhancing eligibility criteria and enabling more inclusive enrollment practices (including, but not limited to, race, ethnicity, and location of residency) in clinical trials. ⁹

We posit that this recent evolution of the global regulatory framework for clinical development is an opportunity for clinical pharmacologists to drive progress through robust scientific positioning of Asia‐inclusive development strategies leveraging the foundational principles of our scientific discipline. In this tutorial, we provide an overview of key enablers for ethnic sensitivity assessment, with a focus on strategies for the prospective application of translational science and quantitative clinical pharmacology tools within the broader context of model‐informed drug development (MIDD). A more holistic view of intrinsic and extrinsic sources of variability in drug disposition and in disease pathophysiology and response using a Totality of Evidence mindset would minimize the need for redundant regional clinical investigation. ¹⁰ , ¹¹ , ¹² In a Totality of Evidence approach, ¹⁰ evidence is substantiated through the confidence gained from consistency across multiple approaches and data sources integrated in a mechanism‐informed manner through modeling and simulation. Quantitative clinical pharmacology tools, such as population disease progression models, quantitative systems pharmacology (QSP) models, and physiologically‐based pharmacokinetic (PBPK) models are valuable in this context. This scientifically guided methodology would take into account recent changes in the worldwide regulatory landscape that emphasize a science‐driven, patient‐centric approach. This approach accounts for growing diversity in the Western region, core principles of the recent ICH E17 guideline for MRCTs, ¹³ , ¹⁴ recent accelerated development and approval measures in Japan, ¹⁵ and regulatory reform in China. ⁸ , ¹⁶

Asia is a heterogeneous continent. In the context of this tutorial, we discuss drug development considerations applicable to the eastern and northern regions of Asia (i.e., Chinese, Korean, and Japanese ethnic populations). This practical grouping of Japanese, Korean, and Chinese populations lends itself to a pragmatic approach for ethnic sensitivity evaluations and MRCT design. Reports of greater similarities than differences in the pharmacokinetics (PK), safety, and efficacy of drugs between these Asian subpopulations lend support to their holistic consideration as a pragmatic starting point for Asia‐inclusive global drug development while being cognizant of the possibility of subtle but potentially important differences between these subpopulations. ¹⁷ , ¹⁸ , ¹⁹ The HUGO Pan‐Asian SNP project has revealed marked genetic diversity between and within these Asian populations (e.g., within Chinese). ²⁰ , ²¹ Nevertheless, when viewed in the context of the multifactorial sources of variability in drug disposition, disease pathophysiology, and treatment response, such consolidation is a practically suitable approach but should be coupled with a commitment to continuously evaluate intrinsic and extrinsic sources of variability throughout the development life cycle.

ETHNO‐BRIDGING IN EARLY CLINICAL DEVELOPMENT

When a global drug development program is initiated in the Western Hemisphere, the need for, timing of, and extent/design of Asian phase I ethno‐bridging evaluations should be based on solid scientific rationale. Not all molecules/therapeutics should be treated the same. A science‐driven, case‐by‐case approach that is informed by ICH E5 scientific principles is crucial. For a complete list of ICH E5 factors associated with greater or lower risks for ethnic sensitivity, refer to Appendix D of the guideline document. ²² Later sections of this tutorial will provide an overview of opportunities for translational research and MIDD enablers for the assessment of ethnic sensitivity.

The following are some examples of useful considerations to guide the need for and design of standalone Asian phase I investigations (Figure  1 ) ²³ , ²⁴ :

• Known or expected sources of interethnic variation based on absorption, distribution, metabolism, and elimination (ADME) or PK,

• Evidence for interethnic variation in the pharmacologic target/mechanism of action,

• Safety profile (e.g., target organs) in the Western population suggestive of an increased risk in Asian populations,

• Narrow therapeutic index,

• Clinically meaningful ethnic sensitivity in safety or efficacy in the drug class under consideration.

Figure 1.

Determinants of ethnic sensitivity and associated uncertainty that inform the timing for assessment of Asian ethnic sensitivity. ADME, absorption, distribution, metabolism, and elimination; MOA, mechanism of action; PD, pharmacodynamic; PK, pharmacokinetic.

In a recent survey of MRCTs supporting the approval of drugs in Japan between 2007 and 2017, the approaches used to evaluate PK differences in Japanese vs. non‐Japanese populations were examined. ¹⁴ Approximately 25% of evaluated Japan‐inclusive MRCTs embedded PK characterization for ethnic sensitivity assessments in the MRCT without standalone phase I PK ethno‐bridging studies. Of note, the cases where dedicated evaluation of PK in the Japanese population was not conducted ahead of initiating a global MRCT were largely those where ethnic sensitivity was expected to be low (e.g., topical or intravenously administered drugs without first‐pass metabolism) and a minority of cases where the indication was a rare disease.

Balanced pragmatism and scientific rigor, in the context of proactive regulatory communications, must guide decisions regarding the content and timing of Asian phase I clinical studies (Figure  1 ).

Timing of Asian phase I ethno‐bridging evaluation

If the timing of inclusion of Asian populations in global MRCT(s) is intended to be at or following proof‐of‐concept (POC), it should suffice to initiate Asian phase I investigation in parallel after a suitable inflection point is reached (e.g., after completion of multiple‐dose safety and tolerability, pharmacodynamic (PD), and PK/PD characterization in support of the likely phase II dose range). This is depicted in scenario A of Figure  2 . However, if an Asia‐inclusive geographic footprint of phase II is desired because Asia is a region of focus or due to an accelerated global development strategy, an Asian phase I evaluation would need to be conducted earlier. This is depicted in scenario B (for a pivotal phase III) and scenario C (for a pivotal phase II) of Figure  2 . Ethno‐bridging data to support inclusion of Asian populations in an MRCT can be generated in a standalone Asian phase I study or through incorporation of Asian cohorts in the Western first‐in‐human (FIH) study. ²⁵ , ²⁶ For accelerated development programs (e.g., oncology), it may even be possible to consider generating such data in a safety/PK lead‐in phase within the first Asia‐inclusive MRCT. The lead‐in phase would specify a minimum number of patients for intensive PK sampling and close monitoring of safety. Review of emerging data from the safety lead‐in phase would inform the decision to trigger full expansion of enrollment of patients in the East Asian region at a common dose in the global pivotal study.

Figure 2.

Illustrative scenarios of timing of ethno‐bridging to enable Asia‐inclusive MRCTs. Scenario A depicts a setting of standard phase I, II, III development with an Asia‐inclusive phase III MRCT and conduct of Asian ethno‐bridging assessments in parallel with the Western phase II trial. Scenario B depicts an alternative scenario of Asia‐inclusive phase II MRCT with dose decision for the East Asian population based on an Asian phase I trial conducted during the latter half of the Western phase I program (e.g., after completion of SRD and in parallel with MRD cohorts). Scenario C depicts the generation of requisite data for assessment of ethnic sensitivity in the timeframe towards the end of the Western phase I trial to enable Asia‐inclusive globalization of a phase II pivotal trial under an accelerated development scenario (e.g., oncology). ^aIn cases where ethnic sensitivity is expected to be low based on ICH E5 principles, it should be possible to justify the Asian dose in an Asia‐inclusive MRCT without a dedicated Asian phase I dose‐finding trial. ^bAsian phase I data can be generated either in the context of an Asian phase I trial conducted in representative East Asian country(ies) or in healthy subjects of Asian race(s) residing in the West, or where appropriate under accelerated development settings, as a lead‐in/safety and PK run‐in phase in the first Asia‐inclusive MRCT. ^cInformative sparse PK sampling for population PK analyses in the MRCT will provide definitive PK in East Asian patient populations (e.g., Japanese and Chinese) in the respective country‐level regulatory filings. ICH, International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use; MRCT, multiregional clinical trial; MRD, multiple rising dose; Ph, phase; POC, proof‐of‐concept; PK, pharmacokinetics; Pop PK, population pharmacokinetics; SRD, single rising dose.

Design considerations for Asian phase I ethno‐bridging assessments in healthy volunteers

When a standalone Asian phase I ethno‐bridging study is part of the strategy and the clinical pharmacology of the molecule can be characterized in healthy volunteers, one important design consideration is whether to perform this study in healthy volunteers enrolled in Asia or to enroll Asian populations outside of Asia as part of ongoing Western clinical development. Traditionally, standalone Japanese phase I studies or Japanese cohorts within Western FIH studies have provided representative East Asian phase I data. Selection of the Japanese population as a representative one is based on typical regulatory expectations by the Pharmaceuticals and Medical Devices Agency (PMDA) for clinical ethno‐bridging data ahead of enrollment in Japan in larger phase II/III trials. Incorporation of representative Asian population(s) as cohort(s) in a healthy volunteer FIH study conducted outside of Asia (e.g., United States) could be more efficient, as this could be achieved without additional clinical trial application filings or supply chain considerations. In contrast, generation of Asian phase I data in an Asian country or countries provides beneficial early experience, which can potentially benefit longer‐term success of Asia‐inclusive development and fast‐to‐registration strategies and provide some degree of trust for larger trials in the region. In addition, generation of phase I data in Asia accounts for the unlikely but potential impact of extrinsic factors. ²⁷ These ethno‐bridging evaluations in Asian populations can also be designed as pan‐Asian studies open to subjects of any major East Asian population through conduct at phase I site(s) with access to such volunteer populations. The value of conducting Pan‐Asian phase I MRCTs engaging a research network of East Asian investigators with expertise in clinical pharmacology and ethno‐bridging science (e.g., Asia Clinical Pharmacology Network) has been discussed. ²⁸ A pan‐Asian ethno‐bridging evaluation (as opposed to only evaluating a single representative Asian population, such as Japanese) has the added advantage of supporting data‐driven positions regarding consistency in PK/PD, safety, and selected dose for the East Asian region at large, thereby more confidently enabling phase II/III MRCT designs, including prospective definition of a pooled East Asian region based on ICH E17 principles. ¹³ With increasing opportunities for Asia‐inclusive phase I MRCTs and efficiencies in regulatory processes in China, the option of including sites in the region in a global FIH study after completion of dose escalation should be considered.

An additional design consideration for Asian phase I studies is the need for multiple‐ or repeat‐dose data. As a base case, it is recommended that single‐dose PK data in healthy subjects over a clinically relevant range of doses (e.g., 3 dose levels informed by dose linearity in the Western population) should suffice for assessment of ethnic sensitivity. Exceptions include cases where overt time‐dependent nonlinearities in PK are noted in the Western multiple ascending dose study and interethnic differences are noted in single‐dose Asian PK data. Additionally, if the expected ethnic sensitivity is in the safety profile following multiple‐dose administration (e.g., when there is potential for ethnic sensitivity in adverse events that may manifest only following multiple dose administration), a multiple‐dose design involving repeat‐dose administration may be needed to adequately characterize repeat‐dose safety and tolerability profile in the Asian population for dose confirmation.

Design considerations for ethno‐bridging evaluations in oncology drug development

In oncology drug development, it is customary to conduct multidose/multicycle phase I studies in patients with advanced malignancies. Pan‐Asian phase I ethno‐bridging study designs have been executed in patients with hematologic and non‐hematologic malignancies. ²⁴ , ²⁶ , ²⁹ , ³⁰ , ³¹ A design for this type of assessment within oncology phase I could specify collection of serial PK data in at least one Japanese and one Chinese subject per cohort during escalation and a minimum number of Japanese and Chinese patients (e.g., 6–12) during expansion, as implemented for the antibody‐drug conjugate TAK‐264. ²⁶

Phase I studies in patients with cancer represent one option for generating relevant ethno‐bridging data and may need to be considered for molecules that either cannot be dosed in healthy volunteers at clinically relevant doses or in cases where interpopulation differences in long‐term safety profile following multidose/ multicycle administration are of specific concern. However, it is important to consider the value of healthy volunteer, single‐dose phase I Asian PK assessments for molecules that can be safely administered to healthy volunteers where the questions of focus for the ethno‐bridging assessment are less related to multidose/multicycle safety and tolerability and more related to establishing consistency in PK/PD properties to support common dosage in an Asia‐inclusive MRCT. Single‐dose healthy volunteer studies can be completed far more efficiently than a typical oncology Asian phase I trial. ³² As an example, timely conduct of a single‐dose ethno‐bridging PK study in Japanese healthy volunteers for the anaplastic lymphoma kinase inhibitor brigatinib enabled enrollment of Asian populations in the pivotal phase II ALTA trial and obviated the prior need for a dedicated Japanese phase I study. ³³ Although the full benefits of the ethno‐bridging data were not leveraged in this particular case, as neither Japan nor China was included in the ALTA trial, the inclusion of other Asian countries in ALTA (e.g., South Korea, Singapore, and Hong Kong) provided valuable clinical experience, enabling efficient regulatory and global development strategies for Asia.

In another example of the development of mobocertinib for non‐small cell lung cancer harboring epidermal growth factor receptor exon 20 insertion mutations, there was a meaningful representation (~15%) of Asian patients in the Western FIH study conducted in the United States, ³⁴ likely explained in part by the higher incidence of epidermal growth factor receptor mutations in non‐small cell lung cancer in Asian populations. ³⁵ PK data on mobocertinib and total pharmacologically active species (molar sum of parent drug and 2 active metabolites with similar potency and plasma free fraction) could thus be evaluated across the dose escalation and expansion phases in the Asian subset in comparison to White subjects, supporting lack of ethnic sensitivity (Figure  3 ). This enabled Asia‐inclusive globalization of the pivotal study of mobocertinib, including mainland China, with informative sparse PK sampling for population PK modeling. ³⁶

Figure 3.

PK data in the first‐in‐human study of mobocertinib for NSCLC with EGFR exon 20 insertion mutations summarized by race. Plasma concentration‐time plots of mobocertinib (160 mg) are shown in panel (a), and panel (b) depicts box plots of systemic exposure of total active species (molar sum of mobocertinib and its 2 equipotent active metabolites with similar free fractions as parent drug). See reference 34 for details of trial design and PK characterization. AUC, area under the curve; EGFR, epidermal growth factor receptor; NSCLC, non‐small cell lung cancer; PK, pharmacokinetic.

FORWARD AND REVERSE TRANSLATIONAL RESEARCH ENABLERS

ADME characterization and pharmacogenomic studies

Qualitative and quantitative understanding of the mechanisms and molecular determinants of ADME of drug candidates is crucial to forecasting risk for interethnic variation in drug exposures. ³⁷ , ³⁸ , ³⁹ This requires comprehensive nonclinical data on expected human clearance mechanisms and relative contributions of specific drug‐metabolizing enzymes and transporters, which expands with emerging human data and includes timely conduct of the human mass balance study. ⁴⁰

If a potentially important role for ADME‐related proteins that display ethnic variation in pharmacogenetics, expression, or activity is identified (e.g., CYP2C19 (Cytochrome P450 Family 2 Subfamily C Member 19), BCRP (Breast Cancer Resistance Protein), and OATP1B1 (Organic Anion Transporting Polypeptide 1B1)), ² , ⁴¹ the totality of evidence must be considered in the overall assessment of risk for ethnic sensitivity (Figure  1 ). In some cases, it may be necessary to ensure that the assays for pharmacogenomic variation provide coverage for identification of allelic variants primarily relevant to Asian populations (e.g., UDP‐glucuronosyltransferase (UGT) 1A1*6, CYP2C19*2, SLCO1B1*15, ABCG2 c.421C>A). ¹⁸ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ For example, one patient of Chinese descent from a Western phase I study experienced an increase in systemic exposures of tivantinib that was associated with grade 4 febrile neutropenia and grade 3 mucositis, which was subsequently linked to the subject being a poor CYP2C19 metabolizer (CYP2C19*2/*2). This finding led to the design of the Japanese phase I study for the same drug to be stratified by CYP2C19 genotype to limit these toxicities, given the more frequent loss of function allele in Asian populations. ⁴⁵ , ⁴⁶

Disease biology/target and mechanism of action considerations

Knowledge of the global molecular epidemiology of the disease is key to inform risk assessment for interethnic variation. This includes knowledge of differences between Asian and Western populations in target expression, genetic/epigenetic and functional variation in the target, and associated pathway genes/proteins.

For biotherapeutics (e.g., monoclonal antibody‐based therapies, including bispecific constructs and antibody drug conjugates) with potential for target‐mediated drug disposition (TMDD), target expression can be a crucial determinant of clearance. As such, early understanding of the potential population differences in target expression under typical disease conditions can aid in forecasting of the risk for interethnic variation in PK and PD using mechanism‐based quantitative translational frameworks. ⁴⁷ Data demonstrating lack of apparent differences can be strong justification to forego dedicated Asian phase I evaluation. ²³

Planning and initiation of this research are not reliant on molecule‐specific data/information and, as such, can be timed well in advance of candidate selection, as it relies mainly on the knowledge of the target, mechanism of action, and therapeutic hypothesis. Biology and health data are important data sources that can inform disease understanding. In the last decade, Asian governments have significantly expanded resources to establish disease‐focused biobanks and emphasize quality management of biological resources. The national database of hospital‐based cancer registries in Japan is an example of infrastructure created to support evidence‐based cancer care and control. ⁴⁸ Collaboration with external partners (hospitals, academic institutions, or governments) on biobank generation and database quality can address population variability in critical disease factors (i.e., disease phenotype and factors that contribute to disease incidence, progression, and severity) and predictors of drug response. With the emerging importance of diversity in the gut microbiome, characterizing the impact of dietary factors and regional variations in the microbiome should be considered. ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ Drug development teams should engage in a robust inquiry around the need for, value of, and strategic objectives of such Asian‐related research as early as possible, given the likely need for organizational investments in external collaborations and development of appropriate analytical methodology and quantitative models to forecast the impact of ethnic variation on underlying biology.

In a recent example, during the development of pevonedistat, examination of the mutational landscape in higher‐risk myelodysplastic syndromes/low‐blast acute myeloid leukemia supported conservation in the molecular pathology of the target diseases between Asian and Western populations and across Asian populations (e.g., Korean and Japanese). This information, together with similarity in pevonedistat PK and safety across these populations, contributed to the rationalization of a global Asia‐inclusive phase III trial and consideration of a pooled East Asian region for assessment of consistency in benefit‐risk, applying principles of ICH E17 for MRCTs. ⁵⁴

Global pharmacoepidemiologic knowledge management

Asian and Western populations can vary in disease etiology, severity, prognostic factors, and pathophysiology. Extrinsic factors may also differ between countries (e.g., Japan vs. China). Furthermore, patient treatment plans may be different due to variations in diagnostic methods (clinical and molecular) and the current local standards of care. All relevant factors demand inquiry and integration as part of the decision to include Asian populations in global MRCTs. For example, in relapsed/refractory multiple myeloma, significant differences in disease severity have been described in China, with patients presenting with more advanced/refractory disease and differences in prior treatments vs. Western populations. Although the treatment effect of ixazomib, when added to lenalidomide and dexamethasone, was statistically and clinically significant in both populations (Figure  4 ), the absolute progression‐free survival in the Chinese population (6.7 months with ixazomib/lenalidomide/dexamethasone and 4 months with lenalidomide/dexamethasone) was substantially shorter than that in the global population (20.6 months with ixazomib/lenalidomide/dexamethasone and 14.7 months with lenalidomide/dexamethasone) in the randomized phase III TOURMALINE‐MM1 trial. ⁵⁵ , ⁵⁶

Figure 4.

Progression‐free survival in Chinese and global populations in the phase III TOURMALINE‐MM1 trial. Dex, dexamethasone; Len, lenalidomide.

Of note, this trial integrated China in a continuation study under a country‐specific protocol extension of the randomized global phase III trial, which allowed robust assessment of the efficacy of ixazomib across both global and Chinese populations without the confounding effects of these differences. However, such imbalances in prognostic factors and clinical outcomes, if not appropriately controlled for (e.g., via stratification by region) or if encountered in single‐arm phase II studies, can compromise interpretation of trial outcomes across regions. This example illustrates the critical importance of epidemiologic knowledge management during the design of MRCTs, considering not only drug‐related but, importantly, also disease‐related intrinsic and extrinsic factors across populations, per ICH E17.

Assessments of global epidemiologic conservation and diversity should be conducted early in development and well in advance of MRCT planning, as these considerations can impact POC strategy if access in Asian populations is a key strategic imperative. For example, combination drug development can pose specific challenges if the combination partner (e.g., standard‐of‐care therapy selected for addition to the investigational agent) is not approved or clinically used for the intended indication in all countries. Of course, selection of combination partners should be driven by mechanism of action and the underlying therapeutic hypothesis, but if the available options do not appear feasible for clinical trial conduct and registrational strategies in Asia, these risks will require early acknowledgment and assessment of alternative strategies for clinical development in Asia. Regional variation in standards of non‐pharmacologic components of patient management (e.g., behavioral modification or surgery) and the use of traditional Asian medicines, if not controlled for and considered in the analysis, can introduce imbalance/bias, inflate placebo response, and compromise interpretability and MRCT success. Comparator selection for randomized controlled trials additionally requires careful consideration as the reference treatment may not be conserved in Asia at large or in certain Asian countries. These considerations further emphasize the need for early cross‐functional knowledge management of the global clinical and regulatory landscape.

MODEL‐INFORMED DRUG DEVELOPMENT ENABLERS

Population pharmacology models

Population pharmacology models are key to quantifying the potential impact of ethnic variation on drug response. Four categories of population pharmacology models, including QSP, PBPK, population PK, and exposure‐response models, are part of the broader MIDD toolbox.

QSP models can provide an understanding of the relative contributions of the drug candidate or its metabolites to efficacy and safety at a pathway level, including insight into how receptor variability, signaling heterogeneity, and genetic variability in xenobiotic metabolism and transport can impact outcomes. QSP modeling can provide predictive value in dose selection, the need for alternative dosing, and the probability of demonstrating an outcome under different intrinsic and extrinsic factor conditions that differentiate Asian and Western populations. ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ Additionally, advances in the ability to create mathematical models of tumor immunology and the cancer‐immune cycle have resulted in the development of QSP frameworks that can incorporate multidimensional sources of variation. ⁶¹ With emerging knowledge of diversity in the microbiome and implications for response to cancer immunotherapy, ⁶² , ⁶³ , ⁶⁴ it is envisioned that QSP models will play an important role in evaluating the impact of regional diversity in the microbiome and immunophenotype on the benefit‐risk profile of immuno‐oncology drug candidates.

QSP modeling can also bring value to reverse translation decision making by validating mechanistic hypotheses that require human outcome data. Although QSP modeling is an emerging science, advances in molecular biology, the ability to measure cellular and functional events, and emerging interest by regulatory authorities will likely result in QSP becoming a common tool in Asian study‐related decision making.

PBPK models aid in forecasting PK in Asian vs. Western populations and can be deployed well in advance of clinical data availability to get an early read on level of risk for ethnic sensitivity in PK. ⁶⁵ , ⁶⁶ , ⁶⁷ , ⁶⁸ , ⁶⁹ , ⁷⁰ These models integrate molecule‐specific information on human clearance mechanisms and population‐specific information, including demographic (e.g., body size), physiologic (e.g., liver weight/blood flow), biochemical (e.g., hepatic abundance of drug‐metabolizing enzymes), and genetic (e.g., frequencies of relevant polymorphisms in ADME genes) characteristics to enable quantitative PK predictions. Population system parameters for Chinese and Japanese populations have been published and integrated into the Simcyp population PBPK simulator, ⁷¹ , ⁷² as has a population model for the Korean population. ⁷³ As data on human PK and clearance mechanisms emerge during clinical development in the Western population, the initial PBPK model can be recalibrated to update PK predictions in Asian populations. Predictions of PK in Asian populations using a PBPK model that has been verified to predict a drug’s PK in non‐Asian populations can be valuable in either determining the need for an Asian phase I study and/or guiding the design and starting dose for the first Asian phase I study. Advances in the ability to quantify drug‐metabolizing enzymes and transporters, leverage “liquid biopsy” approaches to assess ADME variation, and assess the contribution of the gut microbiome to variability in human drug metabolism and disposition should enable continuous improvement in the fidelity of PBPK frameworks for predicting ethnic sensitivity in PK. ⁷⁴ , ⁷⁵ , ⁷⁶

Population PK modeling of data collected in phase I and II studies provide vital knowledge on the sources of variability in clinical PK. The impact of a lower distribution of body weights in Asian populations on drug exposure relative to the Western population can be simulated using allometric principles and can complement predictions from PBPK models. For example, population PK modeling of brigatinib, utilizing data from Japanese healthy volunteers residing in the United States, showed a lack of both ethnic sensitivity and clinically relevant body weight effects, allowing inclusion of some Asian countries into the pivotal trial without a standalone bridging study. ²⁵

In the case of monoclonal antibody‐based therapeutics, when TMDD is evident, knowledge of target expression in the Asian vs. Western populations can provide valuable input for model‐informed assessment of risk for clinically meaningful ethnic sensitivity in PK and dose‐response relationships. If the PK are linear without evidence of TMDD and the allometrically predicted exposure distribution in Asian populations (considering their body weight distribution) does not suggest clinically meaningful differences vs. the Western population, initiation of Asia‐inclusive global clinical development without an Asian phase I trial can be defended with health authorities (e.g., the PMDA). Although body weight in the global adult population is typically not a clinically meaningful source of variability in antibody PK and dosage requirements, ²³ observed variability due to differences in body weight between populations could, in some cases, necessitate a dose adjustment. One example is omalizumab, where dose is adjusted by immunoglobulin E level and body weight. ⁷⁷

Characterization of exposure‐safety, exposure‐PD, and exposure‐efficacy (where available) relationships in the Western population informs the investigational agent’s therapeutic window/index. Data on PK differences between Asian and Western populations cannot be used in isolation to infer the level of risk for ethnic sensitivity or to guide dosing decisions. A common global dose may be appropriate, even with modest PK differences if supported by quantitative understanding of the therapeutic exposure window. However, if data‐driven scientific considerations from population pharmacology characterization (i.e., population PK and exposure‐response analyses) indicate that Asian populations require a different dose to maximize benefit‐risk, one should not infer that an Asia‐inclusive pivotal trial should be deferred. Although the regulatory hurdles will be higher, appropriate positioning of an exposure‐matched dosing strategy should be considered, ²⁴ as provided for in the ICH E17 guideline. ¹³

In the case of simeprevir, development was initiated in Asia due to the high prevalence of hepatitis C virus (HCV). At the time of submission to the FDA, exposure in Asians was 3.4‐fold higher than the overall population in the phase III trials, and was associated with an increased risk of rash and pruritus. ⁷⁸ , ⁷⁹ The initial FDA review concluded that patients of East Asian descent may need a reduced simeprevir dose, and the label stated that a dose recommendation could not be made for patients of East Asian ancestry, with a postmarketing requirement to define the appropriate dose for this patient population. ⁷⁸ , ⁷⁹ A later phase III trial in China and South Korea showed that mean simeprevir plasma exposure in East Asian subjects with HCV was 2.1‐fold higher vs. non‐Asian subjects with HCV, albeit with a similar safety profile. As a result, the same dose was approved in Asian and Western populations. ⁸⁰

In another case, an exposure‐matched dosing strategy was discussed during development for the investigational Aurora A kinase inhibitor alisertib, where a lower regional dose for Asian populations was determined to be necessary to preserve the benefit‐risk profile due to a clinically relevant difference in PK between Asian and Western populations. ³¹ , ⁸¹

These two examples illustrate the importance of evaluating ethnic/regional variation in drug exposures to inform recommended dosage and benefit‐risk profile. When differences in exposures are observed, timely regulatory consultations supported by strong, science‐driven positions regarding dosage recommendations for the Asian population are crucial to a successful Asia‐inclusive MRCT design.

Disease models

Platform models of disease progression dynamics, including model‐based meta‐analyses, can be extremely valuable in quantifying the impact of regional variations in intrinsic or extrinsic factors on disease progression and outcomes (independent of PK or PD differences). For example, longitudinal models of clinical and/or outcome end points that quantify the impact of patient‐specific clinical and demographic factors (e.g., disease stage and age), as well as factors related to medical practice ecosystems (e.g., diagnostic modalities, prior therapies, and standards of care), can forecast the overall risk for ethnic variation in drug response, associated implications for MRCT performance, and the probability of technical and regulatory success. With incorporation of molecular disease‐defining covariates, these models can be valuable in guiding global development of precision medicines. The impact of regional variations in the underlying molecular portraits of patients and their relationships to regional variation in available prior therapy and patterns of resistance are seldom quantitatively characterized, although such understanding is germane to the design of MRCTs. However, these problems provide opportunities for iterative forward and reverse translational research, bolstered by the power of population disease modeling and machine learning. ⁸²

Clinical trial simulations from disease models conditioned on the distribution of relevant covariates in Asian populations can provide in silico probabilities of country‐specific drug‐effect differences on primary and key secondary efficacy end points. Simulations from these models can be used to inform decisions regarding global footprints of confirmatory trials and stratification factors. The risks of Asia‐inclusive globalization under various design scenarios can be quantified via simulation to assess the impact of patient heterogeneity on trial success when Asian patients may represent 5% to 30% of the global trial population. This can minimize redundant regional clinical investigation while mitigating the risk of excessive heterogeneity or imbalance. ⁸³ Importantly, the absence of meaningful differences in expected outcomes for the Asian population and between the constituent East Asian populations can provide support to extrapolate Western data to Asia and use data from across East Asian populations for regulatory review and decision making, leveraging ICH E17 principles.

Development of longitudinal disease progression models ideally requires access to well‐annotated, large‐scale, patient‐level datasets (e.g., contemporaneous real‐world data and/or individual patient data from clinical trials). In one example, a previously developed equation characterizing the risk of progression of chronic kidney disease to kidney failure was independently validated in the Korean population. ⁸⁴ With data sharing and trial transparency on the rise in clinical research, there are now multiple complementary avenues to access patient‐level data. For example, control‐arm data in certain oncology indications are accessible via Project Data Sphere ⁸⁵ and in many chronic diseases via the TransCelerate Historical Trial Data Sharing initiative. ⁸⁶ Data from competitor trials, where applicable, can be requested under the provisions of the European Medicines Agency’s Policy 70. Nevertheless, given that access to patient‐level datasets with adequate annotation is not trivial, planning for development of disease models requires foresight. Model‐based meta‐analytic approaches can also be applied to trial‐level data from systematically curated literature and other public sources to quantify the effects of race/ethnicity or region of enrollment among other covariates. ⁸⁷ It is worth noting that, given their established value in enhancing efficiency of POC trial designs and decision making, the drivers for investment in development of platform disease models go beyond their applicability in informing Asia‐inclusive drug development strategies. ⁸⁸

ICH E17 AND THE POOLED REGION CONCEPT: AN OPPORTUNITY FOR CLINICAL PHARMACOLOGY

Holistic approaches to Asia‐inclusive development are facilitated by the ICH E17 guideline for MRCT design. This recently finalized regulatory guideline and the evolution in the Asian regulatory landscape (e.g., China regulatory reform) are promoting simultaneous global drug development and near‐simultaneous global drug registration. The core principles of the E17 guidelines ¹³ are as follows:

1. Conduct well‐designed MRCTs to increase drug development efficiency and support regulatory decision making across regions,

2. Understand relevant intrinsic and extrinsic factor effects early in MRCT design,

3. Allocate sample size by region to verify consistency in treatment effect while allowing feasibility in recruitment and timely trial conduct,

4. Pool prespecified regions based upon similarities in drug‐ and disease‐related intrinsic and extrinsic factors,

5. Use a single primary analysis supported by structured exploration of consistency,

6. Ensure high‐quality trial design and conduct, and

7. Encourage efficient communication between sponsors and regulatory authorities during MRCT design.

A specific opportunity of direct relevance to clinical pharmacologists is the pooled region concept under ICH E17. Prospectively defining a pooled East Asian region based on scientific rationale will confer advantages in efficiency over a country‐specific strategy. In principle, the MRCT would be designed to demonstrate consistency in treatment effect in an adequately sized pooled East Asian region vs. the global clinical trial population. The pooled region would still be designed with a reasonable representation of the constituent populations (e.g., specific countries) as opposed to requiring obligate minimum sample sizes per country to individually demonstrate consistency at the country level. Pooling justification should be based on a prospective and systematic evaluation of similarity in drug‐ and disease‐related intrinsic and extrinsic factors. ¹³

Figure  5 offers a framework to facilitate cross‐functional and cross‐regional discussions in drug development teams for synthesizing the required body of scientific evidence to design MRCTs with a pooled East Asian region. The framework is comprised of five anchors with the following associated questions. The respective clinical pharmacology enablers for each of these questions are also indicated:

1. Does ethnic sensitivity assessment based on ICH E5 principles support an expected lack of meaningful differences in PK/PD and safety among the subpopulations comprising the pooled East Asian region to support a common dosage?

   • Enablers: ADME; PBPK models; QSP models; phase I ethno‐bridging; population PK models; PK/PD, exposure‐safety and exposure‐efficacy relationships to inform therapeutic index.

2. For the target indication and patient population being investigated in the MRCT, are the subpopulations comprising the pooled East Asian region generally similar with respect to disease epidemiology (e.g., age‐adjusted incidence and prognostic factors) and current standards of care, considering currently approved therapies and local compendial guidelines?

   • Enablers: Systematic literature reviews, model‐based meta‐analyses, simulations from disease progression models with covariates.

3. For approved and/or investigational treatments (especially those with related mechanisms of action) in the target indication under study, are treatment responses and effective doses generally conserved across the subpopulations comprising the pooled East Asian region? If not, can the observed differences be explained by differences in key prognostic factors?

   • Enablers: Model‐based meta‐analyses of completed MRCTs in the target indication designed to evaluate cross‐population/cross‐region consistency in treatment effects, Simulations from disease progression models with covariates.

4. Is the disease pathophysiology conserved across the subpopulations comprising the East Asian region at the clinical phenotype and molecular levels? This assessment should consider factors such as population frequencies of genetic and/or other (e.g., transcriptomic and metabolomic) signatures of disease activity, prognosis, and treatment response (especially to drugs with related mechanisms of action). It should also consider evaluation of similarity in prior treatments that may have implications for treatment response to subsequent lines of therapy (e.g., through disease evolution and resistance mechanisms).

   • Enablers: Simulations from QSP models to assess impact of population variability in disease biology/molecular pathology, simulations from disease progression models with molecular pathology covariates.

5. What sample size of the pooled East Asian region would provide an adequate probability of demonstrating consistency in efficacy relative to the global population under the expected range of treatment effects? This should be prospectively defined at the point of trial design.

   • Enablers: Stochastic clinical trial simulations from population exposure‐response and/or disease progression models with relevant covariate effects.

Figure 5.

Proposed framework to guide design and justification of a Pooled East Asian Region approach in the design of MRCTs leveraging ICH E17 principles. ICH, International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use; MRCT, multiregional clinical trial; PK, pharmacokinetics.

Questions 1 to 4, as written, specifically address conservation of the noted features among subpopulations comprising the pooled East Asian region. Equally important for design of an Asia‐inclusive MRCT is the general assessment of conservation in relation to the global clinical trial population. As discussed earlier, the framework shown in Figure  5 was applied to design an Asia‐inclusive MRCT for the investigational anticancer agent pevonedistat in patients with higher‐risk myelodysplastic syndrome, higher‐risk chronic myelomonocytic leukemia, or low‐blast acute myeloid leukemia. ⁵⁴

Such deep interrogation of disease biology, epidemiology, and drug‐related PK/PD and safety properties coupled with a prospectively defined sample size allocation and statistical analysis plan for consistency in treatment effects requires highly integrated inter‐disciplinary effort. It requires quantitative knowledge management of disease‐related and clinical trial data in the target indication. Recent analyses of completed MRCTs across multiple disease areas and drug treatments (schizophrenia, diabetes, chronic obstructive pulmonary disease, bipolar disorders, attention deficit hyperactivity disorder, and benign prostatic hyperplasia) are yielding valuable insights, generally supporting similarity of treatment outcomes between East Asian populations. ⁸⁹ , ⁹⁰ , ⁹¹ We encourage the conduct and publication of such retrospective analyses of MRCTs, including model‐based analyses applying methods in pharmacometrics to build understanding of population variability in treatment outcomes across diseases and drug classes/mechanisms to inform design of future MRCTs.

DISCUSSION

Recent progressive changes in the global regulatory landscape catalyzed by advances at the intersection of translational, clinical, and regulatory science have enabled a holistic and efficient approach to Asia‐inclusive global drug development. The ICH E17 guideline brings a principled approach, reinforcing opportunities, where appropriate, for a holistic approach to the East Asian region in MRCTs, thereby minimizing the need for redundant clinical investigation at the country and/or regional levels. The principles of clinical pharmacology and translational science are foundational to enabling an Asia‐inclusive development plan informed by risk assessment for ethnic sensitivity. This tutorial outlined the principles for consideration to help guide this cross‐functional process.

It is recommended that opportunities for Asia‐inclusive development and a holistic approach to the region at large be considered early in the development life cycle. Many considerations related to the overall assessment of risk for ethnic sensitivity and the tactical plan for characterization of these risks are mechanism‐ and disease‐related. As such, it is possible for the requisite body of translational and epidemiologic research regarding intrinsic and extrinsic factors of relevance to the mechanism and disease area to be initiated even before a development candidate is selected. This can be supplemented with drug‐specific considerations related to ADME mechanisms following candidate selection.

A rational, science‐guided approach that considers the region at large, not solely driven by historically precedented, country‐level regulatory expectations, is foundational for success. When the risk for ethnic sensitivity is low and a common global dose is supported, one should consider the opportunity for inclusion of East Asian countries in an MRCT without standalone phase I dose‐finding trials. The success of this approach relies on proactive regulatory communication of underlying scientific rationale with prospective integration of population PK and exposure‐response analyses in the pivotal trial(s) to support the global dose for Asian patients. Forward and reverse translational research bolstered by application of an MIDD toolkit of population, systems pharmacology, and disease progression models to quantify regional/ethnic variation will inform strategies to mitigate risk in Asia‐inclusive MRCTs. We trust that systematic integration of these opportunities with a Totality of Evidence mindset and cross‐functional/cross‐regional partnerships will enable more inclusive and efficient global drug development resulting in decreased access lag for Asian populations.

FUNDING

This work was funded by Takeda Pharmaceutical Company, Ltd.

CONFLICT OF INTEREST

K.V. and M.R. are former employees of Takeda Development Center Americas, Inc. K.V. is a current employee of EMD Serono Research & Development Institute, Inc. M.R. is employed with the University of Florida as a research professor. All authors who are current or former employees of Takeda (except K.V.) own stock in Takeda Pharmaceutical Company, Ltd. P.S. and T.L. are employees of Certara, Inc.

DISCLAIMER

As an Associate Editor of Clinical Pharmacology & Therapeutics, Karthik Venkatakrishnan was not involved in the review or decision process for this paper.