The clinical research industry faces persistent challenges in protocol interpretation, metadata inconsistency, manual standards implementation, limited traceability, and prolonged regulatory submission timelines. While CDISC standards provide a structured framework for clinical data management, their implementation often requires significant manual effort across study artifacts such as protocols, CRFs, EDC systems, SDTM datasets, metadata repositories, and validation workflows. This paper presents an end-to-end automation framework based on CDISC360i principles, leveraging Biomedical Concepts (BCs) as reusable metadata assets to establish a digital thread across study design, data collection, standards implementation, and submission preparation. The framework integrates artificial intelligence to support protocol interpretation, metadata mapping, and standards transformation, reducing manual intervention and improving consistency. Open-source technologies enable automated CRF generation, EDC configuration, SDTM specialization, Define-XML creation, and validation processes. By combining metadata-driven automation, AI-assisted workflows, and standards-based validation, the solution enhances traceability, data quality, regulatory compliance, and overall study efficiency while accelerating submission readiness.

CDISC Open Rules is redefining how data conformance validation is performed across the industry. Significant progress has been achieved on both the rules and the CDISC Open Rules Engine (CORE), with major refactoring efforts focused on improving performance, scalability and long-term sustainability. This presentation will provide a comprehensive update on the current state CORE: from the rule authoring process and completion status to the engine and the technical enhancements implemented so far. The roadmap ahead will outline the plans for expanding rule coverage, governance, and innovations that will enable broad adoption through open-source collaboration. Attendees will leave with a clear view of where CORE stands today, a practical understanding of how to integrate open-source CORE validation into existing submission pipelines, and concrete ways to contribute their expertise to shape the future of CDISC Open Rules as the open-source reference implementation for conformance rules.

Join our interactive chatting session, where participants will discuss SCDM Competencies and CDISC-related topics in small groups.

Topics may include RPA/AI adoption, Risk-Based Approaches, open-source solutions, and other emerging trends in clinical data management/CDISC.

Details of the discussion topics and the pre-registration URL will be shared once available

Dr. Takuhiro Yamaguchi is Professor, Division of Biostatistics, Tohoku University Graduate School of Medicine; Representative Director, Japan Society of Clinical Trials and Research; Past Chair, SCDM Japan Steering Committee; SCDM Japan Steering Committee (JSC); and Past J3C (Japan CDISC Coordinating Committee) Member

Dr. Yuki Ando is Principal Senior Scientist for Biostatistics, PMDA; Member, CDISC Board of Directors; Member, Japan CDISC Coordinating Committee (J3C); and Member, SCDM Japan Steering Committee (JSC)

In a significant collaboration, eight leading vaccine companies — AstraZeneca, GlaxoSmithKline, Johnson & Johnson, Merck, Moderna, Pfizer, Takeda and Sanofi — have formed the Vaccines Industry Standards Group (VISG). Over the past three years, this initiative has focused on harmonizing interpretations of regulatory submission guidance and recurrent feedback, as well as CDISC data standards. The group recognizes that aligning the understanding of requirements — such as participant diary data collection and the submission of reactogenicity and efficacy data — accelerates time to market and benefits global health.  
This unified approach could facilitate future collaboration with Health Authorities and CDISC, aiming to update the CDISC Vaccines TAUG to meet current Health Authorities' expectations, thereby ensuring clarity and consistency in submission standards. 
Our collaborative model can serve as a blueprint for other therapeutic areas within the pharmaceutical industry, demonstrating how organizations can work together to streamline regulatory processes while maintaining a competitive edge in product innovation.

FDA review documents capture the analyses and evidence underlying regulatory decision-making, yet much of this knowledge remains locked in individual PDF documents. This presentation summarizes an AI-based systematic analysis of FDA New Drug Application (NDA) review documents published between 2021 and 2026. Thousands of tables and figures were extracted, categorized, and standardized into a searchable knowledge base of regulatory review outputs, which were cross validated by multiple large language models to enhance reliability.This investigation identifies the analyses, tables, and figures most frequently used in FDA reviews, highlights emerging trends in regulatory review, and demonstrates how AI can characterize review practices on a scale. Selected analysis results generated using data from the CDISC Pilot Study demonstrate how ARS-compliant tables and figures align with FDA review tables and figures. These findings illustrate the value of standardized analysis outputs for more consistent and efficient submission preparation and regulatory review.

Background and Objective: Clinical domain knowledge remains insufficiently standardized and structured despite advances in healthcare information systems, limiting interoperability among different platforms, such as EHRs, registries, PHRs, and research systems. We previously evaluated the alignment of a lifestyle related disease self-management item set with CDISC SDTM. Based on this experience, we aim to establish a multi-society framework for developing and governing minimum item sets.Methods: Within a TEAM initiative, participating societies collaboratively developed minimum item sets and a common governance framework.

Results and Conclusions: Ten clinical domains were selected based on healthcare priorities, and a standardized methodology for developing and governing minimum item sets was established. Disease-specific item sets are under development in multiple domains, including heart failure, depression, and perinatal care. International standards alignment was incorporated into data definitions. This framework is expected to improve data quality, interoperability, and reusability, facilitating data-driven research, registry integration, and AI development.

Utilizing real-world data (RWD) from electronic health records (EHRs) for clinical research and post-marketing surveillance is hindered by data-format heterogeneity across institutions and vendors. Moreover, an EDC entered separately from the EHR cannot natively capture routine clinical data in a standard format. This work embeds CDISC standards (SDTM variables, Controlled Terminology, MedDRA, and WHODrug) directly into JASPEHR, a vendor-neutral, HL7 FHIR-based EHR documentation template platform, so that SDTM-conformant data are generated at the moment of EHR entry, eliminating the mapping effort conventionally required after data collection. We specified this architecture and implemented a proof of concept spanning EHR data entry through SDTM dataset generation on a verification environment built with NLM Form Builder and a HAPI FHIR R4 Server, and have delivered sample terminology sets to participating EHR vendors. We will discuss this architecture and our roadmap toward finalizing the FHIR specification, vendor implementation, and closer engagement with CDISC.

GCP Renovation is driving a fundamental redesign of clinical trials and quality management in response to increasing protocol complexity, diversified data sources, and the growing limitations of conventional GCP operations. Reflecting the principles of ICH E8(R1) and ICH E6(R3), this presentation outlines three key shifts: from retrospective, comprehensive checking to Quality by Design focused on Critical-to-Quality factors; from paper- and EDC-centered processes to lifecycle-based management of data derived from eSource and multiple digital sources; and from sponsor-driven, one-way quality oversight to collaborative, open dialogue among sponsors, CROs, investigators, CRCs, patients, and other interested parties. The presentation further discusses how AI, risk-based monitoring, and digital data flow can enhance efficiency, support risk-proportionate decision-making, and enable integrated quality control. GCP Renovation should be understood not as an incremental operational burden, but as a strategic transformation of the entire clinical trial quality management.

The adoption of eSource enables real-time and high-quality data management. However, the true value of these digital tools is maximized only when they are harmonized with clinical workflows and utilized in collaboration with the site. Seamlessly linking study design with site execution is a critical factor in further enhancing data reliability. Since clinical sites represent the most upstream point of data generation, the quality established at this stage serves as the foundation for the entire research.Resolving minor operational frictions through optimized design and improved system usability leads to both the efficient use of resources and higher data quality. This presentation shares insights gained from the perspective of medical institutions to present the value and future prospects for more reliable data collection and the further advancement of clinical research and trials.

As clinical trials grow more complex, success depends on identifying risk earlier, reducing inefficiencies, and embedding quality throughout execution. This session explores how the integration of RBQM, RBDM, and IDD creates a proactive, data-driven operating model that transforms trial oversight. By combining risk assessment, centralized monitoring, targeted data review, and integrated clinical, RBQM, RBDM, and IDD enable faster signal detection, earlier intervention, and effective decision-making focusing on critical data. The result is improved data quality, inspection readiness, optimized resource, accelerated study delivery. Attendees will gain practical insights into how moving from functional driven, reactive delivery to proactive risk prevention and purpose driven delivery results in higher-quality outcomes with greater operational efficiency through risk-based surveillance and issue prioritization.This presentation demonstrates how RBQM and IDD work together to create a scalable, quality-focused framework that improves trial performance while maintaining patient safety and regulatory compliance, leveraging KRIs, QTLs for proactive risk management.

• Hidenobu Kondo, A2 Healthcare
• Shohei Ishii, Keio University Hospital
• Chiharu Shimada, Parexel International
• Hideki Suganami, JSC; Kowa
• Yukikazu Hayashi, Chair, JSC; A2 Healthcare

This session provides an overview of Digital Data Flow with CDISC under the theme “Design once, use everywhere.” The presentation explores how Digital Data Flow can be achieved through the reuse of standardized study design metadata, including the role of CDISC standards such as USDM. By placing standardized study design metadata at the center of clinical trial operations, protocol information can be reused across downstream systems and deliverables, including EDC, CTMS, IRT/RTSM, ePRO, and submission workflows. The presentation introduces industry trends, guiding principles, and practical implementation examples, including OpenStudyBuilder integration. These use cases demonstrate how CDISC standards, APIs, and metadata repositories can automate study setup, reduce duplicate work, support faster protocol amendments, improve interoperability and traceability, and establish a foundation for end-to-end clinical research automation.

Unified Study Definition Model (USDM) is the core component for realizing Digital Data Flow (DDF) and provides foundation for the digitized study protocol. The USDM is also considered to be the standard data model that enables ICH M11 compliance as well as business process automation.This presentation will provide an in-depth overview of the USDM, its architectural principles, and some key components such as the Schedule of Activities and Biomedical Concepts. We will then explore benefits of the digitized study protocol using USDM, such as facilitating clearer study designs and enhancing protocol contents reuse. We will also explore how USDM, together with metadata-driven approaches, standards-based integration technologies, and generative AI, can enable further automation across the study lifecycle.

The pharmaceutical industry is increasingly exploring Digital Data Flow (DDF) as a foundation for transforming clinical development. Enabled by emerging standards such as ICH M11, CDISC USDM, and related interoperability initiatives, DDF promotes a “define once, reuse many times” approach in which study information is reused consistently across downstream processes. The vision extends beyond protocol digitization toward end-to-end automation from study design to regulatory submission.DDF has the potential to improve consistency, traceability, speed and efficiency across the clinical trial lifecycle through structured study definitions governed within a centralized repository.However, significant challenges remain before this vision becomes reality. Achieving DDF at scale will require orchestrating transformation across people, processes, technology, governance, and implementation assets, beyond the adoption of standards alone.This presentation explores the value, future outlook, and implementation challenges of DDF from the perspective of pharmaceutical companies, including business case development, roadmap planning, organizational readiness, and data governance.

• Akira Soma, Oracle Health & Life Sciences
• Kunihito Ebi, Fujitsu Limited
• Hidemi Hasegawa, Boehringer Ingelheim

Edit check programming is a critical yet labor-intensive step in study build, traditionally requiring manual translation of natural-language rule specifications into executable data-validation code. We developed an AI-powered Edit Check Code Generator that automates this process. Using Named Entity Recognition to map rule descriptions to eCRF data items and a large language model to generate validated code, the tool converts entire specifications in batch within minutes. A companion module, Edit Check Unit Testing automatically produces test scripts and configuration files covering open and close cases to streamline validation. Together they improve efficiency, consistency, and scalability while accelerating study start-up. This presentation shares our architecture, measured time savings, and lessons learned, including why human review remains essential and how data management roles are evolving toward AI supervision, prompt design, and quality oversight in an AI-enabled environment.

Clinical trial validation activities depend on vendor-provided test data, which may be delayed, incomplete, or limited in scenario coverage. These challenges can postpone validation activities and delay identification of mapping, transfer, and integration issues during study startup. This solution presents an AI/ML-enabled synthetic data generation framework designed to create clinical trial test datasets using protocol requirements, visit schedules, metadata, and transfer specifications. The framework combines metadata-driven business rules with generative AI capabilities to create scenarios such as missing visits, protocol deviations, out-of-range values, unit conversion cases, reconciliation gaps, and edge cases required for testing. Generated datasets are validated against predefined rules before use. By enabling validation teams to begin testing earlier without waiting for vendor-delivered test data, the approach reduces reliance on vendor timelines, improves scenario coverage, and supports earlier issue detection. Synthetic data does not replace vendor data; it accelerates validation readiness, improves efficiency, and supports proactive risk mitigation.

In clinical development, automation of data conversion and analytical programming is hindered by manual interpretation and manual coding. In this presentation, we would like to introduce, through actual examples, an approach to overcome this bottleneck by fusing "CDISC 360i" with an "AI Coding Agent."

CDISC 360i's four metadata layers—USDM, Biomedical Concepts, Analytics Concepts, and ARS—serve as an important source for AI-understandable specifications. Since the AI Coding Agent can directly utilize this structured common language, hallucination from ambiguous interpretation is eliminated, and high quality code is autonomously generated with complete traceability and a self-correction loop.

By complementing gaps in human-oriented documents (Protocol, SAP, Shell) with CDISC 360i metadata, we generate not merely program files but ""Artifacts"" containing practical verification evidence.
In addition, we introduce new challenges, including high cost of metadata design, review burden of large-volume artifacts, and synchronization difficulty at protocol amendment, together with a realistic governance roadmap.

Clinical Data Management support operations remain largely reactive, with teams resolving recurring issues — codelist misconfigurations, standards version conflicts, edit check failures — without leveraging the operational intelligence these interactions contain. Existing ticketing platforms offer basic AI features but lack cross-study pattern detection, predictive quality alerting, and clinical DM-specific risk intelligence.This presentation introduces a practical four-layer framework for applying AI to CDM support workflows: Intelligent Classification, Cross-Study Pattern Detection, Risk-Informed Prioritization, and Proactive Alerting. Each layer is implementable independently using accessible tools and existing operational data. Early observations reveal that recurring issue categories concentrate resolution time, systemic patterns manifest across studies simultaneously, and leading indicators in early-phase tickets correlate with later-stage escalations. We discuss alignment with ICH E6(R3) quality-by-design principles and PMDA data reliability expectations, providing a replicable roadmap for teams ready to shift from reactive resolution to predictive quality management.

As the clinical development landscape continues to evolve, improving efficiency through the use of standards and data-processing automation is becoming increasingly important. CDISC is also advancing standards with a focus on automation. However, some areas are not fully covered by existing standards, and contributions from user communities have become an important part of standards development. Therapeutic Area User Guides (TAUGs) provide guidance for consistently representing disease-specific clinical concepts within CDISC standards, and their expansion serves as a key foundation for promoting standardization and improving efficiency from data collection through statistical analysis. The CDISC Japan User Group (CJUG) established a Task Force to develop the Ophthalmology TAUG in collaboration with CDISC. Through this initiative, we have developed CJUG drafts of the TAUG and Biomedical Concepts. This presentation will share our activities, achievements, key learnings, and future plans. We hope these experiences will provide insights for future standards development and community-driven initiatives.

BackgroundNew CDISC standards become valuable only when they can be readily implemented by the community. Open-source software provides an effective mechanism for translating specifications into practical tools.MethodsWe developed two SAS packages supporting emerging CDISC standards. The first, sas_dataset_json, provides bidirectional conversion between SAS datasets and Dataset-JSON v1.1 and has been accepted into the COSA Repository Directory. The second, sARDen, converts analysis results into ARS-compliant Analysis Results Data (ARD) and has also been submitted to COSA.ResultsDeveloping these packages revealed practical considerations beyond implementing the specifications, including interpreting evolving standards,  validating outputs, documenting usage, and preparing projects for community contribution. The COSA submission process also provided valuable experience in open-source governance and software quality.ConclusionThis presentation shares lessons learned from developing and contributing open-source implementations of emerging CDISC standards. 

CDISC Open Rules Engine (CORE) is under development in open source. CORE is the tool that provides reference implementation of validation rules. Not only CDISC standard rules, but also regulatory rules are in scope of CORE. Currently, some part of FDA business rules are created, but PMDA rules are not included in this process. Is it practical to author PMDA rules on current version of CORE? All components required by PMDA are ready to use? What are the barriers for Japan users? Is it best timing to start working for PMDA rules? Small and private implementation of PMDA rules on CORE has been done by the authors. This presentation shares results/experiences of the trial and suggestions on how CORE can be applied in Japan.

• Naoki Kato, Otsuka Pharmaceutical Co., Ltd
• Yutaka Morioka, EPS Corporation
• Hajime Shimizu, Individual Contributor

Clinical research is entering a transformative era driven by artificial intelligence (AI), computational biology, and advanced data science. A compelling vision for the next 15–20 years is the emergence of human digital twins that are dynamic computational representations of individuals that integrate genomic, clinical, physiological, and real-world data to simulate health, disease progression, and therapeutic response.

This presentation explores how advances in computational modeling, AI, and digital health may enable future clinical trials to be conducted largely in silico using virtual patient populations. For the clinical research community, this transformation places increasing importance on data quality, interoperability, and predictive analytics. Clinical Data Managers will ensure data quality, governance, metadata standards, and regulatory compliance. Data Engineers will build scalable data architecture and integration pipelines, while Data Scientists and AI specialists will develop and validate predictive models and digital twin platforms.

Although significant scientific and regulatory challenges remain, human digital twins have the potential to make clinical research faster, safer, more personalized, and more inclusive while reshaping the future of evidence generation.