The 2026 CDISC Interchange featured 88 presentations. Like many attendees, I was only able to catch a fraction of the
sessions I wanted to see in person. To bridge the gap, I used AI to analyze every slide deck, extract the main themes,
highlight key takeaways, and create a topic-based directory. If you have access to the
archive, this guide will help you quickly
locate the presentations most relevant to your team. If you missed the event entirely, it offers a comprehensive
overview and a preview of what to expect in the coming year.
The codes in parentheses are the archive's session codes: (4A.2) is session 4, track A, second talk. Codes without a
track letter, such as 1.4, are plenary sessions, and codes with several letters, such as 2ABC.1, are joint sessions.
Each code links to that presentation. The CDISC program ran in tracks A to C and the TMF program in tracks D and E. I
presented one of the talks (5C.2) and summarize it the same way as the rest. The directory of all 88 presentations,
grouped by topic, is at the end.
Recurring themes
- Machine-readable standards. CDISC's goal is that "By the end of 2030, all CDISC standards are digital, connected, and
easily consumable by users and systems" (1.3). SDTMIG v4.0 adds variable groups and root C-codes so that "variables
are now semantic entities" (4A.2). The TMF Standard Model V1 will publish its terms in the CDISC CT library, with
metadata aligned with USDM where applicable (2DE.2). The current SDS lead noted that "Lots of SDTMIG 4.0 labor could
have been avoided with more digitization" (4A.3).
- AI drafting with human sign-off. CDISC titled its own 360i Phase 3 method "AI Drafts, a Named Expert Signs" (1.4).
Merck's SDTM engine team wrote "Nothing an agent produces is self-approving" (7A.2), and Novartis's PAIR team wrote
"AI is the glue, not the foundation" (3C.5). In the TMF track, TransPerfect described keeping the model version,
prompts and thresholds under change control (6E.2), and BASE and BeOne treated prompts and model version as versioned
settings (4D.2).
- Validation alone does not establish correctness. HumanTrue's CTO used the phrase "Valid is not true" to describe USDM
files that pass validation and still encode the wrong schedule (6B.2). A protocol-authoring team warned that an agent
left unchecked "will produce a plausible, well formed, wrong entity" (5B.2). A Pinnacle 21 consultant asked for
"Better Questions, Not More Flags": checks on how the aCRF, SDTM, Define-XML, ADaM and TFLs connect (6A.1).
- Traceability from protocol to analysis. 360i followed one endpoint, progression-free survival, from definition to
analysis (2ABC.1). BMS described embedding biomedical concept IDs across protocol, eCRF, SDTM and ADaM to catch drift
at study build (7C.2). The RWD Lineage talk proposed a Define-XML extension that traces each SDTM data point to its
source (4C.2).
- Planning for test data. 360i Phase 3 adds a "Synthetic test data facility", with "synthetic data generated for each
endpoint, used to test every layer, and shipped with the package" (1.4). The Open Rules team wants "One
comprehensive set of test data per bundle" (7B.1), and a TMF AI program advised: "Plan for access and test data as a
workstream, not a task" (4D.2). AstraZeneca and ClinChoice set up an unblinded shadow team to catch real-data issues
"which don't exist in the dummy data" used for dry runs (4B.3).
Key takeaways
- CDISC released its first production conformance rule packages. It lists Open Rules packages for SDTMIG v3.2, v3.3 and
v3.4, the first FDA Business Rules package and the open-source CORE v1.0 engine under "Release, October 2026", and
states: "The rules are the product; CORE is a reference engine" (1.4).
- SDTMIG v4.0 publication is targeted for Q2 2027, pending resolution of public review comments (4A.2). NS-- datasets
replace SUPP-- (4A.1), and SDS lists "No scheduled updates to SDTM & SDTMIG" after v4.0 (4A.3).
- ADaM+IG v3.0 is mainly a consolidation, with public review estimated for the second half of 2027 (3A.1). The opening
plenary slide (1.4) gives the next release a different name, ADaMIG v4.0.
- Dataset-JSON is not yet a regulatory requirement. An FDA statement quoted in 7B.3, shown as of 17 June 2026, says CDER
and CBER "will conduct further testing", and the adoption decision is marked "Timing not final" (7B.3).
- 360i Phase 3 moves from one endpoint to a therapeutic area. Each endpoint goes through seven gates, with AI drafting
and a named expert signing at every gate (1.4).
- The TMF Standard Model V1 replaces the TMF Reference Model. Community review opens on 19 October 2026, final public
review follows in Q1 2027, and release is planned for April 2027 (2DE.1).
Summary by topic
- "Phase 2 moved 360i from vision to working software: four workstreams, fifty volunteers, one breast-cancer study"
(1.4).
- Phase 2 prototypes:
- the SoA Workbench, which builds study designs as USDM JSON and is available under COSA, the CDISC Open Source
Alliance (1.4)
- a Data Definition Engine that generates Define-XML v2.1 from USDM, biomedical concepts and SDTM dataset
specializations (2ABC.2). CDISC noted that "Value-level metadata and variable origins still need manual
post-processing" (1.4).
- a Digital DTA logical model for data from non-CRF sources (2ABC.3)
- Analysis Concepts, with R and SAS code attached to each method (2ABC.4)
- a Digital TAUG prototype (2ABC.5)
- Phase 3 takes one therapeutic area and its most-used endpoints through seven gates, from scoping to external
release. Its Prove stage calls for "Test packages sponsors can run in their own environments and report back on"
(1.4).
- The CDISC Library held 1,475 biomedical concepts as of July 2026 (1.4).
- CDISC asked the community for "Your voice", "Your data" and "Your time" (1.3).
Conformance rules and CORE (1.4, 7B.1, 6A.1)
- Authoring status for v1.0 (7B.1):
- SDTMIG v3.2 to v3.4: 473 rules, 100% written, 97% verified
- FDA Business Rules: 186 rules, 100% written, 95% verified
- ADaMIG v1.0 to v1.3: 802 rules, 79% written, unverified; CDISC verification starts after the SDTMIG release
- Verisian reported that the SQL-based engine "Successfully runs datasets > 15GB" (7B.1).
- Next come an Open Rules governance group in Q4 2026, an industry pilot on production study data in Q1 2027, and
certified engines in "2027 and beyond" (1.4).
- The Pinnacle 21 talk proposed five questions for connected evidence, under the headings Trace, Transform, Reproduce,
Reconcile and Defend (6A.1).
SDTM v3.0, SDTMIG v4.0 and non-standard variables (4A.1 to 4A.3, 5A.1)
- NS-- datasets are horizontal, with one record per parent record, and merge on --SEQ without a transpose (4A.1, 5A.1).
- New domains (4A.2):
- DC for multiple participations
- LC for lab results in conventional units
- CC for clinical classifications, split from RS
- EA for event adjudication
- morphology and physiology findings domains AU, ED, GI, HM and IG
- Other changes:
- --BLFL is removed, with --LOBXFL in its place (4A.1).
- AGELO and AGEHI replace AGETXT (4A.1).
- SUBJID is required in DM and DC (4A.2).
- IQVIA's list for sponsors: update macro libraries for NS--, validate IDVAR/IDVARVLN links, complete Define-XML (5A.1).
- Further changes will come from agency requests or industry consensus, not a scheduled release cycle (4A.3).
- 4A.1 lists SDTM v3.0 for release in 2026. 4A.2 plans SDTMIG v4.0 publication for Q2 2027.
- ADaM+IG v3.0 merges the model document, the ADaMIG, OCCDS and the NCA, popPK and anti-drug antibody guides under the
heading "Consolidation, Not Reinvention" (3A.1).
- ADaM Conformance Rules v5.0 remains a supporting normative document. Long examples move into separate informative
documents, and the team proposed knowledge-base articles written for AI retrieval (3A.2).
- New content (3A.3):
- PARAMXT in place of PARQUAL
- an Analysis Topic concept
- the NADR variables for nadir
- R2ANRLO and R2ANRHI
- ATOXDSC
- alternate-anchor study days
- Implementation talks:
- DOSExx variables for weight-based dosing (5A.2)
- estimands with an intercurrent-events dataset following PHUSE WP-092 (6A.3)
Digital protocols: USDM, DDF and ICH M11 (5B.1 to 5B.3, 6B.1 to 6B.3, 3B.1 to 3B.3)
- USDM v4.1 aligns with ICH M11. Public review opens at the end of November 2026, and publication is planned for early
2027 (1.4).
- Six teams showed tooling that produces USDM, and most also use it to build the EDC: Zifo (5B.1), an M11 authoring
system (5B.2), EDETEK (5B.3), Evinova (6B.1), HumanTrue (6B.2) and Sycamore (6B.3).
- Sycamore reported protocol-to-EDC in 4 to 5 weeks instead of 8 to 10, from a single pilot it asked readers to
interpret with caution (6B.3).
- Gaps:
- USDM leaves the mapping from timing to target day open, so each consumer writes its own (6B.2).
- HumanTrue wants "A way to compare" two USDM files of the same protocol (6B.2).
- "USDM version, CT package, SDTMIG version and EDC capability rarely move together" (5B.3).
- Evinova proposed extending USDM with costs, allocations, drop-out rates and milestones (6B.1).
- Saama showed biomedical concepts embedded in USDM 4.0 generating an ODM CRF (3B.3).
- Medidata broke endpoints into components mapped to USDM, biomedical concepts and ARS (3B.2). Scientix proposed
NGSI-LD as a linked-data layer for 360i (3B.1).
Biomedical concepts, eSource, external data and RWD (2ABC.3, 5A.3, 6C.1 to 6C.3, 7C.2, 7C.3, 4C.1 to 4C.3)
- The 360i Digital DTA talk opened its problem statement with "UP TO 70% of clinical trial data can originate from
non-CRF sources" (2ABC.3). BeOne showed what a transfer specification needs for vendor biomarker data (5A.3).
- BMS gave examples of drift between protocol and data, such as protocol visit days 1, 15 and 29 against SDTM days 1, 14
and 28 (7C.2).
- Paradigm Health described binding every eCRF field to a biomedical concept and pre-populating it from EHR data over
FHIR (6C.2).
- TCS rated LB, VS and EG as good fits for FHIR, and AE, MH, DS and SV as assist-only (6C.3).
- Axtria:
- warned that forms built for manual entry don't always ingest cleanly (6C.1)
- reported that the structured FHIR path covers "16.9% of data points" (7C.3)
- called for shared reference validation datasets (7C.3)
- RWD Lineage has a draft specification on GitHub and plans a hackathon in November 2026, marked tentative (4C.2).
- AbbVie described a pipeline for high-frequency sensor data to SDTM (4C.1). Hexaware noted there is no standard SDTM
structure yet for continuous, high-frequency data (4C.3).
- Sycamore's Dataset-JSON talk listed the milestones so far (7B.3):
- the CDISC and PHUSE pilot, September 2023 to April 2024
- the Federal Register notice, 9 April 2025
- API and compression standards, 11 December 2025
It proposed three checks (schema, a Define-XML cross-check and round-trip integrity) and a 90-day readiness plan.
- A talk looking ahead to 2041 placed Dataset-JSON and NDJSON in the submission layer and said "Define-XML stays the
metadata" (7C.1).
- AstraZeneca and ClinChoice used dummy data in dry runs, and set up an unblinded shadow team to review real data before
database lock. They listed the issues it may find as including duplicate records, missing values, outliers and vendor
data inconsistent with its transfer specification, "which don't exist in the dummy data" (4B.3). On a pivotal study,
they delivered the eCRT package 18 days after receiving database-lock data, compared with 33 days previously (4B.3).
- Pfizer reported "88-94% less manual effort" drafting reviewer guides with GenAI. One of its rules: "Missing facts are
flagged for the author, never guessed" (5C.1).
- Immanant migrated aCRF annotations to MSG v2.0 and assembled Define-XML from a study's existing sources (4B.2).
AI Innovation Challenge and analysis automation (3C.1 to 3C.6, 5C.2, 5C.3, 7A.1 to 7A.3, 7B.2)
- The three winners and three runners-up:
- Synthetic data for automation testing: Syn2Real (Novartis and AWS, winner, 3C.1) and Leova (BioInformatiCo,
runner-up, 3C.2). Leova reported "1,626 tests across 170 edit checks" and "47 problems, found before a single
participant was enrolled".
- SAP generation: Smart ClinSAP (Saama, winner, 3C.3) and SAP-Genie (Jazz, runner-up, 3C.4). Smart ClinSAP extends
USDM with a Study Analysis Definition Model, and SAP-Genie produces an SAP as ARS JSON along with specifications,
Define-XML 2.1 and SAS program skeletons.
- TFL generation: PAIR (Novartis and AWS, winner, 3C.5) and TLFGenix (Zifo, runner-up, 3C.6), a multi-agent pipeline
from the SAP to TFLs.
- Lilly's analysis-results pilot produced "36 TFLs from 26 ARDS files" with "100% Standard code". It listed "0 count and
missing data handling" first among its issues (5C.3).
- Merck and AWS argued for specifications that can be checked before any code runs: "Agree on the specification, not
the engine" (7A.1). Merck's SDTM engine team wrote: "Drafting fell to minutes. Review did not." (7A.2).
- Alcon and Veramed described reusing standards under governance, with conformance checks built into the pipelines
(7A.3).
- Clymb Clinical showed linked metadata running from protocol to CSR (7B.2).
- My talk proposed seven separately reported axes for judging synthetic study data, with CDISC CORE run inside the
generation loop (5C.2).
TMF (2DE, 3D, 3E, 4D, 4E, 5D, 5E, 6D, 6E, 7DE, 8DE)
- TMF Standard Model V1:
- Structure: 12 zones, 50-plus record groups and 570 record types with 5-digit IDs. The Core and Recommended
designations are "no longer relevant" (2DE.2).
- The program lead shared an LLM prompt for crosswalking a company TMF index to V1 (2DE.1).
- AI validation:
- TransPerfect cited the joint FDA and EMA "Guiding Principles of Good AI Practice in Drug Development" of
14 January 2026, and concluded: "There is not a new kind of validation" (6E.2).
- USDM Life Sciences listed five artifacts to write before choosing an AI tool. It warned of review failures a demo
will not reveal, such as "400 proposals cleared in 11 minutes" (6E.1).
- Arithmos cited the 2 April 2026 warning letter to Purolea Cosmetics Lab as "The first cGMP warning letter to cite
the misuse of AI" (3D.1).
- Programs and results:
- BeOne's TMF AI program started with use cases in 2024 and plans to go live in Q4 2026 (4D.2).
- CSL Behring and Just in Time GCP reviewed site personnel records across 29 sites and found 791 issues (3D.3).
- AuditLog.AI reported verifying 230,253 exported evidence files by hash, with the paper under peer review (6E.3).
- In Montrium's ICH E6(R3) survey, 11% of respondents reported full implementation, and 43% reported audit trails that
"exist but are not routinely reviewed" (6D.2).
Looking ahead
Two reviews are open or about to open: the TMF Standard Model V1 from 19 October, and USDM v4.1 from the end of
November. Where to start, by role:
- If you maintain SDTM macros, start with the NS-- list in 5A.1.
- If you run validation, watch the October rule packages and the Q1 2027 industry pilot.
- If you author or consume protocols, join the USDM v4.1 review.
- If you own a TMF, the V1 community review opens 19 October.
| When |
What |
Source |
| October 2026 |
Open Rules SDTMIG v3.2 to v3.4 and FDA Business Rules packages in production; CORE v1.0 |
1.4 |
| 19 October 2026 |
TMF Standard Model V1 community review |
2DE.1 |
| 29 October 2026 |
Korea CDISC Day, Seoul |
8ABC.2 |
| 4 to 5 November 2026 |
CDISC x SCDM Japan Conference, Tokyo |
8ABC.2 |
| November 2026 (tentative) |
RWD Lineage hackathon |
4C.2 |
| 15 to 18 November 2026 |
PHUSE EU Connect, Glasgow |
8ABC.2 |
| End of November 2026 |
USDM v4.1 public review opens |
1.4 |
| 2 December 2026 |
Webinar: 360i Phase 2 outcomes and the road to Phase 3 |
1.4 |
| Q4 2026 |
Open Rules governance group |
1.4 |
| End of 2026 |
Digital DTA internal review |
1.4 |
| October 2026 to January 2027 |
SDTMIG v4.0 public review comment resolution |
4A.2 |
| 22 January 2027 |
Japan Academic Workshop (virtual) |
8ABC.2 |
| 28 to 30 January 2027 |
CDISC Interchange at PHUSE APAC Connect, Bengaluru |
8ABC.2 |
| Q1 2027 |
Open Rules industry pilot; TMF Standard Model V1 final public review |
1.4; 2DE.1 |
| Early 2027 |
USDM v4.1 publication |
1.4 |
| 14 to 18 March 2027 |
PHUSE US Connect, Dallas |
8ABC.2 |
| Q1 or Q2 2027 |
Digital DTA 60-day public review |
2ABC.3 |
| Q2 2027, pending comments |
SDTMIG v4.0 publication |
4A.2 |
| April 2027 |
TMF Standard Model V1 release with implementation guide |
2DE.1 |
| 26 to 27 May 2027 |
2027 Europe Interchange, Barcelona |
8ABC.2 |
| Second half of 2027 (estimated) |
ADaM+IG v3.0 public review |
3A.1 |
| 2027 and beyond |
Certified rule engines |
1.4 |
| 18 to 19 October 2027 |
2027 US Interchange, San Diego |
8ABC.2 |
This year's Denver meeting drew 514 registrants from more than 220 organizations, a record for the US Interchange and
25% more than in 2025 (8ABC.2). CDISC keeps its standards timeline at
cdisc.org/standards/timeline.
If you read only five decks, read these:
- 1.4 for the overall CDISC plan
- 4A.2 for SDTMIG v4.0
- 7B.1 for the conformance rules
- 6B.2 for what breaks in USDM adoption
- 6E.2 for validating AI in a regulated system
An open question: is synthetic data good enough?
Several talks touched on a question my next post addresses: when is synthetic or dummy data good enough for the testing it
is used for?
- The Syn2Real team opened with "Why code built on synthetic data breaks", and named "three things" that "were missing
from the data all along": conformance, coherence and traceability (3C.1).
- The PAIR pilot ran into "dummy data that were difficult to make representative of real life" (3C.5).
- 360i's sixth gate is meant to produce "A package proven to hold up with synthetic data" (1.4).
My next post, "Is Your Synthetic Data Good Enough?", will address it with the seven assessment axes from my talk (5C.2).
If you would judge synthetic study data differently,
tell me on LinkedIn.
Directory of presentations by topic
All 88 presentations in the archive, grouped by topic. Titles, presenters and organizations are taken from each deck,
with one misspelled name corrected; where a deck prints none, they come from the CDISC program page.
Opening and closing sessions
| Code |
Title |
Presenters |
Organizations |
| 1.1 |
Welcome |
Jennifer Duff, Karen Roy |
NA3C; CDISC |
| 1.2 |
The Rise of Precision Empathy™ |
Jason Binder |
Ember |
| 1.3 |
Industry Pulse… CDISC's 2030 Vision |
Chris Decker |
CDISC |
| 1.4 |
CDISC 360i and the State of the Standards |
Peter Van Reusel |
CDISC |
| 8ABC.1 |
Clinical Trials, Reimagined: What the Next Decade May Look Like |
Ron Fitzmartin, Jonathan Chainey, Bill Illis, Kristi Bronkan, Stacie L. Daugherty |
Decision Analytics LLC; Genentech, a member of the Roche Group; Novartis; Kaiser Permanente |
| 8ABC.2 |
Thank you & Highlights |
Nicole Harmon |
CDISC |
| 8DE.3 |
Closing |
Jamie Toth |
BeOne Medicines |
CDISC strategy and 360i
| Code |
Title |
Presenters |
Organizations |
| 2ABC.1 |
360i Phase 2: Journey of an Endpoint |
Julie Smiley |
CDISC |
| 2ABC.2 |
360i Phase 2: Define-XML Generation |
Darren Moreland |
CDISC |
| 2ABC.3 |
From Fragmentation to Interoperability: The Digital DTA Transformation |
Valerie Sheft |
Merck |
| 2ABC.4 |
Analysis Concepts: A Standard for Executable Clinical Analysis |
Edwin van Stein |
CDISC |
| 2ABC.5 |
Digital Therapeutic Area User Guides: A Modern Approach to Developing CDISC Standards |
Bess LeRoy |
CDISC |
Conformance rules and CORE
| Code |
Title |
Presenters |
Organizations |
| 6A.1 |
Seeking Validation: Verified or Truly Vetted? Challenges on the Path to Regulatory Approval |
Julie Ann Hood |
Pinnacle 21 by Certara |
| 7B.1 |
CDISC Open Rules: Progress, Performance, and Foundations for the Future |
Tomás Sabat Stöfsel |
Verisian |
SDTM, SDTMIG v4.0 and non-standard variables
| Code |
Title |
Presenters |
Organizations |
| 4A.1 |
SDTM 3.0: What's Changed; What's New |
Michael Wise |
Edwards LifeSciences |
| 4A.2 |
What's Coming in SDTMIG v4.0 (and new information since the last interchange) |
Maria Sekac |
Merck & Co., Inc. |
| 4A.3 |
Report from SDS: Future Directions |
Dave Scocca |
Rho, Inc. |
| 5A.1 |
SUPP to NSV: Transforming Data Representation for Improved Reviewer Utility |
Soumya Rajesh |
IQVIA |
ADaM
| Code |
Title |
Presenters |
Organizations |
| 3A.1 |
ADaM +IG v3.0: Consolidating the Model & Implementation Guide |
Tatiana Sotingco, Nancy Brucken |
CDISC ADaM Team |
| 3A.2 |
Curation of Informative Text as Part of ADaM+IG v3.0 Update |
Brian Harris |
CDISC ADaM Team |
| 3A.3 |
ADaM 3.0 Update Part 3: New Content |
Sandra Minjoe, Alyssa Wittle |
ICON plc; Atorus Research |
| 5A.2 |
Around the Data DOSE-y Doe, How Much Fun Can Your Data Can Be! |
Inka Leprince, Richann Jean Watson |
InkaStat Solutions; DataRich Consulting |
| 6A.2 |
Architectural Divergence Across Three Hemophilia Studies: A Standards-Driven Framework for Rare Disease Multi-Study Programs |
Prasanna Kakade, Achin Wadhwa |
Ephicacy Consulting Group |
| 6A.3 |
From Framework to Practice: Practical ADaM Implementation |
Jie Yang |
SGS Pharma Clinical Research |
Digital protocols: USDM, DDF and ICH M11
| Code |
Title |
Presenters |
Organizations |
| 3B.1 |
NGSI-LD as a Context Management Layer for CDISC 360i |
Bo Lora |
Scientix.ai Inc |
| 3B.2 |
Semantic Bridges: Linking Study Design to Analytical Concepts |
Luke Morgan, Bhargav Koduru |
Medidata |
| 3B.3 |
AI-Driven Study Orchestration and Implementation: Powered by CDISC Biomedical Concepts |
Murugesan Vadivel, Chetan Tamballa |
Saama Technologies |
| 5B.1 |
Closing the Digital Data Flow Gap |
Sai Sri Sravya Deevi, Uma Maheswari Adiseshan |
Zifo RnD Solutions |
| 5B.2 |
From Study Intent to USDM-Compliant Digital Protocol |
Achin Wadhwa, Prasanna Kakade |
Ephicacy Consulting Group |
| 5B.3 |
From Digital Protocols to Executable Trials |
Jian Chen |
EDETEK, Inc. |
| 6B.1 |
A Guide to Utilizing USDM from Study Concept to Study Execution |
Chris Damato, Ali Soltani |
Evinova AG |
| 6B.2 |
What It Actually Takes |
Bill Lynch |
HumanTrue |
| 6B.3 |
What Breaks First When You Go DDF? |
Udiksha Jaiswal |
Sycamore Informatics |
Biomedical concepts, eSource, external data and RWD
| Code |
Title |
Presenters |
Organizations |
| 4C.1 |
Bridging the Gap: A Validated Pipeline for Transforming High-Frequency Sensor Data to SDTM |
Aman Thukral, Sanjay Bhardwaj |
AbbVie |
| 4C.2 |
RWD Lineage: A CDISC Standard for Reliable Real-World Data (RWD) |
Anita Umesh, Tasha Nagamine |
Data standards consultant (Anita Umesh); Droice Labs (Tasha Nagamine) |
| 4C.3 |
From Wearables to Regulatory Submissions: Digital Health Technologies in Clinical Research |
Varun Sethuraman, Venkata Suresh Babu Bondu |
Hexaware Technologies |
| 5A.3 |
Transforming Vendor Biomarker Data into CDISC Compliant SDTM Datasets |
Devi Gohimukkula, Saravanan Kandasamy |
BeOne Medicines |
| 6C.1 |
Validating the Site ↔ Vendor ↔ Sponsor Data Workflow |
Heather Marino |
Axtria, Inc. |
| 6C.2 |
Operationalizing CDISC Biomedical Concepts as the Semantic Layer of an eSource Platform |
Miten Mistry, Abhishek Dabral |
Paradigm Health |
| 6C.3 |
Reimagining Clinical Trials with FHIR-to-CDISC Interoperability |
Prasoon Sangwan, Arka Mazumder |
Tata Consultancy Services |
| 7C.2 |
From Intent to Execution: How Biomedical Concepts Quietly Rewired the Safety Data Pipeline |
Chithra Subramaniam, Rohini Goyal |
Bristol Myers Squibb |
| 7C.3 |
Unlocking eSource Potential |
Jo-Zahn Baxter |
Axtria |
Data exchange and submissions
| Code |
Title |
Presenters |
Organizations |
| 4B.2 |
Interrupting the Conversation: AI-Assisted Clinical Submission Artifacts as a Model for Targeted AI Adoption |
Carlo Radovsky |
Immanant |
| 4B.3 |
Advancing Global Regulatory Submissions through Next Generation Submission (NGS) |
Carrie Zhang, Asif Karbhari |
ClinChoice; AstraZeneca |
| 5C.1 |
From Manual Writing to Intelligent Drafting: Leveraging GenAI to Create Data Reviewer's Guides for Regulatory Submissions |
Sirisha Davuluri, Valerie Cadorett |
Pfizer Inc. |
| 7B.3 |
Dataset-JSON Is Coming: What FDA Adoption Could Mean for Standards, Systems, and Submission Readiness |
Bill Qubeck |
Sycamore Informatics |
| 7C.1 |
Beyond XPT: Reimagining the Clinical Submission for 2041 |
Raj Chennamaneni, Sudhir Kedare, Stephen Hamburg |
Jazz Pharmaceuticals |
AI Innovation Challenge and analysis automation
| Code |
Title |
Presenters |
Organizations |
| 3C.1 |
Syn2Real: Traceable, AI-generated synthetic SDTM and ADaM for automation testing |
Thibaud Coroller |
Novartis |
| 3C.2 |
Closing the loop: Testing a study build with synthetic data, from USDM to ADaM (Leova) |
Shirley Collie, Rob Kieser |
BioInformatiCo |
| 3C.3 |
Smart ClinSAP: AI Driven Generation of Statistical Analysis Plans |
Murugesan Vadivel, Chetan Tamballa |
Saama Technologies |
| 3C.4 |
SAP-Genie: AI-Driven Generation of Statistical Analysis Plans |
Raj Chennamaneni, Kumud Kanneganti, Sachin Tiwari, Dilip Nalla, Sudhir Kedare |
Jazz Pharmaceuticals |
| 3C.5 |
PAIR: AI TFL Automated Generation |
Thibaud Coroller, Nicolas Ballarini |
Novartis |
| 3C.6 |
TLF GENIX - AI-Driven Tables, Figures, and Listings (TLF) Generation |
Umamaheswari Adiseshan, Sai Sri Sravya Deevi |
Zifo RnD Solutions |
| 5C.2 |
Agentic AI Meets Data Simulation |
W. Stetson Line |
Clinventive |
| 5C.3 |
What happens when analysis and display are independently produced as Analysis Results DataSets (ARDS) and TFLs? |
Andrew Miskell |
Eli Lilly and Company |
| 7A.1 |
From Human-Readable Mappings to Machine-Verifiable SDTM Transformation Specifications with AI-Assisted Authoring |
Rostislav Markov, Janet Low |
Amazon Web Services, Inc.; Merck & Co., Inc. |
| 7A.2 |
Accelerating SDTM Engine Development and Adoption |
Srini Malipeddi, Proma Debnath |
Merck & Co., Inc. |
| 7A.3 |
Governance Powered CDISC Implementation: Driving Standards Reuse at Scale |
Minda Quiroz, Shyam Ardeshana, Bruce Chambers |
Alcon; Veramed |
| 7B.2 |
From Protocol to CSR: Reimagining Clinical Trials with AI and Next-Gen CDISC Models |
Bhavin Busa |
Clymb Clinical |
TMF
| Code |
Title |
Presenters |
Organizations |
| 2DE.1 |
From Reference Model to Standard Model: How We Got Here |
Donna Dorozinsky |
Just in Time GCP |
| 2DE.2 |
TMF Standard Model V1: Understanding the Architecture |
Gill Gittens |
TransPerfect Life Sciences |
| 2DE.3 |
One Standard, Multiple Subsets |
Paul Carter |
Montrium Inc. |
| 3D.1 |
Same Team, Different Worlds |
Rob Jones |
Arithmos Tech |
| 3D.2 |
Human + AI: Driving TMF Excellence Through Culture, Collaboration, and Connected Systems |
Harinarayan Gopichandran |
Sycamore Informatics |
| 3D.3 |
Transforming TMF Oversight: Applying AI to Address Persistent Site Personnel Record Gaps |
Niamh Smith, Aaron Grant |
CSL Behring; Just in Time GCP |
| 3E.1 |
Elevating the Investigator Site File: Driving Standardization and Data Integrity in Global Clinical Trials |
Guillaume Gérard |
Agatha Inc. |
| 3E.2 |
From Transactional Support to Trusted Partnership: Achieving TMF Excellence Together |
Jared Brooslin, Seth Pearl |
Intellia Therapeutics; Phlexglobal |
| 3E.3 |
From Daily Operations to Audit Spotlight: Standardizing Document Management and Quality Assurance Collaboration |
Cristina Iannaccone |
SGS Pharma Clinical Research |
| 4D.1 |
Smarter TMFs: Solving Today's Challenges While Preparing for an AI-Enabled Future |
Christina Allen, Angie Gill |
Daiichi Sankyo, Inc.; Phlexglobal, Inc. |
| 4D.2 |
From AI Hype to TMF Reality: Early Lessons from Implementing AI Agents |
Pawel Rucki, Jamie Toth, Nisi Nazim |
Base life science; BeOne Medicines |
| 4D.3 |
Redefining TMF Strategy with Agentic AI |
Jim Horstmann |
Veeva Systems |
| 4E.1 |
Leading Change in the Trial Master File |
Rhonda Nolan |
Regeneron Pharmaceuticals, Inc. |
| 4E.2 |
Japanese TMF community update: Recent highlights including TMF RM survey review |
Yuto Kanda, Mai Nishihara |
Chugai Pharmaceutical Co., Ltd.; Chugai Clinical Research Center Co., Ltd. |
| 4E.3 |
TMF Culture: Two TED-Style Provocations |
Nick Hargaden, Rebecca Halbur |
Genmab; Moderna |
| 5D.1 |
The Clinical Study Data Flow Diagram |
Liz Farrell, Lisa Mulcahy |
Agios Pharmaceuticals; Mulcahy Consulting, LLC |
| 5D.2 |
Leverage AI To Accomplish the Impossible: An Inspection Ready TMF |
Marion Mays, Chad Scribner |
Jerion Consulting Group; Epista Life Sciences |
| 5D.3 |
Case Study: The 12-Month TMF Turnaround |
Suzanne Turner |
ICE Consulting LLC |
| 5E.1 |
Your Clinical Systems Are Connected – Your TMF Governance Isn't |
Olga Radyushkina, Tiffani Nelson |
Epista; Gilead |
| 5E.2 |
From Repository to Intelligent Platform: Connected Standards and AI for RWE eTMF |
Chris Huang, Tom Stonehouse |
Oracle Health & Life Sciences; Oracle Life Sciences |
| 5E.3 |
Panel Discussion: End of Study Quality and Archiving Inspection Findings |
Chris Sigley, Ashley Avery, Liz Farrell, Curran Murphy |
Arkivum; AbbVie; Agios Pharmaceuticals; Kura Oncology |
| 6D.1 |
Integrating the Risk Based Principles of ICH E6 R3 to Drive TMF Inspection Steadiness |
Dawn Niccum |
inSeption Group |
| 6D.2 |
The Connected TMF Mindset |
Christina Mantzioros |
Montrium |
| 6D.3 |
Proactive Oversight Means Compliance Confidence |
Jenn Stamper, Carol Radwanski |
Just in Time, GCP |
| 6E.1 |
Leveraging Enterprise AI in TMF Management |
Brittany Walker |
USDM Life Sciences |
| 6E.2 |
Approaches to AI Validation in Regulated Clinical Systems and the eTMF |
Jay Smith |
TransPerfect Life Sciences |
| 6E.3 |
Proving Your TMF Has Not Changed: A Post-Export, Zero-Custody Verification Model Validated in Blinded Multi-Operator Conditions |
Fernando Telles |
AuditLog.AI Research Initiative; Cardiovascular Diagnostic Audit & AI Pty Ltd |
| 7DE.2 |
From Standard to Practice: Bringing TMF SM V1 to Life: Change Management for TMF SM V1 |
Steph Viscomi |
Kymera Therapeutics |
| 8DE.2 |
Moving to a Standard and Beyond |
Paul Carter |
Montrium Inc. |
W. Stetson Line is President of Clinventive, LLC, and the creator of Verdatic, a digital twin data pipeline for clinical studies. He founded and co-led the CDISC SDS Conformance Rules team, whose work formed the basis of the current CORE rules catalog.