Artificial Intelligence in Precision Medicine for Diabetes and Other Chronic Diseases: Challenges and Opportunities in Developing Predictive Models and Biomarkers for Disease Initiation, Progression, Prevention, and Subtypes

Event Details

Meeting Resources

• Pre-Workshop Webinar 1 Flyer (PDF, 301.3 KB)
• Pre-Workshop Webinar 2 Flyer (PDF, 292.55 KB)

Meeting Purpose

In the pursuit of precision medicine for diabetes and other chronic diseases, vast amounts of data and literature have been accumulated over decades. Artificial intelligence and machine learning (AI/ML) have made substantial strides in biomedicine, enhancing biomarker development drug discovery, improving diagnostics, and ultimately leading to more personalized and effective health care. Recent advances in AI/ML including those in generative AI and Large Language Models (LLMs), hold immense potential to further revolutionize biomedicine. This workshop aims to bring together biomedical researchers and AI/ML experts, to discuss the critical challenges, crosscutting gaps, and opportunities for and actionable items in leveraging AI/ML and other recent data science advances in precision medicine.

Meeting Objectives

• Be informed of the unique and transformative opportunities in precision medicine offered by recent advances in AI and other data science areas.
• Discuss the status of AI-based precision medicine for diabetes and other chronic conditions.
• Discuss community needs and gaps in AI-based precision medicine and how the NIH/NIDDK data science programs can help.

Organizing Committee

*External Co-Chairs

*Marcela Brissova, Vanderbilt University
*Jeffrey Grethe, University of California, San Diego
*Wei Wang, University of California, Los Angeles
Eric Brunskill, Ph.D., National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), National Institutes of Health (NIH)
Debbie Gipson, M.D., M.S., NIDDK, NIH
Daniel Gossett, Ph.D., NIDDK, NIH
Carol Haft, Ph.D., NIDDK, NIH
Jia Nie, Ph.D., NIDDK, NIH
Xujing Wang, Ph.D., NIDDK, NIH
Ashley Xia, M.D., Ph.D., NIDDK, NIH

Abstract Submission Deadline

September 15, 2024

Registration Deadline

October 25, 2024

Pre-workshop Speaker Series

The pre-workshop speaker series is intended to set the stage for discussions about the advancement of AI and data science and its applications to and impact upon precision medicine.

Part I: The Bio-Behavioral Dimensions of Diabetes Heterogeneity

The first series will be hosted on Thursday, October 17 from 11 – 1 p.m., EST, and will feature:
Dr. Yao Qin, University of California, Santa Barbara on “Data-driven Machine Learning and Closed-loop Diabetes Care”
Dr. Ashu Sabharwal, Rice University, on “Bio-behavioral Pathways in Diabetes”.

Flyer 1 (PDF, 301.3 KB)

Part II: Advances in AI and Applications in Biomedicine

The second series will be hosted on Thursday, October 24 from 1-3:00 p.m., EST, and will feature:
Dr. James Zou Stanford University, on “AI Agents in Biomedicine”, and
Dr. Eran Halperin, University of California, Los Angeles, on “AI Challenges and Opportunities across Data Modalities in Medicine”.

Flyer 2 (PDF, 292.55 KB)

Agenda

Day 1: Wednesday, October 30, 2024

8:00 a.m. – 8:30 a.m.

Check-in at Registration

8:30 a.m. – 8:35 a.m.

Opening, Welcome by NIDDK Leadership
Robert Star, M.D., Director, Division of Kidney, Urologic, and Hematologic Diseases, NIDDK, NIH, Bethesda, Maryland

8:35 a.m. – 8:40 a.m.

Introduction of Co-chairs and Workshop Overview

8:40 a.m. – 9:20 a.m.

Keynote 1: Advancing Health at the Speed of Artificial Intelligence (AI)
Hoifung Poon, Ph.D., General Manager, Microsoft Health Futures, Microsoft Research, Redmond, Washington
Moderator: Xujing Wang, Ph.D., NIDDK, NIH
Online Questions Moderator: Daniel Rothwell, NIDDK, NIH

9:20 a.m. – 9:35 a.m.

Break

9:35 a.m. – 11:05 a.m.

Session 1: Recent Advances in AI and Applications to Biomedicine
Moderators: Wei Wang, Ph.D., University of California, Los Angeles and Ji (Carl) Yang, Ph.D., Emory University
Online Questions Moderation: Anisa Prasad, The Frank H. Netter MD School of Medicine

9:35 a.m. – 9:50 a.m.

Ethical AI in Biomedicine: Addressing Bias and Promoting Equity
Tina Hernandez-Boussard, Ph.D., M.P.H., M.S., FACMI, Stanford University School of Medicine

9:50 a.m. – 10:05 a.m.

Navigating AI in Biomedicine: Opportunities and Risk of LLMs in Real-World Tasks
Zhiyong Lu, Ph.D., FACMI, FIAHSI, National Library of Medicine (NLM), NIH

10:05 a.m. – 10:20 a.m.

Building Knowledge Graphs Towards Transparent Biomedical AI
Jie Liu, Ph.D., University of Michigan

10:20 a.m. – 10:35 a.m.

A Translational Artificial Intelligence Framework in Advancing Healthcare of Tomorrow
Hongfang Liu, Ph.D., The University of Texas Health Houston

10:35 a.m. – 11:05 a.m.

Session 1 Panel Discussion

11:05 a.m. – 11:20 a.m.

Break

11:20 a.m. – 11:50 a.m.

Trainee Flash Talks
Moderators: Daniel Gossett, Ph.D., NIDDK, NIH and Jia Nie, Ph.D., NIDDK, NIH

11:50 a.m. – 12:50 p.m.

Lunch Break
Trainee scholars networking with Senior Scientists and NIH Program Officers (Rooms 1121-1125)
Lunch Networking (Labeled boxed lunches will be delivered around 11:45 a.m.)
Workshop Planning Committee Meeting (Room TBD)

12:50 p.m. – 2:20 p.m.

Session 2: AI-based Precision Medicine in Diabetes and Other Chronic Conditions
Moderators: Marcela Brissova, Ph.D., Vanderbilt University and Mandana Rezaeiahari, Ph.D., University of Arkansas for Medical Science
Online Questions Moderator: Polina Kukhareva, Ph.D., University of Utah

12:50 p.m. – 1:05 p.m.

Machine Learning Approaches Applied to Genetics of Diabetes to Elucidate Disease Mechanisms
Miriam S. Udler, M.D., Ph.D., Broad Institute

1:05 p.m. – 1:20 p.m.

Heterogeneity of Atypical Diabetes
Hemang Parikh, Ph.D., University of South Florida

1:20 p.m. – 1:35 p.m.

Precision Medicine in Diabetes: Perspective from Asia
Ronald C.W. Ma, M.B. BChir (Cantab), MRCP (UK), FRCP (London), FRCP (Edinburgh), FHKCP, FHKAM (Medicine), The Chinese University of Hong Kong

1:35 p.m. – 1:50 p.m.

Revolutionizing Phenotyping: How AI is Shaping Its Future
Wei-Qi Wei, M.D., Ph.D., Vanderbilt University, Nashville, Tennessee

1:50 p.m. – 2:20 p.m.

Session 2 Panel Discussion

2:20 p.m. – 2:50 p.m.

Break and Poster Setup

2:50 p.m. – 4:00 p.m.

Breakout Group Discussions Part 1: Topics 1 (AI) and 2 (Diabetes)

4:00 p.m. – 5:30 p.m.

Poster Session (Rooms 1175-1185)

5:30 p.m.

Adjournment

Day 2: Thursday, October 31, 2024

8:00 a.m. – 8:30 a.m.

Check-in and View Posters

8:30 a.m. – 8:40 a.m.

Day 1 Recap by Co-chairs

8:40 a.m. – 8:45 a.m.

Remarks and Welcome by NIDDK Leadership
Gregory Germino, M.D., Deputy Director, NIDDK, NIH

8:45 a.m. – 10:15 a.m.

Session 3: The NIH/NIDDK Data Science Programs in Supporting AI-based Precision Medicine, a Dialogue
Moderators: Noël Burtt, Broad Institute and Ashley Xia, M.D., Ph.D., NIDDK, NIH
Online Questions Moderator: Kate Libit, NIDDK, NIH

8:45 a.m. – 9:00 a.m.

Opportunities for Studying Heterogeneity, Uncovering New Biomarkers, and Predicting Progression in Type 2 Diabetes Using the AI-READI Dataset
Cecilia S. Lee, M.D, University of Washington

9:00 a.m. – 9:15 a.m.

dkNET & Its Scalable Cloud-based Computing Platform to Support the Integration of AI/ML Toward Accelerating Discovery and Translation
Chen Li, Ph.D., University of California, Irvine, California

9:15 a.m. – 9:30 a.m.

PanKbase: A New Knowledgebase to Support T1D Research
Kyle Gaulton, Ph.D., University of California, San Diego

9:30 a.m. – 9:45 a.m.

NIDDK-CR Resources for Research: Opportunities for AI-based Data Science
Rebecca Rodriguez, Ph.D., M.S., NIDDK, NIH

9:45 a.m. – 10:15 a.m.

Session 3 Panel Discussion

10:15 a.m. – 10:30 a.m.

Break and Poster Viewing

10:30 a.m. – 11:10 a.m.

Keynote 2
Moderator: Carol Haft, Ph.D., NIDDK, NIH
Online Questions Moderator: Mansi Mehta, NIDDK, NIH
From Glucose Patterns to Health Outcomes: A Generalizable Foundation Model for Continuous Glucose Monitor Data Analysis
Eran Segal, Ph.D., Weizmann Institute of Science

11:10 a.m. – 12:45 p.m.

Breakout Group Discussions, Part II: Topics 3 (Ethics) and 4 (Collaboration) & Working Lunch

11:10 a.m. – 11:40 a.m.

Topic 3 (Ethics)

11:40 a.m. – 12:10 p.m.

Topic 4 (Collaboration)

12:10 p.m. – 12:45 p.m.

Pick up Lunch (Labeled boxed lunches will be delivered around 11:45 a.m.)

Breakout Group Co-leads Prepare and Upload Report Back Slides

12:45 p.m. – 1:45 p.m.

Reports Back by Breakout Groups

1:45 p.m. – 2:45 p.m.

Panelist-led Discussions
Moderators: Shuibing Chen, Ph.D., Cornell University and Katerina Kechris, Ph.D., Colorado University School of Public Health
Online Questions Moderator: Fan Feng, Ph.D., Vanderbilt University

3-5 Minute Summary from each Panelist, Followed by Panelists-led Open Discussions

Topic: Opportunities and Challenges in AI
Wei Wang, Ph.D., University of California, Los Angeles

Topic: Challenges and Opportunities in Precision Medicine of T2D
Michael Snyder, Ph.D., Stanford University School of Medicine

Topic: NIH/NIDDK Data Science Programs
Jeffrey Grethe, Ph.D., University of California San Diego

Topic: Workforce Development
Talitha Washington, Ph.D., Clark Atlanta University

2:45 p.m. – 3:00 p.m.

Closing Remarks
Workshop Co-chairs
Dr. William Cefalu, M.D., Director, Division of Diabetes, Endocrinology, and Metabolic Diseases, NIDDK, NIH

3:00 p.m.

Adjournment

Breakout Group Format, Discussion Topics and Suggested Questions

Group Green

Co-leads: Xiangqin Cui and Pinaki Sarder
Notetaker: Anisa Prasad

Group Orange

Co-leads: John Kaddis (Day 1), Patrick Macdonald (Day 2) and Peipei Ping
Notetaker: Polina Kukhareva

Group Blue

Co-leads: Jake Chen and Anna Gloyn
Notetaker: Elizabeth Healey

Group Purple

Co-leads: Rozalina McCoy and Ashutosh Sabharwal
Notetaker: Valentina Roquemen-Echeverri

Breakout Group Workflow

A diagram representing the breakout group workflow for each group for respective topics.

Discussion topics and suggested questions

Topic 1 (AI): Overcoming Data Integration and Quality Challenges in Precision Medicine

• Objective: Discuss strategies for integrating multimodal data (e.g., genomics, clinical data, imaging) in AI-driven precision medicine while ensuring data quality and reliability.
• Guiding Questions:
  • What are the primary barriers to effective data integration in precision medicine?
  • How can AI/ML models be optimized to handle the diversity and complexity of biomedical data?
  • What best practices or frameworks can be developed to ensure high data quality and interoperability across different datasets? 

Topic 2 (Diabetes): Advancing Predictive Modeling of Diabetes Subtypes and its related complications with AI/ML

• Objective: Focus on the development of predictive models and the identification of disease subtypes in chronic diseases like diabetes using AI/ML.
• Guiding Questions:
  • What are the key challenges in developing accurate predictive models for disease progression and prevention? How can AI/ML help?
  • What are the key challenges in characterizing disease heterogeneity and improving precision in diagnosis and treatment?
  • What/where are the AI-ready datasets that would help with the model design?
• More examples of specific questions for precision medicine:
  • What is the role of mechanistic understanding and prior knowledge in disease?
    How can prior knowledge, particularly from the vast amount of disease-related literature, be effectively integrated into precision medicine approaches?
  • Characterizing disease heterogeneity:
    Given advancements in AI, is the best approach to precision medicine through defining disease subtypes and/or patient subgroups? Are there other methods that might be more effective?
  • Incorporating pre-clinical data:
    How can pre-clinical data be utilized to better characterize clinical heterogeneity and contribute to more accurate precision medicine solutions?

Topic 3 (Ethics): Addressing Ethical and Social Implications of AI in medicine and health.

• Objective: Explore the ethical challenges associated with AI in precision medicine, including (but are not limited to) bias, data privacy and protection, and equitable access to AI-driven healthcare solutions.
• Guiding Questions:
  • What are the most pressing ethical concerns related to AI in precision medicine?
  • How can AI models be designed to minimize bias and ensure fairness across diverse populations?
  • What policies or regulations should be in place to protect patient data privacy and security while promoting innovation?

Topic 4 (Collaboration): Enhancing Collaboration Between Biomedical Researchers and AI/ML Experts

• Objective: Discuss how to foster effective interdisciplinary collaboration between biomedical researchers and AI/ML experts to advance precision medicine.
• Guiding Questions:
  • What are the common challenges in collaboration between biomedical researchers and AI/ML specialists?
  • How can institutions or NIH/NIDDK facilitate and catalyze better communication, identify AI ready datasets and recourses, joint research initiatives between these groups?
  • What training or educational programs are needed to bridge the knowledge gap between these disciplines?
• More examples specific questions:

Moving forward, to better support the community, should the focus of government agencies be:

• Supporting the use of operational AI: Empower domain-specific biomedical researchers to utilize AI without the need for data scientists' involvement; and focus on developing AI algorithms and models tailored to biomedical applications that do not require the active participation of domain experts during development.
• Facilitating collaboration: Encourage collaboration between biomedical; or researchers and data scientists to co-design and co-develop AI algorithms, models, and applications for addressing domain-specific challenges.

Abstracts

*All presenters must register in advance for the conference. Registration is free. *

Submission Deadline

September 15, 2024

All Abstract Submission

All abstracts must be submitted via email to Mark Dennis of The Scientific Consulting Group, Inc., with “NIDDK AI Workshop” in the subject line.

For trainee travel awards, please find details and submit additional documents in here: https://dknet.org/about/ai_workshop_travel_award. The travel awards deadline is August 30, 2024.

Abstract Topics

Abstracts for oral or poster presentations are encouraged to present Advances in AI and applications in biomedicine, and AI-based Precision Medicine for diabetes and other chronic conditions. The topics could be but not limited:

• heterogeneous pathways and trajectories to disease,
• AI-based precision medicine for diabetes and other chronic conditions.
• molecular biomarkers & real-world needs and challenges.

Abstracts from trainees and early-stage investigators are particularly encouraged. 

Submitting Abstracts

Please follow the instructions below to format an abstract. (Note: Submissions will not be edited for spelling or grammar and will be accepted “as is.”)

• The abstract should be a Microsoft Word document that does not exceed 500 words(not including the abstract’s title, name and affiliation of all authors, and one table or figure).
• with 1-inch margins, typed single space, and using Times New Roman font; a 12-point font should be used for everything except the title.
• Please indicate whether you would like your abstract submission to be considered for an oral presentation only, a poster presentation only, or for either an oral or poster presentation.
• The abstract’s title should be typed in Title Case using bold 16-point fontand should clearly represent the nature of the investigation.
• List the author’s first and last names, degree, affiliation, city, and state. If more than one author is listed, include complete information for all authors and underline the presenting author’s name. Please indicate if the presenting author is a trainee or early-stage investigator.
• One table or figure may be included; however, the abstract may not be longer than one page, including the table or figure.
• The abstract file name should follow this format: primary author’s LastName_FirstWordOfTitle (e.g., Zucker Effects).

Presenters are welcome to submit abstracts previously presented in other venues, including work that has been delivered at other conferences or published within the past 2 years. If the work has been published, please include a full citation and a link to the paper.

Abstract Template:  LastName_FirstWordOfTitle.docx (DOCX, 20.42 KB)

All Abstract Acceptance Notifications

Applicants will be notified if their abstract has been accepted for a poster or oral presentation by approximately 10/01/2024. Detailed instructions regarding oral and poster presentations will be provided at the time of acceptance.