Leaders in Pharmaceutical Business Intelligence Group, LLC, Doing Business As LPBI Group, Newton, MA

• 08:30 – 09:30 Gathering & Registration

• 09:30 – 09:40 Opening Remarks and Video

  Collaborations between Academia and HMO & Hospitals and Companies.

  other two speakers

• 09:40 – 10:00 Overview of AI in Health in Israel Today

  • Ziv Katzir, Head of the National Plan for Artificial Intelligence Infrastructure, Israel Innovation Authority, Israel
  • comes from CS not Health: AI Tools for Medical Treatment with Decision Support relay on Experts – future different, Diagnostics, Clinical development, Drug development
  • Extreme multimodality Multi-purpuse Clinical AI <<— produce more data sensing ans sequencing data cost reduced, medical imaging, pathology – data integration still behind
  • clinical development & Clinical data: Genomics, proteomics, metabolomics
  • Medical data: Multimodal Sensory Data
  • Therapeutics: Bio/molecular
  • Israel Health ecosystem – 600 companies
  • Future: more automation in decision support automation and autonomy
  • LLM is successful because it was trained on the Internet: guard from bias

• 10:00 – 10:20 AI in Health, International Perspective and Israel’s Role

  • Prof. Ran Balicer, MD, PhD, MPH, CIO & Deputy-DG, Clalit Health Services; Professor, BGU & Charité Universitätsmedizin Berlin (Hon), Israel
  • AI-Driven Healthcare: AI augmate Physicians
  • Hep C Targeted screening of 477 38 had Hep C
  • Call patient based on Predictive Proactive care and give preventive medicine = care change medication by AI. 50,000 patient had gene sequences – One screen Summary Clinical
  • Deep learning, X-Ray had an error showing location of fracture, super human diagnostics
  • AI-guided Dx, De-skilling
  • Generative AI, –>>> PANDA: Physician AI Navigation Decision Assistant
  • AI-driven transformation: Stay away from Pilots
  • Clalit HC AI Autonomy Scale – triage done by AI: Which film will be read first done by AI
  • AI > MD – Human in the loop. AI is the bigger helper
  • AI > [AI+MD =MD]
  • OPTICA – evaluation of AI Tools
  • AI will allow Physician to augment trust with patients
  • AI – change of Curriculum in the Medical School class
  • Regulatory: If AI makes mistakes – compensation will be paid

• 10:20 – 10:40 AI and Health in Academia: How These Ideas Translate to Industry

  • Prof. Noam Shomron, Professor, Head of Digital Medicine Research Team, Tel Aviv University, Israel
  • AI in Health in Academia, genomics to clinical questions: accelerate by genomics, DNA seq, advance, Vaccine COVIS in 6 months, Understand apply: Gene Editing, Cell Therapy – change DNA and correct it  – early detection. Pre-natal during pregnancy, Non-invasive Prenatal Testing, every letter in the DNA of Mother or of baby by nucleotids at week 5 or 12. Deep Learnin 20 cancer aptient Blood Test: Profiling Cancer vs non-Cancer. Microdoses in use. Digital signature. Identify early for early intervention for PTSD.10% will have it. Identify for early intervention.

• 10:40 – 12:00 Data for AI in Health

• 10:40 – 11:00 Lecture

  • Prof. Gabi Barbash, CEO, Psifas; Director Bench to Bedside Program, Weizmann Institute of Science, Israel
  • Psifas LEAD, BioBank for all universities.                        Genetic variations in Subpopulstion µµ
  • Genetic consultation 2.2% gene 8,000 patients with correctable genes of 54,000 screened.
  • Collaboration of all Hospitals
  • CRO, Reichman Institute, MDS treatment to avoid bone marrow transplant
  • COmpare two groups MI and group of normal coronary
  • Psifas Data base is a collaborative data collection effort by Public funding. Retrospective and prospective. Commercial use will pay. Non-commercial is Public domain.

• 11:00 – 111:20 Ministry of Health

  • Guy Livne, Head of Health Informatics, Data & Innovation Dep., The Directorate of Governmental Medical Centers, Israel
  • Collaboration for AI models – Kineret Data Lake has data from 25 hospitals, 100 subpopulation as categories OMOP – Global standard Patient ID across all Hospitals, all data in the Cloud, Workflow unified
  • 3 month from local to OMOP. 97% data is in OMOP standards of standard data. There is Non- OMOP data, structured and non-structured. Kineret collaborat with all parties
  • CVDLINK – Horizon project – Cardiac data
  • One single Tool for Federated learning, OMOP standard.
  • Lineret OMOP BOT – create cohort, define study, go-no-go study, PI communication, done in 1 hour instead of months.
  • Predictive modeling, Multi-center studies, collaboration with abroad countries: Sleep studies,
  • Data is de-identified
  • CVD, Ichilov Hospital uses Camillion not Kineret
  • Apply to Kineret for data sets

• 11:20 – 11:40 AI and the Future of Health Monitoring: Making Sense of Physiological Data

  • Dr. Joachim A. Behar, Associate Professor, Technion Faculty of Biomedical Engineering and Faculty of Data and Decision Sciences, Israel
  • Physiological data like Vital signs Clinician Decision data actual data: Sleep AI Study – “Sleep Apnea (8 hours of recorded data at Ichilov) data analysis by AI” –>> SleepAI Solution is a start up. ECG 12 lead for 24 hours data recording study for interpretation using AI supporting CVD care moved to Holter ECG in bed wear of belt. Technion-Holter Study on Heart Failure Risk Hospitalization or Death (x2) vs Death (x4). 8AM to 4PM most important time span to signal identification – time window vs any others
  • Circadian A-Fib: risk for Supervised Learning study.
  • Syncope symptom.
  • Lirot.ai – Ophthalmology – Scans of Retina, thickness of layer Diagnosis of Glaucoma – generalizability across domains, 12.8 improvement over benchmark studies.
  • AI vs Human Expert: Senior 10 Ophthalmologist compared with AI 10 years Experience in Retina:  For AMD – AI performed in diagnosis better
  • OCT modeling deployed in 3 Medical Center
  • Digitization od 10 years of Data
  • The Robot with Vision by AIMLAB

• 11:40 – 12:00 Collaborative Longitudinal Data Platforms: The Hidden Infrastructure of AI in Health

  • Steven E. Labkoff, MD, Collaborating Scientist, Beth Israel Deaconess Medical Center, Harvard Medical School, USA

  • Artificial intelligence in healthcare continues to accelerate, yet its real-world clinical impact remains constrained by the limitations of the underlying data. Most AI models are built on fragmented, cross-sectional information that provides only a narrow view of patient health. Truly meaningful clinical insights—early detection, trajectory modeling, treatment optimization, and trial acceleration—require data that follow patients over time. This talk examines why collaborative longitudinal data platforms are emerging as the essential foundation for high-value AI in medicine.

    Longitudinal platforms combine multi-year patient journeys with multi-modal inputs, integrating clinical data, genomics, imaging, digital measures, and patient-reported outcomes. When developed collaboratively across institutions, they provide the scale, heterogeneity, and continuity needed for generalizable and trustworthy AI. These platforms depend on shared standards, reproducible pipelines, data provenance, and modern privacy-preserving approaches such as federated learning.

    Drawing on examples from the Multiple Myeloma Research Foundation and a rare disease patient advocacy organization, as well as work within the Harvard DCI Network, the talk illustrates how longitudinal platforms are already reshaping clinical research, real-world evidence generation, and operational decision-making.

    The session concludes by addressing the unique barriers in the United States—most notably the absence of a national patient identifier, extreme fragmentation across care settings, and persistent interoperability gaps—which collectively hinder the creation of robust longitudinal data and limit the full potential of AI in healthcare.

    • Steven E. Labkoff, MD, Collaborating Scientist, Beth Israel Deaconess Medical Center, Harvard Medical School, USA

    -issues of large vs small models

    -BMS experience – longitudinal multi-modal be governed

    -Limitation of the data: Sparse phynotypes, multi-institutional -expensive to build

    -Longitudinal Out patient, Inpatient, tokanization, fragmented data sets to platform, consent collection, Medical Record – unwilligness to share information

    -sophisticated Use Cases from Longitudinal medical registry: CureCloud – MMRF 1500 patients were recruited.

    -linked with Insurance claim data

    -Federated data model of small data sets from multiple geographies – collaborate between institution is challenging for collaboratinf platforms

    -Culture of Data stuardship, legal aggrements. Biases bulit.

    -DCI Network’s Efforts: Patients want to be on Clinical Trials. Difficulties betweem institutions and Patients – AI is helping screening, high degree matching Recruitment was too hard, low rate ecpensive for Pharma companies

    -longitudinal data in place – use digital-twins in rare diseases as use case

    -inadvertly issues of identification, biases, ability to deal with bias befor LLMs, new drug came to market, Myeloma case, standard of care changed mid way. Data vs algorithms

    -challenge on who own the data

    -wearables are for One patient data, they have a place. Algorithms for data download, reliability, measure handful of parameters not all needed, place a role in data collection

• 12:00 – 13:05 AI & Health Start-Ups: VC’s and Lightning Talks

• 12:00 – 12:15 The American VC Perspective-

  • Bruce Taragin, Managing Director, Blumberg Capital, USA
  • 870MM under management, early investor in CheckPoint
  • Healthcare: AI enablement digital health data system
  • US Health Tech Venture Activity: Data & AI to avoid errors.
  • Data infrastructure compamy Angelo related to Palantir, AI Platforms (like UnitedHealth), medical imaging and personalization: Diagnosis in real time, curation, full stack solutions, surgical intelligence
  • six Ts – teams, tractions, Tech, terrain, Terms, Theme

• 12:15 – 12:30 The European VC Perspective

  • Marc Greuter, General Partner, Planven, Switzerland
  • Scaling AI healthtech in Europe
  • 300M assets under management, Zurich and Tel Aviv
  • Healthtech – investments in Insrael: Cathworks (acq by Metronics), IBEX
  • Healthtech in Europe: Population of 450MM, Universal healthcare coverage in Europe, $100Bn in Europe,
  • Europe had HQS of Big Pharma in Europe, Biotech supplier
  • AI in Health: Drug discovery, Imaging,
  • commertializing: Not technology alone, healthcare system are very complex to adopt innovations into the workflow of Physicians
  • Cloud computing and governing data for access and collaboration OVERregulation of AI in Europe
  • EU AI ACT Market Franfmentation to Harmonization
  • Healthtech – high risk due to patient harm potential: Human oversight of AI and AI software can causes harm and demages due. AI SW in EU is treated as a Manufacturer.
  • Data standardized needed fro commercialization in EU, avoid Bias in data

• 13:05 – 13:55 LUNCH

• 13:55 – 15:00 Accelerating Company/Product Development

  (Parallel Session begins in Naphtali Building-Efter Auditorium)

• 13:55 – 14:15 Designing Genome Editing Solutions using DNA Foundational Models

  • Dr. Yaniv Shmueli, CTO and Co-Founder, Cassidy Bio, Israel

  • Dr. Yair Benita, CTO, AION Labs, Israel

  • The global push toward next-generation therapeutics is accelerating investment in cell and gene editing modalities. Yet designing safe and effective genome-editing strategies remains a major bottleneck: unintended off-target edits, genotoxicity, and inconsistent performance across cell types continue to slow clinical translation. As multiplex editing and applications across diverse cellular contexts become routine, the need for scalable, predictive, data-driven design tools is growing rapidly. In this talk, I will outline the key challenges in developing in silico models that can accurately predict genome-editing outcomes and support hit-to-lead candidate selection. I will discuss how Genomic Foundational Models can help address these challenges, and how such models can be trained on experimental datasets at scale. Finally, I will present results demonstrating predictive performance for editing efficacy and safety, along with strategies for validating these predictions through wet-lab assays and preclinical studies.

    CRISPER & Gene Editing: GuideRNA (gRNS) – marks location

    CRISPER-Cas based Genome Editing: Functionality

    >> Foundational Models in Biology

    > DNA Foundational Models Generate data in the Lab –>> predict and verify Paradigm

    >> On-Target Efficacy Prediction – desirable

    >> Off-Targer Efficacy Prediction – not desirable

    >> Repair Outcome Prediction – Prediction verified to deliver clinical confidence

    >> Reducing Experimental Burden

    7,000 monogenic disorders

    @@@

    AION Labs Portfolio – Co-development, each partner,Technology – AWS, Funding: AMITI, Talent

    –drug discovery portfolio: Principles of AION Labs COmpany Creation: $1MM funding, for a probelm Pharma initiate. Define a probelm with Pharma, problem definition – Proof of Concept

    Small molecule: Pharma came out wiht an idea: Prophet – initiated the solution for a concept offer to them

    Cassidy – Technology existed seeking for owners CRISPR DNA Language, gRNA repair outcome

• 14:15 – 14:30 CytoReason

  • Prof. Shai Shen-Orr, Co-Founder & CSO, CytoReason, Israel
  • Bring Data to life, at Scale
  • Precision medicine vs small molecule $2.3Bn to fund a new drug
  • AI+Data + Drug discovery by AI vs give tools to other people to develop drug
  • A platform to sequence more to drive scientific Decisions at Scale – Pharma R&D for Analytical Computational biology
  • Computational Disease Model – ML Translational MOA, clinical heterogeiniety – Knowledge Treatment biology – knowledgeAI component , ML, for Scalable integration with AI AGENTS
  • In BiomedicineLLM, AI Agent with NVIDIA,
  • CytoReason – Model Factory – Platform across models and within model deep modeling using Reinforced Learning  Human Feedback (RLHF) – Criteria prioritization
  • Disease models – Drug vs customer treatment

• 14:30 – 14:45 8400 The Health Network

  • Adv. Daphna Murvitz, L.L.B., Chief Integrative Medicine Officer, Samueli Integrative Cancer Pioneering Institute, Davidoff Center, RMC; Board Member, 8400 The Health Network , Israel
  • Zero gap between Public sector, investors, Academia, Policy on National infrastructure for AI, Policy on other economic and social initiatives for AI: Military to civilian, Skills, data shared in several industries,  Cybertech, 7,000 start ups 1,600 are in Healthcare
  • HealthTech success HC system, Technology strength, Infrastructure, Human Capital,
  • Ecosystem mission: global connections, acoss sectirs, From Ego system to Ecosystem

• 14:45 – 15:00 AI for Accelerating Product Development Through Better Recruitment and Trial Design

  • Steven E. Labkoff, MD, Collaborating Scientist, Beth Israel Deaconess Medical Center, Harvard Medical School, USA

  • Clinical development is becoming increasingly difficult, costly, and slow, with patient recruitment emerging as one of the most significant barriers to trial success. Many therapeutic areas—especially rare diseases and oncology—now face intense competition for the same limited patient populations. At the same time, site selection remains highly variable and often unreliable, and protocol designs frequently introduce avoidable burdens that slow enrollment or trigger costly amendments. These structural challenges collectively undermine timelines, reduce trial quality, and delay the delivery of new therapies to patients.

    This talk explores how artificial intelligence—particularly large language models and multimodal data analytics—is beginning to transform product development by addressing these core bottlenecks. Use cases include automated protocol evaluation to identify operational risks before trial launch; predictive, data-driven methods for site selection that distinguish true “platinum sites” from historically underperforming centers; and AI-enabled data preparation workflows that significantly accelerate and standardize analytics.

    Drawing on real examples from my work in clinical operations analytics and recent strategy engagements, the talk will highlight how AI can bring earlier predictability, fewer amendments, better recruitment, and more reliable feasibility. Ultimately, AI is reshaping the earliest and most critical decisions in clinical development—helping organizations deliver medicines to patients faster and with greater confidence.

    Clinical Drug development down stream, execution,

    80% of Clinical Trials are longer that planned due to recruiting hardship. COmpetition for same patients, pool of Patient per site per Month (PSM) – decreases. Molecule that shows signal is hard, to design protocols for Clinical Trials is even harder. Trial Ecosystem: Patient willing to join Trials decrease. Site selection endouring weak point. SIte Feasibility Form (SFQ). Protocol Designs a weak point. Transition fro Phase II to Phase III then 17 years of Patent life. WITH AI: Automate Protocol Design: Use Case1: LLM-based Protocol Evaluation, too much patient burden. LLM Protocol Recommendations Use Case 2: SIte Selection using MULTImodal RWD in conjunction iwth Simulations Use Case 3: AI for Reproducible Analytics Use Case 4: Clinical Trial simulation for Protocol.

    Mundane AI: Ai Literacy, Cultural Challenge, Impact of SO WHAT, RISK Over-reliance on AI in Clinical Development

    Everyday AI

    investment 20%-30%

    Research AI Current investment 60%-70%

    Business AI

    investment 20%-30%

• 15:00 – 15:10 BREAK

• 15:10 – 15:30 Gen AI Meets the Complexity of Biology

  • Dr. Michal Rosen-Zvi, Director of Healthcare and Life Sciences, IBM Research; Chief Scientist, CC-IBM Discovery Accelerator, Israel
  • Foundations Models in Health & Medicine: Protein, DNA, RNA, Amino acid representation, small molecules, biology, patients – language and its structured applied to Biomedical & Biology. Algorithm design: Representation of Data and the sequence of amino Acid. Abstractin the problem: How binding a drug molecule to a protein: 3D of molecule data spatial representaion – the abstraction – learn probability density as a differential equation as a representation of one molecule.
  • BMFM: BioMedical Foundation Model: at IBM –>>  Open source Open Science (code is in Github) ->>>>> Small molecules 
  • A Family of Novel Foundation Models: Cell Culture with Transcriptomic FM 

   

• 15:30 – 16:15 NVIDIA Start Up Panel

  (Parallel Session begins in Naphtali Building-Efter Auditorium)

  • Moderator:

    Amit Bleiweiss, Senior Data Scientist, NVIDIA , Israel

  • Perception AI, 2012, Radiology
  • Generative AI, ChatGPT 2022
  • Agentic AI, 2024
  • Physical AI, 2024
  • Data fine tuning SLM
  • Agentic AI:
  • DRY LABS:
    • Dana Sinai, PhD, VP AI, Laguna Health, Israel
    • Use of LLMs on Social workers documentation Text: Comments of edits on documentation is used as Training data
    • What AI can do – this is not a Hype – agent will perform

    • Tomer Ben David, Co-founder and CEO, Vortex Imaging, Israel
    • Use of GPUs Neuro-network libraries
    • Medical device and regulation

    • Eran Miller, Co-Founder & Chief Business Officer, MNDL Bio, Israel
    • AI-based DNA and biophysical
    • Vaccine

    • Shahar Harel, Head of AI, Quris AI, Israel
    • End to end model
    • AI is a Hype, POC to test Chemestry has different matrix not Agents

   

• 16:15 – 16.35  COFFEE BREAK

• 16:35 – 17:20 Regulating AI in Healthcare and Data for Research

  • Moderator:

    Adv. Daphna Murvitz, L.L.B., Chief Integrative Medicine Officer, Samueli Integrative Cancer Pioneering Institute, Davidoff Center, RMC; Board Member, 8400 The Health Network , Israel

• 16:35 – 16:50 Lecture

  • Assaf Parker, Head of Innovation and Technologies, Digital Health Division, Israeli Ministry of Health, Israel
  • AI-enabled technologies is a game changer
  • Ministry Oo Health Initiatives in AI: HMO’s: Mental Health, Rehab, Aging & Hospitals: Documentation, Administration of Therapeutics
• 16:50 – 17:05 

  • Eng. Inbar Blum, Director of Healthcare Innovation, Growth Division, Israel Innovation Authority, Israel
  • Investment 0.5Bn per year
  • Capital raising – ok
  • Authority Investments: Research in Academia, Labs ans DB, Innovation Center, Incubators, Deep-tech Startup Fund: Ideation, Pre-seed, Seed 1st round
  • Health-tech portfolio: ML comapnie, Healthcare sector
  • AI can fill the GapDecision Making Systems, DSS, Prevention Support System
  • AI will mitigate the shortage in health care delivery using Robots

• 17:05 – 17:20 FHIR

  • Interoperability – FHIR – Fast Healthcare Interoperability Resources
  • Benefit of a community Approach: Max reuse, efficiency, Workflow, legislation & Broad Implementation
  • Data Portability Act, IDF, Civil Aviation Authority
  • Supply chain
  • Data Standards For AI makes it easier to use.

• 17:20 – 18:05 AI in Medical Centers: How is AI Used?

17:20 – 17:35 UCI Health Susan Samueli Integrative Health Institute

• Dr. Hilla Vardi Behar, Senior Data Scientist, The Samueli Integrative Cancer Pioneering Institute, Israel
•  Mission & Vision
• Clinical significance Overall survival in Immmunotherapy-Treated Cancer Patient and monitoring the patient: Body composition Metrics – CT Scan with 3D: CHnage in Muscle,
• Model I: Detection from 3D several segment, which ones??? Detection of L3 mid-vertebra
• Model II: create segment axial 2D Train a model by data from segments

Mr. Yotam Hadari, Paramedic

AI in Shiba Hospital

-data management

-Aortic Stenosis Labeling

-model development

-AI-ECG, Ejection fraction for reduction for LVEF Multimodal Model Hospitalization vs non-hospitalization

 

 

• 17:35 – 17:50 Maccabi Healthcare Services

  Alon Yaffe, CIO, Meuhedet

  AI Strategy @Meuhedet: VIsion: “Noah” as the All go to: Staff and Customers

  1.4M Patients, 150 IT Dept., 30 years of data, 10PB

  –AI solutions for staff: MS CoPilot, automation, GitHub CoPilot, Cursor [Vibe coding], Building Agentic AI for Enterprise Data Discovery

  –AI solutions for Customer:

  –Phone Service for Non-Customers schedule an appointment

  –Phone Service for Customers schedule an appointment

  –Medical AI – Supporting Clinical Teams:

  —Risky Pregnancies – Prediction of risks in Pregnancy: Diabetes prescriptions

  —Doctor-Patient Summary

• 17:50 – 18:05 Lecture

  • Oleg Brodt, AI Specialist, INCD, Israel
  • AI is coming not here, yet
  • AI tools built to help with COVID – No success
  • Data privacy problems
  • AI can be unpredictable – what is the harm? In medical advice and FAKE advice, eating disorder
  • Hallucinations – therefore, consult your PCP.
  • Bias in X-Ray classifiers