HCC-CareAgent: A Pilot Multi-Agentic AI System for Future of Cancer Care

Presentation Type: Ignite-Style Talk
Presentation Time: 11:15 AM - 11:25 AM

Abstract Keywords: Generative AI in Clinical Workflow: Ambient Listening, Chart Summarization, Automated Response with LLM, Clinical Decision Support and Care Pathways, Analytical Artificial Intelligence: ML, Digital Pathology, Imaging AI, Predictive Analytics, Governance
Primary Track: Big Data for Health


Artificial intelligence (AI) is rapidly transforming oncology, with generative AI and large language models (LLMs) enabling new forms of automated reasoning and multidisciplinary decision support. Recent advances in agentic AI systems—where multiple specialized AI agents collaborate autonomously—present a promising but underexplored opportunity to model the complex, multi-step workflows inherent to cancer care. In this study, we introduce HCC-CareAgent at the University of Pittsburgh Medical Center (UPMC) Hillman Cancer Center (HCC), a pilot multi-agent AI system designed to simulate interdisciplinary expertise across the continuum of cancer diagnosis, treatment planning, and post-treatment surveillance. Built using the Agent Development Kit (ADK) and Agent2Agent (A2A) communication protocol, HCC-CareAgent integrates seven guideline-grounded agents representing intake, radiology, pathology, surgery, oncology, and surveillance, all orchestrated by a central coordinating agent. Each agent is implemented using a locally deployed LLaMA 3.2 model to ensure data privacy and compatibility with clinical environments.
We evaluated the system using synthetic bone cancer cases to assess workflow coordination, logical consistency, and alignment with authoritative guidelines such as NCCN, CAP, and ASTRO. HCC-CareAgent successfully generated coherent diagnostic pathways, treatment recommendations, and surveillance plans, demonstrating the feasibility of agentic AI for modeling complex oncology workflows. Although preliminary, these results highlight the potential of agentic systems to support future oncology practice by enhancing care coordination, reducing clinician cognitive burden, and enabling scalable, guideline-concordant decision support. Ongoing work includes expanding collaboration across the Hillman Cancer Center, integrating domain expert feedback, and evaluating the system using real-world patient data across multiple cancer types.

Speaker(s):
Yanshan Wang, PhD
University of Pittsburgh

Author(s):
Praz Nagarajan, BS - University of Pittsburgh; SUMIT Kapoor, MD - Baylor College of Medicine; Sumit Kapoor, MD - University of Pittsburgh Medical Center; Yanshan Wang, PhD - University of Pittsburgh;

---

Expedited Laboratory & Medication Governance with Provider–Analyst Dyads

Presentation Type: Ignite-Style Talk
Presentation Time: 11:25 AM - 11:35 AM

Abstract Keywords: Leadership and Strategy, Diagnostics, Change Management
Primary Track: Driving Change at Scale through Effective Leadership and Governance


This Ignite session describes an expedited governance model for laboratory and ambulatory medication content in the EHR, led by a physician informaticist–analyst dyad. Using structured intake, a governance-approved naming standard, and clear routing of routine, coaching, and controversial requests, we improved build turnaround, standardized reference laboratory procedures, and achieved record-fast Aura interface go-lives. We are now extending the same approach to Clinical Pharmacy Practitioner-analyst dyads for ambulatory medication governance.

Speaker(s):
Peter Mack, MD, MS
Novant Health

Author(s):
Peter Mack, MD, MS - Novant Health;

---

From Concept to Clinic: A Collaborative Model for High-Impact Translational Research

Presentation Type: Ignite-Style Talk
Presentation Time: 11:35 AM - 11:45 AM

Abstract Keywords: Innovation Partnerships, Implementation Science, and Learning Health Systems, Outcomes Improvement and Equity, Social Determinants of Health (SDoH)
Primary Track: Innovation through Industry, Public Health, Non-profit and Commercial Partnerships


Healthcare organizations need innovation faster than traditional research pathways allow. Limited funding and operational burden slow the validation of promising technologies, delaying patient impact. Partnerships with startups enable health systems to streamline evidence generation and shorten the time from idea to clinical use. Mission-aligned investment strengthens these collaborations, advancing NIH principles of translational research while rapidly delivering solutions to historically underfunded, high-need communities such as the Gulf South.

Speaker(s):
Grace Williams, PhD
Ochsner Health

Author(s):
Grace Williams, PhD - Ochsner Health; Jason Hill, MD, MSc, MMM - OchsnerHealth; Kevan Simms, MBA, PMP - Ochsner Health; Lisa Birdsall-Fort, MD, MPH - Ochsner Health; Daniel Fort, PhD, MPH - OchsnerHealth; Christopher Girardo, DO - Ochsner Health;

---

Discount Artificial Intelligence Safety and Effectiveness Evaluation (DAISEE)

Presentation Type: Ignite-Style Talk
Presentation Time: 11:45 AM - 11:55 AM

Abstract Keywords: Human Factors and Usability, Generative AI in Clinical Workflow: Ambient Listening, Chart Summarization, Automated Response with LLM, Change Management
Primary Track: Advancing Wellness for Providers and Community with Consideration of Human Factors


Rapid emergence of Artificial Intelligence (AI)-based tools in clinical practice has sparked a surge of evaluation frameworks to ensure tools are safe and effective for healthcare. In practice, however, the push to deploy new tools is like a freight train; efforts to build elaborate guardrails on it can rarely keep up. DAISEE is an applied informatics strategy like Discount Usability, 1 proposed in 1989, to produce benefits of a methodology without overwhelming costs or time.

Speaker(s):
Susan Regli, PhD
University of Pennsylvania Health System

Author(s):
Susan Regli, PhD - University of Pennsylvania Health System;

---

Lifestyle Informatics: An Emerging Field in Clinical Informatics

Presentation Type: Ignite-Style Talk
Presentation Time: 11:55 AM - 12:05 PM

Abstract Keywords: Population Health, Digital Therapeutics, Remote Patient Monitoring (RPM), and Digital Engagement, Wearable Health Technology, Clinical Decision Support and Care Pathways, Social Determinants of Health (SDoH), Standards, Terminology, and Interoperability, TEFCA, FHIR, Precision Medicine, Multi-Omics, and Pharmacology Integration, Outcomes Improvement and Equity, Clinician Well-Being
Primary Track: Implementing Real-World Change, Digital Engagement, and Connected Health


Lifestyle medicine offers powerful tools to prevent and improve chronic disease, yet its impact is limited without strong informatics integration. This presentation highlights how data from daily behaviors like nutrition, movement, sleep, stress and social factors, can be transformed into actionable insights through standards, interoperability and decision support. By connecting personal health data to clinical workflows, informatics enables scalable, personalized, and equitable lifestyle interventions that ultimately improve patient outcomes and reduce healthcare burden.

Speaker(s):
Gagandeep Anand, MD
University of Minnesota & Thrive Wellness Institute

Author(s):
Gagandeep Anand, MD - University of Minnesota & Thrive Wellness Institute;

---

Pets, Patients, and Perspectives: A Survey of Attitudes Toward Veterinary Data Use and One Health Data Linkage

Presentation Type: Ignite-Style Talk
Presentation Time: 12:05 PM - 12:15 PM

Abstract Keywords: Data Privacy, Cybersecurity, Reliability, and Security, Ethics, Analytics, Registries, and the Digital Command Center
Primary Track: Big Data for Health


Veterinary medical data represent an underutilized resource for accelerating translational science and precision medicine. We evaluated animal owners’ perspectives on sharing de-identified veterinary health data for research, including One Health applications and cross-species data linkage. Using a structured survey, we assessed awareness of veterinary data collection, comfort with data sharing across domains, governance priorities, and willingness to link owner and animal health data. Insights will inform ethical frameworks and governance practices for responsible data use.

Speaker(s):
Katie Mullen, DVM, MS
UNC Chapel Hill

Author(s):
Nadia Saklou, DVM, PhD - Colorado State University; Adam Kiehl, Master of Applied Statistics - Colorado State University; Melissa Haendel, PhD - University of North Carolina at Chapel Hill; Tracy Webb, DVM, PhD - Colorado State University;

---

Optimizing Well-Child Visits with Clinical Decision Support Tools

Presentation Type: Ignite-Style Talk
Presentation Time: 12:15 PM - 12:25 PM

Abstract Keywords: Clinical Decision Support and Care Pathways, Population Health, Digital Therapeutics, Remote Patient Monitoring (RPM), and Digital Engagement, Outcomes Improvement and Equity
Primary Track: Implementing Real-World Change, Digital Engagement, and Connected Health


This Ignite talk describes how we improved Medicaid Well-Child Visit completion rates at Waco Family Medicine by developing a Clinical Decision Support (CDS) tool. This tool enables the real-time identification of children who are due or overdue for preventive care, replacing inefficient manual workflows. By automating eligibility detection and supporting both outreach and opportunistic scheduling, this intervention improved visit completion rates, strengthened preventive care delivery, and enhanced the financial sustainability of our organization.

Speaker(s):
Mohammad Mahmoud Mohammad, Dr
Waco Family Medicine

Author(s):
Mohammad Mahmoud Mohammad, Dr - Waco Family Medicine; John Gill, MD, FAAFP, FAMIA - Waco Family Medicine - Institute; Kelly Nieves, MD - Waco Family Medicine;

---