Artificial Intelligence / Machine Learning Demonstration Projects 2025

Crowdsourcing ideas to bring advances in data science, machine learning, and artificial intelligence into real-world clinical practice.

A Learning Healthcare System, as envisioned by the Institute of Medicine and the NIH Collaboratory, aims to bring new knowledge quickly into the healthcare delivery system, including advances in data science, machine learning, and artificial intelligence.   

This is the second call for UCSF’s Artificial Intelligence/Machine Learning Demonstration Projects - “AI Pilots” - aimed at advancing UCSF as a Learning Healthcare System, sponsored by UCSF and UCSF Health Leadership. This also represents the fourth institution-wide Call for Proposals for Learning Health System Demonstration Projects. 

Funded project teams will receive investigator salary support for 1 year (up to 10% effort for project lead(s) starting July 1, 2025), with an option to extend for a second year, depending upon project progress.  

To aid in the development and improvement of proposals, "office hours" will be offered throughout Phase 1 and 2, for open discussions around possible projects and submissions. Office hours will include a data scientist and a data engineer from the Health AI team to help scope and refine project goals and the APex Enabled Research (AER) team to discuss feasibility. Additionally, during Phase 2 (Open improvement), the UCSF community can comment, follow and like proposals while researchers may refine their proposals.

Read important information about this forum

  • March 3rd - April 6th , 11:59pm: Phase 1 - Open Submission
    • Develop and submit a 3-Page Proposal and a Letter of Support
    • Subscribe to email updates to view new proposals and comments
  • April 7th – April 22nd , 11:59pm: Phase 2 - Open Improvement
    • Browse proposals and comment to improve others' proposals 
    • Review comments on your own proposal and revise if you like 
  • May 15th: Announcement of selected projects
  • July 1st: Funding period begins
Review

Commenting is closed.

Latest Announcement

Review of AI Pilots has been completed

Submitted by Zeanid Noor on May 29, 2025 - 10:59pm Last revised by Zeanid Noor on May 29, 2025 - 11:12pm.

Dear AI Pilot Proposers

 As you know, we’ve had a spectacular response to this RFP and have now completed review of the 40 proposals. Congratulations to the following projects that the review committee selected as winners:

We will be in touch with project leads for next steps.

While we could only award a few projects, we do have several ideas in the works to help the remaining projects / PIs continue to make progress on the exciting and innovative work that has been proposed.  We’ll invite all unselected project teams to an information session to be held this Summer to share these ideas and hear feedback about what might be useful.

This was a highly competitive RFP and we’re encouraged by the breadth of ideas & enthusiasm to bring ‘new knowledge quickly into the healthcare delivery system, including advances in data science, machine learning, and artificial intelligence.’  Thank you for your submission.

AI Pilots Review Committee

 

See all announcements

Proposals (40 total)

Displaying 21 - 30

Pages

AI-Driven Endometriosis Symptom and Risk Assessment Tool for Personalized Patient Management

Submitted by Jeannette Lager on April 6, 2025 - 1:26am Last revised by Zeanid Noor on May 29, 2025 - 11:32pm.
Proposal Status: 
Review Complete

The UCSF Health Problem: Endometriosis is a chronic, debilitating condition affecting approximately 10% of reproductive-age women worldwide. It is characterized by the growth of endometrial-like tissue outside the uterus, leading to severe pelvic pain, dysmenorrhea, dyspareunia, bowel and bladder dysfunction, and infertility. Despite its prevalence, endometriosis remains significantly underdiagnosed, resulting in profound patient suffering and substantial economic burden.

A Multimodal Foundation Model for Enhanced Prostate Cancer Care at UCSF Health

Submitted by Seyed Mansour Abtahi on April 5, 2025 - 11:15pm Last revised by Zeanid Noor on May 29, 2025 - 11:33pm.
Proposal Status: 
Review Complete

1.  The UCSF Health Problem

Prostate cancer is the most commonly diagnosed malignancy among men in the United States. Accurate diagnosis, staging, and monitoring are critical for effective treatment planning and improving patient outcomes. Imaging modalities, such as Prostate-Specific Membrane Antigen (PSMA) Positron Emission Tomography (PET), CT, and MRI, provide valuable insights but often lack the integration necessary to fully capture the complexity of prostate cancer progression. Current challenges include:

Implementation of an AI-Powered Platform for Scalable, Real-Time Lab Monitoring in Immunosuppressive Treatment

Submitted by Augusto Garcia Agundez Garcia on April 5, 2025 - 9:36pm Last revised by Zeanid Noor on May 29, 2025 - 11:34pm.
Proposal Status: 
Review Complete

1. The UCSF Health Problem

Optimizing New Patient Self-Scheduling Pathways with AI/ML

Submitted by Nicole Jiam on April 5, 2025 - 8:42am Last revised by Zeanid Noor on May 29, 2025 - 11:35pm.
Proposal Status: 
Review Complete

Section 1: The UCSF Health Problem

At UCSF and other leading academic medical centers, the referral intake and triage process for new patients is strained, leading to long delays and high rates of incomplete referrals.

CLEAR-CHE: Covert Liver Encephalopathy Assessment using Recorded Clinical Health Encounters

Submitted by Jin Ge on April 4, 2025 - 11:29pm Last revised by Zeanid Noor on May 29, 2025 - 11:35pm.
Proposal Status: 
Review Complete

Early-stage or covert hepatic encephalopathy (CHE), which can be present in up to 60% of patients with cirrhosis, often goes unrecognized by both patients and clinicians.  Without timely identification and intervention, such as initiating or titrating lactulose or rifaximin, patients can rapidly progress to overt HE.  This is strong emerging evidence that suggest a patient’s voice can serve as a biomarker for CHE (PMID 35861546, PMID 39264936).  This proposal seeks to leverage ambient AI scribe recordings obtained through routine clinical care as a voice-based biomarker for CHE detection.

LiVersa-CirrhosisRx: Integrating a Liver Disease Specific LLM within Clinical Decision Support System for Cirrhosis Care

Submitted by Jin Ge on April 4, 2025 - 11:04pm Last revised by Zeanid Noor on May 29, 2025 - 11:36pm.
Proposal Status: 
Review Complete

We have previously constructed a liver disease specific Large Language Model (LLM) called “LiVersa” using Retrieval Augmented Generation (PMID 38451962, PMID 38935858).  We have also constructed “CirrhosisRx,” which is a rule-based (non-AI) clinical decision support system for guideline adherent inpatient cirrhosis care, built on the EngageRx platform using SMART-on-FHIR (NCT05967273, PMID 38407255).  In this proposal, we plan to integrate LiVersa into CirrhosisRx via Fast Healthcare Interoperability Resources (FHIR) application progressing interface (API) calls to provide a patient-personalized dynamic information retrieval system to provide guideline-based clinical recommendations to clinician end-users.

SPICE-LD: Supporting Personalized, Inclusive, Culturally-appropriate Eating in Liver Diseases

Submitted by Jin Ge on April 4, 2025 - 10:52pm Last revised by Zeanid Noor on May 29, 2025 - 11:36pm.
Proposal Status: 
Review Complete

Patients with chronic liver diseases, including those with cirrhosis and those awaiting transplantation, have unique and often stringent nutritional requirements.  There is an unmet need to provide treating clinicians with culturally-appropriate, evidence-based, and patient-specific nutritional recommendations that resonates with that person’s dietary norms and preferences.  This proposal seeks to bridge this gap through a culturally appropriate and personalized RAG-LLM assistant that supports nutritional counseling in liver diseases.

Automated Knee Osteoarthritis Grading Decision Support Tool

Submitted by Eugene Ozhinsky on April 4, 2025 - 5:13pm Last revised by Zeanid Noor on May 29, 2025 - 11:37pm.
Primary Author: Yuntong Ma
Proposal Status: 
Review Complete

Plain knee radiographs are a cornerstone in evaluating osteoarthritis (OA), and at UCSF, nearly all such studies include Kellgren-Lawrence (KL) grading to assess severity. While essential, this grading task is repetitive, time-consuming, and inherently subjective—especially in borderline cases. In fact, studies show only moderate inter-reader reliability in KL grading, creating inconsistencies in diagnosis and downstream care decisions. KL grading adds to radiologist workload, reducing time available for complex cases and clinical consults. We propose integrating an automated AI model for KL grading directly into the radiology workflow at UCSF. By deploying this AI system at the point of care, we can support more efficient workflows, better diagnostic consistency, and enhanced trainee learning.

 

Precision Summarization: Knowledge-Grounded AI for High-Impact Specialty Care

Submitted by Vivek Rudrapatna on April 4, 2025 - 11:20am Last revised by Zeanid Noor on May 29, 2025 - 11:38pm.
Proposal Status: 
Review Complete

The UCSF Health Problem

We aim to automate specialty-specific chart review, a high provider burden, error-prone process that costs clinicians 15–45 minutes per patient, leading to 6–12 hours of weekly administrative overhead contributing to pajama time. This burden impairs clinic throughput, contributes to burnout, and compromises care quality, especially in specialties like Gastroenterology where fragmented, longitudinal data including scanned outside records are the norm.

Pages