From bits to bedside: Translating routine clinical data into precision mammography

Dexter Hadley (University of California, San Francisco)

4:00pm–4:40pm Wednesday, September 11, 2019

Location: LL21 E/F

AI Business Summit, Impact of AI on Business and Society

Secondary topics: Health and Medicine, Machine Learning, Mobile Computing, IoT, Edge, Text, Language, and Speech

Average rating:

(5.00, 1 rating)

Who is this presentation for?

CEOs, CTOs, and CMOs of healthcare institutions and startups; clinicians working in healthcare and breast cancer; and academics working on health AI research

Level

Beginner

Description

Typically, large healthcare institutions have large-scale quantities of clinical data to facilitate precision medicine through an AI paradigm. However, this so-called big data is hardly translated into improved patient care, because AI algorithms like deep learning cannot readily ingest or reason over it.

Dexter Hadley details how the University of California, San Francisco (UCSF), uses natural language processing techniques to curate routine clinical data for over 1M mammograms at UCSF for important breast cancer outcomes. It then uses deep learning and proposes blockchain approaches to AI that can learn clinically relevant models to detect breast cancer on mammography. The end point of the work is to directly apply AI on routine clinical data for breast cancer to improve patient outcomes.

What you'll learn

Understand how routine data generation in healthcare is exploding and serves as a valuable resource to define clinically relevant AI today

Dexter Hadley

University of California, San Francisco

Dexter Hadley as an assistant professor of pedatrics, pathology, and laboratory medicine at the University of California, San Francisco (UCSF). His expertise is in translating big data into precision medicine and digital health. His background is in genomics and computational biology, and he has training in clinical pathology. His research generates, annotates, and ultimately reasons over large multimodal data stores to identify novel biomarkers and potential therapeutics for disease. His early work resulted in a successful precision medicine clinical trial for ADHD (Clinicaltrials.gov identifier NCT02286817) for a first-in-class, nonstimulant neuromodulator to be targeted across the neuropsychiatric disease spectrum. More recently, his laboratory was funded by the NIH Big Data to Knowledge Initiative to develop the Stargeo.org online portal to crowdsource annotations of open genomics big data that allows users to discover the functional genes and biological pathways that are defective in disease. He also develops state-of-the-art data-driven models of clinical intelligence that drive clinical applications to more precisely screen, diagnose, and manage disease. Toward this end, he has been recognized by UCSF with various awards including the inaugural UCSF Marcus Award for Precision Medicine to develop a digital learning health system to use smartphones to screen for skin cancer as well as a pilot award in precision imaging to better screen mammograms for invasive breast cancer. In general, the end point of his work is rapid proofs of concept clinical trials in humans that translate into better patient outcomes and reduced morbidity and mortality across the spectrum of disease.