Seminars

Abstract: Sleep and neurodegeneration are deeply intertwined, with a potentially bi-directional relationship that remains challenging to disentangle. In this talk, I will highlight how large-scale datasets—from deeply phenotyped community-based cohorts to real-world electronic health records—are helping us better understand this two-way interplay between sleep and aging brain health. Drawing on findings across multiple longitudinal cohorts and EHR-linked populations, I will show how multidimensional sleep data (including polysomnography, actigraphy, contactless monitoring, and emerging digital biomarkers) can reveal early signatures of Alzheimer’s disease (AD) and Parkinson’s disease (PD), and how sleep disturbances may function both as modifiable risk factors and early prodromal indicators. By integrating epidemiology and AI-driven analytics, this work demonstrates how big data can clarify the temporal dynamics between sleep and neurodegeneration, shed light on underlying mechanisms, and advance precision sleep medicine strategies aimed at reducing risk for AD and PD.

Bio: Yue Leng is an Associate Professor of Psychiatry at the University of California, San Francisco (UCSF) and a Senior Atlantic Fellow at the Global Brain Health Institute. She holds an MPhil and PhD in Epidemiology from the University of Cambridge and completed her postdoctoral training at UCSF. Dr. Leng’s research focuses on elucidating the links between sleep, circadian rhythms, and the risk of Alzheimer's and Parkinson's disease, employing epidemiological methods alongside AI and big data analytics. Her research has been funded by NIH K99/R00, R21, and R01 awards and has received widespread media coverage.

Abstract: A novel representation called the evolutionary s-rep represents anatomic objects by encompassing an unusually rich collection of geometric features, image intensity based features, genetic features, etc.  It can be used for shape-based classification of anatomic objects, hypothesis testing on inter-class shape variations, and object segmentation, using web-resident software. The evolutionary s-rep and example applications will be discussed. This new representation provides better statistics than alternatives because its across-object positional correspondences are based on geometry not just on the object boundary but also in its interior, on second-order and not just positional (0th order) features, and on finding for each case a diffeomorphism from a basic, ellipsoidal object.

Bio: 

My PhD dissertation, in 1967, was the first in medical image computing, and I have been teaching, writing, and doing research in that area continuously since my first summer job at Massachusetts General Hospital in 1962 and since my joining the UNC Computer Science faculty in 1967. I have had many collaborations across the UNC Medical School, in some of whose departments I have had adjunct appointments. I led the committee that led to the formation of the Biomedical Research Imaging Center at UNC. 61 PhDs have been produced with me as principal advisor.

My early research emphases were on image quality restoration and on 2D and 3D display, as well as related models of human vision. I helped advise Charles Metz’s dissertation in that area, and we collaborated for some years, including a paper that was an early component of the EM algorithm. However, for the last 4.5 decades my focus has been on geometric models of anatomic objects especially suited for statistics, statistical analyses of these, and the applications of that in diagnosis, treatment planning, and object segmentation and registration, with clinical targets all over the body and with many image sources. Major medical targets were radiation oncology, neuroscience, and recently colonoscopy. The form of model I have developed is skeletal, and in particular what we call the evolutionary s-rep, with its advantages over object boundary based models of also locally capturing interior properties such as object curvature and cross-object width and of providing an object-relative coordinate system important in accessing image intensities as they are used for segmentation and registration. Towards diagnosis and other statistical objectives, my statistics professor colleague JS Marron and I have made important contributions in methods of statistics of shape that recognize that object geometry cannot be directly analyzed by Euclidean methods because abstractly it resides on a curved manifold. Most especially, the method called Principal Nested Spheres (CPNS), allowing statistical analysis of directional data, was developed in our laboratory. Collaborations with Kitware, Inc. in software development and tutorials related to shape analysis in the salt.slicer.org toolkit including s-reps uses, especially by Jared Vicory, have been and continue to be important. Successes of a variety of types for statistics on s-reps include the commercial success of the company, Morphormics, now part of Accuray, that we spun off and whose main product at that time, built upon statistics of skeletal models, focused on segmentation of male pelvic organs from CT for radiation treatment planning. Other work showed registrations and segmentations of mobile structures across medical imaging modalities. Our work over the last decade or so shows that our latest form of s-reps that evolve from ellipsoids while according to rich geometric properties of the object interior and boundary yield notable improvements over boundary point distribution models and models based on smooth deformations of means  for object representation in classification, hypothesis testing  and production of correspondence across a population of neuroanatomic objects.

Abstract: Understanding tumor heterogeneity remains one of the greatest challenges in cancer biology and precision medicine. My research integrates high-throughput single-cell analysis, microfluidics, and artificial intelligence to dissect cellular heterogeneity, cancer cell migration, and therapeutic resistance. We developed microfluidics capable of analyzing limited cell numbers for single-cell RNA-Seq and profiled 666 circulating tumor cells (CTCs) from 21 breast cancer patients, revealing distinct CTC subpopulations with potential implications for treatment selection. In parallel, we engineered high-throughput microfluidic systems that enable quantitative analysis of tens of thousands of migrating cancer cells per chip. Using scalable injection molding fabrication and robotic automation, we performed large-scale screening of 2,726 compounds, identifying promising inhibitors of cancer cell migration and metastasis. To address therapy resistance, we established an image-based single-cell morphological profiling pipeline to characterize polyploid giant cancer cells (PGCCs), a subpopulation that survives chemotherapy and regenerates tumor cells. High-throughput screening identified several anti-PGCC compounds that overcome treatment-induced resistance in breast cancer models. Building upon these experimental datasets, we developed deep learning–based virtual screening models that integrate chemical, morphological, and literature-derived features to perform in silico prediction of >24,000 small molecules. This AI framework enables efficient exploration and prioritization of potential candidates within a vast search space. Together, these technologies establish a powerful multidisciplinary platform to decode cancer cell heterogeneity, uncover therapeutic vulnerabilities, and accelerate the discovery of precision treatments.

Bio: Yu-Chih Chen received his dual bachelor degrees in Electrical Engineering and Law from the National Taiwan University, Taipei in 2008, his Ph.D. degree in Electrical & Computer Engineering at the University of Michigan, Ann Arbor in 2014, where he continued to work as a research faculty in both Electrical & Computer Engineering Department and Forbes Institute for Cancer Discovery. He is currently an assistant professor in the Department of Computational and Systems Biology and UPMC Hillman Cancer Center at the University of Pittsburgh School of Medicine. He is also affiliated with Joint CMU-Pitt Ph.D. Program in Computational Biology. He is the recipient of Taiwan Semiconductor Manufacturing Company (TSMC) Outstanding Student Research Award (2008), Orenstein Ph.D. Fellowship (2009), Best Post-Doctoral Speaker Award in Microfluidics in Biomedical Sciences Training Program, UMich (2015), Emerging Forbes Scholar selected by Forbes Institute for Cancer Discovery (2017), and Hillman Early Career Fellow for Innovative Cancer Research (2021). His current research focuses on high-throughput single-cell analysis and deep learning for cancer precision medicine.

Abstract: TBA

Bio: TBA

Abstract: The dream of precision health is to develop a data-driven, continuous learning system where new health information is instantly incorporated to optimize care delivery and accelerate biomedical discovery. The confluence of technological advances and social policies has led to rapid digitization of multimodal, longitudinal patient journeys, such as electronic health records (EHRs), imaging, and multiomics. Our overarching research agenda lies in advancing multimodal generative AI for precision health, where we harness real-world data to pretrain a virtual patient model as digital twins for patients in forecasting disease progression and treatment response. This enables us to synthesize multimodal, longitudinal information for millions of cancer patients, and apply the population-scale real-world evidence to advancing precision oncology, in deep partnerships with real-world stakeholders such as large health systems and life sciences companies.

Bio: Hoifung Poon is the General Manager of Real-World Evidence at Microsoft Research and an affiliated faculty at the University of Washington Medical School. He leads biomedical AI research and incubation, with the overarching goal of structuring medical data to optimize delivery and accelerate discovery for precision health. His team and collaborators are among the first to explore large language models (LLMs) and multimodal generative AI in health applications, producing popular open-source foundation models such as PubMedBERT, BioGPT, BiomedCLIP, LLaVA-Med, BiomedParse, with tens of millions of downloads. His latest publications in Nature and Cell features groundbreaking digital pathology and spatial proteomics foundation models such as GigaPath and GigaTIME. He has led successful research partnerships with large health providers and life science companies, creating AI systems in daily use for applications such as molecular tumor board and clinical trial matching. His prior work has been recognized with Best Paper Awards from premier AI venues such as NAACL, EMNLP, and UAI, and he was named the "Technology Champion" by the Puget Sound Business Journal in the 2024 Health Care Leadership Awards. He received his PhD in Computer Science and Engineering from the University of Washington, specializing in machine learning and NLP.

Abstract: Periodontitis is a prevalent chronic inflammatory disease and a leading cause of tooth loss worldwide, with substantial implications for systemic health and health care utilization. Despite its burden, early identification of individuals at high risk for disease onset and progression remains challenging in routine clinical practice. This seminar will describe a data-driven framework for developing, testing, and validating prediction models for periodontitis using real-world electronic dental record (EDR) and linked health data. The talk will cover key methodological considerations, including cohort construction from longitudinal clinical data, outcome definition using contemporary periodontal classifications, feature engineering from structured dental and medical records, and model development using machine-learning approaches. Emphasis will be placed on internal and external validation strategies, transportability across heterogeneous populations and care settings, and evaluation of model performance, calibration, and clinical utility. The seminar will also discuss challenges related to data quality, missingness, bias, and fairness, as well as strategies such as federated learning to enable multi-institutional validation without sharing patient-level data. Together, this work highlights how predictive modeling can support earlier risk stratification, inform personalized preventive strategies, and advance the integration of clinical decision support tools into dental practice, while illustrating broader lessons for applying biomedical data science methods to oral health research.

Bio: Jay Patel (BDS, MS, PhD in Informatics) is an Assistant Professor and Director of the Center for Artificial Intelligence (AI), Data Science, and Informatics at Temple University Kornberg School of Dentistry. He is among the few in the US with formal training in both clinical dentistry and a doctorate in informatics, an interdisciplinary background that uniquely positions him to lead transformative AI-driven dental research. Dr. Patel’s work focuses on developing AI models and software applications using large-scale electronic health record data to predict disease onset and progression, with the ultimate goal of improving prevention strategies. His research also emphasizes the integration of oral and systemic health through the linkage of medical and dental records to enable real-time health information exchange and support comprehensive data-driven studies. He has developed over 40 natural language processing pipelines and electronic dental record quality metrics to extract valuable insights from unstructured clinical notes. Dr. Patel has been an invited speaker and panelist at leading institutions, including the National Institute of Dental and Craniofacial Research and the National Institutes of Health, where he has presented his pioneering work in dental AI. He has received several NIH-NIDCR-funded awards as both Principal Investigator and Co-Investigator, along with prestigious foundation awards including the William Buttler Award, Robert Wood Johnson Foundation, New Jersey Health Foundation, and the CareQuest Institute. Dr. Patel has published over 50 peer-reviewed papers and is the recipient of multiple national honors, including a U.S. patent for innovation in dental diagnostics.