A.I. Research
Studies that use AI, machine learning, and deep learning to assess human health, discover trends in health risks, and predict disease.
Ongoing Projects
PAC-AID — Pacific Center for AI and Data Science in Medicine
Founded at the University of Hawaiʻi Cancer Center as the AI Precision Health Institute (AI PHI) and now an NIH/NIGMS COBRE, PAC-AID uses advanced technology — AI, machine learning, and deep learning — to assess human health and predict disease risk across Hawaiʻi and the Pacific.
PAC-AID ↗ · About the transition
The Makawalu Study — Breast Cancer Screening in the Pacific using Portable Ultrasound
Addresses elevated advanced-stage breast-cancer rates in the Pacific relative to the US mainland, particularly in areas lacking mammography services. Study details →
Deep Learning and Total-Body DXA Scans (TBDXA.I.) (Completed)
Applies deep learning to total-body DXA scan data to generate predictive algorithms for health outcomes. Study details →
Three-Compartment Breast Lesion Detection (3CB / q3CB)
Examines computer-aided diagnostic systems, compositional analysis (protein, lipid, water), and radiomics in digital breast tomosynthesis. Supported by NCI R01CA257652. Study details →
NCI Federated Learning Network (P30 Supplement)
Privacy-preserving collaborative artificial intelligence across NCI-designated Cancer Centers. Evaluates distributed machine-learning models for medical imaging without transferring raw patient data or Protected Health Information across institutional boundaries.
AIM-AHEAD Hub Project
NIH AIM-AHEAD Consortium initiative developing multimodal deep-learning risk models for high-disparity cancers (breast, liver/HCC, and lung) using longitudinal Electronic Health Record (EHR) data and clinical imaging from diverse Pacific populations.
Related Publications
Recent AI/ML work from the lab spans breast-lesion detection in ultrasound images, mammographic breast-density assessment, mortality prediction from DXA imaging, and skin-cancer diagnosis in multiethnic populations.