About
Theo Dapamede is a physician-scientist and postdoctoral fellow in the HITI Lab, Department of Radiology and Imaging Sciences, Emory University. His research applies artificial intelligence to medical imaging, with a focus on opportunistic screening: using images already taken for one purpose to uncover hidden signs of other diseases.
His lead work shows that AI-based quantification of breast arterial calcification on routine screening mammograms predicts cardiovascular disease and mortality in women. The study was published in the European Heart Journal (2026) and covered by more than 110 news outlets, including The Washington Post (plain-language summary). He has also developed a deep learning model that screens for chronic kidney disease from chest radiographs (eClinicalMedicine, 2026), and is extending the approach to heart failure and other cardiovascular conditions.
Alongside this, Theo studies how clinical AI performs after it leaves the lab: evaluating commercial AI tools in real-world use and building large language model pipelines for automated post-deployment monitoring. He also works on the image preprocessing that makes models generalise, from showing the importance of DICOM look-up tables (LUT) to co-developing RadReader, a Python package that reads DICOM images as a radiologist would see them.
Theo earned his medical degrees from Universitas Brawijaya, Indonesia, a Master of Health Technology Innovation from the University of Sydney, and a PhD from the University of Otago, New Zealand, where he built constrained neural networks for MARS spectral photon-counting CT. His work has been recognised with the SIIM Best Research Paper Award (2025) and the Emory School of Medicine High Impact Research Award (2026), and he serves on the 2026โ2027 Early Career Editorial Board of Radiology: Cardiothoracic Imaging.
