Cross-Institutional Structured Radiology Reporting for Lung Cancer Screening

We developed an open-source LLM to generate structured Level-2 radiology reports from free-text descriptions, addressing formatting errors, hallucinations, and privacy concerns. Using our dynamic template-constrained decoding scheme, our method improved LLAMA-3.1 by 10.42% and outperformed GPT-4o by 17.19%.

Chuang Niu, Parisa Kaviani, Qing Lyu, Mannudeep K. Kalra, Christopher T. Whitlow, Ge Wang

arXiv, 2409.18319, 2024

Review Article: Physics-Inspired Generative Models in Medical Imaging

Physics-inspired generative models, in particular diffusion and Poisson flow models, enhance Bayesian methods and promise great utilities in medical imaging. This review examines the transformative role of such generative methods. This review will hopefully give peers and learners a timely snapshot of this new family of physics-driven generative models.

Dennis Hein, Afshin Bozorgpour, Dorit Merhof, Ge Wang

arXiv, 2407.10856, 2024 (to appear in Ann Rev BME, IF=12.8)

Call for Collaboration: Towards a Nomenclature Consensus for Diverse Intelligent Systems

Disagreements about language use are common both between and within fields. Where interests require multidisciplinary collaboration or the field of research has the potential to impact society at large, it becomes critical to minimize these disagreements where possible.

Brett Kagan, Michael Mahlis, Anjali Bhat, Josh Bongard, Victor M. Cole, Phillip Corlett, Christopher Gyngell, Thomas Hartung, Bianca Jupp, Michael Levin, Tamra Lysaght, Nicholas Opie, Adeel Razi, Lena Smirnova, Ian Tennant, Peter Thestrup Wade, Ge Wang

Innovation 5(5):100658, 2024 (IF=33.1)

Specialty-Oriented Generalist Medical AI for Chest CT Screening

Modern medical records include a vast amount of multimodal free text clinical data and imaging data. Fully mining such big data requires multitasking; otherwise, occult but important aspects may be overlooked. Here we propose the first-of-its-kind medical multimodal-multitask foundation model (M3FM) with application in lung cancer screening and related tasks.

Chuang Niu, Qing Lyu, Christopher D. Carothers, Parisa Kaviani, Josh Tan, Pingkun Yan, Mannudeep K. Kalra, Christopher Whitlow, Ge Wang

arXiv, 2024.100658, 2024 (pending review)

Learned Alternating Minimization Algorithm for Dual-domain Sparse-View CT Reconstruction

We propose a Learned Alternating Minimization Algorithm (LAMA) for dual-domain sparse-view CT image reconstruction. LAMA is induced by a variational model with learnable nonsmooth nonconvex regularizers, which are parameterized as composite functions of deep networks. We prove that LAMA is convergent for reliable image reconstruction.

Chi Ding, Qingchao Zhang, Ge Wang, Xiaojing Ye, Yunmei Chen

Proc. International Conference on MICCAI, 173-183, 2023

Medical Imaging in Increasing Dimensions

In this popular science article, the historical development of medical imaigng technologie is discussed along with the bright perspective of this critical healthcare field. The story is presented through the lens of increasing image dimentionality from two to many and from real to imaginary.

Ge Wang

American Scientist 111:294-301, 2023

Development of Metaverse for Intelligent Healthcare

In this perspective, we present metaverse use cases, including virtual comparative scanning, raw data sharing, augmented regulatory science and metaversed medical intervention. We also identify specific action items for coordinated efforts to build the MeTAI metaverse.

Ge Wang, Andreu Badal, Xun Jia, Jonathan S. Maltz, Klaus Mueller, Kyle J. Myers, Chuang Niu, Michael Vannier, Pingkun Yan, Zhou Yu, Rongping Zeng

Nature Machine Intelligence 4:922-929, 2022

Deep Learning for Tomographic Image Reconstruction

In this review, we provide a general background, highlight representative results with an emphasis on medical imaging, and discuss key issues that need to be addressed in this emerging field. In particular, tomographic imaging is an integral part of modern medicine, and will play a key role in personalized, preventive and precision medicine and make it intelligent, inexpensive and indiscriminate.

Ge Wang, Jong Chul Ye and Bruno De Man

Nature Machine Intelligence 2:737-748, 2020

Machine Learning for Tomographic Imaging

We present a unified overview of deep-learning-based tomographic imaging. Key concepts, including classic reconstruction ideas and human vision inspired insights, are introduced as a foundation for a thorough examination of artificial neural networks and deep tomographic reconstruction. X-ray CT and MRI reconstruction methods are covered in detail, and other medical imaging applications are discussed as well.

Ge Wang, Yi Zhang, Xiaojing Ye and Xuanqin Mou

IOP Publishing, 2020

Competitive Performance of a Modularized Deep Neural Network Compared to Commercial Algorithms for Low-dose CT Image Reconstruction

This study confirms that our deep learning approach performs either favourably or comparably in terms of noise suppression and structural fidelity, and is much faster than commercial iterative reconstruction algorithms.

Hongming Shan, Atul Padole, Fatemeh Homayounieh, Uwe Kruger, Ruhani Doda Khera, Chayanin Nitiwarangkul, Mannudeep K. Kalra, and Ge Wang

Nature Machine Intelligence 1:269–276, 2019

NIH NIBIB News

Image Reconstruction is a New Frontier of Machine Learning

This is the first special issue dedicated to the theme of “Machine Learning for Image Reconstruction”. In addition to well-known analytic and iterative methods for tomographic image reconstruction, machine learning is an emerging approach for image reconstruction, and likewise image reconstruction is a new frontier of machine learning. There are exciting research and application opportunities ahead for smart imaging and precision medicine.

Ge Wang, Jong Chu Ye, Klaus Mueller and Jeffrey A. Fessler

IEEE Trans. on Medical Imaging 37:1289-1296, 2018

Machine Learning Will Transform Radiology Significantly within the Next 5 Years

Top 10 most downloaded articles of Med. Phys.

Ge Wang, Mannudeep K. Kalra and Colin G. Orton

Medical physics 44:2041-2044, 2017

Perspective on Deep Imaging

First perspective on machine learning / deep learning for tomographic imaging, as a roadmap for the new area of “deep reconstruction” / “deep imaging” and a basis for the 1st special journal issue on this theme (IEEE TMI)

Ge Wang

IEEE Access 4:8914-8924, 2016

Featured by IEEE Access, 2018