Brain Metastases Detection The detection of brain metastases (BM) in their early stages may significantly improve the patients' outcomes. Thus, we perform research and development of ML algorithms for the automated detection of particularly small BM (<15mm) in T1-weighted 3D MRI data. Publication(s): • Automated Brain Metastases Detection Framework for T1-Weighted Contrast-Enhanced 3D MRI, IEEE Journal of Biomedical and Health Informatics, 2020. Link • Augmented Networks for Faster Brain Metastases Detection in T1-Weighted Contrast-Enhanced 3D MRI, Comp. Med. Imaging and Graphics, 2022. Link • Advancing Brain Metastases Detection in T1-Weighted Contrast-Enhanced 3D MRI using Noisy Student-based Training, MDPI Diagnostics, 2022. Link • Prediction of Model Generalizability for Unseen Data: Methodology and Case Study in Brain Metastases Detection in T1-Weighted Contrast-Enhanced 3D MRI, Computers in Biology and Medicine, 2023. Link • Patent App: Systems for Automated Lesion Detection and Related Methods, Application No.: 17/401,536, Filing Date: August 13, 2021 • Patent App: System and Method for Prediction of Artificial Intelligence Model Generalizability, Application No.: 63/380,419, Filing Date: October 21, 2022 Grant(s): • Automated Intracranial Metastasis Detection Algorithm. Link Key people: E. Dikici, X. V. Nguyen, M. Bigelow, and L. M. Prevedello |
|
Predicting Cognitive Outcomes in Dementia Dementia represents a heterogeneous collection of disease processes, but AI tools may offer opportunities for a patient’s prognosis and management to be individually customized based on MRI images or other clinical data. We have begun evaluating deep learning models for predicting future rates of cognitive decline in Mild Cognitive Impairment, Alzheimer disease, and other dementias by applying convolutional neural networks to thin-section T1-weighted MRI images obtained at or near baseline, in combination with other clinical variables. Publication(s): • Predicting Rate of Cognitive Decline at Baseline Using a Deep Neural Network with Multidata Analysis, J Med Imag, 2020. Link • Predicting Mental Decline Rates in Mild Cognitive Impairment from Baseline MRI Volumetric Data, Alzheimer Dis Assoc Disord, 2021. Link Presentation(s): • NRI Neuroimaging Symposium 2022: Link Key people: S. Candemir, X. V. Nguyen, and L. M. Prevedello |
|
Synthetic Medical Data Generation Sharing medical images between institutions, or even inside the same institution, is restricted by various laws and regulations; research projects requiring large datasets may suffer as a result. These limitations might be addressed by an abundant supply of synthetic data that (1) is representative and (2) does not closely resemble the original images. Hence, we research&develop specialized GAN formulations to generate data addressing these needs. Publication(s): • Constrained Generative Adversarial Network Ensembles for Sharable Synthetic Medical Images, SPIE Journal of Med. Imaging, 2021. Link • Patent App: Methods for Creating Privacy-Protecting Synthetic Data Leveraging a Constrained Generative Ensemble Model, Application No.: 17/401,543, Filing Date: August 13, 2021 Key people: E. Dikici, M. Bigelow, and L. M. Prevedello |
|
Coronary Artery Atherosclerosis Detection Coronary Artery Disease results from the accumulation of atherosclerotic plaque within the walls of the coronary artery tree. While this may cause restricted blood flow to the heart muscle from significant degrees of narrowing of the coronary artery lumen, even mildly stenotic plaque presents a significant risk to the affected patient. Accordingly, we work on the research and development of AI-based systems that automatically extract and help to analyze coronary arteries/branches in CCTA data. Publication(s): • Automated Coronary Artery Atherosclerosis Detection and Weakly Supervised Localization on Coronary CT Angiography with a Deep 3-Dimensional Convolutional Neural Network, Comp. Med. Imaging and Graphics, 2020. Link Key people: S. Candemir, M. Bigelow, and L. M. Prevedello |
|
Machine Intelligence Algorithms from Multi-Modal, Multi-Institutional COVID-19 Data Supported by the Medical Imaging Data Resource Center initiative (http://midrc.org) this collaboration between Ohio State University and Emory University aims to develop new machine learning algorithms that combine imaging data with electronic health record information to diagnose and monitor important clinical outcomes in patients with Covid-19 patients. Publication(s): • Mortality Prediction Analysis among COVID-19 Inpatients Using Clinical Variables and Deep Learning Chest Radiography Imaging Features, MDPI Tomography, 2022. Link Key people: L. M. Prevedello, E. Dikici, S. Candemir, M. Bigelow, and X. V. Nguyen |
|
Integration of AI into Radiology Workflows The integration of AI into existing radiology workflows is significant as (1) radiologists can significantly benefit from enhanced automation in various imaging tasks due to AI, and (2) radiologists’ feedback is utilized to further improve the AI application. Here in LAI², we investigate integration methodologies and workflow adaptations. Publication(s): • Integrating AI into Radiology Workflow: Levels of Research, Production, and Feedback Maturity, SPIE Journal of Med. Imaging, 2021. Link Key people: M. Bigelow, E. Dikici, and L. M. Prevedello |