Bulletin of the American Physical Society
APS March Meeting 2021
Volume 66, Number 1
Monday–Friday, March 15–19, 2021; Virtual; Time Zone: Central Daylight Time, USA
Session F15: Artificial Intelligence, Machine Learning, and Data Science in Medicine and BiomedicineLive
|
Hide Abstracts |
Sponsoring Units: GMED GDS Chair: Michael Boss, American College of Radiology; Jie Ren, Merck & Co. |
Tuesday, March 16, 2021 11:30AM - 12:06PM Live |
F15.00001: Integrating machine learning and multiscale modeling: perspectives, challenges, and opportunities in the biological, biomedical, and behavioral sciences Invited Speaker: Ellen Kuhl Prof. Ellen Kuhl has been the champion in integrating machine learning techniques with multiscale physical modeling, particularly in the area of biomedicine. She leads multiple efforts in building multi-scale modeling for the human heart and human brian, and more recently has been leading machine learning models to understand and predict COVID-19 infections. |
Tuesday, March 16, 2021 12:06PM - 12:18PM Live |
F15.00002: Comparison of statistical parametric mapping method and scaled subprofile model for functional neuroimage analysis Lara hocurscak, Tadej Tomanic, Maja Trost, Urban Simoncic Two common analytic techniques for functional neuroimages are Statistical parametric mapping (SPM) and Scaled subprofile model based on principle component analysis (SSM/PCA). SPM is based on mass univariate testing in which signals from each region are compared between two groups. SSM/PCA is a multivariate PCA-based algorithm that construct the disease-related pattern that discriminate between patients and controls. |
Tuesday, March 16, 2021 12:18PM - 12:30PM Live |
F15.00003: A Deep Learning Network for Disease Classification with Longitudinal Data Alison Deatsch, Robert Jeraj Deep learning approaches to medical image analysis have achieved success in modeling disease progression. However, their application to longitudinal analysis is low and most fail to leverage time-relevant patient data. This work aims to develop a deep learning model to predict disease status and investigate the influence of longitudinal data on model performance. We train a 2-class convolutional neural network (CNN) both with and without a cascaded recurrent neural network (RNN) to investigate the impact of longitudinal features on Alzheimer’s Disease (AD) classification. We use a dataset of 1,260 18F-FDG PET scans containing brain images from 87 normal (NC) patients, 30 patients with stable mild cognitive impairment (sMCI), 81 with progressive MCI (pMCI), and 65 with AD across 3 timepoints with 1 year gaps. Performance is also evaluated using a similar dataset of 822 MRI scans. The CNN+RNN model shows improvement over the CNN alone for both (AD vs NC) and (sMCI vs pMCI) classification. Thus adding longitudinal data leads to better identification of the stages of AD than single-timepoint data alone. Saliency maps were also analyzed to explore the most important regions for the network’s decisions. |
Tuesday, March 16, 2021 12:30PM - 12:42PM Live |
F15.00004: Developing Dynamical Models to Characterize Stroke Gait Impairments Taniel Winner, Trisha Kesar, Gordon Berman, Lena Ting Understanding the dynamics that generate human gait is essential for designing tailored rehabilitative therapies for stroke survivors. Most purely biomechanical models, however, are highly sensitive to parameters, and individuals often express a large degree of cycle-to-cycle variability, preventing robust and accurate inference of the underlying dynamical system that generates an individual’s gait. Here, we present a Recurrent Neural Network (RNN)-based model that produces a robust kinematic gait signature for visualizing and comparing gait kinematics between able-bodied individuals and stroke-survivors. Extracting the model’s internal activations, we identify unique gait signatures for each individual. These signatures reliably distinguish stroke from unaffected individuals, as well as different gait types from each other. Moreover, these metrics are speed invariant - which can be potentially useful for determining the best gait rehabilitation for stroke-survivors without evaluating them at some optimal speed. Building from these dynamical characterizations, it may be possible to build generative models of gait function, allowing us to generate individual-specific rehabilitative therapies. |
Tuesday, March 16, 2021 12:42PM - 1:18PM Live |
F15.00005: The Predictive Value of Deep Learning and Radiomics in Medical Imaging Invited Speaker: Jayashree Kalpathy-Cramer As discussed by GDS and GMED conference committees |
Tuesday, March 16, 2021 1:18PM - 1:30PM Live |
F15.00006: Estimation of Radiobiological Indices in Radiotherapy of Lung Cancer using an Artificial Neural Network Mukunda Pudasaini, Theodora Leventouri, Silvia Pella, Dr. Wazir Muhammad The purpose of this study is to develop an artificial neural network (ANN) to predict radiobiological indices in radiotherapy of lung cancer. A total of 100 lung cancer patients’ treatment plans were selected for this study. The target outputs for ANN, Normal tissue complication probability (NTCP) of organs at risk (OAR), and tumor control probability (TCP) of the tumor were calculated. The inputs were planning target volume (PTV), treatment modality, location of the tumor, prescribed dose, number of fractions, the maximum dose to the tumor, and mean doses to the OARs. The ANN is based on a Scaled Conjugate gradient algorithm with one hidden layer having 11 inputs and 5 outputs. 70% of the data was used for training, and 30% for testing the ANN. The ANN predicts NTCP for OARs and TCP for the target with an overall regression value of 0.94. The fitted ANN has mean square error values 0.007 and 0.024 for training and testing respectively. The regression values are 0.97 and 0.91 for training and testing respectively. The results show that ANN can be designed to predict the radiobiological parameters within a 5% error as indicated by the regression value. To validate the full performance of the neural network in case of a lung cancer treatment plan, further research is in progress. |
Tuesday, March 16, 2021 1:30PM - 1:42PM Live |
F15.00007: Superspreading k-cores at the center of COVID-19 pandemic persistence matteo serafino, Higor S. Monteiro, Shaojun Luo, Saulo D. S. Reis, Carles Igual, Antonio S. Lima Neto, Matias Travizan, Jose Soares De Andrade Jr, Hernan A Makse The spread of COVID-19 caused by the recently discovered SARS-CoV-2 virus has become a worldwide problem with devastating consequences. To slow down the spread of the pandemic, mass quarantines have been implemented globally, provoking further social and economic disruptions. Here, we implement a comprehensive contact tracing network analysis to find the optimal quarantine protocol to dismantle the chain of transmission of coronavirus with minimal disruptions to society. Our analysis indicates that superspreading k-core structures persist in the transmission network to prolong the pandemic. Once the k-cores are identified, the optimal strategy to break the chain of transmission is to quarantine a minimal number of 'weak links' with high betweenness centrality connecting the large k-cores. |
Tuesday, March 16, 2021 1:42PM - 1:54PM On Demand |
F15.00008: Framework for Assessing the Impact of CNN-based Image Segmentation on Multi-step Biomarker Extraction Daniel Huff, Zan Klanecek, Andrej Studen, Robert Jeraj Convolutional neural networks (CNN) have become endemic to medical image analysis. However, little work has assessed the effect of CNN error on multi-step biomarker extraction. Here, we describe a framework for assessing CNN impact on final biomarker extraction in the clinical application of detecting colitis on 18F-FDG PET/CT. |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700