Document text
Principal Investigator: Joseph Frank
Organization: CLINICAL CENTER
Fiscal Year: 2022
Funding agency: NIH Clinical Center
A CNN architecture was used to create a deep learning-based segmentation model of the liver seen on CT scans of a nonhuman primate (NHP). The CNN architecture was optimized by implementing a feature pyramid network (FPN). This FPN model was then utilized to generalize liver segmentation to other organs such as the spleen and then the whole lung and lung lesions in NHP models of SARS-CoV-2 infection, as conducted at the IRF. Lung lobe segmentation was developed in order to better correlate with histopathologic findings of COVID-19. This is a work-in-progress as the number of lung lobe annotations on CT scans is at a minimum due to the difficulty in creating this ground truth data. Currently, methods that function with low numbers of training data are being explored.
To make these deep learning-based image segmentation methods available for use by the imaging community at the IRF, an automated pipeline process was developed. This pipeline utilizes the NIAID high performance computing environment (Locus) to allow for enhanced parallel processing with graphical processing units (GPUs). These methods are being used by other imaging researchers at IRF to post-process images acquired during infectious disease imaging studies of SARS-CoV-2, Marburg, Ebola, Nipah and Lassa viruses. In addition, these segmentations provide regions of interest for radiomic feature extraction. These hundreds of radiomic features have been input into feature selections methods, such as maximum relevance minimum redundancy (mRMR), to reduce and optimize features for predictive analyses. In addition, a novel feature selection method was developed (mRMR-permute), which uses permutation testing to automatically limit the number of features chosen. The predictive analyses are implemented with conventional machine learning methods that were explored to determine optimal use for SARS-CoV-2 imaging experiments.
To further the specificity of image segmentations, a lesion phenotype classification algorithm is being developed. As an initial implementation, lung lesions seen in CT scans are to be classified as ground glass opacities or consolidations. This classification will help with severity assessment during longitudinal quantification.
Terms: <2019 novel corona virus><2019 novel coronavirus><2019-nCoV><AI system><Architecture><Artificial Intelligence><CAT scan><COVID-19><COVID-19 infection><COVID-19 virus><COVID19><COVID19 infection><COVID19 virus><CT X Ray><CT Xray><CT imaging><CT scan><CV-19><CV19><Classification><CoV-2><CoV2><Communicable Diseases><Communities><Computed Tomography><Computer Reasoning><Data><Disease Progression><EBOV><Ebola virus><Ebola-like Viruses><Engineering / Architecture><Environment><Frankfurt-Marburg Syndrome Virus><Glass><High Performance Computing><Image><Infectious Disease Pathway><Infectious Diseases><Infectious Diseases / Laboratory><Infectious Diseases Research><Infectious Disorder><Investigators><Lassa fever virus><Lassa virus><Lesion><Liver><Machine Intelligence><Marburg><Marburg virus><Marburg-like Viruses><Marburgvirus><Medical Imaging><Methods><Modeling><NIAID><National Institute of Allergy and Infectious Disease><Nipah Virus><Organ><Pathology><Phenotype><Process><Research Personnel><Researchers><SARS corona virus 2><SARS-CO-V2><SARS-COVID-2><SARS-CoV-2><SARS-CoV-2 infection><SARS-CoV2><SARS-CoV2 infection><SARS-associated corona virus 2><SARS-associated coronavirus 2><SARS-coronavirus-2><SARS-related corona virus 2><SARS-related coronavirus 2><SARSCoV2><Severe Acute Respiratory Coronavirus 2><Severe Acute Respiratory Distress Syndrome CoV 2><Severe Acute Respiratory Distress Syndrome Corona Virus 2><Severe Acute Respiratory Distress Syndrome Coronavirus 2><Severe Acute Respiratory Syndrome CoV 2><Severe Acute Respiratory Syndrome-associated coronavirus 2><Severe Acute Respiratory Syndrome-related coronavirus 2><Severe acute respiratory syndrome associated corona virus 2><Severe acute respiratory syndrome corona virus 2><Severe acute respiratory syndrome coronavirus 2><Severe acute respiratory syndrome coronavirus 2 infection><Severe acute respiratory syndrome related corona virus 2><Severities><Severity of illness><Specificity><Spleen><Spleen Reticuloendothelial System><Systematics><Testing><Tomodensitometry><Training><Wuhan coronavirus><X-Ray CAT Scan><X-Ray Computed Tomography><X-Ray Computerized Tomography><Xray CAT scan><Xray Computed Tomography><Xray computerized tomography><base><catscan><classification algorithm><computed axial tomography><computer based prediction><computer tomography><computerized axial tomography><computerized tomography><corona virus disease 2019><coronavirus disease 2019><coronavirus disease 2019 infection><coronavirus disease 2019 virus><coronavirus disease-19><coronavirus disease-19 virus><coronavirus infectious disease-19><deep learning><disease severity><ebolavirus><experiment><experimental research><experimental study><feature extraction><feature selection><hCoV19><hepatic body system><hepatic organ system><high-end computing><imaging><imaging Segmentation><imaging study><infected with COVID-19><infected with COVID19><infected with SARS-CoV-2><infected with SARS-CoV2><infected with coronavirus disease 2019><infected with severe acute respiratory syndrome coronavirus 2><interest><lung lesion><lung lobe><machine learning based classification><machine learning based method><machine learning classification><machine learning method><machine learning methodologies><medical countermeasure><multi-modality><multimodality><nCoV2><network models><non-contrast CT><non-human primate><noncontrast CT><noncontrast computed tomography><nonhuman primate><novel><parallel processing><prediction model><predictive modeling><pulmonary lesion><radiomics><research facility>