eduzhai > Applied Sciences > Engineering >

Predicting Mechanical Ventilation Requirement and Mortality in COVID-19 using Radiomics and Deep Learning on Chest Radiographs A Multi-Institutional Study

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 17 pages

Abstract: Objectives: To predict mechanical ventilation requirement and mortality usingcomputational modeling of chest radiographs (CXR) for coronavirus disease 2019(COVID-19) patients. We also investigate the relative advantages of deeplearning (DL), radiomics, and DL of radiomic-embedded feature maps inpredicting these outcomes.Methods: This two-center, retrospective study analyzed deidentified CXRstaken from 514 patients suspected of COVID-19 infection on presentation atStony Brook University Hospital (SBUH) and Newark Beth Israel Medical Center(NBIMC) between the months of March and June 2020. A DL segmentation pipelinewas developed to generate masks for both lung fields and artifacts for eachCXR. Machine learning classifiers to predict mechanical ventilation requirementand mortality were trained and evaluated on 353 baseline CXRs taken fromCOVID-19 positive patients. A novel radiomic embedding framework is alsoexplored for outcome prediction.Results: Classification models for mechanical ventilation requirement (testN=154) and mortality (test N=190) had AUCs of up to 0.904 and 0.936,respectively. We also found that the inclusion of radiomic-embedded mapsimproved DL model predictions of clinical outcomes.Conclusions: We demonstrate the potential for computerized analysis ofbaseline CXR in predicting disease outcomes in COVID-19 patients. Our resultsalso suggest that radiomic embedding improves DL models in medical imageanalysis, a technique that might be explored further in other pathologies. Themodels proposed in this study and the prognostic information they provide,complementary to other clinical data, might be used to aid physician decisionmaking and resource allocation during the COVID-19 pandemic.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×