+1 Recommend
1 collections
      • Record: found
      • Abstract: found
      • Article: not found

      Deep learning-based model for detecting 2019 novel coronavirus pneumonia on high-resolution computed tomography: a prospective study in 27 patients


      Read this article at

          There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.


          Background: Computed tomography (CT) is the preferred imaging method for evaluating lung infection in new coronavirus infection caused pneumonia. Our research aimed to construct a system based on deep learning for detecting 2019-nCoV pneumonia on high resolution CT, and relieve the working pressure of radiologists and potentially improve the efficiency of diagnosis. Methods: 21,661 CT scan images from 40 2019-nCoV pneumonia patients and 5,100 CT scan images from 24 non-2019-nCoV infected patients were collected to train convolutional neural network model to detect 2019-nCoV pneumonia. Eleven and thirty-one patients with and without 2019-nCoV pneumonia were collected for retrospectively testing. Twenty-seven consecutive patients undergoing CT scans in Feb, 5, 2020 in Renmin Hospital of Wuhan University were prospectively collected to evaluate and compare the efficiency of radiologists against 2019-CoV pneumonia with that of the model. Findings: The model achieved a per-patient sensitivity of 100%, specificity of 93.55%, accuracy of 95.24%, PPV of 84.62%, and NPV of 100%; a per-image sensitivity of 94.34%, specificity of 99.16%, accuracy of 98.85%, PPV of 88.37%, and NPV of 99.61% in retrospective dataset. For 27 prospective patients, the model achieved a comparable performance to that of expert radiologist with much shorter reading time (41.34s [IQR 39.76-44.48] vs. 115.50s [IQR 85.69-118.17] per patient). Interpretation: We have developed a deep learning model showing a comparable performance with expert radiologist using much shorter time. It holds great potential to improve the efficiency of diagnosis and relieve the pressure of radiologists in clinical practice.

          Related collections

          Author and article information

          February 26 2020
          © 2020


          Comment on this article