In this study, we used XR images of 227 patients from publicly available cohorts, and hence, no institutional review board approval was needed. At first, we used three cohorts to discover potential biomarkers for COVID-19, namely the discovery set, and the other two cohorts as an independent validation set.

The discovery set was composed of 195 patients:

29 patients with COVID-19 that had images and clinical data provided by the Italian Society of Medical and Interventional Radiology [17];

127 cases of pneumonia non-related to COVID-19 from the Spanish chest XR cohort (PadChest dataset) with image-associated reports from patients that attended the San Juan de Alicante Hospital, University of Alicante, Spain [18];

39 patients with pneumonia non-related to COVID-19 from the National Library of Medicine, National Institutes of Health (OpenI dataset), who attended various hospitals of the Indiana University School of Medicine, USA [19].

Physicians performed image labeling for the cases from PadChest and OpenI datasets. The validation set was composed of 32 patients from around the world. In this set, images and clinical data were initially collected from the literature by researchers of the University of Montreal [20]. Then, it was completed with follow-up data (survival time, imaging and event dates) by us to perform the radiomic analysis. From those 32 patients, 20 had the diagnosis confirmed for COVID-19 by RT-PCR, and 12 from a different etiology distributed as following: one caused by Pneumocystis fungal pathogen, two by the bacteria Streptococcus, four of them had ARDS, and five had SARS. Table 1 describes demographic data from the cohorts.

Description of the patients

NA, not available

* Mean ± standard deviation (min–max)

Note: The content above has been extracted from a research article, so it may not display correctly.

Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.

We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.