PT  - JOURNAL ARTICLE
AU  - Duchesne, S.
AU  - Gourdeau, D.
AU  - Archambault, P.
AU  - Chartrand-Lefebvre, C.
AU  - Dieumegarde, L.
AU  - Forghani, R.
AU  - Gagné, C.
AU  - Hains, A.
AU  - Hornstein, D.
AU  - Le, H.
AU  - Lemieux, S.
AU  - Lévesque, M.H.
AU  - Martin, D.
AU  - Rosenbloom, L.
AU  - Tang, A.
AU  - Vecchio, F.
AU  - Potvin, O.
AU  - Duchesne, N.
TI  - TRACKING AND PREDICTING COVID-19 RADIOLOGICAL TRAJECTORY USING DEEP LEARNING ON CHEST X-RAYS: INITIAL ACCURACY TESTING
AID  - 10.1101/2020.05.01.20086207
DP  - 2020 Jan 01
TA  - medRxiv
PG  - 2020.05.01.20086207
4099  - http://medrxiv.org/content/early/2020/05/05/2020.05.01.20086207.short
4100  - http://medrxiv.org/content/early/2020/05/05/2020.05.01.20086207.full
AB  - Background Decision scores and ethically mindful algorithms are being established to adjudicate mechanical ventilation in the context of potential resources shortage due to the current onslaught of COVID-19 cases. There is a need for a reproducible and objective method to provide quantitative information for those scores.Purpose Towards this goal, we present a retrospective study testing the ability of a deep learning algorithm at extracting features from chest x-rays (CXR) to track and predict radiological evolution.Materials and Methods We trained a repurposed deep learning algorithm on the CheXnet open dataset (224,316 chest X-ray images of 65,240 unique patients) to extract features that mapped to radiological labels. We collected CXRs of COVID-19-positive patients from two open-source datasets (last accessed on April 9, 2020)(Italian Society for Medical and Interventional Radiology and MILA). Data collected form 60 pairs of sequential CXRs from 40 COVID patients (mean age ± standard deviation: 56 ± 13 years; 23 men, 10 women, seven not reported) and were categorized in three categories: “Worse”, “Stable”, or “Improved” on the basis of radiological evolution ascertained from images and reports. Receiver operating characteristic analyses, Mann-Whitney tests were performed.Results On patients from the CheXnet dataset, the area under ROC curves ranged from 0.71 to 0.93 for seven imaging features and one diagnosis. Deep learning features between “Worse” and “Improved” outcome categories were significantly different for three radiological signs and one diagnostic (“Consolidation”, “Lung Lesion”, “Pleural effusion” and “Pneumonia”; all P < 0.05). Features from the first CXR of each pair could correctly predict the outcome category between “Worse” and “Improved” cases with 82.7% accuracy.Conclusion CXR deep learning features show promise for classifying the disease trajectory. Once validated in studies incorporating clinical data and with larger sample sizes, this information may be considered to inform triage decisions.Competing Interest StatementThe authors have declared no competing interest.Funding StatementThis study has been financed by a COVID-19 Pilot Project grant from the Quebec Bio-Imaging Network as well as concurrent funding from a Discovery Award to the primary investigator (S.D.) from the National Science and Engineering Research Council of Canada.Author DeclarationsAll relevant ethical guidelines have been followed; any necessary IRB and/or ethics committee approvals have been obtained and details of the IRB/oversight body are included in the manuscript.YesAll necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived.YesI understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).YesI have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable.YesAll data are open access (references in manuscript) https://stanfordmlgroup.github.io/competitions/chexpert/ https://www.sirm.org/category/senza-categoria/covid-19/ https://github.com/ieee8023/covid-chestxray-dataset/ https://github.com/medicslab/COVID-19-public AP: anterior-posteriorAUC: area under the (receiver operating) curveCOVID-19: coronavirus diseaseCXR: chest X-rayICU: intensive care unit