tetano
Editor, Senior Moderator
Eur J Radiol
. 2021 Feb 5;109583.
doi: 10.1016/j.ejrad.2021.109583. Online ahead of print.
A multi-center study of COVID-19 patient prognosis using deep learning-based CT image analysis and electronic health records
Kuang Gong[SUP] 1 [/SUP], Dufan Wu[SUP] 1 [/SUP], Chiara Daniela Arru[SUP] 1 [/SUP], Fatemeh Homayounieh[SUP] 1 [/SUP], Nir Neumark[SUP] 2 [/SUP], Jiahui Guan[SUP] 3 [/SUP], Varun Buch[SUP] 2 [/SUP], Kyungsang Kim[SUP] 1 [/SUP], Bernardo Canedo Bizzo[SUP] 2 [/SUP], Hui Ren[SUP] 1 [/SUP], Won Young Tak[SUP] 4 [/SUP], Soo Young Park[SUP] 4 [/SUP], Yu Rim Lee[SUP] 4 [/SUP], Min Kyu Kang[SUP] 5 [/SUP], Jung Gil Park[SUP] 5 [/SUP], Alessandro Carriero[SUP] 6 [/SUP], Luca Saba[SUP] 7 [/SUP], Mahsa Masjedi[SUP] 8 [/SUP], Hamidreza Talari[SUP] 8 [/SUP], Rosa Babaei[SUP] 9 [/SUP], Hadi Karimi Mobin[SUP] 9 [/SUP], Shadi Ebrahimian[SUP] 1 [/SUP], Ning Guo[SUP] 1 [/SUP], Subba R Digumarthy[SUP] 1 [/SUP], Ittai Dayan[SUP] 2 [/SUP], Mannudeep K Kalra[SUP] 10 [/SUP], Quanzheng Li[SUP] 11 [/SUP]
Affiliations
Abstract
Purpose: As of August 30th, there were in total 25.1 million confirmed cases and 845 thousand deaths caused by coronavirus disease of 2019 (COVID-19) worldwide. With overwhelming demands on medical resources, patient stratification based on their risks is essential. In this multi-center study, we built prognosis models to predict severity outcomes, combining patients' electronic health records (EHR), which included vital signs and laboratory data, with deep learning- and CT-based severity prediction.
Method: We first developed a CT segmentation network using datasets from multiple institutions worldwide. Two biomarkers were extracted from the CT images: total opacity ratio (TOR) and consolidation ratio (CR). After obtaining TOR and CR, further prognosis analysis was conducted on datasets from INSTITUTE-1, INSTITUTE-2 and INSTITUTE-3. For each data cohort, generalized linear model (GLM) was applied for prognosis prediction.
Results: For the deep learning model, the correlation coefficient of the network prediction and manual segmentation was 0.755, 0.919, and 0.824 for the three cohorts, respectively. The AUC (95 % CI) of the final prognosis models was 0.85(0.77,0.92), 0.93(0.87,0.98), and 0.86(0.75,0.94) for INSTITUTE-1, INSTITUTE-2 and INSTITUTE-3 cohorts, respectively. Either TOR or CR exist in all three final prognosis models. Age, white blood cell (WBC), and platelet (PLT) were chosen predictors in two cohorts. Oxygen saturation (SpO2) was a chosen predictor in one cohort.
Conclusion: The developed deep learning method can segment lung infection regions. Prognosis results indicated that age, SpO2, CT biomarkers, PLT, and WBC were the most important prognostic predictors of COVID-19 in our prognosis model.
Keywords: COVID-19; Computed tomography; Deep learning; Electronic health records; Prognosis.
. 2021 Feb 5;109583.
doi: 10.1016/j.ejrad.2021.109583. Online ahead of print.
A multi-center study of COVID-19 patient prognosis using deep learning-based CT image analysis and electronic health records
Kuang Gong[SUP] 1 [/SUP], Dufan Wu[SUP] 1 [/SUP], Chiara Daniela Arru[SUP] 1 [/SUP], Fatemeh Homayounieh[SUP] 1 [/SUP], Nir Neumark[SUP] 2 [/SUP], Jiahui Guan[SUP] 3 [/SUP], Varun Buch[SUP] 2 [/SUP], Kyungsang Kim[SUP] 1 [/SUP], Bernardo Canedo Bizzo[SUP] 2 [/SUP], Hui Ren[SUP] 1 [/SUP], Won Young Tak[SUP] 4 [/SUP], Soo Young Park[SUP] 4 [/SUP], Yu Rim Lee[SUP] 4 [/SUP], Min Kyu Kang[SUP] 5 [/SUP], Jung Gil Park[SUP] 5 [/SUP], Alessandro Carriero[SUP] 6 [/SUP], Luca Saba[SUP] 7 [/SUP], Mahsa Masjedi[SUP] 8 [/SUP], Hamidreza Talari[SUP] 8 [/SUP], Rosa Babaei[SUP] 9 [/SUP], Hadi Karimi Mobin[SUP] 9 [/SUP], Shadi Ebrahimian[SUP] 1 [/SUP], Ning Guo[SUP] 1 [/SUP], Subba R Digumarthy[SUP] 1 [/SUP], Ittai Dayan[SUP] 2 [/SUP], Mannudeep K Kalra[SUP] 10 [/SUP], Quanzheng Li[SUP] 11 [/SUP]
Affiliations
- PMID: 33846041
- DOI: 10.1016/j.ejrad.2021.109583
Abstract
Purpose: As of August 30th, there were in total 25.1 million confirmed cases and 845 thousand deaths caused by coronavirus disease of 2019 (COVID-19) worldwide. With overwhelming demands on medical resources, patient stratification based on their risks is essential. In this multi-center study, we built prognosis models to predict severity outcomes, combining patients' electronic health records (EHR), which included vital signs and laboratory data, with deep learning- and CT-based severity prediction.
Method: We first developed a CT segmentation network using datasets from multiple institutions worldwide. Two biomarkers were extracted from the CT images: total opacity ratio (TOR) and consolidation ratio (CR). After obtaining TOR and CR, further prognosis analysis was conducted on datasets from INSTITUTE-1, INSTITUTE-2 and INSTITUTE-3. For each data cohort, generalized linear model (GLM) was applied for prognosis prediction.
Results: For the deep learning model, the correlation coefficient of the network prediction and manual segmentation was 0.755, 0.919, and 0.824 for the three cohorts, respectively. The AUC (95 % CI) of the final prognosis models was 0.85(0.77,0.92), 0.93(0.87,0.98), and 0.86(0.75,0.94) for INSTITUTE-1, INSTITUTE-2 and INSTITUTE-3 cohorts, respectively. Either TOR or CR exist in all three final prognosis models. Age, white blood cell (WBC), and platelet (PLT) were chosen predictors in two cohorts. Oxygen saturation (SpO2) was a chosen predictor in one cohort.
Conclusion: The developed deep learning method can segment lung infection regions. Prognosis results indicated that age, SpO2, CT biomarkers, PLT, and WBC were the most important prognostic predictors of COVID-19 in our prognosis model.
Keywords: COVID-19; Computed tomography; Deep learning; Electronic health records; Prognosis.