tetano
Editor, Senior Moderator
Eur Radiol
. 2022 Feb 19.
doi: 10.1007/s00330-022-08588-8. Online ahead of print.
COVID-19 mortality prediction in the intensive care unit with deep learning based on longitudinal chest X-rays and clinical data
Jianhong Cheng[SUP] #[/SUP][SUP] 1 [/SUP], John Sollee[SUP] #[/SUP][SUP] 2 3 [/SUP], Celina Hsieh[SUP] #[/SUP][SUP] 2 3 [/SUP], Hailin Yue[SUP] 1 [/SUP], Nicholas Vandal[SUP] 4 [/SUP], Justin Shanahan[SUP] 4 [/SUP], Ji Whae Choi[SUP] 2 3 [/SUP], Thi My Linh Tran[SUP] 2 3 [/SUP], Kasey Halsey[SUP] 2 3 [/SUP], Franklin Iheanacho[SUP] 2 3 [/SUP], James Warren[SUP] 5 [/SUP], Abdullah Ahmed[SUP] 2 3 [/SUP], Carsten Eickhoff[SUP] 6 [/SUP], Michael Feldman[SUP] 7 [/SUP], Eduardo Mortani Barbosa Jr[SUP] 4 [/SUP], Ihab Kamel[SUP] 8 [/SUP], Cheng Ting Lin[SUP] 8 [/SUP], Thomas Yi[SUP] 2 3 [/SUP], Terrance Healey[SUP] 2 3 [/SUP], Paul Zhang[SUP] 4 [/SUP], Jing Wu[SUP] 1 [/SUP], Michael Atalay[SUP] 2 3 [/SUP], Harrison X Bai[SUP] 9 [/SUP], Zhicheng Jiao[SUP] 10 11 [/SUP], Jianxin Wang[SUP] 12 [/SUP]
Affiliations
Abstract
Objectives: We aimed to develop deep learning models using longitudinal chest X-rays (CXRs) and clinical data to predict in-hospital mortality of COVID-19 patients in the intensive care unit (ICU).
Methods: Six hundred fifty-four patients (212 deceased, 442 alive, 5645 total CXRs) were identified across two institutions. Imaging and clinical data from one institution were used to train five longitudinal transformer-based networks applying five-fold cross-validation. The models were tested on data from the other institution, and pairwise comparisons were used to determine the best-performing models.
Results: A higher proportion of deceased patients had elevated white blood cell count, decreased absolute lymphocyte count, elevated creatine concentration, and incidence of cardiovascular and chronic kidney disease. A model based on pre-ICU CXRs achieved an AUC of 0.632 and an accuracy of 0.593, and a model based on ICU CXRs achieved an AUC of 0.697 and an accuracy of 0.657. A model based on all longitudinal CXRs (both pre-ICU and ICU) achieved an AUC of 0.702 and an accuracy of 0.694. A model based on clinical data alone achieved an AUC of 0.653 and an accuracy of 0.657. The addition of longitudinal imaging to clinical data in a combined model significantly improved performance, reaching an AUC of 0.727 (p = 0.039) and an accuracy of 0.732.
Conclusions: The addition of longitudinal CXRs to clinical data significantly improves mortality prediction with deep learning for COVID-19 patients in the ICU.
Key points: • Deep learning was used to predict mortality in COVID-19 ICU patients. • Serial radiographs and clinical data were used. • The models could inform clinical decision-making and resource allocation.
Keywords: Artificial intelligence; Coronavirus; Hospital mortality; Machine learning; Prognosis.
. 2022 Feb 19.
doi: 10.1007/s00330-022-08588-8. Online ahead of print.
COVID-19 mortality prediction in the intensive care unit with deep learning based on longitudinal chest X-rays and clinical data
Jianhong Cheng[SUP] #[/SUP][SUP] 1 [/SUP], John Sollee[SUP] #[/SUP][SUP] 2 3 [/SUP], Celina Hsieh[SUP] #[/SUP][SUP] 2 3 [/SUP], Hailin Yue[SUP] 1 [/SUP], Nicholas Vandal[SUP] 4 [/SUP], Justin Shanahan[SUP] 4 [/SUP], Ji Whae Choi[SUP] 2 3 [/SUP], Thi My Linh Tran[SUP] 2 3 [/SUP], Kasey Halsey[SUP] 2 3 [/SUP], Franklin Iheanacho[SUP] 2 3 [/SUP], James Warren[SUP] 5 [/SUP], Abdullah Ahmed[SUP] 2 3 [/SUP], Carsten Eickhoff[SUP] 6 [/SUP], Michael Feldman[SUP] 7 [/SUP], Eduardo Mortani Barbosa Jr[SUP] 4 [/SUP], Ihab Kamel[SUP] 8 [/SUP], Cheng Ting Lin[SUP] 8 [/SUP], Thomas Yi[SUP] 2 3 [/SUP], Terrance Healey[SUP] 2 3 [/SUP], Paul Zhang[SUP] 4 [/SUP], Jing Wu[SUP] 1 [/SUP], Michael Atalay[SUP] 2 3 [/SUP], Harrison X Bai[SUP] 9 [/SUP], Zhicheng Jiao[SUP] 10 11 [/SUP], Jianxin Wang[SUP] 12 [/SUP]
Affiliations
- PMID: 35184218
- DOI: 10.1007/s00330-022-08588-8
Abstract
Objectives: We aimed to develop deep learning models using longitudinal chest X-rays (CXRs) and clinical data to predict in-hospital mortality of COVID-19 patients in the intensive care unit (ICU).
Methods: Six hundred fifty-four patients (212 deceased, 442 alive, 5645 total CXRs) were identified across two institutions. Imaging and clinical data from one institution were used to train five longitudinal transformer-based networks applying five-fold cross-validation. The models were tested on data from the other institution, and pairwise comparisons were used to determine the best-performing models.
Results: A higher proportion of deceased patients had elevated white blood cell count, decreased absolute lymphocyte count, elevated creatine concentration, and incidence of cardiovascular and chronic kidney disease. A model based on pre-ICU CXRs achieved an AUC of 0.632 and an accuracy of 0.593, and a model based on ICU CXRs achieved an AUC of 0.697 and an accuracy of 0.657. A model based on all longitudinal CXRs (both pre-ICU and ICU) achieved an AUC of 0.702 and an accuracy of 0.694. A model based on clinical data alone achieved an AUC of 0.653 and an accuracy of 0.657. The addition of longitudinal imaging to clinical data in a combined model significantly improved performance, reaching an AUC of 0.727 (p = 0.039) and an accuracy of 0.732.
Conclusions: The addition of longitudinal CXRs to clinical data significantly improves mortality prediction with deep learning for COVID-19 patients in the ICU.
Key points: • Deep learning was used to predict mortality in COVID-19 ICU patients. • Serial radiographs and clinical data were used. • The models could inform clinical decision-making and resource allocation.
Keywords: Artificial intelligence; Coronavirus; Hospital mortality; Machine learning; Prognosis.