Machine learning-based prognostic modeling using clinical data and quantitative radiomic features from chest CT images in COVID-19 patients

• Published in: Computers in biology and medicine 2021-05, Vol.132, p.104304-104304, Article 104304
• Main Authors: Shiri, Isaac, Sorouri, Majid, Geramifar, Parham, Nazari, Mostafa, Abdollahi, Mohammad, Salimi, Yazdan, Khosravi, Bardia, Askari, Dariush, Aghaghazvini, Leila, Hajianfar, Ghasem, Kasaeian, Amir, Abdollahi, Hamid, Arabi, Hossein, Rahmim, Arman, Radmard, Amir Reza, Zaidi, Habib
• Format: Article
• Language: English
• Subjects: Algorithms; Artificial intelligence; Blood pressure; Chest; Computed tomography; Computed tomography (CT); Coronaviruses; COVID-19; Datasets; Demography; Disease; Feature extraction; Hospitals; Laboratories; Laboratory tests; Learning algorithms; Lesions; Lungs; Machine learning; Medical imaging; Medical prognosis; Modeling; Modelling; Multivariate analysis; Oxygen content; Patients; Pneumonia; Prognosis; Radiomics; Redundancy; Severe acute respiratory syndrome coronavirus 2; Software; Survival; Training; Tuberculosis; Urea
• ISSN: 0010-4825; 1879-0534
• DOI: 10.1016/j.compbiomed.2021.104304

To develop prognostic models for survival (alive or deceased status) prediction of COVID-19 patients using clinical data (demographics and history, laboratory tests, visual scoring by radiologists) and lung/lesion radiomic features extracted from chest CT images. Overall, 152 patients were enrolled in this study protocol. These were divided into 106 training/validation and 46 test datasets (untouched during training), respectively. Radiomic features were extracted from the segmented lungs and infectious lesions separately from chest CT images. Clinical data, including patients’ history and demographics, laboratory tests and radiological scores were also collected. Univariate analysis was first performed (q-value reported after false discovery rate (FDR) correction) to determine the most predictive features among all imaging and clinical data. Prognostic modeling of survival was performed using radiomic features and clinical data, separately or in combination. Maximum relevance minimum redundancy (MRMR) and XGBoost were used for feature selection and classification. The receiver operating characteristic (ROC) curve and the area under the ROC curve (AUC), sensitivity, specificity, and accuracy were used to assess the prognostic performance of the models on the test datasets. For clinical data, cancer comorbidity (q-value