Machine learning-based prognostic modeling using clinical data and quantitative radiomic features from chest CT images in COVID-19 patients. (May 2021)

• Record Type: Journal Article
• 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
• Is Part Of: Computers in biology and medicine. Volume 132(2021)
• Publication Date: 2021-05
• Subjects: COVID-19 -- Computed tomography (CT) -- Radiomics -- Prognosis -- Modeling; Medicine -- Data processing -- Periodicals; Biology -- Data processing -- Periodicals; 610.285
• DOI: 10.1016/j.compbiomed.2021.104304
• Languages: English
• ISSNs: 0010-4825

Abstract

Abstract: Objective: 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. Methods: 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. Results: For clinical data, cancer comorbidity (q-value < 0.01), consciousness level (q-value < 0.05) and radiological score involved zone (q-value < 0.02) were found to have high correlated features with outcome. Oxygen saturation (AUC = 0.73, q-value < 0.01) and Blood Urea Nitrogen (AUC = 0.72, q-value = 0.72) were identified as high clinical features. For lung radiomic features, SAHGLE (AUC = 0.70) and HGLZE (AUC = 0.67) from GLSZM were identified as most prognostic features. Amongst lesion radiomic features, RLNU from GLRLM (AUC = 0.73), HGLZE from GLSZM (AUC = 0.73) had the highest performance. In multivariate analysis, combining lung, lesion and clinical features was determined to provide the most accurate prognostic model (AUC = 0.95 ± 0.029 (95%CI: 0.95–0.96), accuracy = 0.88 ± 0.046 (95% CI: 0.88–0.89), sensitivity = 0.88 ± 0.066 (95% CI = 0.87–0.9) and specificity = 0.89 ± 0.07 (95% CI = 0.87–0.9)). Conclusion: Combination of radiomic features and clinical data can effectively predict outcome in COVID-19 patients. The developed model has significant potential for improved management of COVID-19 patients. Graphical abstract: Image 1 Highlights: Radiomics–based machine learning models predict the survival of ccRCC patients. Clinical/imaging features could potentially be used for COVID-19 patients management. Radiomic and clinical features combination can effectively enhance outcome prediction. Radiomic features of the lung provided more information than COVID-19 lesions.