Machine Learning to Predict In-Hospital Mortality in COVID-19 Patients Using Computed Tomography-Derived Pulmonary and Vascular Features

Abstract

Pulmonary parenchymal and vascular damage are frequently reported in COVID-19 patients and can be assessed with unenhanced chest computed tomography (CT), widely used as a triaging exam. Integrating clinical data, chest CT features, and CT-derived vascular metrics, we aimed to build a predictive model of in-hospital mortality using univariate analysis (Mann–Whitney U test) and machine learning models (support vectors machines (SVM) and multilayer perceptrons (MLP)). Patients with RT-PCR-confirmed SARS-CoV-2 infection and unenhanced chest CT performed on emergency department admission were included after retrieving their outcome (discharge or death), with an 85/15% training/test dataset split. Out of 897 patients, the 229 (26%) patients who died during hospitalization had higher median pulmonary artery diameter (29.0 mm) than patients who survived (27.0 mm, p < 0.001) and higher median ascending aortic diameter (36.6 mm versus 34.0 mm, p < 0.001). SVM and MLP best models considered the same ten input features, yielding a 0.747 (precision 0.522, recall 0.800) and 0.844 (precision 0.680, recall 0.567) area under the curve, respectively. In this model integrating clinical and radiological data, pulmonary artery diameter was the third most important predictor after age and parenchymal involvement extent, contributing to reliable in-hospital mortality prediction, highlighting the value of vascular metrics in improving patient stratification.

Machine Learning Algorithms to Predict In-Hospital Mortality in Patients with Traumatic Brain Injury

“…The nine training subgroups were then combined, leaving one training and one testing cohort. The training and testing datasets were constructed using a pure random 10-fold cross-validation [38]. The following statistical classification measures were used for the evaluations: precision, recall, F1 score, true positive rate (TPR), false positive rate (FPR), area under the receiver operating characteristic curve (AUC), accuracy [33], and average success rate of classification [36].…”

Chest imaging in patients with acute respiratory failure because of coronavirus disease 2019

“…The frequent involvement of the pulmonary circulation in COVID-19, with diffuse pathophysiologic drivers of hypoxemia, capillary microthrombosis, and thromboembolism [ 4 , 35 ], has been extensively described. At CT scan, vascular alterations are seen as vascular thickening, pulmonary artery enlargement, and vascular congestion [ 36 , 37 ].…”

A Promising and Challenging Approach: Radiologists’ Perspective on Deep Learning and Artificial Intelligence for Fighting COVID-19

“…The model successfully predicted the risk of progression on the third, fifth, and seventh day, which suggested that a multimodal combination of CT radiomics characteristics and clinical variables could predict the severity and progression of COVID-19 to critical illness with high accuracy. Moreover, Schaffino et al [43] used machine learning models to assess lung parenchyma and vascular damage. Their results showed that the best models of SVM and MLP considered the same ten input features, yielding an AUC of 0.747 and 0.844, respectively.…”