Eleven routine clinical features predict COVID-19 severity uncovered by machine learning of longitudinal measurements.

Severity prediction of COVID-19 remains one of the major clinical challenges for the ongoing pandemic. Here, we have recruited a 144 COVID-19 patient cohort, resulting in a data matrix containing 3,065 readings for 124 types of measurements over 52 days. A machine learning model was established to predict the disease progression based on the cohort consisting of training, validation, and internal test sets. A panel of eleven routine clinical factors constructed a classifier for COVID-19 severity prediction, achieving accuracy of over 98% in the discovery set. Validation of the model in an independent cohort containing 25 patients achieved accuracy of 80%. The overall sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were 0.70, 0.99, 0.93, and 0.93, respectively. Our model captured predictive dynamics of lactate dehydrogenase (LDH) and creatine kinase (CK) while their levels were in the normal range. This model is accessible at https://www.guomics.com/covidAI/ for research purpose.

Laboratory indicators in COVID-19 and other pneumonias: Analysis for differential diagnosis and comparison of dynamic changes during 400-day follow-up.

BACKGROUND: COVID-19 is spreading rapidly all over the world, the patients' symptoms can be easily confused with other pneumonia types. Therefore, it is valuable to seek a laboratory differential diagnostic protocol of COVID-19 and other pneumonia types on admission, and to compare the dynamic changes in laboratory indicators during follow-up. METHODS: A total of 143 COVID-19, 143 bacterial pneumonia and 145 conventional viral pneumonia patients were included. The model group consisted of 140 COVID-19, 80 bacterial pneumonia and 60 conventional viral pneumonia patients, who were age and sex matched. We established a differential diagnostic model based on the laboratory results of the model group on admission via a nomogram, which was validated in an external validation group. We also compared the 400-day dynamic changes of the laboratory indicators among groups. RESULTS: LASSO regression and multivariate logistic regression showed that eosinophils (Eos), total protein (TP), prealbumin (PA), potassium (K), high-density lipoprotein cholesterol (HDLC), and low-density lipoprotein cholesterol (LDLC) could differentiate COVID-19 from other pneumonia types. The C-index of the nomogram model was 0.922. Applying the nomogram to the external validation group showed an area under the curve (AUC) of 0.902. The 400-day change trends of the laboratory indexes varied among subgroups divided by sex, age, oxygenation index (OI), and pathogen. CONCLUSION: The laboratory model was highly accurate at providing a new method to identify COVID-19 in pneumonia patients. The 400-day dynamic changes in laboratory indicators revealed that the recovery time of COVID-19 patients was not longer than that of other pneumonia types.

Predicting the COVID-19 infection with fourteen clinical features using machine learning classification algorithms.

While the RT-PCR is the silver bullet test for confirming the COVID-19 infection, it is limited by the lack of reagents, time-consuming, and the need for specialized labs. As an alternative, most of the prior studies have focused on Chest CT images and Chest X-Ray images using deep learning algorithms. However, these two approaches cannot always be used for patients' screening due to the radiation doses, high costs, and the low number of available devices. Hence, there is a need for a less expensive and faster diagnostic model to identify the positive and negative cases of COVID-19. Therefore, this study develops six predictive models for COVID-19 diagnosis using six different classifiers (i.e., BayesNet, Logistic, IBk, CR, PART, and J48) based on 14 clinical features. This study retrospected 114 cases from the Taizhou hospital of Zhejiang Province in China. The results showed that the CR meta-classifier is the most accurate classifier for predicting the positive and negative COVID-19 cases with an accuracy of 84.21%. The results could help in the early diagnosis of COVID-19, specifically when the RT-PCR kits are not sufficient for testing the infection and assist countries, specifically the developing ones that suffer from the shortage of RT-PCR tests and specialized laboratories.

Prediction of severe illness due to COVID-19 based on an analysis of initial Fibrinogen to Albumin Ratio and Platelet count.

Concomitant coagulation disorder can occur in severe patients withCOVID-19, but in-depth studies are limited. This study aimed to describe the parameters of coagulation function of patients with COVID-19 and reveal the risk factors of developing severe disease. This study retrospectively analyzed 113patients with SARS-CoV-2 infection in Taizhou Public Health Center. Clinical characteristics and indexes of coagulation function were collected. A multivariate Cox analysis was performed to identify potential biomarkers for predicting disease progression. Based on the results of multivariate Cox analysis, a Nomogram was built and the predictive accuracy was evaluated through the calibration curve, decision curve, clinical impact curve, and Kaplan-Meier analysis. Sensitivity, specificity, predictive values were calculated to assess the clinical value. The data showed that Fibrinogen, FAR, and D-dimer were higher in the severe patients, while PLTcount, Alb were much lower. Multivariate Cox analysis revealed that FAR and PLT count were independent risk factors for disease progression. The optimal cutoff values for FAR and PLT count were 0.0883 and 135*109/L, respectively. The C-index [0.712 (95% CI = 0.610-0.814)], decision curve, clinical impact curve showed that Nomogram could be used to predict the disease progression. In addition, the Kaplan-Meier analysis revealed that potential risk decreased in patients with FAR<0.0883 and PLT count>135*109/L.The model showed a good negative predictive value [(0.9474 (95%CI = 0.845-0.986)].This study revealed that FAR and PLT count were independent risk factors for severe illness and the severity of COVID-19 might be excluded when FAR<0.0883 and PLT count>135*109/L.

The hemocyte counts as a potential biomarker for predicting disease progression in COVID-19: a retrospective study.

Objectives In December 2019, there was an outbreak of coronavirus disease 2019 (COVID-19) in Wuhan, China, and since then, the disease has been increasingly spread throughout the world. Unfortunately, the information about early prediction factors for disease progression is relatively limited. Therefore, there is an urgent need to investigate the risk factors of developing severe disease. The objective of the study was to reveal the risk factors of developing severe disease by comparing the differences in the hemocyte count and dynamic profiles in patients with severe and non-severe COVID-19. Methods In this retrospectively analyzed cohort, 141 confirmed COVID-19 patients were enrolled in Taizhou Public Health Medical Center, Taizhou Hospital, Zhejiang Province, China, from January 17, 2020 to February 26, 2020. Clinical characteristics and hemocyte counts of severe and non-severe COVID patients were collected. The differences in the hemocyte counts and dynamic profiles in patients with severe and non-severe COVID-19 were compared. Multivariate Cox regression analysis was performed to identify potential biomarkers for predicting disease progression. A concordance index (C-index), calibration curve, decision curve and the clinical impact curve were calculated to assess the predictive accuracy. Results The data showed that the white blood cell count, neutrophil count and platelet count were normal on the day of hospital admission in most COVID-19 patients (87.9%, 85.1% and 88.7%, respectively). A total of 82.8% of severe patients had lymphopenia after the onset of symptoms, and as the disease progressed, there was marked lymphopenia. Multivariate Cox analysis showed that the neutrophil count (hazard ratio [HR] = 4.441, 95% CI = 1.954-10.090, p = 0.000), lymphocyte count (HR = 0.255, 95% CI = 0.097-0.669, p = 0.006) and platelet count (HR = 0.244, 95% CI = 0.111-0.537, p = 0.000) were independent risk factors for disease progression. The C-index (0.821 [95% CI, 0.746-0.896]), calibration curve, decision curve and the clinical impact curve showed that the nomogram can be used to predict the disease progression in COVID-19 patients accurately. In addition, the data involving the neutrophil count, lymphocyte count and platelet count (NLP score) have something to do with improving risk stratification and management of COVID-19 patients. Conclusions We designed a clinically predictive tool which is easy to use for assessing the progression risk of COVID-19, and the NLP score could be used to facilitate patient stratification management.