Building a predictive model to identify clinical indicators for COVID-19 using machine learning method

Searching for New Biomarkers to Assess COVID-19 Patients: A Pilot Study

During the initial diagnosis of urgent medical conditions, which include acute infectious diseases, it is important to assess the severity of the patient's clinical state as quickly as possible. Unlike individual biochemical or physiological indicators, derived indices make it possible to better characterize a complex syndrome as a set of symptoms, and therefore quickly take a set of adequate measures. Recently, we reported on novel diagnostic indices containing butyrylcholinesterase (BChE) activity, which is decreased in COVID-19 patients. Also, in these patients, the secretion of von Willebrand factor (vWF) increases, which leads to thrombosis in the microvascular bed. The objective of this study was the determination of the concentration and activity of vWF in patients with COVID-19, and the search for new diagnostic indices. One of the main objectives was to compare the prognostic values of some individual and newly derived indices. Patients with COVID-19 were retrospectively divided into two groups: survivors (n = 77) and deceased (n = 24). According to clinical symptoms and computed tomography (CT) results, the course of disease was predominantly moderate in severity. The first blood sample (first point) was taken upon admission to the hospital, the second sample (second point)-within 4-6 days after admission. Along with the standard spectrum of biochemical indicators, BChE activity (BChEa or BChEb for acetylthiocholin or butyrylthiocholin, respectively), malondialdehyde (MDA), and vWF analysis (its antigen level, AGFW, and its activity, ActWF) were determined and new diagnostic indices were derived. The pooled sensitivity, specificity, and area under the receiver operating curve (AUC), as well as Likelihood ratio (LR) and Odds ratio (OR) were calculated. The level of vWF antigen in the deceased group was 1.5-fold higher than the level in the group of survivors. Indices that include vWF antigen levels are superior to indices using vWF activity. It was found that the index [Urea] × [AGWF] × 1000/(BChEb × [ALB]) had the best discriminatory power to predict COVID-19 mortality (AUC = 0.91 [0.83, 1.00], p < 0.0001; OR = 72.0 [7.5, 689], p = 0.0002). In addition, [Urea] × 1000/(BChEb × [ALB]) was a good predictor of mortality (AUC = 0.95 [0.89, 1.00], p < 0.0001; OR = 31.5 [3.4, 293], p = 0.0024). The index [Urea] × [AGWF] × 1000/(BChEb × [ALB]) was the best predictor of mortality associated with COVID-19 infection, followed by [Urea] × 1000/(BChEb × [ALB]). After validation in a subsequent cohort, these two indices could be recommended for diagnostic laboratories.

Albumin Is a Component of the Esterase Status of Human Blood Plasma

The esterase status of blood plasma can claim to be one of the universal markers of various diseases; therefore, it deserves attention when searching for markers of the severity of COVID-19 and other infectious and non-infectious pathologies. When analyzing the esterase status of blood plasma, the esterase activity of serum albumin, which is the major protein in the blood of mammals, should not be ignored. The purpose of this study is to expand understanding of the esterase status of blood plasma and to evaluate the relationship of the esterase status, which includes information on the amount and enzymatic activity of human serum albumin (HSA), with other biochemical parameters of human blood, using the example of surviving and deceased patients with confirmed COVID-19. In experiments in vitro and in silico, the activity of human plasma and pure HSA towards various substrates was studied, and the effect of various inhibitors on this activity was tested. Then, a comparative analysis of the esterase status and a number of basic biochemical parameters of the blood plasma of healthy subjects and patients with confirmed COVID-19 was performed. Statistically significant differences have been found in esterase status and biochemical indices (including albumin levels) between healthy subjects and patients with COVID-19, as well as between surviving and deceased patients. Additional evidence has been obtained for the importance of albumin as a diagnostic marker. Of particular interest is a new index, [Urea] × [MDA] × 1000/(BChEb × [ALB]), which in the group of deceased patients was 10 times higher than in the group of survivors and 26 times higher than the value in the group of apparently healthy elderly subjects.

An artificial intelligence approach for predicting death or organ failure after hospitalization for COVID-19: development of a novel risk prediction tool and comparisons with ISARIC-4C, CURB-65, qSOFA, and MEWS scoring systems

BACKGROUND

We applied machine learning (ML) algorithms to generate a risk prediction tool [Collaboration for Risk Evaluation in COVID-19 (CORE-COVID-19)] for predicting the composite of 30-day endotracheal intubation, intravenous administration of vasopressors, or death after COVID-19 hospitalization and compared it with the existing risk scores.

METHODS

This is a retrospective study of adults hospitalized with COVID-19 from March 2020 to February 2021. Patients, each with 92 variables, and one composite outcome underwent feature selection process to identify the most predictive variables. Selected variables were modeled to build four ML algorithms (artificial neural network, support vector machine, gradient boosting machine, and Logistic regression) and an ensemble model to generate a CORE-COVID-19 model to predict the composite outcome and compared with existing risk prediction scores. The net benefit for clinical use of each model was assessed by decision curve analysis.

RESULTS

Of 1796 patients, 278 (15%) patients reached primary outcome. Six most predictive features were identified. Four ML algorithms achieved comparable discrimination (P > 0.827) with c-statistics ranged 0.849-0.856, calibration slopes 0.911-1.173, and Hosmer-Lemeshow P > 0.141 in validation dataset. These 6-variable fitted CORE-COVID-19 model revealed a c-statistic of 0.880, which was significantly (P < 0.04) higher than ISARIC-4C (0.751), CURB-65 (0.735), qSOFA (0.676), and MEWS (0.674) for outcome prediction. The net benefit of the CORE-COVID-19 model was greater than that of the existing risk scores.

CONCLUSION

The CORE-COVID-19 model accurately assigned 88% of patients who potentially progressed to 30-day composite events and revealed improved performance over existing risk scores, indicating its potential utility in clinical practice.

MSCCov19Net: multi-branch deep learning model for COVID-19 detection from cough sounds

Coronavirus has an impact on millions of lives and has been added to the important pandemics that continue to affect with its variants. Since it is transmitted through the respiratory tract, it has had significant effects on public health and social relations. Isolating people who are COVID positive can minimize the transmission, therefore several exams are proposed to detect the virus such as reverse transcription-polymerase chain reaction (RT-PCR), chest X-Ray, and computed tomography (CT). However, these methods suffer from either a low detection rate or high radiation dosage, along with being expensive. In this study, deep neural network-based model capable of detecting coronavirus from only coughing sound, which is fast, remotely operable and has no harmful side effects, has been proposed. The proposed multi-branch model takes M el Frequency Cepstral Coefficients (MFCC), S pectrogram, and C hromagram as inputs and is abbreviated as MSCCov19Net. The system is trained on publicly available crowdsourced datasets, and tested on two unseen (used only for testing) clinical and non-clinical datasets. Experimental outcomes represent that the proposed system outperforms the 6 popular deep learning architectures on four datasets by representing a better generalization ability. The proposed system has reached an accuracy of 61.5 % in Virufy and 90.4 % in NoCoCoDa for unseen test datasets.

Machine learning insights concerning inflammatory and liver-related risk comorbidities in non-communicable and viral diseases

The liver is a key organ involved in a wide range of functions, whose damage can lead to chronic liver disease (CLD). CLD accounts for more than two million deaths worldwide, becoming a social and economic burden for most countries. Among the different factors that can cause CLD, alcohol abuse, viruses, drug treatments, and unhealthy dietary patterns top the list. These conditions prompt and perpetuate an inflammatory environment and oxidative stress imbalance that favor the development of hepatic fibrogenesis. High stages of fibrosis can eventually lead to cirrhosis or hepatocellular carcinoma (HCC). Despite the advances achieved in this field, new approaches are needed for the prevention, diagnosis, treatment, and prognosis of CLD. In this context, the scientific com-munity is using machine learning (ML) algorithms to integrate and process vast amounts of data with unprecedented performance. ML techniques allow the integration of anthropometric, genetic, clinical, biochemical, dietary, lifestyle and omics data, giving new insights to tackle CLD and bringing personalized medicine a step closer. This review summarizes the investigations where ML techniques have been applied to study new approaches that could be used in inflammatory-related, hepatitis viruses-induced, and coronavirus disease 2019-induced liver damage and enlighten the factors involved in CLD development.