Machine learning and artificial intelligence within pediatric autoimmune diseases: applications, challenges, future perspective.

INTRODUCTION: Autoimmune disorders affect 4.5% to 9.4% of children, significantly reducing their quality of life. The diagnosis and prognosis of autoimmune diseases are uncertain because of the variety of onset and development. Machine learning can identify clinically relevant patterns from vast amounts of data. Hence, its introduction has been beneficial in the diagnosis and management of patients. AREAS COVERED: This narrative review was conducted through searching various electronic databases, including PubMed, Scopus, and Web of Science. This study thoroughly explores the current knowledge and identifies the remaining gaps in the applications of machine learning specifically in the context of pediatric autoimmune and related diseases. EXPERT OPINION: Machine learning algorithms have the potential to completely change how pediatric autoimmune disorders are identified, treated, and managed. Machine learning can assist physicians in making more precise and fast judgments, identifying new biomarkers and therapeutic targets, and personalizing treatment strategies for each patient by utilizing massive datasets and powerful analytics.

Texture analysis on routine MRI sequences to differentiate between focal nodular hyperplasia and hepatocellular adenoma.

Purpose: We investigated the diagnostic power of texture analysis (TA) performed on MRI (T2-weighted, gadolinium-enhanced, and diffusion-weighted images) to differentiate between focal nodular hyperplasia (FNH) and hepatocellular adenoma (HCA). Material and methods: This was a retrospective single-centre study. Patients referred for liver lesion characterization, who had a definitive pathological diagnosis, were included. MRI images were taken by a 3-Tesla scanner. The values of TA parameters were obtained using the ImageJ platform by an observer blinded to the clinical and pathology judgments. A non-parametric Mann-Whitney U test was applied to compare parameters between the 2 groups. With receiver operating characteristic (ROC) analysis, the area under the curve (AUC), sensitivity, and specificity were calculated. Finally, we performed a binary logistic regression analysis. A p-value <0.05 was reported as statistically significant. Results: A total of 62 patients with 106 lesions were enrolled. T2 hyperintensity, Atoll sign, and intralesional fat were encountered more in HCAs, and central scars were more frequent in FNHs. Multiple TA features showed statistically significant differences between FNHs and HCAs, including skewness on T2W and entropy on all sequences. Skewness on T2W revealed the most significant AUC (0.841, good, p < 0.0001). The resultant model from binary logistic regression was statistically significant (p < 0.0001) and correctly predicted 84.1% of lesions. The corresponding AUC was 0.942 (excellent, 95% CI: 0.892-0.992, p < 0.0001). Conclusion: Multiple first-order TA parameters significantly differ between these lesions and have almost fair to good diagnostic power. They have differentiation potential and can add diagnostic value to routine MRI evaluations.

Predicting Efficacy of 5-Fluorouracil Therapy via a Mathematical Model with Fuzzy Uncertain Parameters.

Background: Due to imprecise/missing data used for parameterization of ordinary differential equations (ODEs), model parameters are uncertain. Uncertainty of parameters has hindered the application of ODEs that require accurate parameters. Methods: We extended an available ODE model of tumor-immune system interactions via fuzzy logic to illustrate the fuzzification procedure of an ODE model. The fuzzy ODE (FODE) model assigns a fuzzy number to the parameters, to capture parametric uncertainty. We used the FODE model to predict tumor and immune cell dynamics and to assess the efficacy of 5-fluorouracil (5-FU) chemotherapy. Result: FODE model investigates how parametric uncertainty affects the uncertainty band of cell dynamics in the presence and absence of 5-FU treatment. In silico experiments revealed that the frequent 5-FU injection created a beneficial tumor microenvironment that exerted detrimental effects on tumor cells by enhancing the infiltration of CD8+ T cells, and natural killer cells, and decreasing that of myeloid-derived suppressor cells. The global sensitivity analysis was proved model robustness against random perturbation to parameters. Conclusion: ODE models with fuzzy uncertain kinetic parameters cope with insufficient/imprecise experimental data in the field of mathematical oncology and can predict cell dynamics uncertainty band.

Agent-based Modeling of Tumor and Immune System Interactions in Combinational Therapy with Low-dose 5-fluorouracil and Dendritic Cell Vaccine in Melanoma B16F10.

This study is designed to present an agent-based model (ABM) to simulate the interactions between tumor cells and the immune system in the melanoma model. The Myeloid-derived Suppressor Cells (MDSCs) and dendritic cells (DCs) are considered in this model as immunosuppressive and antigen-presenting agents respectively. The animal experiment was performed on 68 B16F10 melanoma tumor-bearing C57BL/6 female mice to collect dynamic data for ABM implementation and validation. Animals were divided into 4 groups; group 1 was control (no treatment) while groups 2 and 3 were treated with DC vaccine and low-dose 5- fluorouracil (5-FU) respectively and group 4 was treated with both DC Vaccine and low-dose of 5-FU. The tumor growth rate, number of MDSC, and presence of CD8+/CD107a+ T cells in the tumor microenvironment were evaluated in each group. Firstly, the tumor cells, the effector immune cells, DCs, and the MDSCs have been considered as the agents of the ABM model and their interaction methods have been extracted from the literature and implemented in the model. Then, the model parameters were estimated by the dynamic data collected from animal experiments.  To validate the ABM model, the simulation results were compared with the real data. The results show that the dynamics of the model agents can mimic the relations among considered immune system components to an emergent outcome compatible with real data. The simplicity of the proposed model can help to understand the results of the combinational therapy and make this model a useful tool for studying different scenarios and assessing the combinational results. Determining the role of each component helps to find critical times during tumor progression and change the tumor and immune system balance in favor of the immune system.

Predicting COVID-19 Cases Among Nurses Using Artificial Neural Network Approach.

We designed a forecasting model to determine which frontline health workers are most likely to be infected by COVID-19 among 220 nurses. We used multivariate regression analysis and different classification algorithms to assess the effect of several covariates, including exposure to COVID-19 patients, access to personal protective equipment, proper use of personal protective equipment, adherence to hand hygiene principles, stressfulness, and training on the risk of a nurse being infected. Access to personal protective equipment and training were associated with a 0.19- and 1.66-point lower score in being infected by COVID-19. Exposure to COVID-19 cases and being stressed of COVID-19 infection were associated with a 0.016- and 9.3-point higher probability of being infected by COVID-19. Furthermore, an artificial neural network with 75.8% (95% confidence interval, 72.1-78.9) validation accuracy and 76.6% (95% confidence interval, 73.1-78.6) overall accuracy could classify normal and infected nurses. The neural network can help managers and policymakers determine which frontline health workers are most likely to be infected by COVID-19.

A COVID-19 forecasting system for hospital needs using ANFIS and LSTM models: A graphical user interface unit.

Background: Centers for Disease Control and Prevention data showed that about 40% of coronavirus disease 2019 (COVID-19) patients had been suffering from at least one underlying medical condition were hospitalized; in which nearly 33% of them needed to be admitted to the intensive care unit (ICU) to receive specialized medical services. Our study aimed to find a proper machine learning algorithm that can predict confirmed COVID-19 hospital admissions with high accuracy. Methods: We obtained data on daily COVID-19 cases in regular medical inpatient units, emergency department, and ICU in the time window between 21 July 2020 and 21 November 2021. Data for the first 183 days (training data set) were used for long short-term memory (LSTM) network, adaptive neuro-fuzzy inference system (ANFIS), support vector regression (SVR) and decision tree model training, whilst the remaining data for the last 60 days (test data set) were used for model validation. To predict the number of ICU and non-ICU patients, we used these models. Finally, a user-friendly graphical user interface unit was designed to load any time series data (here the trend of population of COVID-19 patients) and train LSTM, ANFIS, SVR or tree models for the prediction of COVID-19 cases for one week ahead. Results: All models predicted the dynamics of COVID-19 cases in ICU and non- wards. The values of root-mean-square error and R 2 as model assessment metrics showed that ANFIS model had better predictive power among all models. Conclusion: Artificial intelligence-based forecasting models such as ANFIS system or deep learning approach based on LSTM or regression models including SVR or tree regression play a key role in forecasting the required number of beds or other types of medical facilities during the coronavirus pandemic. Thus, the designed graphical user interface of the present study can be used for optimum management of resources by health care systems amid COVID-19 pandemic.

Effect of Statins on Serum level of hs-CRP and CRP in Patients with Cardiovascular Diseases: A Systematic Review and Meta-Analysis of Randomized Controlled Trials.

BACKGROUND: Several studies have reported that statins have anti-inflammatory effects. Nevertheless, results of clinical trials concerning the effect of statins on the levels of C-reactive protein (CRP) and high-sensitivity CRP (hs-CRP) have been inconsistent. Therefore, we performed a systematic review and meta-analysis of randomized clinical trials (RCTs) evaluating the effect of statins on CRP and hs-CRP levels in patients with cardiovascular diseases (CVDs). METHODS: Literature search of the major databases was performed to find eligible RCTs assessing the effect of statins on serum levels of CRP and hs-CRP from the inception until the last week of April 2021. The effect sizes were determined for weighted mean difference (WMD) and 95% confidence intervals (CI). RESULTS: 26 studies were identified (3010 patients and 2968 controls) for hs-CRP and 20 studies (3026 patients and 2968 controls) for CRP. Statins reduced the serum levels of hs-CRP (WMD = -0.97 mg/L; 95% CI: -1.26 to -0.68 mg/L; P < 0.001) and CRP (WMD = -3.05 mg/L; 95% CI: -4.86 to -1.25 mg/L; P < 0.001) in patients with CVDs. Statins decreased the serum levels of hs-CRP in patients receiving both high-intensity and moderate/low-intensity treatments with these drugs. In addition, the duration of treatment longer than 10 weeks decreased hs-CRP levels. Only high-intensity statin treatment could marginally decrease serum levels of CRP in CVDs patients. CONCLUSIONS: This meta-analysis showed the efficacy of statins to reduce the concentrations of CRP and hs-CRP in patients with different types of CVDs.

Machine learning-based COVID-19 diagnosis by demographic characteristics and clinical data.

INTRODUCTION: To facilitate rapid and effective diagnosis of COVID-19, effective screening can alleviate the challenges facing healthcare systems. We aimed to develop a machine learning-based prediction of COVID-19 diagnosis and design a graphical user interface (GUI) to diagnose COVID-19 cases by recording their symptoms and demographic features. METHODS: We implemented different classification models including support vector machine (SVM), Decision tree (DT), Naïve Bayes (NB) and K-nearest neighbor (KNN) to predict the result of COVID-19 test for individuals. We trained these models by data of 16973 individuals (90% of all individuals included in data gathering) and tested by 1885 individuals (10% of all individuals). Maximum relevance minimum redundancy (MRMR) algorithms used to score features for prediction of result of COVID-19 test. A user-friendly GUI was designed to predict COVID-19 test results in individuals. RESULTS: Study results revealed that coughing had the highest positive correlation with the positive results of COVID-19 test followed by the duration of having COVID-19 signs and symptoms, exposure to infected individuals, age, muscle pain, recent infection by COVID-19 virus, fever, respiratory distress, loss of smell or taste, nausea, anorexia, headache, vertigo, CT symptoms in lung scans, diabetes and hypertension. The values of accuracy, precision, recall, F1-score, specificity and area under receiver operating curve (AUROC) of different classification models computed in different setting of features scored by MRMR algorithm. Finally, our designed GUI by receiving each of the 42 features and symptoms from the users and through selecting one of the SVM, KNN, Naïve Bayes and decision tree models, predict the result of COVID-19 test. The accuracy, AUROC and F1-score of SVM model as the best model for diagnosis of COVID-19 test were 0.7048 (95% CI: 0.6998, 0.7094), 0.7045 (95% CI: 0.7003, 0.7104) and 0.7157 (95% CI: 0.7043, 0.7194), respectively. CONCLUSION: In this study we implemented a machine learning approach to facilitate early clinical decision making during COVID-19 outbreak and provide a predictive model of COVID-19 diagnosis capable of categorizing populations in to infected and non-infected individuals the same as an efficient screening tool.

Crosstalk between Plk1, p53, cell cycle, and G2/M DNA damage checkpoint regulation in cancer: computational modeling and analysis.

Different cancer cell lines can have varying responses to the same perturbations or stressful conditions. Cancer cells that have DNA damage checkpoint-related mutations are often more sensitive to gene perturbations including altered Plk1 and p53 activities than cancer cells without these mutations. The perturbations often induce a cell cycle arrest in the former cancer, whereas they only delay the cell cycle progression in the latter cancer. To study crosstalk between Plk1, p53, and G2/M DNA damage checkpoint leading to differential cell cycle regulations, we developed a computational model by extending our recently developed model of mitotic cell cycle and including these key interactions. We have used the model to analyze the cancer cell cycle progression under various gene perturbations including Plk1-depletion conditions. We also analyzed mutations and perturbations in approximately 1800 different cell lines available in the Cancer Dependency Map and grouped lines by genes that are represented in our model. Our model successfully explained phenotypes of various cancer cell lines under different gene perturbations. Several sensitivity analysis approaches were used to identify the range of key parameter values that lead to the cell cycle arrest in cancer cells. Our resulting model can be used to predict the effect of potential treatments targeting key mitotic and DNA damage checkpoint regulators on cell cycle progression of different types of cancer cells.

Prediction of anti-CD25 and 5-FU treatments efficacy for pancreatic cancer using a mathematical model.

BACKGROUND: Pancreatic ductal adenocarcinoma (PDAC) is a highly lethal disease with rising incidence and with 5-years overall survival of less than 8%. PDAC creates an immune-suppressive tumor microenvironment to escape immune-mediated eradication. Regulatory T (Treg) cells and myeloid-derived suppressor cells (MDSC) are critical components of the immune-suppressive tumor microenvironment. Shifting from tumor escape or tolerance to elimination is the major challenge in the treatment of PDAC. RESULTS: In a mathematical model, we combine distinct treatment modalities for PDAC, including 5-FU chemotherapy and anti- CD25 immunotherapy to improve clinical outcome and therapeutic efficacy. To address and optimize 5-FU and anti- CD25 treatment (to suppress MDSCs and Tregs, respectively) schedule in-silico and simultaneously unravel the processes driving therapeutic responses, we designed an in vivo calibrated mathematical model of tumor-immune system (TIS) interactions. We designed a user-friendly graphical user interface (GUI) unit which is configurable for treatment timings to implement an in-silico clinical trial to test different timings of both 5-FU and anti- CD25 therapies. By optimizing combination regimens, we improved treatment efficacy. In-silico assessment of 5-FU and anti- CD25 combination therapy for PDAC significantly showed better treatment outcomes when compared to 5-FU and anti- CD25 therapies separately. Due to imprecise, missing, or incomplete experimental data, the kinetic parameters of the TIS model are uncertain that this can be captured by the fuzzy theorem. We have predicted the uncertainty band of cell/cytokines dynamics based on the parametric uncertainty, and we have shown the effect of the treatments on the displacement of the uncertainty band of the cells/cytokines. We performed global sensitivity analysis methods to identify the most influential kinetic parameters and simulate the effect of the perturbation on kinetic parameters on the dynamics of cells/cytokines. CONCLUSION: Our findings outline a rational approach to therapy optimization with meaningful consequences for how we effectively design treatment schedules (timing) to maximize their success, and how we treat PDAC with combined 5-FU and anti- CD25 therapies. Our data revealed that a synergistic combinatorial regimen targeting the Tregs and MDSCs in both crisp and fuzzy settings of model parameters can lead to tumor eradication.

Simulating ATO Mechanism and EGFR Signaling with Fuzzy Logic and Petri Net.

BACKGROUND: Interactions of many key proteins or genes in signalling pathway have been studied qualitatively in the literature, but only little quantitative information is available. OBJECTIVE: Although much has been done to clarify the biochemistry of transcriptional dynamics in signalling pathway, it remains difficult to find out and predict quantitative responses. The aim of this study is to construct a computational model of epidermal growth factor receptor (EGFR) signalling pathway as one of hallmarks of cancer so as to predict quantitative responses. MATERIAL AND METHODS: In this analytical study, we presented a computational model to investigate EGFR signalling pathway. Interaction of Arsenic trioxide (ATO) with EGFR signalling pathway factors has been elicited by systematic search in data bases, as ATO is one of the mysterious chemotherapy agents that control EGFR expression in cancer. ATO has dichotomous manner in vivo, dependent on its concentration. According to fuzzy rules based upon qualitative knowledge and Petri Net, we can construct a quantitative model to describe ATO mechanism in EGFR signalling pathway. RESULTS: By Fuzzy Logic models that have the potential to trade with the loss of quantitative information on how different species interact, along with Petri net quantitatively describe the dynamics of EGFR signalling pathway. By this model the dynamic of different factors in EGFR signalling pathway is achieved. CONCLUSION: The use of Fuzzy Logic and PNs in biological network modelling causes a deeper understanding and comprehensive analysis of the biological networks.

Dynamical analysis of the fission yeast cell cycle via Markov chain.

The cell cycle is a complex network involved in the regulation of cell growth and proliferation. Intrinsic molecular noise in gene expression in the cell cycle network can generate fluctuations in protein concentration. How the cell cycle network maintains its robust transitions between cell cycle phases in the presence of these fluctuations remains unclear. To understand the complex and robust behavior of the cell cycle system in the presence of intrinsic noise, we developed a Markov model for the fission yeast cell cycle system. We quantified the effect of noise on gene and protein activity and on the probability of transition between different phases of the cell cycle. Our analysis shows how network perturbations decide the fate of the cell. Our model predicts that the cell cycle pathway (subsequent transitions from [Formula: see text]) is the most robust and probable pathway among all possible trajectories in the cell cycle network. We performed a sensitivity analysis to find correlations between protein interaction weights and transition probabilities between cell cycle phases. The sensitivity analysis predicts how network perturbations affect the transition probability between different cell cycle phases and, consequently, affect different cell fates, thus, forming testable in vitro/in vivo hypotheses. Our simulation results agree with published experimental findings and reveal how noise in the cell cycle regulatory network can affect cell cycle progression.

Interleukin-10 Gene Promoter Polymorphisms and Susceptibility to Asthma: Systematic Review and Meta-analysis.

Several studies have previously assessed the association between interleukin (IL)-10 gene polymorphisms and the risk of asthma, leading to conflicting results. To resolve the incongruent outcomes yielded from different single studies, we conducted the most up-to-date meta-analysis of the IL-10 gene rs1800896, rs1800871, and rs1800872 single-nucleotide polymorphisms (SNPs) and susceptibility to asthma. A systematic literature search performed until April 2020, and the pooled odds ratio (OR) and their corresponding 95% confidence interval (CI) were calculated to determine the association strength. Thirty articles comprising 5678 asthmatic patients and 6079 controls met the inclusion criteria. No significant association was found between rs1800872 SNP and susceptibility to asthma across all genetic models in the overall and subgroup analyses. The rs1800871 SNP had only significant association with a decreased risk of asthma in Europeans (OR 0.66, CI 0.53-0.82, P < 0.001). However, rs1800896 SNP was significantly associated with a decreased risk of asthma by dominant (OR 0.67, CI 0.50-0.90, P < 0.001) and heterozygote (OR 0.66, CI 0.49-0.88, P < 0.001) models in the overall analysis. Subgroup analyses indicated significant association of rs1800896 SNP by dominant (OR 0.45, CI 0.28-0.72, P < 0.001) and heterozygote (OR 0.43, CI 0.26-0.70, P < 0.001) models in the African population. The IL-10 rs1800896 SNP confers protection against the risk of asthma, especially in Africans. Additionally, rs1800871 SNP has a protective role against asthma in Europeans.

Ipsilateral lung normal tissue complication probability parameters for different dose calculation algorithms in radiotherapy of breast cancer.

PURPOSE: Different dose calculation algorithms (DCAs) predict different dose distributions for the same treatment. Awareness of optimal model parameters is vital for estimating normal tissue complication probability (NTCP) for different algorithms. The aim is to determine the NTCP parameter values for different DCAs in left-sided breast radiotherapy, using the Lyman-Kutcher-Burman (LKB) model. MATERIALS AND METHODS: First, the methodology recommended by International Atomic Energy Agency TEC-DOC 1583 was used to establish the accuracy of dose calculations of different DCAs including: Monte Carlo (MC) and collapsed cone algorithms implemented in Monaco, pencil beam convolution (PBC) and analytical anisotropic algorithm (AAA) implemented in Eclipse, and superposition and Clarkson algorithms implemented in PCRT3D treatment planning systems (TPSs). Then, treatment planning of 15 patients with left-sided breast cancer was performed by the mentioned DCAs and NTCP of the left-lung normal tissue were calculated for each patient individually, using the LKB model. For the PB algorithm, the NTCP parameters were taken from previously published values and new model parameters obtained for each DCA, using the iterative least squares methods. RESULTS: For all cases and DCAs, NTCP computation with the same model parameters resulted in >15% deviation in NTCP values. The new NTCP model parameters were classified according to the algorithm type. Thus, the discrepancy of NTCP computations was reduced up to 5% after utilizing adjusted model parameters. CONCLUSIONS: This paper confirms that the NTCP values for a given treatment type are different for the different DCAs. Thus, it is essential to introduce appropriate NTCP parameter values according to DCA adopted in TPS, to obtain a more precise estimation of lung NTCP. Hence, new parameter values, classified according to the DCAs, must be determined before introducing NTCP estimation in clinical practice.

Extended Robust Boolean Network of Budding Yeast Cell Cycle.

BACKGROUND: How to explore the dynamics of transition probabilities between phases of budding yeast cell cycle (BYCC) network based on the dynamics of protein activities that control this network? How to identify the robust structure of protein interactions of BYCC Boolean network (BN)? Budding yeast allows scientists to put experiments into effect in order to discover the intracellular cell cycle regulating structures which are well simulated by mathematical modeling. METHODS: We extended an available deterministic BN of proteins responsible for the cell cycle to a Markov chain model containing apoptosis besides G1, S, G2, M, and stationary G1. Using genetic algorithm (GA), we estimated the kinetic parameters of the extended BN model so that the subsequent transition probabilities derived using Markov chain model of cell states as normal cell cycle becomes the maximum while the structure of chemical interactions of extended BN of cell cycle becomes more stable. RESULTS: Using kinetic parameters optimized by GA, the probability of the subsequent transitions between cell cycle phases is maximized. The relative basin size of stationary G1 increased from 86% to 96.48% while the number of attractors decreased from 7 in the original model to 5 in the extended one. Hence, an increase in the robustness of the system has been achieved. CONCLUSION: The structure of interacting proteins in cell cycle network affects its robustness and probabilities of transitions between different cell cycle phases. Markov chain and BN are good approaches to study the stability and dynamics of the cell cycle network.