Machine learning-based clinical decision support using laboratory data

Clinical Chemistry and Laboratory Medicine (CCLM), Journal Year: 2023, Volume and Issue: 62(5), P. 793 - 823

Published: Nov. 28, 2023

Abstract Artificial intelligence (AI) and machine learning (ML) are becoming vital in laboratory medicine the broader context of healthcare. In this review article, we summarized development ML models how they contribute to clinical workflow improve patient outcomes. The process model involves data collection, cleansing, feature engineering, development, optimization. These models, once finalized, subjected thorough performance assessments validations. Recently, due complexity inherent automated tools were also introduced streamline process, enabling non-experts create models. Clinical Decision Support Systems (CDSS) use techniques on large datasets aid healthcare professionals test result interpretation. They revolutionizing medicine, labs work more efficiently with less human supervision across pre-analytical, analytical, post-analytical phases. Despite contributions at all analytical phases, their integration presents challenges like potential uncertainties, black-box algorithms, deskilling professionals. Additionally, acquiring diverse is hard, models’ can limit use. conclusion, ML-based CDSS greatly enhance decision-making. However, successful adoption demands collaboration among stakeholders, utilizing hybrid intelligence, external validation, assessments.

Language: Английский

---

Hemogram‐based decision tree models for discriminating COVID ‐19 from RSV in infants

Journal of Clinical Laboratory Analysis, Journal Year: 2023, Volume and Issue: 37(6)

Published: March 1, 2023

Decision trees are efficient and reliable decision-making algorithms, medicine has reached its peak of interest in these methods during the current pandemic. Herein, we reported several decision tree algorithms for a rapid discrimination between coronavirus disease (COVID-19) respiratory syncytial virus (RSV) infection infants.A cross-sectional study was conducted on 77 infants: 33 infants with novel betacoronavirus (SARS-CoV-2) 44 RSV infection. In total, 23 hemogram-based instances were used to construct models via 10-fold cross-validation method.The Random forest model showed highest accuracy (81.8%), while terms sensitivity (72.7%), specificity (88.6%), positive predictive value (82.8%), negative (81.3%), optimized most superior one.Random might have significant clinical applications, helping speed up when SARS-CoV-2 suspected, prior molecular genome sequencing and/or antigen testing.

Language: Английский

---

Platelet distribution width (PDW) as a significant correlate of COVID-19 infection severity and mortality

Clinical Chemistry and Laboratory Medicine (CCLM), Journal Year: 2023, Volume and Issue: 62(3), P. 385 - 395

Published: Sept. 19, 2023

Abstract SARS-CoV-2 infection may cause a wide spectrum of symptoms, from asymptomatic, to mild respiratory symptoms and life-threatening sepsis. Among the clinical laboratory biomarkers analyzed during COVID-19 pandemic, platelet indices have raised great interest, due critical involvement platelets in COVID-19-related thromboinflammation. Through an electronic literature search on MEDLINE, CINAHL, PubMed, EMBASE, Web Science, preprint servers we performed updated systematic review aimed at providing detailed analysis studies addressing potential utility distribution width, width (PDW), medicine, exploring possible association between increased PDW levels, disease severity, mortality COVID-19. Our revealed heterogeneity cohorts examined lack homogenous expression indices. We found that 75 % reported significantly elevated values infected compared healthy/non-COVID-19 controls, 40 patients with severe showed than those less-than-severe illness. Interestingly, 71.4 demonstrated significant non survivors vs. survivors. Overall, these results suggest are critically involved as major players process immunothrombosis COVID-19, reactivity morphofunctional alterations mirrored by PDW, indicator heterogeneity. confirm use prognostic sepsis still remains debated limited number draw conclusion, but new opportunities investigate crucial role thrombo-inflammation warranted.

Language: Английский

---

The Application of Artificial Intelligence in Thyroid Nodules: A Systematic Review Based on Bibliometric Analysis

Endocrine Metabolic & Immune Disorders - Drug Targets, Journal Year: 2024, Volume and Issue: 24(11), P. 1280 - 1290

Published: Jan. 5, 2024

Thyroid nodules are common lesions in benign and malignant thyroid diseases. More more studies have been conducted on the feasibility of artificial intelligence (AI) detection, diagnosis, evaluation nodules. The aim this study was to use bibliometric methods analyze predict hot spots frontiers AI

Language: Английский

---

Clinical Hematochemical Parameters in Differential Diagnosis between Pediatric SARS-CoV-2 and Influenza Virus Infection: An Automated Machine Learning Approach

Children, Journal Year: 2023, Volume and Issue: 10(5), P. 761 - 761

Published: April 22, 2023

The influenza virus and the novel beta coronavirus (SARS-CoV-2) have similar transmission characteristics, it is very difficult to distinguish them clinically. With development of information technologies, opportunities arisen for application intelligent software systems in disease diagnosis patient triage.A cross-sectional study was conducted on 268 infants: 133 infants with a SARS-CoV-2 infection 135 an infection. In total, 10 hematochemical variables were used construct automated machine learning model.An accuracy range from 53.8% 60.7% obtained by applying support vector machine, random forest, k-nearest neighbors, logistic regression, neural network models. Alternatively, model convincingly outperformed other models 98.4%. proposed algorithm recommended tree model, randomization-based ensemble method, as most appropriate given dataset.The clinical practice can contribute more objective, accurate, rapid infections children.

Language: Английский

---

Rapid Triage of Children with Suspected COVID-19 Using Laboratory-Based Machine-Learning Algorithms

Viruses, Journal Year: 2023, Volume and Issue: 15(7), P. 1522 - 1522

Published: July 8, 2023

In order to limit the spread of novel betacoronavirus (SARS-CoV-2), it is necessary detect positive cases as soon possible and isolate them. For this purpose, machine-learning algorithms, a field artificial intelligence, have been recognized promising tool. The aim study was assess utility most common algorithms in rapid triage children with suspected COVID-19 using easily accessible inexpensive laboratory parameters. A cross-sectional conducted on 566 treated for respiratory diseases: 280 PCR-confirmed SARS-CoV-2 infection 286 symptoms who were PCR-negative (control group). Six based blood data, tested: random forest, support vector machine, linear discriminant analysis, neural network, k-nearest neighbors, decision tree. training set validated through stratified cross-validation, while performance each algorithm confirmed by an independent test set. Random forest machine models demonstrated highest accuracy 85% 82.1%, respectively. better sensitivity than specificity negative predictive value value. F1 score higher model, 85.2% 82.3%, This might significant clinical applications, helping healthcare providers identify early stage, prior PCR and/or antigen testing. Additionally, could improve overall testing efficiency no extra costs facility.

Language: Английский

---

Multiomics as instrument to promote 3P medical approaches for the overall management of respiratory syncytial viral infections

The EPMA Journal, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 16, 2025

Language: Английский

---

Investigation of the Systemic Immune Inflammation (SII) Index as an Indicator of Morbidity and Mortality in Type 2 Diabetic Retinopathy Patients in a 4-Year Follow-Up Period

Medicina, Journal Year: 2024, Volume and Issue: 60(6), P. 855 - 855

Published: May 24, 2024

Background and Objectives: This study aimed to investigate the relationship between systemic immune inflammation (SII) index development of micro macro complications mortality within first year following three years in type 2 diabetic retinopathy patients. Materials Methods: The retrospective included 523 patients seen endocrinology outpatient clinic our hospital January December 2019. Their demographic clinical characteristics were analyzed using descriptive statistics. normal distribution quantitative data was assessed by Shapiro–Wilk test. Mann–Whitney U, McNemar–Chi-square, Cochran’s Q tests used analyze SII values complication rates over time. An ROC analysis determined sensitivity specificity SII. A multiple linear regression examined variables SII, while Spearman’s test correlation CRP p < 0.05 accepted as significant. Results: mean age 63.5 ± 9.3 years, with 821.4 1010.8. Higher significantly associated acute–chronic renal failure, peripheral arterial disease, hospitalization both (p for all). Significant cut-off found micro- macrovascular death curve identified an optimal value >594.0 predicting near-term (1-year) mortality, a 73.8% 49.4% (area under curve: 0.629, = 0.001). Multiple indicated that smoking at least 20 pack-years had significant positive effect on Spearman showed weak CRP. Conclusions: High predict early late hospitalizations retinopathy. also shows high may microvascular DM risk period In addition, comorbidities inflammatory habits, such long-term smoking, should be considered use

Language: Английский

---

Children and Young People with Long COVID—Comparing Those Seen in Post-COVID Services with a Non-Hospitalised National Cohort: A Descriptive Study

Children, Journal Year: 2023, Volume and Issue: 10(11), P. 1750 - 1750

Published: Oct. 28, 2023

Post-COVID services have been set up in England to treat children with ongoing symptoms of Long COVID. To date, the characteristics seeking treatment from these has not described.

Language: Английский

---

Thrombosis and Bleeding Risk Scores Are Strongly Associated with Mortality in Hospitalized Patients with COVID-19: A Multicenter Cohort Study

Journal of Clinical Medicine, Journal Year: 2024, Volume and Issue: 13(5), P. 1437 - 1437

Published: March 1, 2024

Background: Internationally established guidelines mention pharmacological prophylaxis for all hospitalized COVID-19 patients. However, there are concerns regarding the efficacy and safety of anticoagulants. This study investigated associations between thrombosis/bleeding risk scores clinical outcomes. Methods: We conducted a retrospective review adult patients admitted to two hospitals 2021 2022. analyzed data, laboratory results, low molecular weight heparin (LMWH) use, thrombosis, bleeding, 30-day survival. Results: Of 160 patients, 69.4% were female, median age was 59 years. The rates thrombotic complications mortality 12.5% 36.3%, respectively. LMWH administered 73 (45.6%). with high Padua prediction (PPS) IMPROVEVTE had significantly higher venous thromboembolism (VTE) compared those (30.8% vs. 9.0%, p = 0.006 25.6% 7.7%, 0.006). Similarly, elevated IMPROVEBRS associated increased (hazard ratios 7.49 6.27, respectively; < 0.001). Interestingly, use not decreased incidence VTE when stratified by groups. Conclusions: this suggests that thrombosis bleeding have rate.

Language: Английский

---