Springer eBooks, Год журнала: 2024, Номер unknown, С. 165 - 192
Опубликована: Янв. 1, 2024
Язык: Английский
Pharmaceutics, Год журнала: 2021, Номер 13(9), С. 1514 - 1514
Опубликована: Сен. 18, 2021
Effective therapies for COVID-19 are still lacking, and drug repositioning is a promising approach to address this problem. Here, we adopted medical informatics repositioning. We leveraged large prospective cohort, the UK-Biobank (UKBB, N ~ 397,000), studied associations of prior use all level-4 ATC categories (N = 819, including vaccines) with diagnosis severity. Effects drugs on risk infection, disease severity, mortality were investigated separately. Logistic regression was conducted, controlling main confounders. observed strong highly consistent protective statins. Many top-listed also cardiovascular medications, such as angiotensin-converting enzyme inhibitors (ACEI), angiotensin receptor blockers (ARB), calcium channel blocker (CCB), beta-blockers. Some other showing included biguanides (metformin), estrogens, thyroid hormones, proton pump inhibitors, testosterone-5-alpha reductase among others. by influenza, pneumococcal, several vaccines. Subgroup interaction analyses which revealed differences in effects various subgroups. For example, flu/pneumococcal vaccines weaker obese individuals, while protection statins stronger patients. To conclude, our analysis many candidates, example medications. Further studies required validation.
Язык: Английский
Процитировано
17
American Journal of Transplantation, Год журнала: 2023, Номер 23(12), С. 1908 - 1921
Опубликована: Авг. 30, 2023
Язык: Английский
Процитировано
7
The World Journal of Men s Health, Год журнала: 2023, Номер 41(3), С. 466 - 466
Опубликована: Янв. 1, 2023
Since the beginning of coronavirus disease 19 (COVID-19) pandemic, efforts in defining risk factors and associations between severe acute respiratory syndrome 2 (SARS-CoV-2), clinical, molecular features have initiated. After three years it became evident that men higher adverse outcomes. Such evidence provided impetus for biological fundaments such a gender disparity. Our objective was to analyze most recent literature with aim relationship COVID-19 fertility, particular, we assessed interplay SARS-CoV-2 testosterone systematic review from December 2019 (first novel Hubei province) until March 2022. As fundamental basis understanding, articles pertaining preclinical aspects explaining disparity (n=9) were included. The main categories analyzed being infected according levels (n=5), impact serum on outcomes (n=23), after infection (n=19). Preclinical studies mainly evaluated relation angiotensin-converting enzyme (ACE2) its androgen-mediated regulation, exploring few. Although publications evaluating effect fertility found low infection, follow-up short, some also suggesting no alterations during recovery. More conclusive findings observed levels, generally at experiencing worse (i.e., admission intensive care units, longer hospitalization, death). Interestingly, an inverse women, where associated finding may provide meaningful insights better patient counselling individualization pathways hypogonadism.
Язык: Английский
Процитировано
5
Diagnostics, Год журнала: 2023, Номер 13(10), С. 1749 - 1749
Опубликована: Май 16, 2023
Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), causing a disease called COVID-19, is class of acute respiratory syndrome that has considerably affected the global economy and healthcare system. This virus diagnosed using traditional technique known as Reverse Transcription Polymerase Chain Reaction (RT-PCR) test. However, RT-PCR customarily outputs lot false-negative incorrect results. Current works indicate COVID-19 can also be imaging resolutions, including CT scans, X-rays, blood tests. Nevertheless, X-rays scans cannot always used for patient screening because high costs, radiation doses, an insufficient number devices. Therefore, there requirement less expensive faster diagnostic model to recognize positive negative cases COVID-19. Blood tests are easily performed cost than Since biochemical parameters in routine vary during infection, they may supply physicians with exact information about diagnosis study reviewed some newly emerging artificial intelligence (AI)-based methods diagnose We gathered research resources inspected 92 articles were carefully chosen from variety publishers, such IEEE, Springer, Elsevier, MDPI. Then, these studies classified into two tables which contain use machine Learning deep models while test datasets. In studies, diagnosing Random Forest logistic regression most widely learning performance metrics accuracy, sensitivity, specificity, AUC. Finally, we conclude by discussing analyzing datasets detection. survey starting point novice-/beginner-level researcher perform on classification.
Язык: Английский
Процитировано
5
Machine Learning and Knowledge Extraction, Год журнала: 2023, Номер 5(4), С. 1266 - 1281
Опубликована: Сен. 25, 2023
Machine learning (ML) has been used in different ways the fight against COVID-19 disease. ML models have developed, e.g., for diagnostic or prognostic purposes and using various modalities of data (e.g., textual, visual, structured). Due to many specific aspects this disease its evolution over time, there is still not enough understanding all relevant factors influencing course particular patients. In our work, was a strong involvement medical expert following human-in-the-loop principle. This very important but usually neglected part knowledge extraction (KE) process. Our research shows that explainable artificial intelligence (XAI) may significantly support KE. focused on two scenarios. first scenario, we aimed discover whether adding information about predominant variant impacts performance models. second classification concerning need an intensive care unit given patient connection with explainability AI methods. We nine algorithms, namely XGBoost, CatBoost, LightGBM, logistic regression, Naive Bayes, random forest, SGD, SVM-linear, SVM-RBF. measured resulting precision, accuracy, AUC metrics. Subsequently, from best-performing approaches as follows: (a) features extracted automatically by forward stepwise selection (FSS); (b) attributes their interactions discovered model Both were compared selected experts advance based domain expertise. experiments showed did influence It also turned out much more precise identification significant than FSS. Explainability methods identified almost same interesting among them, which discussed point view. The results consequences are discussed.
Язык: Английский
Процитировано
4
Ophthalmology and Therapy, Год журнала: 2024, Номер 13(12), С. 3025 - 3034
Опубликована: Окт. 10, 2024
Язык: Английский
Процитировано
1
Data Intelligence, Год журнала: 2024, Номер unknown, С. 1 - 34
Опубликована: Апрель 16, 2024
Abstract Background: During the COVID-19 pandemic, clinical trial recruitment cannot be carried out due to travel restrictions, transmission risks and other factors, resulting in stagnation of a large number ongoing or upcoming trials. Objective: An intelligent screening app was developed using artificial intelligence technology rapidly pre-screen potential patients for phase I solid tumor drug Methods: A total 429 process records were collected from 27 trials at First Affiliated Hospital Bengbu Medical College April 2018 May 2021. Features experimental data analyzed, collinearity (principal component analysis) strong correlation (χ2 test) among features eliminated. XGBoost, Random Forest, Naive Bayes used sort weight importance features. Finally, pre-screening models constructed classification machine learning algorithm, optimal model selected. Results: Among records, 33 generated by repeated subject participation different trials, remaining 396 246 (62.12%) screened successfully. The gold standard success is final judgment made principal investigator (PI) based on protocol. Venn diagram identify important feature intersections algorithms. After intersecting top 15 characteristic variables models, 9 common obtained: age, sex, distance residence central institution, histology, stage, tumorectomy, interval diagnosis/postoperative screening, chemotherapy, ECOG (Eastern Cooperative Oncology Group, ECOG) score. To select subset, expanded 12 subsets, performance subsets under validated. results showed that performance, accuracy practicability achieved XGBoost with subset. could accurately predict rates both internal (AUC =3D 0.895) external 0.796) validation, has been transformed into convenient tool facilitate its application settings. Subjects probability exceeding equaling threshold had higher successfully screened. Conclusion: Based model, we created an online prediction calculator visualization -- ISSP (Intelligent Screening Service Platform), which can screen effectively solve problem space time interval. On mobile terminal, it realizes matching between projects patients, completes rapid subjects, so as obtain more subjects. As auxiliary tool, optimizes provides services investigators patients.
Язык: Английский
Процитировано
1
Springer optimization and its applications, Год журнала: 2023, Номер unknown, С. 111 - 128
Опубликована: Янв. 1, 2023
Язык: Английский
Процитировано
3
Frontiers in Public Health, Год журнала: 2023, Номер 11
Опубликована: Май 26, 2023
Background Most existing prognostic models of COVID-19 require imaging manifestations and laboratory results as predictors, which are only available in the post-hospitalization period. Therefore, we aimed to develop validate a model assess in-hospital death risk patients using routinely predictors at hospital admission. Methods We conducted retrospective cohort study with Healthcare Cost Utilization Project State Inpatient Database 2020. Patients hospitalized Eastern United States (Florida, Michigan, Kentucky, Maryland) were included training set, those Western (Nevada) validation set. Discrimination, calibration, clinical utility evaluated model's performance. Results A total 17 954 deaths occurred set ( n = 168 137), 1,352 12 577). The final prediction 15 variables readily admission, including age, sex, 13 comorbidities. This showed moderate discrimination an area under curve (AUC) 0.726 (95% confidence interval [CI]: 0.722—0.729) good calibration (Brier score 0.090, slope 1, intercept 0) set; similar predictive ability was observed Conclusion An easy-to-use based on admission developed validated for early identification high death. can be decision-support tool triage optimize resource allocation.
Язык: Английский
Процитировано
2
Human-Centric Intelligent Systems, Год журнала: 2023, Номер 3(4), С. 508 - 520
Опубликована: Авг. 31, 2023
Abstract Diseases are increasing with exponential rate worldwide. Its detection is challenging task due to unavailability of the experts. Machine learning models provide automated mechanism detect diseases once trained. It has been used predict and many such as cancer, heart attack, liver infections, kidney infections. The new coronavirus become one deadliest diseases. case escalated in unexpected ways. In literature, machine Extreme Gradient Boosting (XGBoosting), Support Vector (SVM), regression, Logistic regression have used. observed that these can COVID cases early but unable find peak point deadline disease. Hence, mathematical designed dead-line disease prediction. These use integral calculus-based Ordinary Differential Equations (ODEs) cases. Governments dependent on models’ pre- diction for preparation hospitalization, medicines, more. higher prediction accuracy required. found literature fractional more accurate detection. Fractional provides choose order derivative value which information processing capability increases. present work, model using calculus devised model, quarantine, symptomatic asymptomatic incorporated proposed not only predicts accurately also gives
Язык: Английский
Процитировано
2