Development and validation of a nomogram-based risk prediction model for unfavorable outcomes in pediatric traumatic brain injury: a retrospective study

Frontiers in Pediatrics, Journal Year: 2025, Volume and Issue: 13

Published: April 11, 2025

Introduction Pediatric traumatic brain injury (PTBI) is linked to significant disability and mortality. This study aimed identify risk factors for unfavorable outcomes in patients with PTBI develop a predictive model. Methods A retrospective analysis was conducted on treated at the 900th Hospital from September 2021 June 2023. Univariate multivariate regression analyses identified adverse facilitated creation of nomogram. The model's accuracy assessed using Receiver Operating Characteristic (ROC) curves, calibration Decision Curve Analysis (DCA). External validation performed Fujian Children's Hospital. Results Key findings indicated that Glasgow Coma Scale (GCS) score ≤8, subdural hematoma, subarachnoid hemorrhage, coagulopathy were independent factors. nomogram achieved an area under ROC curve 0.947 development cohort 0.834 external cohort, demonstrating good fit. DCA results confirmed enhanced prediction outcomes. Conclusions model offers high early identification outcomes, enabling timely interventions improve quality life PTBI.

Language: Английский

---

Artificial intelligence (AI) for neurologists: do digital neurones dream of electric sheep?

Practical Neurology, Journal Year: 2023, Volume and Issue: 23(6), P. 476 - 488

Published: Nov. 17, 2023

Artificial intelligence (AI) is routinely mentioned in journals and newspapers, non-technical outsiders may have difficulty distinguishing hyperbole from reality. We present a practical guide to help neurologists understand healthcare AI. AI being used support clinical decisions treating neurological disorders. introduce basic concepts of AI, such as machine learning natural language processing, explain how healthcare, giving examples its benefits challenges. also cover performance measured, regulatory aspects healthcare. An important theme that general-purpose technology like medical statistics, with broad utility applicable various scenarios, niche approaches are outpaced by broadly many disease areas specialties. By understanding basics potential applications, can make informed when evaluating their practice. This article was written four humans, generative helping formatting image generation.

Language: Английский

---

Machine learning‐based prediction of clinical outcomes after traumatic brain injury: Hidden information of early physiological time series

CNS Neuroscience & Therapeutics, Journal Year: 2024, Volume and Issue: 30(7)

Published: July 1, 2024

To assess the predictive value of early-stage physiological time-series (PTS) data and non-interrogative electronic health record (EHR) signals, collected within 24 h ICU admission, for traumatic brain injury (TBI) patient outcomes.

Language: Английский

---

Fluid-Based Protein Biomarkers in Traumatic Brain Injury: The View from the Bedside

International Journal of Molecular Sciences, Journal Year: 2023, Volume and Issue: 24(22), P. 16267 - 16267

Published: Nov. 13, 2023

There has been an explosion of research into biofluid (blood, cerebrospinal fluid, CSF)-based protein biomarkers in traumatic brain injury (TBI) over the past decade. The availability very large datasets, such as CENTRE-TBI and TRACK-TBI, allows for correlation blood- CSF-based molecular (protein), radiological (structural) clinical (physiological) marker data to adverse outcomes. quality a given biomarker often framed relation predictive power on outcome quantified from area under Receiver Operating Characteristic (ROC) curve. However, this does not itself provide utility but reflects statistical association any population between one or more variables outcome. It is currently established how incorporate integrate biofluid-based patient management because there no standardized role decision making. We review current status discuss we can existing markers practice what additional do need improve diagnoses guide therapy assess treatment efficacy. Furthermore, argue employing machine learning (ML) capabilities with other established, routinely used diagnostic tools, clinician actionable information medical intervention.

Language: Английский

---

Convolutional neural networks for traumatic brain injury classification and outcome prediction

Health Sciences Review, Journal Year: 2023, Volume and Issue: 9, P. 100126 - 100126

Published: Oct. 5, 2023

The detection and classification of traumatic brain injury (TBI) by medical professionals can vary due to subjectivity differences in experience. Thus, a computational approach for detecting classifying TBI would be invaluable an objective diagnosis this injury. In review paper, various machine learning algorithms used detect, classify, predict the severity outcomes clinical setting are discussed. most promising these is convolutional neural network (CNN), which highlighted review.

Language: Английский

---

Traumatic Brain Injury in the Long-COVID Era

Neurotrauma Reports, Journal Year: 2024, Volume and Issue: 5(1), P. 81 - 94

Published: Jan. 1, 2024

Major determinants of the biological background or reserve, such as age, sex, comorbidities (diabetes, hypertension, obesity, etc.), and medications (e.g., anticoagulants), are known to affect outcome after traumatic brain injury (TBI). With unparalleled data richness coronavirus disease 2019 (COVID-19; ∼375,000 counting!) well chronic form, long-COVID, also called post-acute sequelae SARS-CoV-2 infection (PASC), publications (∼30,000 counting) covering virtually every aspect diseases, pathomechanisms, biomarkers, phases, symptomatology, etc., have provided a unique opportunity better understand appreciate holistic nature interconnectivity between organ systems, importance in modifying trajectories affecting outcomes. Such approach is badly needed TBI-induced conditions their totality. Here, I briefly review what about long-COVID/PASC, its underlying—suspected—pathologies, pathobiological changes induced by TBI, other words, TBI endophenotypes, discuss intersection long-COVID/PASC pathobiologies, how considering some factors person's inclusion mechanistic molecular biomarkers can help improve clinical management patients.

Language: Английский

---

Predictive value of the Trauma Rating Index in Age, Glasgow Coma Scale, Respiratory rate and Systolic blood pressure score (TRIAGES) for the short-term mortality of older patients with isolated traumatic brain injury: a retrospective cohort study

BMJ Open, Journal Year: 2024, Volume and Issue: 14(3), P. e082770 - e082770

Published: March 1, 2024

This study aimed to evaluate the effectiveness of Trauma Rating Index in Age, Glasgow Coma Scale, Respiratory rate and Systolic blood pressure score (TRIAGES) predicting 24-hour in-hospital mortality among patients aged 65 years older with isolated traumatic brain injury (TBI).

Language: Английский

---

Federated Convolutional Neural Networks for Predictive Analysis of Traumatic Brain Injury: Advancements in Decentralized Health Monitoring

International Journal of Advanced Computer Science and Applications, Journal Year: 2024, Volume and Issue: 15(4)

Published: Jan. 1, 2024

Traumatic Brain Injury (TBI) is a significant global health concern, often leading to long-term disabilities and cognitive impairments. Accurate timely diagnosis of TBI crucial for effective treatment management. In this paper, we propose novel federated convolutional neural network (FedCNN) framework predictive analysis in decentralized monitoring. The implemented Python, leveraging three diverse datasets: CQ500, RSNA, CENTER-TBI, each containing annotated brain CT images associated with TBI. methodology encompasses data preprocessing, feature extraction using gray level co-occurrence matrix (GLCM), selection employing the Grasshopper Optimization Algorithm (GOA), classification FedCNN. Our approach achieves superior performance compared existing methods such as DANN, RF DT, LSTM, an accuracy 99.2%, surpassing other approaches by 1.6%. FedCNN offers privacy-preserving training across individual networks while sharing model parameters central server, ensuring privacy decentralization Evaluation metrics including accuracy, precision, recall, F1-score demonstrate effectiveness our accurately classifying normal abnormal ROC further validates discriminative ability framework, highlighting its potential advanced tool diagnosis. study contributes field monitoring providing reliable efficient management, offering advancements patient care healthcare Future research could explore extending incorporate additional modalities datasets, well integrating deep learning architectures optimization algorithms improve scalability applications.

Language: Английский

---

Performance Metrics, Algorithms, and Applications of Artificial Intelligence in Vascular and Interventional Neurology

Neurologic Clinics, Journal Year: 2024, Volume and Issue: 42(3), P. 633 - 650

Published: May 18, 2024

Language: Английский

---

Development and Validation of a Novel Classification System and Prognostic Model for Open Traumatic Brain Injury: A Multicenter Retrospective Study

Neurology and Therapy, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 4, 2024

Open traumatic brain injury (OTBI) is associated with high mortality and morbidity; however, the classification of these injuries determination patient prognosis remain uncertain, hindering selection optimal treatment strategies. This study aimed to develop validate a novel OTBI system prognostic model for poor prognosis. retrospective included patients isolated who received at three large medical centers in China between January 2020 June 2022 as training set. Data on collected Fuzong Clinical Medical College Fujian University July 2023 were used validation parameters, including clinical data admission, radiological laboratory findings, details surgical methods, collected. Prognosis was assessed through dichotomized Glasgow Outcome Scale (GOS). A proposed, categorizing based combination intracranial hematoma midline shift observed imaging, logistic regression analyses performed identify risk factors investigate association Finally, nomogram suitable application established validated. Multivariable analysis identified type C (p < 0.001), Coma score (GCS) ≤ 8 subarachnoid hemorrhage (SAH) = 0.004), subdural (SDH) 0.011), coagulopathy 0.020) independent The addition containing all other improved predictive ability (Z 1.983; p 0.047). In set, achieved an area under curve (AUC) 0.917 [95% confidence interval (CI) 0.864–0.970]. calibration closely approximated ideal curve, indicating strong performance model. implementation our proposed its use alongside here may improve prediction aid most

Language: Английский

---