Prediction model for intracranial hypertension demonstrates robust performance during external validation on the CENTER-TBI dataset

Carra, Giorgia; Güiza, Fabian; Depreitere, Bart; Meyfroidt, Geert; Participants, Center-Tbi; Investigators,

doi:10.1007/s00134-020-06247-4

Cited by 10 publications

(20 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, Carra and colleagues undertook the validation testing of Gaussian processes (GPs)-based predictive modeling using the high-resolution CENTER-TBI data set. 10 These algorithms demonstrated good intercenter robustness, with the model achieving an accuracy of 88%, sensitivity of 83%, and specificity of 91% in providing a 30-min forewarning of an ICP crisis (defined as an ICP >30 mm Hg lasting at least 10 consecutive min). However, using GPs, though promising with retrospective data, is computationally intensive and requires 4 h of input data to allow it to make a prediction.…”

Section: Discussionmentioning

confidence: 97%

Prediction of Intracranial Hypertension and Brain Tissue Hypoxia Utilizing High-Resolution Data from the BOOST-II Clinical Trial

Lazaridis

Ajith

Mansour

et al. 2022

Neurotrauma Reports

View full text Add to dashboard Cite

The current approach to intracranial hypertension and brain tissue hypoxia is reactive, based on fixed thresholds. We used statistical machine learning on high-frequency intracranial pressure (ICP) and partial brain tissue oxygen tension (PbtO 2 ) data obtained from the BOOST-II trial with the goal of constructing robust quantitative models to predict ICP/PbtO 2 crises. We derived the following machine learning models: logistic regression (LR), elastic net, and random forest. We split the data set into 70–30% for training and testing and utilized a discrete-time survival analysis framework and 5-fold hyperparameter optimization strategy for all models. We compared model performances on discrimination between events and non-events of increased ICP or low PbtO 2 with the area under the receiver operating characteristic (AUROC) curve. We further analyzed clinical utility through a decision curve analysis (DCA). When considering discrimination, the number of features, and interpretability, we identified the RF model that combined the most recent ICP reading, episode number, and longitudinal trends over the preceding 30 min as the best performing for predicting ICP crisis events within the next 30 min (AUC 0.78). For PbtO 2 , the LR model utilizing the most recent reading, episode number, and longitudinal trends over the preceding 30 min was the best performing (AUC, 0.84). The DCA showed clinical usefulness for wide risk of thresholds for both ICP and PbtO 2 predictions. Acceptable alerting thresholds could range from 20% to 80% depending on a patient-specific assessment of the benefit-risk ratio of a given intervention in response to the alert.

show abstract

Section: Discussionmentioning

confidence: 97%

Prediction of Intracranial Hypertension and Brain Tissue Hypoxia Utilizing High-Resolution Data from the BOOST-II Clinical Trial

Lazaridis

Ajith

Mansour

et al. 2022

Neurotrauma Reports

View full text Add to dashboard Cite

show abstract

“…Their model had a classification accuracy of 77%, a sensitivity of 82% and a specificity of 75% in the first validation cohort. The model was later validated on two other cohorts of adult TBI patients; 151 subjects part of a multicenter study including five centers 21 and 257 other subjects recruited as part of the CENTER-TBI high-resolution ICU monitoring cohort 22 . The precision varies somewhat between the cohorts (sensitivity/specificity 70/90% and 83/91% respectively) but are overall in the same range as the ones found with our model.…”

Section: Discussionmentioning

confidence: 99%

Probabilistic prediction of increased intracranial pressure in patients with severe traumatic brain injury

Wijayatunga

Koskinen

Sundström

2022

Sci Rep

View full text Add to dashboard Cite

Traumatic brain injury (TBI) causes alteration in brain functions. Generally, at intensive care units (ICU), intracranial pressure (ICP) is monitored and treated to avoid increases in ICP with associated poor clinical outcome. The aim was to develop a model which could predict future ICP levels of individual patients in the ICU, to warn treating clinicians before secondary injuries occur. A simple and explainable, probabilistic Markov model was developed for the prediction task ICP ≥ 20 mmHg. Predictions were made for 10-min intervals during 60 min, based on preceding hour of ICP. A prediction enhancement method was developed to compensate for data imbalance. The model was evaluated on 29 patients with severe TBI. With random data selection from all patients (80/20% training/testing) the specificity of the model was high (0.94–0.95) and the sensitivity good to high (0.73–0.87). Performance was similar (0.90–0.95 and 0.73–0.89 respectively) when the leave-one-out cross-validation was applied. The new model could predict increased levels of ICP in a reliable manner and the enhancement method further improved the predictions. Further advantages are the straightforward expandability of the model, enabling inclusion of other time series data and/or static parameters. Next step is evaluation on more patients and inclusion of parameters other than ICP.

show abstract

“…In this and all subsequent studies by Güiza and associates, a tIH event was classified as a rise in ICP ≥30 mm Hg that lasted for at least 10 min. 59 , 69 , 70 …”

Section: Tih Prediction Algorithmsmentioning

confidence: 99%

Development of Traumatic Brain Injury Associated Intracranial Hypertension Prediction Algorithms: A Narrative Review

McNamara

Meka

Anstey

et al. 2023

Journal of Neurotrauma

View full text Add to dashboard Cite

Traumatic intracranial hypertension (tIH) is a common and potentially lethal complication of moderate to severe traumatic brain injury (m-sTBI). It often develops with little warning and is managed reactively with the tiered application of intracranial pressure (ICP)-lowering interventions administered in response to an ICP rising above a set threshold. For over 45 years, a variety of research groups have worked toward the development of technology to allow for the preemptive management of tIH in the hope of improving patient outcomes. In 2022, the first operationalizable tIH prediction system became a reality. With such a system, ICP lowering interventions could be administered prior to the rise in ICP, thus protecting the patient from potentially damaging tIH episodes and limiting the overall ICP burden experienced. In this review, we discuss related approaches to ICP forecasting and IH prediction algorithms, which collectively provide the foundation for the successful development of an operational tIH prediction system. We also discuss operationalization and the statistical assessment of tIH algorithms. This review will be of relevance to clinicians and researchers interested in development of this technology as well as those with a general interest in the bedside application of machine learning (ML) technology.

show abstract

Prediction model for intracranial hypertension demonstrates robust performance during external validation on the CENTER-TBI dataset

Cited by 10 publications

References 5 publications

Prediction of Intracranial Hypertension and Brain Tissue Hypoxia Utilizing High-Resolution Data from the BOOST-II Clinical Trial

Prediction of Intracranial Hypertension and Brain Tissue Hypoxia Utilizing High-Resolution Data from the BOOST-II Clinical Trial

Probabilistic prediction of increased intracranial pressure in patients with severe traumatic brain injury

Development of Traumatic Brain Injury Associated Intracranial Hypertension Prediction Algorithms: A Narrative Review

Contact Info

Product

Resources

About