RESUMO
INTRODUCTION: Multimodal monitoring is the use of data from multiple physiological sensors combined in a way to provide individualized patient management. It is becoming commonplace in the civilian care of traumatic brain-injured patients. We hypothesized we could bring the technology to the battlefield using a noninvasive sensor suite and an artificial intelligence-based patient management guidance system. METHODS: Working with military medical personnel, we gathered requirements for a hand-held system that would adapt to the rapidly evolving field of neurocritical care. To select the optimal sensors, we developed a method to evaluate both the value of the sensor's measurement in managing brain injury and the burden to deploy that sensor in the battlefield. We called this the Value-Burden Analysis which resulted in a score weighted by the Role of Care. The Value was assessed using 7 criteria, 1 of which was the clinical value as assessed by a consensus of clinicians. The Burden was assessed using 16 factors such as size, weight, and ease of use. We evaluated and scored 17 sensors to test the assessment methodology. In addition, we developed a design for the guidance system, built a prototype, and tested the feasibility. RESULTS: The resulting architecture of the system was modular, requiring the development of an interoperable description of each component including sensors, guideline steps, medications, analytics, resources, and the context of care. A Knowledge Base was created to describe the interactions of the modules. A prototype test set-up demonstrated the feasibility of the system in that simulated physiological inputs would mimic the guidance provided by the current Clinical Practice Guidelines for Traumatic Brain Injury in Prolonged Care (CPG ID:63). The Value-Burden analysis yielded a ranking of sensors as well as sensor metadata useful in the Knowledge Base. CONCLUSION: We developed a design and tested the feasibility of a system that would allow the use of physiological biomarkers as a management tool in forward care. A key feature is the modular design that allows the system to adapt to changes in sensors, resources, and context as well as to updates in guidelines as they are developed. Continued work consists of further validation of the concept with simulated scenarios.
Assuntos
Biomarcadores , Lesões Encefálicas Traumáticas , Região de Recursos Limitados , Humanos , Biomarcadores/análise , Lesões Encefálicas Traumáticas/terapia , Lesões Encefálicas Traumáticas/diagnóstico , Militares/estatística & dados numéricos , Monitorização Fisiológica/métodos , Monitorização Fisiológica/instrumentação , Monitorização Fisiológica/normasRESUMO
Objective.The Root SedLine device is used for continuous electroencephalography (cEEG)-based sedation monitoring in intensive care patients. The cEEG traces can be collected for further processing and calculation of relevant metrics not already provided. Depending on the device settings during acquisition, the acquired traces may be distorted by max/min value cropping or high digitization errors. We aimed to systematically assess the impact of these distortions on metrics used for clinical research in the field of neuromonitoring.Approach.A 16 h cEEG acquired using the Root SedLine device at the optimal screen settings was analyzed. Cropping and digitization error effects were simulated by consecutive reduction of the maximum cEEG amplitude by 2µV or by reducing the vertical resolution. Metrics were calculated within ICM+ using minute-by-minute data, including the total power, alpha delta ratio (ADR), and 95% spectral edge frequency. Data were analyzed by creating violin- or box-plots.Main Results.Cropping led to a continuous reduction in total and band power, leading to corresponding changes in variability thereof. The relative power and ADR were less affected. Changes in resolution led to relevant changes. While the total power and power of low frequencies were rather stable, the power of higher frequencies increased with reducing resolution.Significance.Care must be taken when acquiring and analyzing cEEG waveforms from Root SedLine for clinical research. To retrieve good quality metrics, the screen settings must be kept within the central vertical scale, while pre-processing techniques must be applied to exclude unacceptable periods.
Assuntos
Cuidados Críticos , Eletroencefalografia , Humanos , Eletroencefalografia/métodos , Cuidados Críticos/métodos , Processamento de Sinais Assistido por Computador , Monitorização Fisiológica/métodos , Monitorização Fisiológica/instrumentação , MasculinoRESUMO
We have developed a mouse DNA methylation array that contains 296,070 probes representing the diversity of mouse DNA methylation biology. We present a mouse methylation atlas as a rich reference resource of 1,239 DNA samples encompassing distinct tissues, strains, ages, sexes, and pathologies. We describe applications for comparative epigenomics, genomic imprinting, epigenetic inhibitors, patient-derived xenograft assessment, backcross tracing, and epigenetic clocks. We dissect DNA methylation processes associated with differentiation, aging, and tumorigenesis. Notably, we find that tissue-specific methylation signatures localize to binding sites for transcription factors controlling the corresponding tissue development. Age-associated hypermethylation is enriched at regions of Polycomb repression, while hypomethylation is enhanced at regions bound by cohesin complex members. Apc Min/+ polyp-associated hypermethylation affects enhancers regulating intestinal differentiation, while hypomethylation targets AP-1 binding sites. This Infinium Mouse Methylation BeadChip (version MM285) is widely accessible to the research community and will accelerate high-sample-throughput studies in this important model organism.
RESUMO
Continuous multimodal monitoring in neurocritical care provides valuable insights into the dynamics of the injured brain. Unfortunately, the "readiness" of this data for robust artificial intelligence (AI) and machine learning (ML) applications is low and presents a significant barrier for advancement. Harmonization standards and tools to implement those standards are key to overcoming existing barriers. Consensus in our professional community is essential for success.
Assuntos
Inteligência Artificial , Aprendizado de Máquina , HumanosRESUMO
BACKGROUND: Most trials in critical care have been neutral, in part because between-patient heterogeneity means not all patients respond identically to the same treatment. The Precision Care in Cardiac Arrest: Influence of Cooling duration on Efficacy in Cardiac Arrest Patients (PRECICECAP) study will apply machine learning to high-resolution, multimodality data collected from patients resuscitated from out-of-hospital cardiac arrest. We aim to discover novel biomarker signatures to predict the optimal duration of therapeutic hypothermia and 90-day functional outcomes. In parallel, we are developing a freely available software platform for standardized curation of intensive care unit-acquired data for machine learning applications. METHODS: The Influence of Cooling duration on Efficacy in Cardiac Arrest Patients (ICECAP) study is a response-adaptive, dose-finding trial testing different durations of therapeutic hypothermia. Twelve ICECAP sites will collect data for PRECICECAP from multiple modalities routinely used after out-of-hospital cardiac arrest, including ICECAP case report forms, detailed medication data, cardiopulmonary and electroencephalographic waveforms, and digital imaging and communications in medicine files (DICOMs). We partnered with Moberg Analytics to develop a freely available software platform to allow high-resolution critical care data to be used efficiently and effectively. We will use an autoencoder neural network to create low-dimensional representations of all raw waveforms and derivative features, censored at rewarming to ensure clinical usability to guide optimal duration of hypothermia. We will also consider simple features that are historically considered to be important. Finally, we will create a supervised deep learning neural network algorithm to directly predict 90-day functional outcome from large sets of novel features. RESULTS: PRECICECAP is currently enrolling and will be completed in late 2025. CONCLUSIONS: Cardiac arrest is a heterogeneous disease that causes substantial morbidity and mortality. PRECICECAP will advance the overarching goal of titrating personalized neurocritical care on the basis of robust measures of individual need and treatment responsiveness. The software platform we develop will be broadly applicable to hospital-based research after acute illness or injury.