Correction: Expert-augmented automated machine learning optimizes hemodynamic predictors of spinal cord injury outcome.

[This corrects the article DOI: 10.1371/journal.pone.0265254.].

Safety and comparative efficacy of initiating low-molecular-weight heparin within 24 hours of injury or surgery for venous thromboembolism prophylaxis in patients with spinal cord injury: a prospective TRACK-SCI registry study.

OBJECTIVE: Venous thromboembolism (VTE) following traumatic spinal cord injury (SCI) is a significant clinical concern. This study sought to determine the incidence of VTE and hemorrhagic complications among patients with SCI who received low-molecular-weight heparin (LMWH) within 24 hours of injury or surgery and identify variables that predict VTE using the prospective Transforming Research and Clinical Knowledge in SCI (TRACK-SCI) database. METHODS: The TRACK-SCI database was queried for individuals with traumatic SCI from 2015 to 2022. Primary outcomes of interest included rates of VTE (including deep vein thrombosis [DVT] and pulmonary embolism [PE]) and in-hospital hemorrhagic complications that occurred after LWMH administration. Secondary outcomes included intensive care unit and hospital length of stay, discharge location type, and in-hospital mortality. RESULTS: The study cohort consisted of 162 patients with SCI. Fifteen of the 162 patients withdrew from the study, leading to loss of data for certain variables for these patients. One hundred thirty patients (87.8%) underwent decompression and/or fusion surgery for SCI. DVT occurred in 11 (7.4%) of 148 patients, PE in 9 (6.1%) of 148, and any VTE in 18 (12.2%) of 148 patients. The analysis showed that admission lower-extremity motor score (p = 0.0408), injury at the thoracic level (p = 0.0086), admission American Spinal Injury Association grade (p = 0.0070), and younger age (p = 0.0372) were significantly associated with VTE. There were 3 instances of postoperative spine surgery-related bleeding (2.4%) in the 127 patients who had spine surgery with bleeding complication data available, with one requiring return to surgery (0.8%). Thirteen (8.8%) of 147 patients had a bleeding complication not related to spine surgery. There were 2 gastrointestinal bleeds associated with nasogastric tube placement, 3 cases of postoperative non-spine-related surgery bleeding, and 8 cases of other bleeding complications (5.4%) not related to any surgery. CONCLUSIONS: Initiation of LMWH within 24 hours was associated with a low rate of spine surgery-related bleeding. Bleeding complications unrelated to SCI surgery still occur with LMWH administration. Because neurosurgical intervention is typically the limiting factor in initializing chemical DVT prophylaxis, many of these bleeding complications would have likely occurred regardless of the protocol.

Hypotension requiring vasopressor treatment and increased cardiac complications in elderly spinal cord injury patients: a prospective TRACK-SCI registry study.

OBJECTIVE: Increasing life expectancy has led to an older population. In this study, the authors analyzed complications and outcomes in elderly patients following spinal cord injury (SCI) using the established multi-institutional prospective study Transforming Research and Clinical Knowledge in SCI (TRACK-SCI) database collected in the Department of Neurosurgical Surgery at the University of California, San Francisco. METHODS: TRACK-SCI was queried for elderly individuals (≥ 65 years of age) with traumatic SCI from 2015 to 2019. Primary outcomes of interest included total hospital length of stay, perioperative complications, postoperative complications, and in-hospital mortality. Secondary outcomes included disposition location, and neurological improvement based on the American Spinal Injury Association Impairment Scale (AIS) grade at discharge. Descriptive analysis, Fisher's exact test, univariate analysis, and multivariable regression analysis were performed. RESULTS: The study cohort consisted of 40 elderly patients. The in-hospital mortality rate was 10%. Every patient in this cohort experienced at least 1 complication, with a mean of 6.6 separate complications (median 6, mode 4). The most common complication categories were cardiovascular, with a mean of 1.6 complications (median 1, mode 1), and pulmonary, with a mean of 1.3 (median 1, mode 0) complications, with 35 patients (87.5%) having at least 1 cardiovascular complication and 25 (62.5%) having at least 1 pulmonary complication. Overall, 32 patients (80%) required vasopressor treatment for mean arterial pressure (MAP) maintenance goals. The use of norepinephrine correlated with increased cardiovascular complications. Only 3 patients (7.5%) of the total cohort had an improved AIS grade compared with their acute level at admission. CONCLUSIONS: Given the increased frequency of cardiovascular complications associated with vasopressor use in elderly SCI patients, caution is warranted when targeting MAP goals in these patients. A downward adjustment of blood pressure maintenance goals and prophylactic cardiology consultation to select the most appropriate vasopressor agent may be advisable for SCI patients ≥ 65 years of age.

Decision tree-based machine learning analysis of intraoperative vasopressor use to optimize neurological improvement in acute spinal cord injury.

OBJECTIVE: Previous work has shown that maintaining mean arterial pressures (MAPs) between 76 and 104 mm Hg intraoperatively is associated with improved neurological function at discharge in patients with acute spinal cord injury (SCI). However, whether temporary fluctuations in MAPs outside of this range can be tolerated without impairment of recovery is unknown. This retrospective study builds on previous work by implementing machine learning to derive clinically actionable thresholds for intraoperative MAP management guided by neurological outcomes. METHODS: Seventy-four surgically treated patients were retrospectively analyzed as part of a longitudinal study assessing outcomes following SCI. Each patient underwent intraoperative hemodynamic monitoring with recordings at 5-minute intervals for a cumulative 28,594 minutes, resulting in 5718 unique data points for each parameter. The type of vasopressor used, dose, drug-related complications, average intraoperative MAP, and time spent in an extreme MAP range (< 76 mm Hg or > 104 mm Hg) were collected. Outcomes were evaluated by measuring the change in American Spinal Injury Association Impairment Scale (AIS) grade over the course of acute hospitalization. Features most predictive of an improvement in AIS grade were determined statistically by generating random forests with 10,000 iterations. Recursive partitioning was used to establish clinically intuitive thresholds for the top features. RESULTS: At discharge, a significant improvement in AIS grade was noted by an average of 0.71 levels (p = 0.002). The hemodynamic parameters most important in predicting improvement were the amount of time intraoperative MAPs were in extreme ranges and the average intraoperative MAP. Patients with average intraoperative MAPs between 80 and 96 mm Hg throughout surgery had improved AIS grades at discharge. All patients with average intraoperative MAP > 96.3 mm Hg had no improvement. A threshold of 93 minutes spent in an extreme MAP range was identified after which the chance of neurological improvement significantly declined. Finally, the use of dopamine as compared to norepinephrine was associated with higher rates of significant cardiovascular complications (50% vs 25%, p < 0.001). CONCLUSIONS: An average intraoperative MAP value between 80 and 96 mm Hg was associated with improved outcome, corroborating previous results and supporting the clinical verifiability of the model. Additionally, an accumulated time of 93 minutes or longer outside of the MAP range of 76-104 mm Hg is associated with worse neurological function at discharge among patients undergoing emergency surgical intervention for acute SCI.

Expert-augmented automated machine learning optimizes hemodynamic predictors of spinal cord injury outcome.

Artificial intelligence and machine learning (AI/ML) is becoming increasingly more accessible to biomedical researchers with significant potential to transform biomedicine through optimization of highly-accurate predictive models and enabling better understanding of disease biology. Automated machine learning (AutoML) in particular is positioned to democratize artificial intelligence (AI) by reducing the amount of human input and ML expertise needed. However, successful translation of AI/ML in biomedicine requires moving beyond optimizing only for prediction accuracy and towards establishing reproducible clinical and biological inferences. This is especially challenging for clinical studies on rare disorders where the smaller patient cohorts and corresponding sample size is an obstacle for reproducible modeling results. Here, we present a model-agnostic framework to reinforce AutoML using strategies and tools of explainable and reproducible AI, including novel metrics to assess model reproducibility. The framework enables clinicians to interpret AutoML-generated models for clinical and biological verifiability and consequently integrate domain expertise during model development. We applied the framework towards spinal cord injury prognostication to optimize the intraoperative hemodynamic range during injury-related surgery and additionally identified a strong detrimental relationship between intraoperative hypertension and patient outcome. Furthermore, our analysis captured how evolving clinical practices such as faster time-to-surgery and blood pressure management affect clinical model development. Altogether, we illustrate how expert-augmented AutoML improves inferential reproducibility for biomedical discovery and can ultimately build trust in AI processes towards effective clinical integration.

Topological network analysis of patient similarity for precision management of acute blood pressure in spinal cord injury.

Background: Predicting neurological recovery after spinal cord injury (SCI) is challenging. Using topological data analysis, we have previously shown that mean arterial pressure (MAP) during SCI surgery predicts long-term functional recovery in rodent models, motivating the present multicenter study in patients. Methods: Intra-operative monitoring records and neurological outcome data were extracted (n = 118 patients). We built a similarity network of patients from a low-dimensional space embedded using a non-linear algorithm, Isomap, and ensured topological extraction using persistent homology metrics. Confirmatory analysis was conducted through regression methods. Results: Network analysis suggested that time outside of an optimum MAP range (hypotension or hypertension) during surgery was associated with lower likelihood of neurological recovery at hospital discharge. Logistic and LASSO (least absolute shrinkage and selection operator) regression confirmed these findings, revealing an optimal MAP range of 76-[104-117] mmHg associated with neurological recovery. Conclusions: We show that deviation from this optimal MAP range during SCI surgery predicts lower probability of neurological recovery and suggest new targets for therapeutic intervention. Funding: NIH/NINDS: R01NS088475 (ARF); R01NS122888 (ARF); UH3NS106899 (ARF); Department of Veterans Affairs: 1I01RX002245 (ARF), I01RX002787 (ARF); Wings for Life Foundation (ATE, ARF); Craig H. Neilsen Foundation (ARF); and DOD: SC150198 (MSB); SC190233 (MSB).

Spinal cord injury is a devastating condition that involves damage to the nerve fibers connecting the brain with the spinal cord, often leading to permanent changes in strength, sensation and body functions, and in severe cases paralysis. Scientists around the world work hard to find ways to treat or even repair spinal cord injuries but few patients with complete immediate paralysis recover fully. Immediate paralysis is caused by direct damage to neurons and their extension in the spinal cord. Previous research has shown that blood pressure regulation may be key in saving these damaged neurons, as spinal cord injuries can break the communication between nerves that is involved in controlling blood pressure. This can lead to a vicious cycle of dysregulation of blood pressure and limit the supply of blood and oxygen to the damaged spinal cord tissue, exacerbating the death of spinal neurons. Management of blood pressure is therefore a key target for spinal cord injury care, but so far, the precise thresholds to enable neurons to recover are poorly understood. To find out more, Torres-Espin, Haefeli et al. used machine learning software to analyze previously recorded blood pressure and heart rate data obtained from 118 patients that underwent spinal cord surgery after acute spinal cord injury. The analyses revealed that patients who suffered from either low or high blood pressure during surgery had poorer prospects of recovery. Statistical models confirming these findings showed that the optimal blood pressure range to ensure recovery lies between 76 to 104-117 mmHg. Any deviation from this narrow window would dramatically worsen the ability to recover. These findings suggests that dysregulated blood pressure during surgery affects to odds of recovery in patients with a spinal cord injury. Torres-Espin, Haefeli et al. provide specific information that could improve current clinical practice in trauma centers. In the future, such machine learning tools and models could help develop real-time models that could predict the likelihood of a patient's recovery following spinal cord injury and related neurological conditions.

Reproducible analysis of disease space via principal components using the novel R package syndRomics.

Biomedical data are usually analyzed at the univariate level, focused on a single primary outcome measure to provide insight into systems biology, complex disease states, and precision medicine opportunities. More broadly, these complex biological and disease states can be detected as common factors emerging from the relationships among measured variables using multivariate approaches. 'Syndromics' refers to an analytical framework for measuring disease states using principal component analysis and related multivariate statistics as primary tools for extracting underlying disease patterns. A key part of the syndromic workflow is the interpretation, the visualization, and the study of robustness of the main components that characterize the disease space. We present a new software package, syndRomics, an open-source R package with utility for component visualization, interpretation, and stability for syndromic analysis. We document the implementation of syndRomics and illustrate the use of the package in case studies of neurological trauma data.

Diagnostic blood RNA profiles for human acute spinal cord injury.

Diagnosis of spinal cord injury (SCI) severity at the ultra-acute stage is of great importance for emergency clinical care of patients as well as for potential enrollment into clinical trials. The lack of a diagnostic biomarker for SCI has played a major role in the poor results of clinical trials. We analyzed global gene expression in peripheral white blood cells during the acute injury phase and identified 197 genes whose expression changed after SCI compared with healthy and trauma controls and in direct relation to SCI severity. Unsupervised coexpression network analysis identified several gene modules that predicted injury severity (AIS grades) with an overall accuracy of 72.7% and included signatures of immune cell subtypes. Specifically, for complete SCIs (AIS A), ROC analysis showed impressive specificity and sensitivity (AUC: 0.865). Similar precision was also shown for AIS D SCIs (AUC: 0.938). Our findings indicate that global transcriptomic changes in peripheral blood cells have diagnostic and potentially prognostic value for SCI severity.

Injury volume extracted from MRI predicts neurologic outcome in acute spinal cord injury: A prospective TRACK-SCI pilot study.

Conventional MRI measures of traumatic spinal cord injury severity largely rely on 2-dimensional injury characteristics such as intramedullary lesion length and cord compression. Recent advances in spinal cord (SC) analysis have led to the development of a robust anatomic atlas incorporated into an open-source platform called the Spinal Cord Toolbox (SCT) that allows for quantitative volumetric injury analysis. In the current study, we evaluate the prognostic value of volumetric measures of spinal cord injury on MRI following registration of T2-weighted (T2w) images and segmented lesions from acute SCI patients with a standardized atlas. This IRB-approved prospective cohort study involved the image analysis of 60 blunt cervical SCI patients enrolled in the TRACK-SCI clinical research protocol. Axial T2w MRI data obtained within 24 h of injury were processed using the SCT. Briefly, SC MRIs were automatically segmented using the sct_deepseg_sc tool in the SCT and segmentations were manually corrected by a neuro-radiologist. Lesion volume data were used as predictor variables for correlation with lower extremity motor scores at discharge. Volumetric MRI measures of T2w signal abnormality comprising the SCI lesion accurately predict lower extremity motor scores at time of patient discharge. Similarly, MRI measures of injury volume significantly correlated with motor scores to a greater degree than conventional 2-D metrics of lesion size. The volume of total injury and of injured spinal cord motor regions on T2w MRI is significantly and independently associated with neurologic outcome at discharge after injury.

Transforming Research and Clinical Knowledge in Spinal Cord Injury (TRACK-SCI): an overview of initial enrollment and demographics.

OBJECTIVE: Traumatic spinal cord injury (SCI) is a dreaded condition that can lead to paralysis and severe disability. With few treatment options available for patients who have suffered from SCI, it is important to develop prospective databases to standardize data collection in order to develop new therapeutic approaches and guidelines. Here, the authors present an overview of their multicenter, prospective, observational patient registry, Transforming Research and Clinical Knowledge in SCI (TRACK-SCI). METHODS: Data were collected using the National Institute of Neurological Disorders and Stroke (NINDS) common data elements (CDEs). Highly granular clinical information, in addition to standardized imaging, biospecimen, and follow-up data, were included in the registry. Surgical approaches were determined by the surgeon treating each patient; however, they were carefully documented and compared within and across study sites. Follow-up visits were scheduled for 6 and 12 months after injury. RESULTS: One hundred sixty patients were enrolled in the TRACK-SCI study. In this overview, basic clinical, imaging, neurological severity, and follow-up data on these patients are presented. Overall, 78.8% of the patients were determined to be surgical candidates and underwent spinal decompression and/or stabilization. Follow-up rates to date at 6 and 12 months are 45% and 36.3%, respectively. Overall resources required for clinical research coordination are also discussed. CONCLUSIONS: The authors established the feasibility of SCI CDE implementation in a multicenter, prospective observational study. Through the application of standardized SCI CDEs and expansion of future multicenter collaborations, they hope to advance SCI research and improve treatment.

Effects of inflammation on stem cells: together they strive?

Inflammation entails a complex set of defense mechanisms acting in concert to restore the homeostatic balance in organisms after damage or pathogen invasion. This immune response consists of the activity of various immune cells in a highly complex manner. Inflammation is a double-edged sword as it is reported to have both detrimental and beneficial consequences. In this review, we discuss the effects of inflammation on stem cell activity, focusing primarily on neural stem/progenitor cells in mammals and zebrafish. We also give a brief overview of the effects of inflammation on other stem cell compartments, exemplifying the positive and negative role of inflammation on stemness. The majority of the chronic diseases involve an unremitting phase of inflammation due to improper resolution of the initial pro-inflammatory response that impinges on the stem cell behavior. Thus, understanding the mechanisms of crosstalk between the inflammatory milieu and tissue-resident stem cells is an important basis for clinical efforts. Not only is it important to understand the effect of inflammation on stem cell activity for further defining the etiology of the diseases, but also better mechanistic understanding is essential to design regenerative therapies that aim at micromanipulating the inflammatory milieu to offset the negative effects and maximize the beneficial outcomes.

Neuroinflammation and central nervous system regeneration in vertebrates.

Injuries in the central nervous system (CNS) are one of the leading causes of mortality or persistent disabilities in humans. One of the reasons why humans cannot recover from neuronal loss is the limited regenerative capacity of their CNS. By contrast, non-mammalian vertebrates exhibit widespread regeneration in diverse tissues including the CNS. Understanding those mechanisms activated during regeneration may improve the regenerative outcome in the severed mammalian CNS. Of those mechanisms, recent evidence suggests that inflammation may be important in regeneration. In this review we compare the different events following acute CNS injury in mammals and non-mammalian vertebrates. We also discuss the involvement of the immune response in initiating regenerative programs and how immune cells and neural stem/progenitor cells (NSPCs) communicate.

Acute inflammation initiates the regenerative response in the adult zebrafish brain.

The zebrafish regenerates its brain after injury and hence is a useful model organism to study the mechanisms enabling regenerative neurogenesis, which is poorly manifested in mammals. Yet the signaling mechanisms initiating such a regenerative response in fish are unknown. Using cerebroventricular microinjection of immunogenic particles and immunosuppression assays, we showed that inflammation is required and sufficient for enhancing the proliferation of neural progenitors and subsequent neurogenesis by activating injury-induced molecular programs that can be observed after traumatic brain injury. We also identified cysteinyl leukotriene signaling as an essential component of inflammation in the regenerative process of the adult zebrafish brain. Thus, our results demonstrate that in zebrafish, in contrast to mammals, inflammation is a positive regulator of neuronal regeneration in the central nervous system.

Regenerative neurogenesis from neural progenitor cells requires injury-induced expression of Gata3.

The adult zebrafish brain, unlike mammalian counterparts, can regenerate after injury owing to the neurogenic capacity of stem cells with radial glial character. We hypothesized that injury-induced regenerative programs might be turned on after injury in zebrafish brain and enable regenerative neurogenesis. Here we identify one such gene-the transcription factor gata3-which is expressed only after injury in different zebrafish organs. Gata3 is required for reactive proliferation of radial glia cells, subsequent regenerative neurogenesis, and migration of the newborn neurons. We found that these regeneration-specific roles of Gata3 are dependent on the injury because Gata3 overexpression in the unlesioned adult zebrafish brain is not sufficient to induce neurogenesis. Thus, gata3 acts as a specific injury-induced proregenerative factor that is essential for the regenerative capacity in vertebrates.

The chemokine receptor cxcr5 regulates the regenerative neurogenesis response in the adult zebrafish brain.

BACKGROUND: Unlike mammals, zebrafish exhibits extensive neural regeneration after injury in adult stages of its lifetime due to the neurogenic activity of the radial glial cells. However, the genes involved in the regenerative neurogenesis response of the zebrafish brain are largely unknown. Thus, understanding the underlying principles of this regeneration capacity of the zebrafish brain is an interesting research realm that may offer vast clinical ramifications. RESULTS: In this paper, we characterized the expression pattern of cxcr5 and analyzed the function of this gene during adult neurogenesis and regeneration of the zebrafish telencephalon. We found that cxcr5 was upregulated transiently in the RGCs and neurons, and the expression in the immune cells such as leukocytes was negligible during both adult neurogenesis and regeneration. We observed that the transgenic misexpression of cxcr5 in the ventricular cells using dominant negative and full-length variants of the gene resulted in altered proliferation and neurogenesis response of the RGCs. When we knocked down cxcr5 using antisense morpholinos and cerebroventricular microinjection, we observed outcomes similar to the overexpression of the dominant negative cxcr5 variant. CONCLUSIONS: Thus, based on our results, we propose that cxcr5 imposes a proliferative permissiveness to the radial glial cells and is required for differentiation of the RGCs to neurons, highlighting novel roles of cxcr5 in the nervous system of vertebrates. We therefore suggest that cxcr5 is an important cue for ventricular cell proliferation and regenerative neurogenesis in the adult zebrafish telencephalon. Further studies on the role of cxcr5 in mediating neuronal replenishment have the potential to produce clinical ramifications in efforts for regenerative therapeutic applications for human neurological disorders or acute injuries.