Neurological recovery after traumatic spinal cord injury: what is meaningful? A patients' and physicians' perspective

Efficacy of a machine learning-based approach in predicting neurological prognosis of cervical spinal cord injury patients following urgent surgery within 24 h after injury

We aimed to develop a machine learning (ML) model for predicting the neurological outcomes of cervical spinal cord injury (CSCI). We retrospectively analyzed 135 patients with CSCI who underwent surgery within 24 h after injury. Patients were assessed with the American Spinal Injury Association Impairment Scale (AIS; grades A to E) 6 months after injury. A total of 34 features extracted from demographic variables, surgical factors, laboratory variables, neurological status, and radiological findings were analyzed. The ML model was created using Light GBM, XGBoost, and CatBoost. We evaluated Shapley Additive Explanations (SHAP) values to determine the variables that contributed most to the prediction models. We constructed multiclass prediction models for the five AIS grades and binary classification models to predict more than one-grade improvement in AIS 6 months after injury. Of the ML models used, CatBoost showed the highest accuracy (0.800) for the prediction of AIS grade and the highest AUC (0.90) for predicting improvement in AIS. AIS grade at admission, intramedullary hemorrhage, longitudinal extent of intramedullary T2 hyperintensity, and HbA1c were identified as important features for these prediction models. The ML models successfully predicted neurological outcomes 6 months after injury following urgent surgery in patients with CSCI.

Machine learning in clinical diagnosis, prognostication, and management of acute traumatic spinal cord injury (SCI): A systematic review

Background

Machine learning has been applied to improve diagnosis and prognostication of acute traumatic spinal cord injury. We investigate potential for clinical integration of machine learning in this patient population to navigate variability in injury and recovery.

Materials and methods

We performed a systematic review using PRISMA guidelines through PubMed database to identify studies that use machine learning algorithms for clinical application toward improvements in diagnosis, management, and predictive modeling.

Results

Of the 132 records identified, a total of 13 articles met inclusion criteria and were included in final analysis. Of the 13 articles, 5 focused on diagnostic accuracy and 8 were related to prognostication or management of traumatic spinal cord injury. Across studies, 1983 patients with spinal cord injury were evaluated with most classifying as ASIA C or D. Retrospective designs were used in 10 of 13 studies and 3 were prospective. Studies focused on MRI evaluation and segmentation for diagnostic accuracy and prognostication, investigation of mean arterial pressure in acute care and intraoperative settings, prediction of ambulatory and functional ability, chronic complication prevention, and psychological quality of life assessments. Decision tree, random forests (RF), support vector machines (SVM), hierarchical cluster tree analysis (HCTA), artificial neural networks (ANN), convolutional neural networks (CNN) machine learning subtypes were used.

Conclusions

Machine learning represents a platform technology with clinical application in traumatic spinal cord injury diagnosis, prognostication, management, rehabilitation, and risk prevention of chronic complications and mental illness. SVM models showed improved accuracy when compared to other ML subtypes surveyed. Inherent variability across patients with SCI offers unique opportunity for ML and personalized medicine to drive desired outcomes and assess risks in this patient population.

The impact of surgical timing on motor level lowering in motor complete traumatic spinal cord injury patients

Patients with motor complete traumatic spinal cord injury (tSCI) have a low potential to recover ambulation. Motor level recovery, adjacent to the level of injury, could influence functional independency. This study addresses whether surgical timing influences motor level recovery in patients with sensorimotor complete (AIS A) and motor complete sensory incomplete (AIS B) tSCI. A retrospective cohort study was performed in the Netherlands in patients with motor complete tSCI (C2-L2), who consecutively underwent surgery between January 2010 and April 2020. Neurological examination was performed directly at presentation to the ER and at discharge from the rehabilitation facility. Motor level lowering, AIS grade and Upper and Lower Extremity Motor Score (UEMS and LEMS) recovery were calculated for patients who underwent early (<24h) and late (24h+) surgery. A total of 96 patients met the inclusion criteria. In the multivariate analysis late surgical decompression (24h+) was negatively associated with ≥1 motor level lowering and ≥2 AIS grade improvement (Odds Ratio (OR) 0.11 (95% CI: 0.01, 0.67) p=0.046, OR 0.06 (95% CI: 0.00, 047) p=0.030). The presence of sacral sparing (AIS B) at initial examination, and cervical level of the tSCI were associated with ≥1 motor level lowering. In addition, AOSpine C-type injuries were negatively associated with any type of neurological recovery, except motor level lowering. Although sensorimotor complete injuries as well as thoracolumbar injuries negatively influence neurological recovery, early surgical decompression (<24h) appears independently associated with enhanced neurological recovery in patients with traumatic spinal cord injury despite level and severity of injury.

A NEW SCORE BASED ON THE INTERNATIONAL STANDARDS FOR NEUROLOGICAL CLASSIFICATION OF SPINAL CORD INJURY FOR INTEGRATIVE EVALUATION OF CHANGES IN SENSORIMOTOR FUNCTIONS

Sensorimotor function of patients with spinal cord injury (SCI) is commonly assessed according to the International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI). From the ISNCSCI segmental motor and sensory assessments, upper and lower extremity motor scores (UEMS and LEMS), sum scores of pin prick (PP) and light touch (LT) sensation, the neurological level of injury (NLI) and the classification of lesion severity according to the American Spinal Injury Association Impairment Scale (AIS) grade are derived. Changes of these parameters over time are widely used to evaluate neurological recovery. However, evaluating recovery based on a single ISNCSCI scoring or classification variable may misestimate overall recovery. Here, we propose an Integrated Neurological Change Score (INCS) based on the combination of normalized changes between two-time points of UEMS, LEMS, and total PP and LT scores. To assess the agreement of INCS with clinical judgement of meaningfulness of neurological changes, changes of ISNCSCI variables between two time-points of 88 patients from an independent cohort were rated by 20 clinical experts according to a 5-categories Likert Scale. As for individual ISNCSCI variables, neurological change measured by INCS is associated to severity (AIS grade), age and time since injury, but INCS better reflects clinical judgment about meaningfulness of neurological changes than individual ISNCSCI variables. In addition, INCS is related with changes in functional independence measured by the Spinal Cord Independence Measure (SCIM) in patients with tetraplegia. INCS may be a useful measure of overall neurological change in clinical studies.

Equitable partnerships between scientists and persons living with spinal cord injury will strengthen research scope, quality, and outcomes

PURPOSE OF REVIEW

Partnerships across all stakeholders in the research process strengthen the outcomes and ultimate usability of research. The purpose of this review is to discuss the current level of inclusion of people living with spinal cord injury (SCI) in the research process, the science of engagement and benefits of partnerships in research, and emerging resources available to help promote ethical and effective partnerships in SCI research.

RECENT FINDINGS

Significant strides have been made in interacting with people living with SCI to help identify the problem(s) that are important to study (i.e. the first step in the research process). The SCI research field is lagging in partnering with people living with SCI throughout the rest of the research process despite a plethora of evidence-based principles and strategies for effective partnerships in the broader context of research. There are several emerging resources specific to SCI to help researchers and the community begin to build meaningful partnerships throughout the entire cycle of research.

SUMMARY

The SCI research field already values partnerships with clinicians and promotes the concept of 'bench-to-bedside and back again'. Now is the time to take it a step further to 'bench-to-bedside-to-community and back again'.

Wavelet coherence as a measure of trunk stabilizer muscle activation in wheelchair fencers

BACKGROUND

Intermuscular synchronization constitutes one of the key aspects of effective sport performance and activities of daily living. The aim of the study was to assess the synchronization of trunk stabilizer muscles in wheelchair fencers with the use of wavelet analysis.

METHODS

Intermuscular synchronization and antagonistic EMG-EMG coherence were evaluated in the pairs of the right and the left latissimus dorsi/external oblique abdominal (LD/EOA) muscles. The study group consisted of 16 wheelchair fencers, members of the Polish Paralympic Team, divided into two categories of disability (A and B). Data analysis was carried out in three stages: (1) muscle activation recording using sEMG; (2) wavelet coherence analysis; and (3) coherence density analysis.

RESULTS

In the Paralympic wheelchair fencers, regardless of their disability category, the muscles were activated at low frequency levels: 8-20 Hz for category A fencers, and 5-15 Hz for category B fencers.

CONCLUSIONS

The results demonstrated a clear activity of the trunk muscles in the wheelchair fencers, including those with spinal cord injury, which can be explained as an outcome of their intense training. EMG signal processing application have great potential for performance improvement and diagnosis of wheelchair athletes.