Machine learning to predict sports-related concussion recovery using clinical data

Diffusion tensor analysis of white matter tracts is prognostic of persisting post-concussion symptoms in collegiate athletes

BACKGROUND AND OBJECTIVES

After a concussion diagnosis, the most important issue for patients and loved ones is how long it will take them to recover. The main objective of this study is to develop a prognostic model of concussion recovery. This model would benefit many patients worldwide, allowing for early treatment intervention.

METHODS

The Concussion Assessment, Research and Education (CARE) consortium study enrolled collegiate athletes from 30 sites (NCAA athletic departments and US Department of Defense service academies), 4 of which participated in the Advanced Research Core, which included diffusion-weighted MRI (dMRI) data collection. We analyzed the dMRI data of 51 injuries of concussed athletes scanned within 48 h of injury. All athletes were cleared to return-to-play by the local medical staff following a standardized, graduated protocol. The primary outcome measure is days to clearance of unrestricted return-to-play. Injuries were divided into early (return-to-play < 28 days) and late (return-to-play >= 28 days) recovery based on the return-to-play clinical records. The late recovery group meets the standard definition of Persisting Post-Concussion Symptoms (PPCS). Data were processed using automated, state-of-the-art, rigorous methods for reproducible data processing using brainlife.io. All processed data derivatives are made available at https://brainlife.io/project/63b2ecb0daffe2c2407ee3c5/dataset. The microstructural properties of 47 major white matter tracts, 5 callosal, 15 subcortical, and 148 cortical structures were mapped. Fractional Anisotropy (FA) and Mean Diffusivity (MD) were estimated for each tract and structure. Correlation analysis and Receiver Operator Characteristic (ROC) analysis were then performed to assess the association between the microstructural properties and return-to-play. Finally, a Logistic Regression binary classifier (LR-BC) was used to classify the injuries between the two recovery groups.

RESULTS

The mean FA across all white matter volume was negatively correlated with return-to-play (r = -0.38, p = 0.00001). No significant association between mean MD and return-to-play was found, neither for FA nor MD for any other structure. The mean FA of 47 white matter tracts was negatively correlated with return-to-play (rμ = -0.27; rσ = 0.08; rmin = -0.1; rmax = -0.43). Across all tracts, a large mean ROC Area Under the Curve (AUCFA) of 0.71 ± 0.09 SD was found. The top classification performance of the LR-BC was AUC = 0.90 obtained using the 16 statistically significant white matter tracts.

DISCUSSION

Utilizing a free, open-source, and automated cloud-based neuroimaging pipeline and app (https://brainlife.io/docs/tutorial/using-clairvoy/), a prognostic model has been developed, which predicts athletes at risk for slow recovery (PPCS) with an AUC=0.90, balanced accuracy = 0.89, sensitivity = 1.0, and specificity = 0.79. The small number of participants in this study (51 injuries) is a significant limitation and supports the need for future large concussion dMRI studies and focused on recovery.

The Future of Artificial Intelligence in Sports Medicine and Return to Play

Artificial intelligence (AI) has shown tremendous growth over the last decade, with the more recent development of clinical applications in health care. The ability of AI to synthesize large amounts of complex data automatically allows health care providers to access previously unavailable metrics and thus enhance and personalize patient care. These innovations include AI-assisted diagnostic tools, prediction models for each treatment pathway, and various tools for workflow optimization. The extension of AI into sports medicine is still early, but numerous AI-driven algorithms, devices, and research initiatives have delved into predicting and preventing athlete injury, aiding in injury assessment, optimizing recovery plans, monitoring rehabilitation progress, and predicting return to play.

Development of a Machine-Learning-Based Tool for Overnight Orthokeratology Lens Fitting

Purpose

Orthokeratology (ortho-K) is widely used to control myopia. Overnight ortho-K lens fitting with the selection of appropriate parameters is an important technique for achieving successful reductions in myopic refractive error. In this study, we developed a machine-learning model that could select ortho-K lens parameters at an expert level.

Methods

Machine-learning models were established to predict the optimal ortho-K parameters, including toric lens option (toric or non-toric), overall diameter (OAD; 10.5 or 11.0 mm), base curve (BC), return zone depth (RZD), landing zone angle (LZA), and lens sagittal depth (LensSag). The analysis included 547 eyes of 297 Korean adolescents with myopia or astigmatism. The dataset was randomly divided into training (80%, n = 437 eyes) and validation (20%, n = 110 eyes) sets at the patient level. The model was trained based on clinical ortho-K lens fitting performed by highly experienced experts and ophthalmic measurements.

Results

The final machine-learning models showed accuracies of 92.7% and 86.4% for predicting the toric lens option and OAD, respectively. The mean absolute errors for the BC, RZD, LZA, and LensSag predictions were 0.052 mm, 2.727 µm, 0.118°, and 5.215 µm, respectively. The machine-learning model outperformed the manufacturer's conventional initial lens selector in predicting BC and RZD.

Conclusions

We developed an expert-level machine-learning-based model for determining comprehensive ortho-K lens parameters. We also created a web-based application.

Translational Relevance

This model may provide more accurate fitting parameters for lenses than those of conventional calculations, thus reducing the need to rely on trial and error.

Efficient clinical data analysis for prediction of coal workers' pneumoconiosis using machine learning algorithms

PURPOSE

The purpose of this study is to propose an efficient coal workers' pneumoconiosis (CWP) clinical prediction system and put it into clinical use for clinical diagnosis of pneumoconiosis.

METHODS

Patients with CWP and dust-exposed workers who were enrolled from August 2021 to December 2021 were included in this study. Firstly, we chose the embedded method through using three feature selection approaches to perform the prediction analysis. Then, we performed the machine learning algorithms as the model backbone and combined them with three feature selection methods, respectively, to determine the optimal predictive model for CWP.

RESULTS

Through applying three feature selection approaches based on machine learning algorithms, it was found that AaDO₂ and some pulmonary function indicators played an important role in prediction for identifying CWP of early stage. The support vector machine (SVM) algorithm was proved as the optimal machine learning model for predicting CWP, with the ROC curves obtained from three feature selection methods using SVM algorithm whose AUC values of 97.78%, 93.7%, and 95.56%, respectively.

CONCLUSION

We developed the optimal model (SVM algorithm) through comparisons and analyses among the performances of different models for the prediction of CWP as a clinical application.

Vestibular/ocular motor screening (VOMS) score for identification of concussion in cases of non-severe head injury: A systematic review

Background

and importance Each year, 1.4 million patients attend a UK ED with a head injury. Mild traumatic brain injury affects up to 300/100 000 admitted patients/year and a greater number of non-admitted patients. Identifying those patients with a head injury that have concussion, and of those, which will have a prolonged recovery, is critical for discharge planning. The Vestibular/Ocular Motor Screening test (VOMS) has been reported as a useful “sideline tool” to evaluate for sports-related concussion (SRC). VOMS has been assessed for utility primarily for predicting in head-injured, which cases will have concussion, and secondarily in predicting in concussed patients, which will have prolonged recovery. Originally described in 2014, VOMS has not been subject to systematic review or meta-analysis, with regard to its predictive performance for concussion.

Objective

To assess the state of VOMS evidence for dichotomously classifying concussion status in patients with non-severe head injury

Design

Systematic review.

Setting and participants

Studies comprising the review enrolled ambulatory head-injured adults and children, usually from sports-related settings, in Europe or the USA.

Exposure

VOMS.

Outcome measures

Presence of concussion, presence of prolonged recovery in concussed patients

Main results

The review identified 17 studies, characterized by a wide variety of specific approaches to administering and scoring VOMS. While VOMS showed promise as a screening tool for concussion, marked study heterogeneity precluded generation of a pooled effect estimate for VOMS performance.

Conclusion

VOMS is potentially useful as a concussion screening tool. Available evidence from the SRC arena suggests sensitivity ranging from 58–96%, with specificity 46−92%. Directions for future VOMS research should include evaluation of standardized administration and scoring, potentially of a simpler VOMS (with fewer components), in a general head-injured population. Further analysis of precisely defined VOMS application may be useful to determine the proper place of VOMS screening for the head-injured.

A machine learning approach for predicting suicidal ideation in post stroke patients

Currently, the identification of stroke patients with an increased suicide risk is mainly based on self‐report questionnaires, and this method suffers from a lack of objectivity. This study developed and validated a suicide ideation (SI) prediction model using clinical data and identified SI predictors. Significant variables were selected through traditional statistical analysis based on retrospective data of 385 stroke patients; the data were collected from October 2012 to March 2014. The data were then applied to three boosting models (Xgboost, CatBoost, and LGBM) to identify the comparative and best performing models. Demographic variables that showed significant differences between the two groups were age, onset, type, socioeconomic, and education level. Additionally, functional variables also showed a significant difference with regard to ADL and emotion (p < 0.05). The CatBoost model (0.900) showed higher performance than the other two models; and depression, anxiety, self-efficacy, and rehabilitation motivation were found to have high importance. Negative emotions such as depression and anxiety showed a positive relationship with SI and rehabilitation motivation and self-efficacy displayed an inverse relationship with SI. Machine learning-based SI models could augment SI prevention by helping rehabilitation and medical professionals identify high-risk stroke patients in need of SI prevention intervention.

Comparison of Prescribed Physical Therapy to a Home Exercise Program for Pediatric Sports-Related Concussion Patients

The purpose of this retrospective chart review was to compare sports-related concussion (SRC) recovery time in protracted recovery (≥28 days) patients who were prescribed physical therapy (PPT) with those who were only provided a home exercise program (HEP). We hypothesized PPT would be associated with shorter recovery times relative to HEP. Associations were evaluated with multivariable zero-truncated negative binomial regressions. Among the 48 (30.2%) PPT and 111 (69.8%) HEP patients, the majority were female (57.9%), the mean age was 15.3 ± 1.4 (PPT) and 14.2 ± 2.8 (HEP), and time to clinic was a median 6.0 (IQR = 3.0–27.0; PPT) and 7.0 (IQR = 3.0–23.0; HEP) days. After adjusting for demographic (age, sex) and clinical measures (concussion history, convergence, VOMS, PCSS score, and days to clinic), PPT unexpectedly was associated with 1.21 (95% CI: 1.05, 1.41) additional recovery days compared with HEP. One reason for this could be related to patients adhering to the number of a priori prescribed PT sessions which may or may not have aligned with the patient’s symptom resolution. Future research should explore this hypothesis while aiming to evaluate the effect of PPT versus HEP using a randomized design. If confirmed, these findings are encouraging for patients who could not otherwise access or afford specialty rehabilitation.

Health Information Prediction System of Infant Sports Based on Deep Learning Network

The sensed data from infant sports and training programs are useful in analyzing their health conditions and forecasting any disorders or abnormalities. The sensed information is processed for providing errorless predictions for infant diseases/disorders, coupled with artificial intelligence and sophisticated healthcare technologies. The problem of noncongruent sensed data impacting the forecast occurs due to errors between consecutive training iterations. This problem is addressed using the deep learning (PEST-DL) proposed perceptible error segregation technique. The training process is halted between two consecutive iterations generating errors until a similarity verification based on infant history is performed. The similarity output determines the errors due to mismatching data observations, and therefore, the data augmentation is performed. The first perceptible error is mitigated by training the learning paradigm with all possible infant history data in the learning process. This prevents prediction lag and data omissions due to discrete availability. The learning is trained from the identified error with the precise detected disorder/abnormality data previously detected. Therefore, the first and consecutive training data segregate error instances from the actual training iterations. This improves the prediction accuracy and precision with controlled error and time complexity.

Towards defining biomarkers to evaluate concussions using virtual reality and a moving platform (BioVRSea)

Current diagnosis of concussion relies on self-reported symptoms and medical records rather than objective biomarkers. This work uses a novel measurement setup called BioVRSea to quantify concussion status. The paradigm is based on brain and muscle signals (EEG, EMG), heart rate and center of pressure (CoP) measurements during a postural control task triggered by a moving platform and a virtual reality environment. Measurements were performed on 54 professional athletes who self-reported their history of concussion or non-concussion. Both groups completed a concussion symptom scale (SCAT5) before the measurement. We analyzed biosignals and CoP parameters before and after the platform movements, to compare the net response of individual postural control. The results showed that BioVRSea discriminated between the concussion and non-concussion groups. Particularly, EEG power spectral density in delta and theta bands showed significant changes in the concussion group and right soleus median frequency from the EMG signal differentiated concussed individuals with balance problems from the other groups. Anterior-posterior CoP frequency-based parameters discriminated concussed individuals with balance problems. Finally, we used machine learning to classify concussion and non-concussion, demonstrating that combining SCAT5 and BioVRSea parameters gives an accuracy up to 95.5%. This study is a step towards quantitative assessment of concussion.