Using Deep-Learning-Based Artificial Intelligence Technique to Automatically Evaluate the Collateral Status of Multiphase CTA in Acute Ischemic Stroke.

BACKGROUND: Collateral status is an important predictor for the outcome of acute ischemic stroke with large vessel occlusion. Multiphase computed-tomography angiography (mCTA) is useful to evaluate the collateral status, but visual evaluation of this examination is time-consuming. This study aims to use an artificial intelligence (AI) technique to develop an automatic AI prediction model for the collateral status of mCTA. METHODS: This retrospective study enrolled subjects with acute ischemic stroke receiving endovascular thrombectomy between January 2015 and June 2020 in a tertiary referral hospital. The demographic data and images of mCTA were collected. The collateral status of all mCTA was visually evaluated. Images at the basal ganglion and supraganglion levels of mCTA were selected to produce AI models using the convolutional neural network (CNN) technique to automatically predict the collateral status of mCTA. RESULTS: A total of 82 subjects were enrolled. There were 57 cases randomly selected for the training group and 25 cases for the validation group. In the training group, there were 40 cases with a positive collateral result (good or intermediate) and 17 cases with a negative collateral result (poor). In the validation group, there were 21 cases with a positive collateral result and 4 cases with a negative collateral result. During training for the CNN prediction model, the accuracy of the training group could reach 0.999 ± 0.015, whereas the prediction model had a performance of 0.746 ± 0.008 accuracy on the validation group. The area under the ROC curve was 0.7. CONCLUSIONS: This study suggests that the application of the AI model derived from mCTA images to automatically evaluate the collateral status is feasible.

Prediction of Cognitive Degeneration in Parkinson's Disease Patients Using a Machine Learning Method.

This study developed a predictive model for cognitive degeneration in patients with Parkinson's disease (PD) using a machine learning method. The clinical data, plasma biomarkers, and neuropsychological test results of patients with PD were collected and utilized as model predictors. Machine learning methods comprising support vector machines (SVMs) and principal component analysis (PCA) were applied to obtain a cognitive classification model. Using 32 comprehensive predictive parameters, the PCA-SVM classifier reached 92.3% accuracy and 0.929 area under the receiver operating characteristic curve (AUC). Furthermore, the accuracy could be increased to 100% and the AUC to 1.0 in a PCA-SVM model using only 13 carefully chosen features.

Plasma total tau predicts executive dysfunction in Parkinson's disease.

OBJECTIVES: Cognitive impairment is an important non-motor aspect of Parkinson's disease (PD). Amyloid-ß and tau pathologies are well-established in Alzheimer's disease and commonly coexist with synucleinopathy in PD. However, the levels of these biomarkers in the plasma of patients with PD and their relationship with specific cognition domains remain to be clarified. The current study compared the motor severity and neuropsychological assessment of general and specific cognition, with plasma levels of α-synuclein (α-syn), amyloid-ß 42 (Aß42), and total tau (t-tau) in PD subjects. METHODS: Plasma levels of α-syn, Aß42, and t-tau were measured in 55 participants with PD through immunomagnetic reduction assay. The evaluation of motor severity and comprehensive neuropsychological assessment was performed in all participants. RESULTS: The level of plasma α-syn was negatively correlated with the scores of Unified Parkinson's Disease Rating Scale part III [r = (-.352), p = .008]. The level of plasma t-tau was negatively correlated with the scores of digits recall forwards and digits recall backwards [r = (-.446), p = .001; r = (-.417), p = .002, respectively]. No correlations were found between the levels of α-syn and Aß42 and any neuropsychological tests. CONCLUSIONS: This study concluded a lower level of plasma α-syn was correlated with motor dysfunction in PD patients, and a higher level of plasma t-tau was correlated with lower cognitive performance, especially for attention and executive function. These results propose the possibility of using plasma biomarkers to predict specific cognitive performance in PD subjects.

Do cognitive performance and physical function differ between individuals with motoric cognitive risk syndrome and those with mild cognitive impairment?

BACKGROUND: Motoric cognitive risk syndrome (MCR) is defined by slow gait speed combined with subjective cognitive complaint. MCR is a predementia syndrome, similar to mild cognitive impairment (MCI). However, there is currently no study comparing the differences in cognitive performance and physical function between these two types of cognitive impairment. Thus, the aim of this study is to compare cognitive performance and physical function in individuals with MCR versus MCI. METHODS: A total of 77 participants, free of dementia, were recruited from the neurological outpatient clinic of a medical center in Taiwan. Participants were separated into 2 groups, MCR (n = 33) and MCI (n = 44) groups, based on definition criteria from previous studies. The priority was to assign a diagnosis of MCR first, followed by MCI. Hence, "pure" MCI had no overlap with MCR syndrome. Cognitive performance, including executive function, attention, working memory, episode memory, visuospatial function, and language, were measured. Physical functions such as activities in daily living, the Tinetti Assessment Scale, and the Timed Up and Go test were also measured. RESULTS: Executive function, attention, working memory, episodic memory and language were all significantly lower in the MCR group than the MCI group. Abilities related to physical function, including those measured by the Tinetti Assessment Scale and the Timed Up and Go test, were significantly lower in the MCR group than the MCI group. CONCLUSIONS: We noted that cognitive performance and physical function were lower in MCR individuals than MCI but without MCR syndrome. However, the conclusions were based on the enrollment procedure of participants prioritizes the MCR syndrome. Because of the overlap of MCR and MCI, future studies should use different enrollment strategies to further clarify the status of these two populations.

Predicting Cognitive Decline in Parkinson's Disease with Mild Cognitive Impairment: A One-Year Observational Study.

We conducted an observational study to investigate clinical predictors of cognitive decline in patients with mild cognitive impairment (MCI), with a focus on patients with Parkinson's disease (PD) and Alzheimer's disease (AD). The study was performed with detailed neuropsychological testing, a portable device for gait analysis, and a comprehensive geriatric assessment for patients with MCI. Cognitive decline was defined as subjective cognitive impairment with an objective decline in the Mini-Mental State Examination (MMSE) ≥2 points at the one-year follow-up. Participants (n = 74) had a median age of 70 (interquartile range 60-79) years, and 45.9% of them were women. At the end of the study, 17.6% of the patients with MCI had a cognitive decline. Although no differences were observed between groups at the baseline cognitive study, patients with PD-MCI demonstrated more cognitive decline than patients with AD-MCI (28.6% vs. 7.7% p = 0.03). Patients with PD-MCI had more physical disabilities, including scores of instrumental activities of daily living (IADL), Tinetti balance, and gait scores, and some Timed Up and Go components. Initial Clinical Dementia Rating-Sum of Boxes score was a better predictor of future cognitive decline than MMSE in PD-MCI. For predicting the occurrence of cognitive decline in PD-MCI, the prediction accuracy increased from the reduced model (AUC = 0.822, p < 0.001) to the full model (a total of five independent variables, AUC = 0.974, p < 0.001). Given the potentially modifiable predictor, our findings also highlight the importance of identifying sleep quality and the ability to perform IADL.

Gait-Based Machine Learning for Classifying Patients with Different Types of Mild Cognitive Impairment.

Mild cognitive impairment (MCI) may be caused by Alzheimer's disease, Parkinson's disease (PD), cerebrovascular accident, nutritional or metabolic disorders, or mental disorders. It is important to determine the cause and treatment of dementia as early as possible because dementia may appear in remission. Decline in MCI cognitive function may affect a patient's walking performance. Therefore, all participants in this study participated in an experiment using a portable gait analysis system to perform walk, time up and go, and jump tests. The collected gait parameters are used in a machine learning classification model based on a support vector machine (SVM) and principal component analysis (PCA). The aim of the study is to predict different types of MCI patients based on gait information. It is shown that the machine learning classification model can predict different types of MCI patients. Specifically, the PCA-SVM model demonstrated better classification performance with 91.67% accuracy and 0.9714 area under the receiver operating characteristic curve (ROC AUC) using the polynomial kernel function in classifying PD-MCI and non-PD-MCI patients.

The effect of adjusting the stroke value of jet dispenser on the performance of glucose biosensor with gold-plated electrode.

This study investigates the effect of dispensing glucose oxidase enzyme onto the reaction zone of a golden electrode via a jetting dispenser. Each droplet of enzyme solution dispensed onto the reaction zone of golden electrode weighs exactly the same at around 0.4 mg. This study shows that the spring and needle assembly can be controlled by adjusting the stroke value of the stroke adjustment knob to generate different jet dispensing effects, thus affecting the change of droplets of glucose oxidase enzyme solution within the reaction zone of the golden electrode. This study performs experiments using three stroke values, which are 0.8, 1.2, and 1.6 mm. The experimental results show that adjusting the stroke value to change the droplet of the dispensing liquid will significantly affect the accuracy of the test strip reading value. When the stroke value is adjusted to 1.5 mm, the standard deviation of the test strip is 3.8 mg/dL and the coefficient of variation is 4.1%-6.1%. The study suggested that adjusting the stroke value can stabilize the dispensed droplet to increase the stability of test strip reading, improving the accuracy of the test strip reading. This adjustment method can also be applied to the process of other biochemical sensors.

Validation of a computer-based bronchoscopy simulator developed in Taiwan.

BACKGROUND/PURPOSE: Conventional training in bronchoscopy may increase patient's discomfort and procedure-related morbidity. Computer-based bronchoscopy simulator (CBBS) permits the acquisition and evaluation of the necessary skills through a realistic bronchoscopic experience. This study was conducted to validate the use of a CBBS system developed in Taiwan as a learning and assessment tool. METHODS: Twenty novice bronchoscopists and 10 expert bronchoscopists were enrolled as subjects in this prospective study. The 20 novice bronchoscopists were randomized into two groups, which received conventional bronchoscopic training or CBBS training and then completed a satisfaction survey. Subsequently, the novices who received CBBS training underwent an observational performance trial and the results were compared with those of expert bronchoscopists. All 10 expert bronchoscopists completed a realism survey and observational trial after CBBS performance. RESULTS: The satisfaction survey showed that the CBBS training program significantly increased participants' satisfaction (p = 0.002) and interest in learning (p < 0.001). The realism survey by the 10 expert bronchoscopists indicated that CBBS provides a favorable degree of realism with regard to the mechanical and visual parameters examined. Analysis of the performance results showed that the following parameters were capable of differentiating the participants by level of expertise: total procedure time (p = 0.002), percentage of bronchial segments entered (p = 0.012), percentage of bronchial segments identified (p < 0.001), percentage of repeated bronchial segments entered (p = 0.004), percentage of pathologies identified (p < 0.001), number of times that the bronchoscope tip collided with airway walls (p = 0.013), and number of times oral instruction was needed (p = 0.01). CONCLUSION: CBBS is a valid training method that increases interest in learning and provides a favorable degree of virtual realism. It can also distinguish various levels of competence at actual bronchoscopy and may have a useful role in the bronchoscopic training curriculum.