Addressing Binary Classification over Class Imbalanced Clinical Datasets Using Computationally Intelligent Techniques

Improving clinical abbreviation sense disambiguation using attention-based Bi-LSTM and hybrid balancing techniques in imbalanced datasets

RATIONALE

Clinical abbreviations pose a challenge for clinical decision support systems due to their ambiguity. Additionally, clinical datasets often suffer from class imbalance, hindering the classification of such data. This imbalance leads to classifiers with low accuracy and high error rates. Traditional feature-engineered models struggle with this task, and class imbalance is a known factor that reduces the performance of neural network techniques.

AIMS AND OBJECTIVES

This study proposes an attention-based bidirectional long short-term memory (Bi-LSTM) model to improve clinical abbreviation disambiguation in clinical documents. We aim to address the challenges of limited training data and class imbalance by employing data generation techniques like reverse substitution and data augmentation with synonym substitution.

METHOD

We utilise a Bi-LSTM classification model with an attention mechanism to disambiguate each abbreviation. The model's performance is evaluated based on accuracy for each abbreviation. To address the limitations of imbalanced data, we employ data generation techniques to create a more balanced dataset.

RESULTS

The evaluation results demonstrate that our data balancing technique significantly improves the model's accuracy by 2.08%. Furthermore, the proposed attention-based Bi-LSTM model achieves an accuracy of 96.09% on the UMN dataset, outperforming state-of-the-art results.

CONCLUSION

Deep neural network methods, particularly Bi-LSTM, offer promising alternatives to traditional feature-engineered models for clinical abbreviation disambiguation. By employing data generation techniques, we can address the challenges posed by limited-resource and imbalanced clinical datasets. This approach leads to a significant improvement in model accuracy for clinical abbreviation disambiguation tasks.

Boosting efficiency in a clinical literature surveillance system with LightGBM

Given the suboptimal performance of Boolean searching to identify methodologically sound and clinically relevant studies in large bibliographic databases, exploring machine learning (ML) to efficiently classify studies is warranted. To boost the efficiency of a literature surveillance program, we used a large internationally recognized dataset of articles tagged for methodological rigor and applied an automated ML approach to train and test binary classification models to predict the probability of clinical research articles being of high methodologic quality. We trained over 12,000 models on a dataset of titles and abstracts of 97,805 articles indexed in PubMed from 2012-2018 which were manually appraised for rigor by highly trained research associates and rated for clinical relevancy by practicing clinicians. As the dataset is unbalanced, with more articles that do not meet the criteria for rigor, we used the unbalanced dataset and over- and under-sampled datasets. Models that maintained sensitivity for high rigor at 99% and maximized specificity were selected and tested in a retrospective set of 30,424 articles from 2020 and validated prospectively in a blinded study of 5253 articles. The final selected algorithm, combining a LightGBM (gradient boosting machine) model trained in each dataset, maintained high sensitivity and achieved 57% specificity in the retrospective validation test and 53% in the prospective study. The number of articles needed to read to find one that met appraisal criteria was 3.68 (95% CI 3.52 to 3.85) in the prospective study, compared with 4.63 (95% CI 4.50 to 4.77) when relying only on Boolean searching. Gradient-boosting ML models reduced the work required to classify high quality clinical research studies by 45%, improving the efficiency of literature surveillance and subsequent dissemination to clinicians and other evidence users.

Binary classification with fuzzy logistic regression under class imbalance and complete separation in clinical studies

BACKGROUND

In binary classification for clinical studies, an imbalanced distribution of cases to classes and an extreme association level between the binary dependent variable and a subset of independent variables can create significant classification problems. These crucial issues, namely class imbalance and complete separation, lead to classification inaccuracy and biased results in clinical studies.

METHOD

To deal with class imbalance and complete separation problems, we propose using a fuzzy logistic regression framework for binary classification. Fuzzy logistic regression incorporates combinations of triangular fuzzy numbers for the coefficients, inputs, and outputs and produces crisp classification results. The fuzzy logistic regression framework shows strong classification performance due to fuzzy logic's better handling of imbalance and separation issues. Hence, classification accuracy is improved, mitigating the risk of misclassified conditions and biased insights for clinical study patients.

RESULTS

The performance of the fuzzy logistic regression model is assessed on twelve binary classification problems with clinical datasets. The model has consistently high sensitivity, specificity, F1, precision, and Mathew's correlation coefficient scores across all clinical datasets. There is no evidence of impact from the imbalance or separation that exists in the datasets. Furthermore, we compare the fuzzy logistic regression classification performance against two versions of classical logistic regression and six different benchmark sources in the literature. These six sources provide a total of ten different proposed methodologies, and the comparison occurs by calculating the same set of classification performance scores for each method. Either imbalance or separation impacts seven out of ten methodologies. The remaining three produce better classification performance in their respective clinical studies. However, these are all outperformed by the fuzzy logistic regression framework.

CONCLUSION

Fuzzy logistic regression showcases strong performance against imbalance and separation, providing accurate predictions and, hence, informative insights for classifying patients in clinical studies.

A Study on the Robustness and Stability of Explainable Deep Learning in an Imbalanced Setting: The Exploration of the Conformational Space of G Protein-Coupled Receptors

G-protein coupled receptors (GPCRs) are transmembrane proteins that transmit signals from the extracellular environment to the inside of the cells. Their ability to adopt various conformational states, which influence their function, makes them crucial in pharmacoproteomic studies. While many drugs target specific GPCR states to exert their effects-thereby regulating the protein's activity-unraveling the activation pathway remains challenging due to the multitude of intermediate transformations occurring throughout this process, and intrinsically influencing the dynamics of the receptors. In this context, computational modeling, particularly molecular dynamics (MD) simulations, may offer valuable insights into the dynamics and energetics of GPCR transformations, especially when combined with machine learning (ML) methods and techniques for achieving model interpretability for knowledge generation. The current study builds upon previous work in which the layer relevance propagation (LRP) technique was employed to interpret the predictions in a multi-class classification problem concerning the conformational states of the β2-adrenergic (β2AR) receptor from MD simulations. Here, we address the challenges posed by class imbalance and extend previous analyses by evaluating the robustness and stability of deep learning (DL)-based predictions under different imbalance mitigation techniques. By meticulously evaluating explainability and imbalance strategies, we aim to produce reliable and robust insights.

Involuntary psychiatric hospitalisation - differences and similarities between patients detained under the mental health act and according to the legal guardianship legislation

BACKGROUND

Involuntary psychiatric hospitalisation occurs under different legal premises. According to German law, detention under the Mental Health Act (MHA) is possible in cases of imminent danger of self-harm or harm to others, while detention according to the legal guardianship legislation (LGL) serves to prevent self-harm if there is considerable but not necessarily imminent danger. This study aims to compare clinical, sociodemographic and environmental socioeconomic differences and similarities between patients hospitalised under either the MHA or LGL.

METHODS

We conducted a retrospective health records analysis of all involuntarily hospitalised cases in the four psychiatric hospitals of the city of Cologne, Germany, in 2011. Of the 1,773 cases, 87.3% were detained under the MHA of the federal state of North Rhine-Westphalia and 6.4% were hospitalised according to the federal LGL. Another 6.3% of the cases were originally admitted under the MHA, but the legal basis of detention was converted to LGL during the inpatient psychiatric stay (MHA→LGL cases). We compared sociodemographic, clinical, systemic and environmental socioeconomic (ESED) variables of the three groups by means of descriptive statistics. We also trained and tested a machine learning-based algorithm to predict class membership of the involuntary modes of psychiatric inpatient care.

RESULTS

Cases with an admission under the premises of LGL lived less often on their own, and they were more often retired compared to MHA cases. They more often had received previous outpatient or inpatient treatment than MHA cases, they were more often diagnosed with a psychotic disorder and they lived in neighbourhoods that were on average more socially advantaged. MHA→LGL cases were on average older and more often retired than MHA cases. More often, they had a main diagnosis of an organic mental disorder compared to both MHA and LGL cases. Also, they less often received previous psychiatric inpatient treatment compared to LGL cases. The reason for detention (self-harm or harm to others) did not differ between the three groups. The proportion of LGL and MHA cases differed between the four hospitals. Effect sizes were mostly small and the balanced accuracy of the Random Forest was low.

CONCLUSION

We found some plausible differences in patient characteristics depending on the legal foundation of the involuntary psychiatric hospitalisation. The differences relate to clinical, sociodemographic and socioeconomical issues. However, the low effect sizes and the limited accuracy of the machine learning models indicate that the investigated variables do not sufficiently explain the respective choice of the legal framework. In addition, we found some indication for possibly different interpretation and handling of the premises of the law in practice. Our findings pose the need for further research in this field.

Novel Multiplexed Plasma Biomarker Panel Has Diagnostic and Prognostic Potential in Children With Hypertrophic Cardiomyopathy

BACKGROUND

Hypertrophic cardiomyopathy (HCM) is defined clinically by pathological left ventricular hypertrophy. We have previously developed a plasma proteomics biomarker panel that correlates with clinical markers of disease severity and sudden cardiac death risk in adult patients with HCM. The aim of this study was to investigate the utility of adult biomarkers and perform new discoveries in proteomics for childhood-onset HCM.

METHODS

Fifty-nine protein biomarkers were identified from an exploratory plasma proteomics screen in children with HCM and augmented into our existing multiplexed targeted liquid chromatography-tandem/mass spectrometry-based assay. The association of these biomarkers with clinical phenotypes and outcomes was prospectively tested in plasma collected from 148 children with HCM and 50 healthy controls. Machine learning techniques were used to develop novel pediatric plasma proteomic biomarker panels.

RESULTS

Four previously identified adult HCM markers (aldolase fructose-bisphosphate A, complement C3a, talin-1, and thrombospondin 1) and 3 new markers (glycogen phosphorylase B, lipoprotein a and profilin 1) were elevated in pediatric HCM. Using supervised machine learning applied to training (n=137) and validation cohorts (n=61), this 7-biomarker panel differentiated HCM from healthy controls with an area under the curve of 1.0 in the training data set (sensitivity 100% [95% CI, 95-100]; specificity 100% [95% CI, 96-100]) and 0.82 in the validation data set (sensitivity 75% [95% CI, 59-86]; specificity 88% [95% CI, 75-94]). Reduced circulating levels of 4 other peptides (apolipoprotein L1, complement 5b, immunoglobulin heavy constant epsilon, and serum amyloid A4) found in children with high sudden cardiac death risk provided complete separation from the low and intermediate risk groups and predicted mortality and adverse arrhythmic outcomes (hazard ratio, 2.04 [95% CI, 1.0-4.2]; P=0.044).

CONCLUSIONS

In children, a 7-biomarker proteomics panel can distinguish HCM from controls with high sensitivity and specificity, and another 4-biomarker panel identifies those at high risk of adverse arrhythmic outcomes, including sudden cardiac death.

Development of a Real-time Force-based Algorithm for Infusion Failure Detection

BACKGROUND

Continuous subcutaneous insulin infusion (CSII) is a common treatment option for people with diabetes (PWD), but insulin infusion failures pose a significant challenge, leading to hyperglycemia, diabetes burnout, and increased hospitalizations. Current CSII pumps' occlusion alarm systems are limited in detecting infusion failures; therefore, a more effective detection method is needed.

METHODS

We conducted five preclinical animal studies to collect data on infusion failures, utilizing both insulin and non-insulin boluses. Data were captured using in-line pressure and flow rate sensors, with additional force data from CSII pumps' onboard sensors in one study. A novel classifier model was developed using this dataset, aimed at detecting different types of infusion failures through direct utilization of force sensor data. Performance was compared against various occlusion alarm thresholds from commercially available CSII pumps.

RESULTS

The testing dataset included 251 boluses. The Bagging classifier model showed the highest performance metrics among the models tested, exhibiting high accuracy (96%), sensitivity (94%), and specificity (98%), with lower false-positive and false-negative rate compared with traditional occlusion alarm pressure thresholds.

CONCLUSIONS

Our study developed a novel non-threshold classifier that outperforms current occlusion alarm systems in CSII pumps in detecting infusion failures. This advancement has the potential to reduce the risk of hyperglycemia and hospitalizations due to undetected infusion failures, offering a more reliable and effective CSII therapy for PWD. Further studies involving human participants are recommended to validate these findings and assess the classifier's performance in a real-world setting.

Construction and validation of a progression prediction model for locally advanced rectal cancer patients received neoadjuvant chemoradiotherapy followed by total mesorectal excision based on machine learning

Background

We attempted to develop a progression prediction model for local advanced rectal cancer(LARC) patients who received preoperative neoadjuvant chemoradiotherapy(NCRT) and operative treatment to identify high-risk patients in advance.

Methods

Data from 272 LARC patients who received NCRT and total mesorectal excision(TME) from 2011 to 2018 at the Fourth Hospital of Hebei Medical University were collected. Data from 161 patients with rectal cancer (each sample with one target variable (progression) and 145 characteristic variables) were included. One Hot Encoding was applied to numerically represent some characteristics. The K-Nearest Neighbor (KNN) filling method was used to determine the missing values, and SmoteTomek comprehensive sampling was used to solve the data imbalance. Eventually, data from 135 patients with 45 characteristic clinical variables were obtained. Random forest, decision tree, support vector machine (SVM), and XGBoost were used to predict whether patients with rectal cancer will exhibit progression. LASSO regression was used to further filter the variables and narrow down the list of variables using a Venn diagram. Eventually, the prediction model was constructed by multivariate logistic regression, and the performance of the model was confirmed in the validation set.

Results

Eventually, data from 135 patients including 45 clinical characteristic variables were included in the study. Data were randomly divided in an 8:2 ratio into a data set and a validation set, respectively. Area Under Curve (AUC) values of 0.72 for the decision tree, 0.97 for the random forest, 0.89 for SVM, and 0.94 for XGBoost were obtained from the data set. Similar results were obtained from the validation set. Twenty-three variables were obtained from LASSO regression, and eight variables were obtained by considering the intersection of the variables obtained using the previous four machine learning methods. Furthermore, a multivariate logistic regression model was constructed using the data set; the ROC indicated its good performance. The ROC curve also verified the good predictive performance in the validation set.

Conclusions

We constructed a logistic regression model with good predictive performance, which allowed us to accurately predict whether patients who received NCRT and TME will exhibit disease progression.

Application of single wrist-wearable accelerometry for objective motor diary assessment in fluctuating Parkinson's disease

Advanced Parkinson's disease (PD) is characterized by motor fluctuations including unpredictable oscillations remarkably impairing quality of life. Effective management and development of novel therapies for these response fluctuations largely depend on clinical rating instruments such as the widely-used PD home diary, which are associated with biases and errors. Recent advancements in digital health technologies provide user-friendly wearables that can be tailored for continuous monitoring of motor fluctuations. Their criterion validity under real-world conditions using clinical examination as the gold standard remains to be determined. We prospectively examined this validity of a wearable accelerometer-based digital Parkinson's Motor Diary (adPMD) using the Parkinson's Kinetigraph (PKG®) in an alternative application by converting its continuous data into one of the three motor categories of the PD home diary (Off, On and Dyskinetic state). Sixty-three out of 91 eligible participants with fluctuating PD (46% men, average age 66) had predefined sufficient adPMD datasets (>70% of half-hour periods) from 2 consecutive days. 92% of per-protocol assessments were completed. adPMD monitoring of daily times in motor states showed moderate validity for Off and Dyskinetic state (ICC = 0.43-0.51), while inter-rating methods agreements on half-hour-level can be characterized as poor (median Cohen's κ = 0.13-0.21). Individualization of adPMD thresholds for transferring accelerometer data into diary categories improved temporal agreements up to moderate level for Dyskinetic state detection (median Cohen's κ = 0.25-0.41). Here we report that adPMD real-world-monitoring captures daily times in Off and Dyskinetic state in advanced PD with moderate validities, while temporal agreement of adPMD and clinical observer diary data is limited.

Machine learning (ML) techniques to predict breast cancer in imbalanced datasets: a systematic review

Knowledge discovery in databases (KDD) is crucial in analyzing data to extract valuable insights. In medical outcome prediction, KDD is increasingly applied, particularly in diseases with high incidence, mortality, and costs, like cancer. ML techniques can develop more accurate predictive models for cancer patients' clinical outcomes, aiding informed healthcare decision-making. However, cancer prediction modeling faces challenges because of the unbalanced nature of the datasets, where there is a small minority category of patients with a cancer diagnosis compared to a majority category of cancer-free patients. Imbalanced datasets pose statistical hurdles like bias and overfitting when developing accurate prediction models. This systematic review focuses on breast cancer prediction articles published from 2008 to 2023. The objective is to examine ML methods used in three critical steps of KDD: preprocessing, data mining, and interpretation which address the imbalanced data problem in breast cancer prediction. This work synthesizes prior research in ML methods for breast cancer prediction. The findings help identify effective preprocessing strategies, including balancing and feature selection methods, robust predictive models, and evaluation metrics of those models. The study aims to inform healthcare providers and researchers about effective techniques for accurate breast cancer prediction.

A schizophrenia study based on multi-frequency dynamic functional connectivity analysis of fMRI

At present, fMRI studies mainly focus on the entire low-frequency band (0. 01-0.08 Hz). However, the neuronal activity is dynamic, and different frequency bands may contain different information. Therefore, a novel multi-frequency-based dynamic functional connectivity (dFC) analysis method was proposed in this study, which was then applied to a schizophrenia study. First, three frequency bands (Conventional: 0.01-0.08 Hz, Slow-5: 0.0111-0.0302 Hz, and Slow-4: 0.0302-0.0820 Hz) were obtained using Fast Fourier Transform. Next, the fractional amplitude of low-frequency fluctuations was used to identify abnormal regions of interest (ROIs) of schizophrenia, and dFC among these abnormal ROIs was implemented by the sliding time window method at four window-widths. Finally, recursive feature elimination was employed to select features, and the support vector machine was applied for the classification of patients with schizophrenia and healthy controls. The experimental results showed that the proposed multi-frequency method (Combined: Slow-5 and Slow-4) had a better classification performance compared with the conventional method at shorter sliding window-widths. In conclusion, our results revealed that the dFCs among the abnormal ROIs varied at different frequency bands and the efficiency of combining multiple features from different frequency bands can improve classification performance. Therefore, it would be a promising approach for identifying brain alterations in schizophrenia.

Data sharing in the context of community-engaged research partnerships

Over the past 20 years, the National Institutes for Health (NIH) has implemented several policies designed to improve sharing of research data, such as the NIH public access policy for publications, NIH genomic data sharing policy, and National Cancer Institute (NCI) Cancer Moonshot public access and data sharing policy. In January 2023, a new NIH data sharing policy has gone into effect, requiring researchers to submit a Data Management and Sharing Plan in proposals for NIH funding (NIH. Supplemental information to the, 2020b; NIH. Final policy for data, 2020a). These policies are based on the idea that sharing data is a key component of the scientific method, as it enables the creation of larger data repositories that can lead to research questions that may not be possible in individual studies (Alter and Gonzalez, 2018; Jwa and Poldrack, 2022), allows enhanced collaboration, and maximizes the federal investment in research. Important questions that we must consider as data sharing is expanded are to whom do benefits of data sharing accrue and to whom do benefits not accrue? In an era of growing efforts to engage diverse communities in research, we must consider the impact of data sharing for all research participants and the communities that they represent. We examine the issue of data sharing through a community-engaged research lens, informed by a long-standing partnership between community-engaged researchers and a key community health organization (Kruse et al., 2022). We contend that without effective community engagement and rich contextual knowledge, biases resulting from data sharing can remain unchecked. We provide several recommendations that would allow better community engagement related to data sharing to ensure both community and researcher understanding of the issues involved and move toward shared benefits. By identifying good models for evaluating the impact of data sharing on communities that contribute data, and then using those models systematically, we will advance the consideration of the community perspective and increase the likelihood of benefits for all.

Hospital mortality prediction in traumatic injuries patients: comparing different SMOTE-based machine learning algorithms

BACKGROUND

Trauma is one of the most critical public health issues worldwide, leading to death and disability and influencing all age groups. Therefore, there is great interest in models for predicting mortality in trauma patients admitted to the ICU. The main objective of the present study is to develop and evaluate SMOTE-based machine-learning tools for predicting hospital mortality in trauma patients with imbalanced data.

METHODS

This retrospective cohort study was conducted on 126 trauma patients admitted to an intensive care unit at Besat hospital in Hamadan Province, western Iran, from March 2020 to March 2021. Data were extracted from the medical information records of patients. According to the imbalanced property of the data, SMOTE techniques, namely SMOTE, Borderline-SMOTE1, Borderline-SMOTE2, SMOTE-NC, and SVM-SMOTE, were used for primary preprocessing. Then, the Decision Tree (DT), Random Forest (RF), Naive Bayes (NB), Artificial Neural Network (ANN), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost) methods were used to predict patients' hospital mortality with traumatic injuries. The performance of the methods used was evaluated by sensitivity, specificity, Positive Predictive Value (PPV), Negative Predictive Value (NPV), accuracy, Area Under the Curve (AUC), Geometric Mean (G-means), F1 score, and P-value of McNemar's test.

RESULTS

Of the 126 patients admitted to an ICU, 117 (92.9%) survived and 9 (7.1%) died. The mean follow-up time from the date of trauma to the date of outcome was 3.98 ± 4.65 days. The performance of ML algorithms is not good with imbalanced data, whereas the performance of SMOTE-based ML algorithms is significantly improved. The mean area under the ROC curve (AUC) of all SMOTE-based models was more than 91%. F1-score and G-means before balancing the dataset were below 70% for all ML models except ANN. In contrast, F1-score and G-means for the balanced datasets reached more than 90% for all SMOTE-based models. Among all SMOTE-based ML methods, RF and ANN based on SMOTE and XGBoost based on SMOTE-NC achieved the highest value for all evaluation criteria.

CONCLUSIONS

This study has shown that SMOTE-based ML algorithms better predict outcomes in traumatic injuries than ML algorithms. They have the potential to assist ICU physicians in making clinical decisions.

KHGCN: Knowledge-Enhanced Recommendation with Hierarchical Graph Capsule Network

Knowledge graphs as external information has become one of the mainstream directions of current recommendation systems. Various knowledge-graph-representation methods have been proposed to promote the development of knowledge graphs in related fields. Knowledge-graph-embedding methods can learn entity information and complex relationships between the entities in knowledge graphs. Furthermore, recently proposed graph neural networks can learn higher-order representations of entities and relationships in knowledge graphs. Therefore, the complete presentation in the knowledge graph enriches the item information and alleviates the cold start of the recommendation process and too-sparse data. However, the knowledge graph's entire entity and relation representation in personalized recommendation tasks will introduce unnecessary noise information for different users. To learn the entity-relationship presentation in the knowledge graph while effectively removing noise information, we innovatively propose a model named knowledge-enhanced hierarchical graph capsule network (KHGCN), which can extract node embeddings in graphs while learning the hierarchical structure of graphs. Our model eliminates noisy entities and relationship representations in the knowledge graph by the entity disentangling for the recommendation and introduces the attentive mechanism to strengthen the knowledge-graph aggregation. Our model learns the presentation of entity relationships by an original graph capsule network. The capsule neural networks represent the structured information between the entities more completely. We validate the proposed model on real-world datasets, and the validation results demonstrate the model's effectiveness.

Optimizing the predictive power of depression screenings using machine learning

Objective

Mental health self-report and clinician-rating scales with diagnoses defined by sum-score cut-offs are often used for depression screening. This study investigates whether machine learning (ML) can detect major depressive episodes (MDE) based on screening scales with higher accuracy than best-practice clinical sum-score approaches.

Methods

Primary data was obtained from two RCTs on the treatment of depression. Ground truth were DSM 5 MDE diagnoses based on structured clinical interviews (SCID) and PHQ-9 self-report, clinician-rated QIDS-16, and HAM-D-17 were predictors. ML models were trained using 10-fold cross-validation. Performance was compared against best-practice sum-score cut-offs. Primary outcome was the Area Under the Curve (AUC) of the Receiver Operating Characteristic curve. DeLong's test with bootstrapping was used to test for differences in AUC. Secondary outcomes were balanced accuracy, precision, recall, F1-score, and number needed to diagnose (NND).

Results

A total of k = 1030 diagnoses (no diagnosis: k = 775; MDE: k = 255) were included. ML models achieved an AUCQIDS-16 = 0.94, AUCHAM-D-17 = 0.88, and AUCPHQ-9 = 0.83 in the testing set. ML AUC was significantly higher than sum-score cut-offs for QIDS-16 and PHQ-9 (ps ≤ 0.01; HAM_D-17: p = 0.847). Applying optimal prediction thresholds, QIDS-16 classifier achieved clinically relevant improvements (Δbalanced accuracy = 8%, ΔF1-score = 14%, ΔNND = 21%). Differences for PHQ_9 and HAM-D-17 were marginal.

Conclusions

ML augmented depression screenings could potentially make a major contribution to improving MDE diagnosis depending on questionnaire (e.g., QIDS-16). Confirmatory studies are needed before ML enhanced screening can be implemented into routine care practice.

Association between the Systolic Blood Pressure Trajectory and Risk of Stroke in a Health-Management Population in Jiaozuo, China

The trajectories of systolic blood pressure (SBP) in a screening population in Jiaozuo were examined, and the association between the different types of SBP trajectories and the risk of stroke was evaluated. Data of a fixed cohort population from the Jiaozuo Stroke Prevention and Control Project Management Special Database System that underwent community screening in 2015, 2017, 2019, and 2021 were collected. Ultimately, a total of 1,451 participants who met the inclusion criteria for this study were included in the analysis, which was performed using group trajectory modeling. The baseline SBP for each trajectory subgroup was characterized at follow-up. Kaplan-Meier analysis for each trajectory group was also performed, and the relationship between the SBP trajectory and risk of stroke onset during follow-up was validated using a Cox proportional hazards model. Based on the SBP from 2015 to 2021, this cohort population was divided into three groups based on the trajectory development patterns: the low-stable group (37.6%), the moderate-increasing group (53.4%), and the high-acutely increasing group (9%). Gender, age, body mass index, diastolic blood pressure, and fasting blood glucose level were predictive factors for the SBP trajectory group. The cumulative survival risk in the high-acutely increasing group was higher than that of the other two groups. After adjusting for potential confounding factors and using the low-stable group as a reference, the hazard ratios (95% confidence interval) for the risk of stroke onset in the moderate-increasing and high-acutely increasing groups were 1.38 (0.91-2.07) and 1.51 (0.82-2.76), respectively. The results of the analysis demonstrate that higher blood pressure trajectories are associated with a higher risk of stroke and that the risk of stroke can be reduced by better control and management of the SBP.

Predicting graft failure in pediatric liver transplantation based on early biomarkers using machine learning models

The early detection of graft failure in pediatric liver transplantation is crucial for appropriate intervention. Graft failure is associated with numerous perioperative risk factors. This study aimed to develop an individualized predictive model for 90-days graft failure in pediatric liver transplantation using machine learning methods. We conducted a single-center retrospective cohort study. A total of 87 liver transplantation cases performed in patients aged < 12 years at the Severance Hospital between January 2010 and September 2020 were included as data samples. Preoperative conditions of recipients and donors, intraoperative care, postoperative serial laboratory parameters, and events observed within seven days of surgery were collected as features. A least absolute shrinkage and selection operator (LASSO) -based method was used for feature selection to overcome the high dimensionality and collinearity of variables. Among 146 features, four variables were selected as the resultant features, namely, preoperative hepatic encephalopathy, sodium level at the end of surgery, hepatic artery thrombosis, and total bilirubin level on postoperative day 7. These features were selected from different times and represent distinct clinical aspects. The model with logistic regression demonstrated the best prediction performance among various machine learning methods tested (area under the receiver operating characteristic curve (AUROC) = 0.898 and area under the precision-recall curve (AUPR) = 0.882). The risk scoring system developed based on the logistic regression model showed an AUROC of 0.910 and an AUPR of 0.830. Together, the prediction of graft failure in pediatric liver transplantation using the proposed machine learning model exhibited superior discrimination power and, therefore, can provide valuable information to clinicians for their decision making during the postoperative management of the patients.

Effects of Data Augmentation with the BNNSMOTE Algorithm in Seizure Detection Using 1D-MobileNet

Automatic seizure detection technology has important implications for reducing the workload of neurologists for epilepsy diagnosis and treatment. Due to the unpredictable nature of seizures, the imbalanced classification of seizure and nonseizure data continues to be challenging. In this work, we first propose a novel algorithm named the borderline nearest neighbor synthetic minority oversampling technique (BNNSMOTE) to address the imbalanced classification problem and improve seizure detection performance. The algorithm uses the nearest neighbor notion to generate nonseizure samples near the boundary, then determines the seizure samples that are difficult to learn at the boundary, and lastly selects seizure samples at random to be used in the synthesis of new samples. In view of the characteristic that electroencephalogram (EEG) signals are one-dimensional signals, we then develop a 1D-MobileNet model to validate the algorithm's performance. Results demonstrate that the proposed algorithm outperforms previous seizure detection methods on the CHB-MIT dataset, achieving an average accuracy of 99.40%, a recall value of 87.46%, a precision of 97.17%, and an F1-score of 91.90%, respectively. We also had considerable success when we used additional datasets for verification at the same time. Our algorithm's data augmentation effects are more pronounced and perform better at seizure detection than the existing imbalanced techniques. Besides, the model's parameters and calculation volume have been significantly reduced, making it more suitable for mobile terminals and embedded devices.

Racial differences in prevalence and anatomical distribution of tarsal coalition

Previous studies have reported a prevalence of tarsal coalition of 0.03-13%. Calcaneonavicular coalition is known as main anatomical type, and the bilateral occurrence of tarsal coalition is known to be 50% or more. These are the results of studies on Caucasians, there have been few studies targeting large number of East Asians so far. We hypothesized that the prevalence and characteristics of tarsal coalition in East Asians might differ from those in Caucasians. The medical records of 839 patients who underwent bilateral computed tomography on foot and ankle in our hospital from January 2012 to April 2021 were retrospectively reviewed. The overall prevalence was 6.0%, talocalcaneal coalition was the most common anatomical type. The overall bilateral occurrence was 56.5%, talocalcaneal coalition had the highest bilateral occurrence (76.0%) among anatomical types. Isolated union of the posterior facet was the most common subtype of talocalcaneal coalition (43.2%). Talocalcaneal coalition had a significantly higher proportion of coalition-related symptomatic patients than calcaneonavicular coalition (p = 0.019). Our study showed a similar trend to other East Asian studies, confirming the existence of racial differences. The possibility of tarsal coalition in foot and ankle patients in East Asians should always be considered, and bilateral examination is essential for diagnosis.

An efficient computer vision-based approach for acute lymphoblastic leukemia prediction

Leukemia (blood cancer) diseases arise when the number of White blood cells (WBCs) is imbalanced in the human body. When the bone marrow produces many immature WBCs that kill healthy cells, acute lymphocytic leukemia (ALL) impacts people of all ages. Thus, timely predicting this disease can increase the chance of survival, and the patient can get his therapy early. Manual prediction is very expensive and time-consuming. Therefore, automated prediction techniques are essential. In this research, we propose an ensemble automated prediction approach that uses four machine learning algorithms K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Random Forest (RF), and Naive Bayes (NB). The C-NMC leukemia dataset is used from the Kaggle repository to predict leukemia. Dataset is divided into two classes cancer and healthy cells. We perform data preprocessing steps, such as the first images being cropped using minimum and maximum points. Feature extraction is performed to extract the feature using pre-trained Convolutional Neural Network-based Deep Neural Network (DNN) architectures (VGG19, ResNet50, or ResNet101). Data scaling is performed by using the MinMaxScaler normalization technique. Analysis of Variance (ANOVA), Recursive Feature Elimination (RFE), and Random Forest (RF) as feature Selection techniques. Classification machine learning algorithms and ensemble voting are applied to selected features. Results reveal that SVM with 90.0% accuracy outperforms compared to other algorithms.