Random Forest Algorithm for the Classification of Neuroimaging Data in Alzheimer's Disease: A Systematic Review

Prediction of lead (Pb) adsorption on attapulgite clay using the feasibility of data intelligence models

This study investigates the performance of support vector machine (SVM), multivariate adaptive regression spline (MARS), and random forest (RF) models for predicting the lead (Pb) adsorption by attapulgite clay. Models are constructed using batch stochastic data of heavy metal (HM) concentrations under different physicochemical conditions. Implementation of auto-hyper-parameter tuning using grid-search approach and comparative analysis is performed against the benchmark artificial intelligence (AI) models. Models are constructed based on Pb concentration (IC), the dosage of attapulgite clay (dose), contact time (CT), pH, and NaNO3 (SN). Principle component analysis (PCA) and correlation analysis (CA) methods are integrated to assess the importance of the applied predictors and their relationship with the target. Research findings approved the potential of the grid-RF model as a marginal superior predictive model against the grid-SVM in terms of MAE, i.e., 3.29 and 3.34, respectively; moreover, the md scored the same, i.e., 0.93, which reveals the potential predictability for both. Nonetheless, grid-MARS and standalone MARS models remained likewise in their predictability. IC parameter demonstrated the highest influential among all the predictors with the highest value of importance in the case of all three evaluators. The solution pH and dose stands together with marginal differences in case of PCA method; however, solution pH and CT appeared with similarity impact using the PCA method.

Impaired Cytotoxic Response in PBMCs From Patients With COVID-19 Admitted to the ICU: Biomarkers to Predict Disease Severity

Infection by novel coronavirus SARS-CoV-2 causes different presentations of COVID-19 and some patients may progress to a critical, fatal form of the disease that requires their admission to ICU and invasive mechanical ventilation. In order to predict in advance which patients could be more susceptible to develop a critical form of COVID-19, it is essential to define the most adequate biomarkers. In this study, we analyzed several parameters related to the cellular immune response in blood samples from 109 patients with different presentations of COVID-19 who were recruited in Hospitals and Primary Healthcare Centers in Madrid, Spain, during the first pandemic peak between April and June 2020. Hospitalized patients with the most severe forms of COVID-19 showed a potent inflammatory response that was not translated into an efficient immune response. Despite the high levels of effector cytotoxic cell populations such as NK, NKT and CD8+ T cells, they displayed immune exhaustion markers and poor cytotoxic functionality against target cells infected with pseudotyped SARS-CoV-2 or cells lacking MHC class I molecules. Moreover, patients with critical COVID-19 showed low levels of the highly cytotoxic TCRγδ+ CD8+ T cell subpopulation. Conversely, CD4 count was greatly reduced in association to high levels of Tregs, low plasma IL-2 and impaired Th1 differentiation. The relative importance of these immunological parameters to predict COVID-19 severity was analyzed by Random Forest algorithm and we concluded that the most important features were related to an efficient cytotoxic response. Therefore, efforts to fight against SARS-CoV-2 infection should be focused not only to decrease the disproportionate inflammatory response, but also to elicit an efficient cytotoxic response against the infected cells and to reduce viral replication.

A machine learning framework to predict antibiotic resistance traits and yet unknown genes underlying resistance to specific antibiotics in bacterial strains

Recently, the frequency of observing bacterial strains without known genetic components underlying phenotypic resistance to antibiotics has increased. There are several strains of bacteria lacking known resistance genes; however, they demonstrate resistance phenotype to drugs of that family. Although such strains are fewer compared to the overall population, they pose grave emerging threats to an already heavily challenged area of antimicrobial resistance (AMR), where death tolls have reached ~700 000 per year and a grim projection of ~10 million deaths per year by 2050 looms. Considering the fact that development of novel antibiotics is not keeping pace with the emergence and dissemination of resistance, there is a pressing need to decipher yet unknown genetic mechanisms of resistance, which will enable developing strategies for the best use of available interventions and show the way for the development of new drugs. In this study, we present a machine learning framework to predict novel AMR factors that are potentially responsible for resistance to specific antimicrobial drugs. The machine learning framework utilizes whole-genome sequencing AMR genetic data and antimicrobial susceptibility testing phenotypic data to predict resistance phenotypes and rank AMR genes by their importance in discriminating the resistance from the susceptible phenotypes. In summary, we present here a bioinformatics framework for training machine learning models, evaluating their performances, selecting the best performing model(s) and finally predicting the most important AMR loci for the resistance involved.

Visual Search Efficiency in Mild Cognitive Impairment and Alzheimer's Disease: An Eye Movement Study

BACKGROUND

Visual search abilities are essential to everyday life activities and are known to be affected in Alzheimer's disease (AD). However, little is known about visual search efficiency in mild cognitive impairment (MCI), a transitive state between normal aging and dementia. Eye movement studies and machine learning methods have been recently used to detect oculomotor impairments in individuals with dementia.

OBJECTIVE

The aim of the present study is to investigate the association between eye movement metrics and visual search impairment in MCI and AD.

METHODS

127 participants were tested: 43 healthy controls, 51 with MCI, and 33 with AD. They completed an eyetracking visual search task where they had to find a previously seen target stimulus among distractors.

RESULTS

Both patient groups made more fixations on the screen when searching for a target, with longer duration than controls. MCI and AD fixated the distractors more often and for a longer period of time than the target. Healthy controls were quicker and made less fixations when scanning the stimuli for the first time. Machine-learning methods were able to distinguish between controls and AD subjects and to identify MCI subjects with a similar oculomotor profile to AD with a good accuracy.

CONCLUSION

Results showed that eye movement metrics are useful for identifying visual search impairments in MCI and AD, with possible implications in the early identification of individuals with high-risk of developing AD.

Identifying Mild Cognitive Impairment with Random Forest by Integrating Multiple MRI Morphological Metrics

Mild cognitive impairment (MCI) exhibits a high risk of progression to Alzheimer's disease (AD), and it is commonly deemed as the precursor of AD. It is important to find effective and robust ways for the early diagnosis of MCI. In this paper, a random forest-based method combining multiple morphological metrics was proposed to identify MCI from normal controls (NC). Voxel-based morphometry, deformation-based morphometry, and surface-based morphometry were utilized to extract morphological metrics such as gray matter volume, Jacobian determinant value, cortical thickness, gyrification index, sulcus depth, and fractal dimension. An initial discovery dataset (56 MCI/55 NC) from the ADNI were used to construct classification models and the performances were testified with 10-fold cross validation. To test the generalization of the proposed method, two extra validation datasets including longitudinal ADNI data (30 MCI/16 NC) and collected data from Xuanwu Hospital (27 MCI/32 NC) were employed respectively to evaluate the performance. No matter whether testing was done on the discovery dataset or the extra validation datasets, the accuracies were about 80% with the combined morphological metrics, which were significantly superior to single metric (accuracy: 45% ∼76%) and also displayed good generalization across datasets. Additionally, gyrification index and cortical thickness derived from surface-based morphometry outperformed other features in MCI identification, suggesting they were some key morphological biomarkers for early MCI diagnosis. Combining the multiple morphological metrics together resulted in a significantly better and reliable identification model, which may be helpful to assist in the clinical diagnosis of MCI.

Random forest model for feature-based Alzheimer's disease conversion prediction from early mild cognitive impairment subjects

Alzheimer's Disease (AD) conversion prediction from the mild cognitive impairment (MCI) stage has been a difficult challenge. This study focuses on providing an individualized MCI to AD conversion prediction using a balanced random forest model that leverages clinical data. In order to do this, 383 Early Mild Cognitive Impairment (EMCI) patients were gathered from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Of these patients, 49 would eventually convert to AD (EMCI_C), whereas the remaining 334 did not convert (EMCI_NC). All of these patients were split randomly into training and testing data sets with 95 patients reserved for testing. Nine clinical features were selected, comprised of a mix of demographic, brain volume, and cognitive testing variables. Oversampling was then performed in order to balance the initially imbalanced classes prior to training the model with 1000 estimators. Our results showed that a random forest model was effective (93.6% accuracy) at predicting the conversion of EMCI patients to AD based on these clinical features. Additionally, we focus on explainability by assessing the importance of each clinical feature. Our model could impact the clinical environment as a tool to predict the conversion to AD from a prodromal stage or to identify ideal candidates for clinical trials.

Screening for Core Genes Related to Pathogenesis of Alzheimer's Disease

Alzheimer’s disease (AD), a nervous system disease, lacks effective therapies at present. RNA expression is the basic way to regulate life activities, and identifying related characteristics in AD patients may aid the exploration of AD pathogenesis and treatment. This study developed a classifier that could accurately classify AD patients and healthy people, and then obtained 3 core genes that may be related to the pathogenesis of AD. To this end, RNA expression data of the middle temporal gyrus of AD patients were firstly downloaded from GEO database, and the data were then normalized using limma package following a supplementation of missing data by k-Nearest Neighbor (KNN) algorithm. Afterwards, the top 500 genes of the most feature importance were obtained through Max-Relevance and Min-Redundancy (mRMR) analysis, and based on these genes, a series of AD classifiers were constructed through Support Vector Machine (SVM), Random Forest (RF), and KNN algorithms. Then, the KNN classifier with the highest Matthews correlation coefficient (MCC) value composed of 14 genes in incremental feature selection (IFS) analysis was identified as the best AD classifier. As analyzed, the 14 genes played a pivotal role in determination of AD and may be core genes associated with the pathogenesis of AD. Finally, protein-protein interaction (PPI) network and Random Walk with Restart (RWR) analysis were applied to obtain core gene-associated genes, and key pathways related to AD were further analyzed. Overall, this study contributed to a deeper understanding of AD pathogenesis and provided theoretical guidance for related research and experiments.

Classification Models for COVID-19 Test Prioritization in Brazil: Machine Learning Approach

BACKGROUND

Controlling the COVID-19 outbreak in Brazil is a challenge due to the population's size and urban density, inefficient maintenance of social distancing and testing strategies, and limited availability of testing resources.

OBJECTIVE

The purpose of this study is to effectively prioritize patients who are symptomatic for testing to assist early COVID-19 detection in Brazil, addressing problems related to inefficient testing and control strategies.

METHODS

Raw data from 55,676 Brazilians were preprocessed, and the chi-square test was used to confirm the relevance of the following features: gender, health professional, fever, sore throat, dyspnea, olfactory disorders, cough, coryza, taste disorders, and headache. Classification models were implemented relying on preprocessed data sets; supervised learning; and the algorithms multilayer perceptron (MLP), gradient boosting machine (GBM), decision tree (DT), random forest (RF), extreme gradient boosting (XGBoost), k-nearest neighbors (KNN), support vector machine (SVM), and logistic regression (LR). The models' performances were analyzed using 10-fold cross-validation, classification metrics, and the Friedman and Nemenyi statistical tests. The permutation feature importance method was applied for ranking the features used by the classification models with the highest performances.

RESULTS

Gender, fever, and dyspnea were among the highest-ranked features used by the classification models. The comparative analysis presents MLP, GBM, DT, RF, XGBoost, and SVM as the highest performance models with similar results. KNN and LR were outperformed by the other algorithms. Applying the easy interpretability as an additional comparison criterion, the DT was considered the most suitable model.

CONCLUSIONS

The DT classification model can effectively (with a mean accuracy≥89.12%) assist COVID-19 test prioritization in Brazil. The model can be applied to recommend the prioritizing of a patient who is symptomatic for COVID-19 testing.

Development of machine learning model for diagnostic disease prediction based on laboratory tests

The use of deep learning and machine learning (ML) in medical science is increasing, particularly in the visual, audio, and language data fields. We aimed to build a new optimized ensemble model by blending a DNN (deep neural network) model with two ML models for disease prediction using laboratory test results. 86 attributes (laboratory tests) were selected from datasets based on value counts, clinical importance-related features, and missing values. We collected sample datasets on 5145 cases, including 326,686 laboratory test results. We investigated a total of 39 specific diseases based on the International Classification of Diseases, 10th revision (ICD-10) codes. These datasets were used to construct light gradient boosting machine (LightGBM) and extreme gradient boosting (XGBoost) ML models and a DNN model using TensorFlow. The optimized ensemble model achieved an F1-score of 81% and prediction accuracy of 92% for the five most common diseases. The deep learning and ML models showed differences in predictive power and disease classification patterns. We used a confusion matrix and analyzed feature importance using the SHAP value method. Our new ML model achieved high efficiency of disease prediction through classification of diseases. This study will be useful in the prediction and diagnosis of diseases.

Metabolic connectivity-based single subject classification by multi-regional linear approximation in the rat

Metabolic connectivity patterns on the basis of [¹⁸F]-FDG positron emission tomography (PET) are used to depict complex cerebral network alterations in different neurological disorders and therefore may have the potential to support diagnostic decisions. In this study, we established a novel statistical classification method taking advantage of differential time-dependent states of whole-brain metabolic connectivity following unilateral labyrinthectomy (UL) in the rat and explored its classification accuracy. The dataset consisted of repeated [¹⁸F]-FDG PET measurements at baseline and 1, 3, 7, and 15 days (= maximum of 5 classes) after UL with 17 rats per measurement day. Classification in different stages after UL was performed by determining connectivity patterns for the different classes by Pearson's correlation between uptake values in atlas-based segmented brain regions. Connections were fitted with a linear function, with which different thresholds on the correlation coefficient (r = [0.5, 0.85]) were investigated. Rats were classified by determining the congruence of their PET uptake pattern with the fitted connectivity patterns in the classes. Overall, the classification accuracy with this method was 84.3% for 3 classes, 75.0% for 4 classes, and 54.1% for 5 classes and outperformed random classification as well as machine learning classification on the same dataset. The optimal classification thresholds of the correlation coefficient and distance-to-fit were found to be |r| > 0.65 and d = 4 when using Siegel's slope estimator for fitting. This connectivity-based classification method can compete with machine learning classification and may have methodological advantages when applied to support PET-based diagnostic decisions in neurological network disorders (such as neurodegenerative syndromes).

Diagnostic Classification and Biomarker Identification of Alzheimer's Disease with Random Forest Algorithm

Random Forest (RF) is a bagging ensemble model and has many important advantages, such as robustness to noise, an effective structure for complex multimodal data and parallel computing, and also provides important features that help investigate biomarkers. Despite these benefits, RF is not used actively to predict Alzheimer's disease (AD) with brain MRIs. Recent studies have reported RF's effectiveness in predicting AD, but the test sample sizes were too small to draw any solid conclusions. Thus, it is timely to compare RF with other learning model methods, including deep learning, particularly with large amounts of data. In this study, we tested RF and various machine learning models with regional volumes from 2250 brain MRIs: 687 normal controls (NC), 1094 mild cognitive impairment (MCI), and 469 AD that ADNI (Alzheimer's Disease Neuroimaging Initiative database) provided. Three types of features sets (63, 29, and 22 features) were selected, and classification accuracies were computed with RF, Support vector machine (SVM), Multi-layer perceptron (MLP), and Convolutional neural network (CNN). As a result, RF, MLP, and CNN showed high performances of 90.2%, 89.6%, and 90.5% with 63 features. Interestingly, when 22 features were used, RF showed the smallest decrease in accuracy, -3.8%, and the standard deviation did not change significantly, while MLP and CNN yielded decreases in accuracy of -6.8% and -4.5% with changes in the standard deviation from 3.3% to 4.0% for MLP and 2.1% to 7.0% for CNN, indicating that RF predicts AD more reliably with fewer features. In addition, we investigated the importance of the features that RF provides, and identified the hippocampus, amygdala, and inferior lateral ventricle as the major contributors in classifying NC, MCI, and AD. On average, AD showed smaller hippocampus and amygdala volumes and a larger volume of inferior lateral ventricle than those of MCI and NC.

The performance of artificial intelligence-driven technologies in diagnosing mental disorders: An umbrella review (Preprint).. [DOI: 10.2196/preprints.29235] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

BACKGROUND

Diagnosing mental disorders is usually not an easy task and requires a large amount of time and effort given the complex nature of mental disorders. Artificial intelligence (AI) has been successfully exploited in diagnosing many mental disorders. Numerous systematic reviews summarize the evidence on the accuracy of AI models in diagnosing different mental disorders.

OBJECTIVE

This umbrella review aims to synthesize results of previous systematic reviews on the performance of AI models in diagnosing mental disorders.

METHODS

To identify relevant systematic reviews, we searched 11 electronic databases, checked the reference list of the included reviews, and checked the reviews that cited the included reviews. Two reviewers independently selected the relevant reviews, extracted the data from them, and appraised their quality. We synthesized the extracted data using the narrative approach. Specifically, results of the included reviews were grouped based on the target mental disorders that the AI classifiers distinguish.

RESULTS

We included 15 systematic reviews of 852 citations identified by searching all databases. The included reviews assessed the performance of AI models in diagnosing Alzheimer’s disease (n=7), mild cognitive impairment (n=6), schizophrenia (n=3), bipolar disease (n=2), autism spectrum disorder (n=1), obsessive-compulsive disorder (n=1), post-traumatic stress disorder (n=1), and psychotic disorders (n=1). The performance of the AI models in diagnosing these mental disorders ranged between 21% and 100%.

CONCLUSIONS

AI technologies offer great promise in diagnosing mental health disorders. The reported performance metrics paint a vivid picture of a bright future for AI in this field. To expedite progress towards these technologies being incorporated into routine practice, we recommend that healthcare professionals in the field cautiously and consciously begin to explore the opportunities of AI-based tools for their daily routine. It would also be encouraging to see a greater number of meta-analyses and further systematic reviews on performance of AI models in diagnosing other common mental disorders such as depression and anxiety.

CLINICALTRIAL

CRD42021231558

Using Machine Learning to Unravel the Value of Radiographic Features for the Classification of Bone Tumors

Objectives

To build and validate random forest (RF) models for the classification of bone tumors based on the conventional radiographic features of the lesion and patients' clinical characteristics, and identify the most essential features for the classification of bone tumors.

Materials and Methods

In this retrospective study, 796 patients (benign bone tumors: 412 cases, malignant bone tumors: 215 cases, intermediate bone tumors: 169 cases) with pathologically confirmed bone tumors from Nanfang Hospital of Southern Medical University, Foshan Hospital of TCM, and University of Hong Kong-Shenzhen Hospital were enrolled. RF models were built to classify tumors as benign, malignant, or intermediate based on conventional radiographic features and potentially relevant clinical characteristics extracted by three musculoskeletal radiologists with ten years of experience. SHapley Additive exPlanations (SHAP) was used to identify the most essential features for the classification of bone tumors. The diagnostic performance of the RF models was quantified using receiver operating characteristic (ROC) curves.

Results

The features extracted by the three radiologists had a satisfactory agreement and the minimum intraclass correlation coefficient (ICC) was 0.761 (CI: 0.686-0.824, P < .001). The binary and tertiary models were built to classify tumors as benign, malignant, or intermediate based on the imaging and clinical features from 627 and 796 patients. The AUC of the binary (19 variables) and tertiary (22 variables) models were 0.97 and 0.94, respectively. The accuracy of binary and tertiary models were 94.71% and 82.77%, respectively. In descending order, the most important features influencing classification in the binary model were margin, cortex involvement, and the pattern of bone destruction, and the most important features in the tertiary model were margin, high-density components, and cortex involvement.

Conclusions

This study developed interpretable models to classify bone tumors with great performance. These should allow radiographers to identify imaging features that are important for the classification of bone tumors in the clinical setting.

Development of a Safety Management System Tracking the Weight of Heavy Objects Carried by Construction Workers Using FSR Sensors

It has been pointed out that the act of carrying a heavy object that exceeds a certain weight by a worker at a construction site is a major factor that puts physical burden on the worker’s musculoskeletal system. However, due to the nature of the construction site, where there are a large number of workers simultaneously working in an irregular space, it is difficult to figure out the weight of the object carried by the worker in real time or keep track of the worker who carries the excess weight. This paper proposes a prototype system to track the weight of heavy objects carried by construction workers by developing smart safety shoes with FSR (Force Sensitive Resistor) sensors. The system consists of smart safety shoes with sensors attached, a mobile device for collecting initial sensing data, and a web-based server computer for storing, preprocessing and analyzing such data. The effectiveness and accuracy of the weight tracking system was verified through the experiments where a weight was lifted by each experimenter from +0 kg to +20 kg in 5 kg increments. The results of the experiment were analyzed by a newly developed machine learning based model, which adopts effective classification algorithms such as decision tree, random forest, gradient boosting algorithm (GBM), and light GBM. The average accuracy classifying the weight by each classification algorithm showed similar, but high accuracy in the following order: random forest (90.9%), light GBM (90.5%), decision tree (90.3%), and GBM (89%). Overall, the proposed weight tracking system has a significant 90.2% average accuracy in classifying how much weight each experimenter carries.

A Machine Learning-Based Investigation of Gender-Specific Prognosis of Lung Cancers

Background and Objective: Primary lung cancer is a lethal and rapidly-developing cancer type and is one of the most leading causes of cancer deaths. Materials and Methods: Statistical methods such as Cox regression are usually used to detect the prognosis factors of a disease. This study investigated survival prediction using machine learning algorithms. The clinical data of 28,458 patients with primary lung cancers were collected from the Surveillance, Epidemiology, and End Results (SEER) database. Results: This study indicated that the survival rate of women with primary lung cancer was often higher than that of men (p < 0.001). Seven popular machine learning algorithms were utilized to evaluate one-year, three-year, and five-year survival prediction The two classifiers extreme gradient boosting (XGB) and logistic regression (LR) achieved the best prediction accuracies. The importance variable of the trained XGB models suggested that surgical removal (feature “Surgery”) made the largest contribution to the one-year survival prediction models, while the metastatic status (feature “N” stage) of the regional lymph nodes was the most important contributor to three-year and five-year survival prediction. The female patients’ three-year prognosis model achieved a prediction accuracy of 0.8297 on the independent future samples, while the male model only achieved the accuracy 0.7329. Conclusions: This data suggested that male patients may have more complicated factors in lung cancer than females, and it is necessary to develop gender-specific diagnosis and prognosis models.

Quantitative Longitudinal Predictions of Alzheimer's Disease by Multi-Modal Predictive Learning

BACKGROUND

Quantitatively predicting the progression of Alzheimer's disease (AD) in an individual on a continuous scale, such as the Alzheimer's Disease Assessment Scale-cognitive (ADAS-cog) scores, is informative for a personalized approach as opposed to qualitatively classifying the individual into a broad disease category.

OBJECTIVE

To evaluate the hypothesis that the multi-modal data and predictive learning models can be employed for future predicting ADAS-cog scores.

METHODS

Unimodal and multi-modal regression models were trained on baseline data comprised of demographics, neuroimaging, and cerebrospinal fluid based markers, and genetic factors to predict future ADAS-cog scores for 12, 24, and 36 months. We subjected the prediction models to repeated cross-validation and assessed the resulting mean absolute error (MAE) and cross-validated correlation (ρ) of the model.

RESULTS

Prediction models trained on multi-modal data outperformed the models trained on single modal data in predicting future ADAS-cog scores (MAE12, 24 & 36 months= 4.1, 4.5, and 5.0, ρ12, 24 & 36 months= 0.88, 0.82, and 0.75). Including baseline ADAS-cog scores to prediction models improved predictive performance (MAE12, 24 & 36 months= 3.5, 3.7, and 4.6, ρ12, 24 & 36 months= 0.89, 0.87, and 0.80).

CONCLUSION

Future ADAS-cog scores were predicted which could aid clinicians in identifying those at greater risk of decline and apply interventions at an earlier disease stage and inform likely future disease progression in individuals enrolled in AD clinical trials.

Development of Random Forest Algorithm Based Prediction Model of Alzheimer's Disease Using Neurodegeneration Pattern

OBJECTIVE

Alzheimer's disease (AD) is the most common type of dementia and the prevalence rapidly increased as the elderly population increased worldwide. In the contemporary model of AD, it is regarded as a disease continuum involving preclinical stage to severe dementia. For accurate diagnosis and disease monitoring, objective index reflecting structural change of brain is needed to correctly assess a patient's severity of neurodegeneration independent from the patient's clinical symptoms. The main aim of this paper is to develop a random forest (RF) algorithm-based prediction model of AD using structural magnetic resonance imaging (MRI).

METHODS

We evaluated diagnostic accuracy and performance of our RF based prediction model using newly developed brain segmentation method compared with the Freesurfer's which is a commonly used segmentation software.

RESULTS

Our RF model showed high diagnostic accuracy for differentiating healthy controls from AD and mild cognitive impairment (MCI) using structural MRI, patient characteristics, and cognitive function (HC vs. AD 93.5%, AUC 0.99; HC vs. MCI 80.8%, AUC 0.88). Moreover, segmentation processing time of our algorithm (<5 minutes) was much shorter than of Freesurfer's (6-8 hours).

CONCLUSION

Our RF model might be an effective automatic brain segmentation tool which can be easily applied in real clinical practice.

Identification of Immunological Parameters as Predictive Biomarkers of Relapse in Patients with Chronic Myeloid Leukemia on Treatment-Free Remission

BCR-ABL is an aberrant tyrosine kinase responsible for chronic myeloid leukemia (CML). Tyrosine kinase inhibitors (TKIs) induce a potent antileukemic response mostly based on the inhibition of BCR-ABL, but they also increase the activity of Natural Killer (NK) and CD8+ T cells. After several years, patients may interrupt treatment due to sustained, deep molecular response. By unknown reasons, half of the patients relapse during treatment interruption, whereas others maintain a potent control of the residual leukemic cells for several years. In this study, several immunological parameters related to sustained antileukemic control were analyzed. According to our results, the features more related to poor antileukemic control were as follows: low levels of cytotoxic cells such as NK, (Natural Killer T) NKT and CD8±TCRγβ+ T cells; low expression of activating receptors on the surface of NK and NKT cells; impaired synthesis of proinflammatory cytokines or proteases from NK cells; and HLA-E*0103 homozygosis and KIR haplotype BX. A Random Forest algorithm predicted 90% of the accuracy for the classification of CML patients in groups of relapse or non-relapse according to these parameters. Consequently, these features may be useful as biomarkers predictive of CML relapse in patients that are candidates to initiate treatment discontinuation.

A machine learning model to identify early stage symptoms of SARS-Cov-2 infected patients

The recent outbreak of the respiratory ailment COVID-19 caused by novel coronavirus SARS-Cov2 is a severe and urgent global concern. In the absence of effective treatments, the main containment strategy is to reduce the contagion by the isolation of infected individuals; however, isolation of unaffected individuals is highly undesirable. To help make rapid decisions on treatment and isolation needs, it would be useful to determine which features presented by suspected infection cases are the best predictors of a positive diagnosis. This can be done by analyzing patient characteristics, case trajectory, comorbidities, symptoms, diagnosis, and outcomes. We developed a model that employed supervised machine learning algorithms to identify the presentation features predicting COVID-19 disease diagnoses with high accuracy. Features examined included details of the individuals concerned, e.g., age, gender, observation of fever, history of travel, and clinical details such as the severity of cough and incidence of lung infection. We implemented and applied several machine learning algorithms to our collected data and found that the XGBoost algorithm performed with the highest accuracy (>85%) to predict and select features that correctly indicate COVID-19 status for all age groups. Statistical analyses revealed that the most frequent and significant predictive symptoms are fever (41.1%), cough (30.3%), lung infection (13.1%) and runny nose (8.43%). While 54.4% of people examined did not develop any symptoms that could be used for diagnosis, our work indicates that for the remainder, our predictive model could significantly improve the prediction of COVID-19 status, including at early stages of infection.

Neuropsychological assessment could distinguish among different clinical phenotypes of progressive supranuclear palsy: A Machine Learning approach

Progressive supranuclear palsy (PSP) is a rare, rapidly progressive neurodegenerative disease. Richardson's syndrome (PSP-RS) and predominant parkinsonism (PSP-P) are characterized by wide range of cognitive and behavioural disturbances, but these variants show similar cognitive pattern of alterations, leading difficult differential diagnosis. For this reason, we explored with an Artificial Intelligence approach, whether cognitive impairment could differentiate the phenotypes. Forty Parkinson's disease (PD) patients, 25 PSP-P, 40 PSP-RS, and 34 controls were enrolled following the consensus criteria diagnosis. Participants were evaluated with neuropsychological battery for cognitive domains. Random Forest models were used for exploring the discriminant power of the cognitive tests in distinguishing among the four groups. The classifiers for distinguishing diseases from controls reached high accuracies (86% for PD, 95% for PSP-P, 99% for PSP-RS). Regarding the differential diagnosis, PD was discriminated from PSP-P with 91% (important variables: HAMA, MMSE, JLO, RAVLT_I, BDI-II) and from PSP-RS with 92% (important variables: COWAT, JLO, FAB). PSP-P was distinguished from PSP-RS with 84% (important variables: JLO, WCFST, RAVLT_I, Digit span_F). This study revealed that PSP-P, PSP-RS and PD had peculiar cognitive deficits compared with healthy subjects, from which they were discriminated with optimal accuracies. Moreover, high accuracies were reached also in differential diagnosis. Most importantly, Machine Learning resulted to be useful to the clinical neuropsychologist in choosing the most appropriate neuropsychological tests for the cognitive evaluation of PSP patients.

Tunable Q wavelet transform based emotion classification in Parkinson's disease using Electroencephalography

Parkinson's disease (PD) is a severe incurable neurological disorder. It is mostly characterized by non-motor symptoms like fatigue, dementia, anxiety, speech and communication problems, depression, and so on. Electroencephalography (EEG) play a key role in the detection of the true emotional state of a person. Various studies have been proposed for the detection of emotional impairment in PD using filtering, Fourier transforms, wavelet transforms, and non-linear methods. However, these methods require a selection of basis and are confined in terms of accuracy. In this paper, tunable Q wavelet transform (TQWT) is proposed for the classification of emotions in PD and normal controls (NC). EEG signals of six emotional states namely happiness, sadness, fear, anger, surprise, and disgust are studied. Power, entropy, and statistical moments based features are elicited from the highpass and lowpass sub-bands of TQWT. Six features selected by statistical analysis are classified with a k-nearest neighbor, probabilistic neural network, random forest, decision tree, and extreme learning machine. Three performance measures are obtained, maximum mean accuracy, sensitivity, and specificity of 96.16%, 97.59%, and 88.51% for NC and 93.88%, 96.33%, and 81.67% for PD are achieved with a probabilistic neural network. The proposed method proved to be very effective such that it classifies emotions in PD and could be used as a potential tool for diagnosing emotional impairment in hospitals.

Machine Learning Methods in Drug Discovery

The advancements of information technology and related processing techniques have created a fertile base for progress in many scientific fields and industries. In the fields of drug discovery and development, machine learning techniques have been used for the development of novel drug candidates. The methods for designing drug targets and novel drug discovery now routinely combine machine learning and deep learning algorithms to enhance the efficiency, efficacy, and quality of developed outputs. The generation and incorporation of big data, through technologies such as high-throughput screening and high through-put computational analysis of databases used for both lead and target discovery, has increased the reliability of the machine learning and deep learning incorporated techniques. The use of these virtual screening and encompassing online information has also been highlighted in developing lead synthesis pathways. In this review, machine learning and deep learning algorithms utilized in drug discovery and associated techniques will be discussed. The applications that produce promising results and methods will be reviewed.

Prediction of Clinical Outcome at Discharge After Rupture of Anterior Communicating Artery Aneurysm Using the Random Forest Technique

Background: Aneurysmal subarachnoid hemorrhage (SAH) is a devastating disease. Anterior communicating artery (ACoA) aneurysm is the most frequent location of intracranial aneurysms. The purpose of this study is to predict the clinical outcome at discharge after rupture of ACoA aneurysms using the random forest machine learning technique.

Methods: A total of 607 patients with ruptured ACoA aneurysms were included in this study between December 2007 and January 2016. In addition to basic clinical variables, 12 aneurysm morphologic parameters were evaluated. A multivariate logistic regression analysis was performed to determine the independent predictors of poor outcome. Of the 607 patients, 485 patients were randomly selected for training and the remaining for internal testing. The random forest model was developed using the training data set. An additional 202 patients from February 2016 to December 2017 were collected for externally validating the model. The prediction performance of the random forest model was compared with two radiologists.

Results: Patients' age (odds ratio [OR] = 1.04), ventilated breathing status (OR = 4.23), World Federation of Neurosurgical Societies (WFNS) grade (OR = 2.13), and Fisher grade (OR = 1.50) are significantly associated with poor outcome. None of the investigated morphological parameters of ACoA aneurysm is an independent predictor of poor outcome. The developed random forest model achieves sensitivities of 78.3% for internal test and 73.8% for external test. The areas under receiver operating characteristic (ROC) curve of the random forest model were 0.90 for the internal test and 0.84 for the external test. Both sensitivities and areas under ROC curves of our model are superior to those of two raters in both internal and external tests.

Conclusions: The random forest model presents good performance in predicting the outcome after rupture of ACoA aneurysms, which may aid in clinical decision making.

Personalized prediction of delayed graft function for recipients of deceased donor kidney transplants with machine learning

Machine learning (ML) has shown its potential to improve patient care over the last decade. In organ transplantation, delayed graft function (DGF) remains a major concern in deceased donor kidney transplantation (DDKT). To this end, we harnessed ML to build personalized prognostic models to predict DGF. Registry data were obtained on adult DDKT recipients for model development (n = 55,044) and validation (n = 6176). Incidence rates of DGF were 25.1% and 26.3% for the development and validation sets, respectively. Twenty-six predictors were identified via recursive feature elimination with random forest. Five widely-used ML algorithms-logistic regression (LR), elastic net, random forest, artificial neural network (ANN), and extreme gradient boosting (XGB) were trained and compared with a baseline LR model fitted with previously identified risk factors. The new ML models, particularly ANN with the area under the receiver operating characteristic curve (ROC-AUC) of 0.732 and XGB with ROC-AUC of 0.735, exhibited superior performance to the baseline model (ROC-AUC = 0.705). This study demonstrates the use of ML as a viable strategy to enable personalized risk quantification for medical applications. If successfully implemented, our models may aid in both risk quantification for DGF prevention clinical trials and personalized clinical decision making.

Predicting the progression of mild cognitive impairment using machine learning: A systematic, quantitative and critical review

We performed a systematic review of studies focusing on the automatic prediction of the progression of mild cognitive impairment to Alzheimer's disease (AD) dementia, and a quantitative analysis of the methodological choices impacting performance. This review included 172 articles, from which 234 experiments were extracted. For each of them, we reported the used data set, the feature types, the algorithm type, performance and potential methodological issues. The impact of these characteristics on the performance was evaluated using a multivariate mixed effect linear regressions. We found that using cognitive, fluorodeoxyglucose-positron emission tomography or potentially electroencephalography and magnetoencephalography variables significantly improved predictive performance compared to not including them, whereas including other modalities, in particular T1 magnetic resonance imaging, did not show a significant effect. The good performance of cognitive assessments questions the wide use of imaging for predicting the progression to AD and advocates for exploring further fine domain-specific cognitive assessments. We also identified several methodological issues, including the absence of a test set, or its use for feature selection or parameter tuning in nearly a fourth of the papers. Other issues, found in 15% of the studies, cast doubts on the relevance of the method to clinical practice. We also highlight that short-term predictions are likely not to be better than predicting that subjects stay stable over time. These issues highlight the importance of adhering to good practices for the use of machine learning as a decision support system for the clinical practice.

A robust multiplex immunofluorescence and digital pathology workflow for the characterisation of the tumour immune microenvironment

Multiplex immunofluorescence is a powerful tool for the simultaneous detection of tissue-based biomarkers, revolutionising traditional immunohistochemistry. The Opal methodology allows up to eight biomarkers to be measured concomitantly without cross-reactivity, permitting identification of different cell populations within the tumour microenvironment. In this study, we aimed to validate a multiplex immunofluorescence workflow in two complementary multiplex panels and evaluate the tumour immune microenvironment in colorectal cancer (CRC) formalin-fixed paraffin-embedded tissue. We stained CRC and tonsil samples using Opal multiplex immunofluorescence on a Leica BOND RX immunostainer. We then acquired images on an Akoya Vectra Polaris and performed multispectral unmixing using inform. Antibody panels were validated on tissue microarray sections containing cores from six normal tissue types, using qupath for image analysis. Comparisons between chromogenic immunohistochemistry and multiplex immunofluorescence on consecutive sections from the same tissue microarray showed significant correlation (rs  > 0.9, P-value < 0.0001), validating both panels. We identified many factors that influenced the quality of the acquired fluorescent images, including biomarker co-expression, staining order, Opal-antibody pairing, sample thickness, multispectral unmixing and biomarker detection order during image analysis. Overall, we report the optimisation and validation of a multiplex immunofluorescence process, from staining to image analysis, ensuring assay robustness. Our multiplex immunofluorescence protocols permit the accurate detection of multiple immune markers in various tissue types, using a workflow that enables rapid processing of samples, above and beyond previous workflows.

Medical Informatics Platform (MIP): A Pilot Study Across Clinical Italian Cohorts

Introduction: With the shift of research focus to personalized medicine in Alzheimer's Dementia (AD), there is an urgent need for tools that are capable of quantifying a patient's risk using diagnostic biomarkers. The Medical Informatics Platform (MIP) is a distributed e-infrastructure federating large amounts of data coupled with machine-learning (ML) algorithms and statistical models to define the biological signature of the disease. The present study assessed (i) the accuracy of two ML algorithms, i.e., supervised Gradient Boosting (GB) and semi-unsupervised 3C strategy (Categorize, Cluster, Classify-CCC) implemented in the MIP and (ii) their contribution over the standard diagnostic workup. Methods: We examined individuals coming from the MIP installed across 3 Italian memory clinics, including subjects with Normal Cognition (CN, n = 432), Mild Cognitive Impairment (MCI, n = 456), and AD (n = 451). The GB classifier was applied to best discriminate the three diagnostic classes in 1,339 subjects, and the CCC strategy was used to refine the classical disease categories. Four dementia experts provided their diagnostic confidence (DC) of MCI conversion on an independent cohort of 38 patients. DC was based on clinical, neuropsychological, CSF, and structural MRI information and again with addition of the outcome from the MIP tools. Results: The GB algorithm provided a classification accuracy of 85% in a nested 10-fold cross-validation for CN vs. MCI vs. AD discrimination. Accuracy increased to 95% in the holdout validation, with the omission of each Italian clinical cohort out in turn. CCC identified five homogeneous clusters of subjects and 36 biomarkers that represented the disease fingerprint. In the DC assessment, CCC defined six clusters in the MCI population used to train the algorithm and 29 biomarkers to improve patients staging. GB and CCC showed a significant impact, evaluated as +5.99% of increment on physicians' DC. The influence of MIP on DC was rated from "slight" to "significant" in 80% of the cases. Discussion: GB provided fair results in classification of CN, MCI, and AD. CCC identified homogeneous and promising classes of subjects via its semi-unsupervised approach. We measured the effect of the MIP on the physician's DC. Our results pave the way for the establishment of a new paradigm for ML discrimination of patients who will or will not convert to AD, a clinical priority for neurology.

Machine learning method using position-specific mutation based classification outperforms one hot coding for disease severity prediction in haemophilia 'A'

Haemophilia is an X-linked genetic disorder in which A and B types are the most common that occur due to absence or lack of protein factors VIII and IX, respectively. Severity of the disease depends on mutation. Available Machine Learning (ML) methods that predict the mutational severity by using traditional encoding approaches, generally have high time complexity and compromised accuracy. In this study, Haemophilia 'A' patient mutation dataset containing 7784 mutations was processed by the proposed Position-Specific Mutation (PSM) and One-Hot Encoding (OHE) technique to predict the disease severity. The dataset processed by PSM and OHE methods was analyzed and trained for classification of mutation severity level using various ML algorithms. Surprisingly, PSM outperformed OHE, both in terms of time efficiency and accuracy, with training and prediction time improvement in the range of approximately 91 to 98% and 80 to 99% respectively. The severity prediction accuracy also improved by using PSM with different ML algorithms.

The Use of Random Forests to Identify Brain Regions on Amyloid and FDG PET Associated With MoCA Score

PURPOSE

The aim of this study was to evaluate random forests (RFs) to identify ROIs on F-florbetapir and F-FDG PET associated with Montreal Cognitive Assessment (MoCA) score.

MATERIALS AND METHODS

Fifty-seven subjects with significant white matter disease presenting with either transient ischemic attack/lacunar stroke or mild cognitive impairment from early Alzheimer disease, enrolled in a multicenter prospective observational trial, had MoCA and F-florbetapir PET; 55 had F-FDG PET. Scans were processed using the MINC toolkit to generate SUV ratios, normalized to cerebellar gray matter (F-florbetapir PET), or pons (F-FDG PET). SUV ratio data and MoCA score were used for supervised training of RFs programmed in MATLAB.

RESULTS

F-Florbetapir PETs were randomly divided into 40 training and 17 testing scans; 100 RFs of 1000 trees, constructed from a random subset of 16 training scans and 20 ROIs, identified ROIs associated with MoCA score: right posterior cingulate gyrus, right anterior cingulate gyrus, left precuneus, left posterior cingulate gyrus, and right precuneus. Amyloid increased with decreasing MoCA score. F-FDG PETs were randomly divided into 40 training and 15 testing scans; 100 RFs of 1000 trees, each tree constructed from a random subset of 16 training scans and 20 ROIs, identified ROIs associated with MoCA score: left fusiform gyrus, left precuneus, left posterior cingulate gyrus, right precuneus, and left middle orbitofrontal gyrus. F-FDG decreased with decreasing MoCA score.

CONCLUSIONS

Random forests help pinpoint clinically relevant ROIs associated with MoCA score; amyloid increased and F-FDG decreased with decreasing MoCA score, most significantly in the posterior cingulate gyrus.

Optimization of parallel random forest algorithm based on distance weight

In order to overcome the problems of long execution time and low parallelism of existing parallel random forest algorithms, an optimization method for parallel random forest algorithm based on distance weights is proposed. The concept of distance weights is introduced to optimize the algorithm. Firstly, the training sample data are extracted from the original data set by random selection. Based on the extracted results, a single decision tree is constructed. The single decision tree is grouped together according to different grouping methods to form a random forest. The distance weights of the training sample data set are calculated, and then the weighted optimization of the random forest model is realized. The experimental results show that the execution time of the parallel random forest algorithm after optimization is 110 000 ms less than that before optimization, and the operation efficiency of the algorithm is greatly improved, which effectively solves the problems existing in the traditional random forest algorithm.

Machine learning approaches identify male body size as the most accurate predictor of species richness

BACKGROUND

A major challenge in biodiversity science is to understand the factors contributing to the variability of species richness -the number of different species in a community or region - among comparable taxonomic lineages. Multiple biotic and abiotic factors have been hypothesized to have an effect on species richness and have been used as its predictors, but identifying accurate predictors is not straightforward. Spiders are a highly diverse group, with some 48,000 species in 120 families; yet nearly 75% of all species are found within just the ten most speciose families. Here we use a Random Forest machine learning algorithm to test the predictive power of different variables hypothesized to affect species richness of spider genera.

RESULTS

We test the predictive power of 22 variables from spiders' morphological, genetic, geographic, ecological and behavioral landscapes on species richness of 45 genera selected to represent the phylogenetic and biological breath of Araneae. Among the variables, Random Forest analyses find body size (specifically, minimum male body size) to best predict species richness. Multiple Correspondence analysis confirms this outcome through a negative relationship between male body size and species richness. Multiple Correspondence analyses furthermore establish that geographic distribution of congeneric species is positively associated with genus diversity, and that genera from phylogenetically older lineages are species poorer. Of the spider-specific traits, neither the presence of ballooning behavior, nor sexual size dimorphism, can predict species richness.

CONCLUSIONS

We show that machine learning analyses can be used in deciphering the factors associated with diversity patterns. Since no spider-specific biology could predict species richness, but the biologically universal body size did, we believe these conclusions are worthy of broader biological testing. Future work on other groups of organisms will establish whether the detected associations of species richness with small body size and wide geographic ranges hold more broadly.

Spanish Influenza Score (SIS): Usefulness of machine learning in the development of an early mortality prediction score in severe influenza

OBJECTIVE

To develop a mortality prediction score (Spanish Influenza Score [SIS]) for patients with severe influenza considering only variables at ICU admission, and compare its performance respect of Random Forest (RF).

DESIGN

Sub-analysis from the GETGAG/SEMICYUC database.

SCOPE

Intensive Care Medicine.

PATIENTS

Patients admitted to 184 Spanish ICUs (2009-2018) with influenza infection Intervention: None.

VARIABLES

Demographic data, severity of illness, times from symptoms onset until hospital admission (Gap-H), hospital to ICU (Gap-ICU) or hospital to diagnosis (Gap-Dg), antiviral vaccination, number of quadrants infiltrated, acute renal failure, invasive or noninvasive ventilation, shock and comorbidities. The study variable cut-off points and importance were obtained automatically. Logistic regression analysis with cross-validation was performed to develop the SIS score using the output coefficients. Accuracy and discrimination (AUC-ROC) were applied to evaluate SIS and RF. All analyses were performed using R (CRAN-R Project).

RESULTS

A total of 3959 patients were included. The mean age was 55 years (range 43-67), 60% were men, APACHE II 16 (12-21) and SOFA 5 (4-8), with ICU mortality 21.3%. Mechanical ventilation, shock, APACHE II, SOFA, acute renal failure and Gap-ICU were included in the SIS. The latter was generated according to the ORs obtained by logistic regression, and showed an accuracy of 83% with an AUC-ROC of 82%, similar to RF (AUC-ROC 82%).

CONCLUSIONS

The SIS score is easy to apply and shows adequate capacity to stratify the risk of ICU mortality. However, further studies are needed to validate the tool prospectively.

Machine learning improves mortality risk prediction after cardiac surgery: Systematic review and meta-analysis

BACKGROUND

Interest in the usefulness of machine learning (ML) methods for outcomes prediction has continued to increase in recent years. However, the advantage of advanced ML model over traditional logistic regression (LR) remains controversial. We performed a systematic review and meta-analysis of studies comparing the discrimination accuracy between ML models versus LR in predicting operative mortality following cardiac surgery.

METHODS

The present systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis statement. Discrimination ability was assessed using the C-statistic. Pooled C-statistics and its 95% credibility interval for ML models and LR were obtained were obtained using a Bayesian framework. Pooled estimates for ML models and LR were compared to inform on difference between the 2 approaches.

RESULTS

We identified 459 published citations of which 15 studies met inclusion criteria and were used for the quantitative and qualitative analysis. When the best ML model from individual study was used, meta-analytic estimates showed that ML were associated with a significantly higher C-statistic (ML, 0.88; 95% credibility interval, 0.83-0.93 vs LR, 0.81; 95% credibility interval, 0.77-0.85; P = .03). When individual ML algorithms were instead selected, we found a nonsignificant trend toward better prediction with each of ML algorithms. We found no evidence of publication bias (P = .70).

CONCLUSIONS

The present findings suggest that when compared with LR, ML models provide better discrimination in mortality prediction after cardiac surgery. However, the magnitude and clinical influence of such an improvement remains uncertain.

Prediction for cardiovascular diseases based on laboratory data: An analysis of random forest model

Background

To establish a prediction model for cardiovascular diseases (CVD) in the general population based on random forests.

Methods

A retrospective study involving 498 subjects was conducted in Xi'an Medical University between 2011 and 2018. The random forest algorithm was used to screen out the variables that greatly affected the CVD prediction and to establish a prediction model. The important variables were included in the multifactorial logistic regression analysis. The area under the curve (AUC) was compared between logistic regression model and random forest model.

Results

The random forest model revealed the variables, including the age, body mass index (BMI), fasting blood glucose (FBG), diastolic blood pressure (DBP), triglyceride (TG), systolic blood pressure (SBP), total cholesterol (TC), waist circumference, and high‐density lipoprotein‐cholesterol (HDL‐C), were more significant for CVD prediction; the AUC was 0.802 in CVD prediction. Multifactorial logistic regression analysis indicated that the risk factors for CVD included the age [odds ratio (OR): 1.14, 95% confidence intervals (CI): 1.10‐1.17, P < .001], BMI (OR: 1.13, 95% CI: 1.06‐1.20, P < .001), TG (OR: 1.11, 95% CI: 1.02‐1.22, P = .023), and DBP (OR: 1.04, 95% CI: 1.02‐1.06, P = .001); the AUC was 0.843 in CVD prediction. The established logistic regression prediction model was Logit P = Log[P/(1 − P)] = −11.47 + 0.13 × age + 0.12 × BMI + 0.11 × TG + 0.04 × DBP; P = 1/[1 + exp(−Logit P)]. People were prone to develop CVD at the time of P > .51.

Conclusions

A prediction model for CVD is developed in the general population based on random forests, which provides a simple tool for the early prediction of CVD.

Evaluation of the impact of body mass index on venous thromboembolism risk factors

In this paper, we investigate the interaction impacts of body mass index (BMI) on the other important risk factors for venous thromboembolism (VTE), using deep venous thrombosis (DVT) patient data from the International Warfarin Pharmacogenetics Consortium (IWPC). We apply eight machine learning techniques, including naive Bayes classifier (NB), support vector machine (SVM), elastic net regression (ENET), logistic regression (LR), lasso regression (LAR), multivariate adaptive regression splines (MARS), boosted regression tree (BRT) and random forest model (RF). The RF method is selected as the best model for classification. Out of 33 features considered in this study, we identify 12 variables as relatively important risk factors for VTE. Finally, we examine the interaction impacts of BMI on these important VTE risk factors. We conclude that the impacts of risk factors on VTE incidence are varying across different BMI groups, and the variations are different for different risk factors. Therefore the interaction impacts of BMI on the other risk factors have to be taken into account in order to better understand the incidence of VTE.

Nonlinear biomarker interactions in conversion from mild cognitive impairment to Alzheimer's disease

Multiple biomarkers can capture different facets of Alzheimer's disease. However, statistical models of biomarkers to predict outcomes in Alzheimer's rarely model nonlinear interactions between these measures. Here, we used Gaussian Processes to address this, modelling nonlinear interactions to predict progression from mild cognitive impairment (MCI) to Alzheimer's over 3 years, using Alzheimer's Disease Neuroimaging Initiative (ADNI) data. Measures included: demographics, APOE4 genotype, CSF (amyloid‐β42, total tau, phosphorylated tau), [18^F]florbetapir, hippocampal volume and brain‐age. We examined: (a) the independent value of each biomarker; and (b) whether modelling nonlinear interactions between biomarkers improved predictions. Each measured added complementary information when predicting conversion to Alzheimer's. A linear model classifying stable from progressive MCI explained over half the variance (R² = 0.51, p < .001); the strongest independently contributing biomarker was hippocampal volume (R² = 0.13). When comparing sensitivity of different models to progressive MCI (independent biomarker models, additive models, nonlinear interaction models), we observed a significant improvement (p < .001) for various two‐way interaction models. The best performing model included an interaction between amyloid‐β‐PET and P‐tau, while accounting for hippocampal volume (sensitivity = 0.77, AUC = 0.826). Closely related biomarkers contributed uniquely to predict conversion to Alzheimer's. Nonlinear biomarker interactions were also implicated, and results showed that although for some patients adding additional biomarkers may add little value (i.e., when hippocampal volume is high), for others (i.e., with low hippocampal volume) further invasive and expensive examination may be warranted. Our framework enables visualisation of these interactions, in individual patient biomarker ‘space', providing information for personalised or stratified healthcare or clinical trial design.

Detection of Electricity Theft Behavior Based on Improved Synthetic Minority Oversampling Technique and Random Forest Classifier

Effective detection of electricity theft is essential to maintain power system reliability. With the development of smart grids, traditional electricity theft detection technologies have become ineffective to deal with the increasingly complex data on the users’ side. To improve the auditing efficiency of grid enterprises, a new electricity theft detection method based on improved synthetic minority oversampling technique (SMOTE) and improve random forest (RF) method is proposed in this paper. The data of normal and electricity theft users were classified as positive data (PD) and negative data (ND), respectively. In practice, the number of ND was far less than PD, which made the dataset composed of these two types of data become unbalanced. An improved SOMTE based on K-means clustering algorithm (K-SMOTE) was firstly presented to balance the dataset. The cluster center of ND was determined by K-means method. Then, the ND were interpolated by SMOTE on the basis of the cluster center to balance the entire data. Finally, the RF classifier was trained with the balanced dataset, and the optimal number of decision trees in RF was decided according to the convergence of out-of-bag data error (OOB error). Electricity theft behaviors on the user side were detected by the trained RF classifier.

Anterior Hippocampal-Cortical Functional Connectivity Distinguishes Antipsychotic Naïve First-Episode Psychosis Patients From Controls and May Predict Response to Second-Generation Antipsychotic Treatment

BACKGROUND

Converging evidence implicates the anterior hippocampus in the proximal pathophysiology of schizophrenia. Although resting state functional connectivity (FC) holds promise for characterizing anterior hippocampal circuit abnormalities and their relationship to treatment response, this technique has not yet been used in first-episode psychosis (FEP) patients in a manner that distinguishes the anterior from posterior hippocampus.

METHODS

We used masked-hippocampal-group-independent component analysis with dual regression to contrast subregional hippocampal-whole brain FC between healthy controls (HCs) and antipsychotic naïve FEP patients (N = 61, 36 female). In a subsample of FEP patients (N = 27, 15 female), we repeated this analysis following 8 weeks of second-generation antipsychotic treatment and explored whether baseline FC predicted treatment response using random forest.

RESULTS

Relative to HC, untreated FEP subjects displayed reproducibly lower FC between the left anteromedial hippocampus and cortical regions including the anterior cingulate and insular cortex (P < .05, corrected). Anteromedial hippocampal FC increased in FEP patients following treatment (P < .005), and no longer differed from HC. Random forest analysis showed baseline anteromedial hippocampal FC with four brain regions, namely the insular-opercular cortex, superior frontal gyrus, precentral gyrus, and postcentral gyrus predicted treatment response (area under the curve = 0.95).

CONCLUSIONS

Antipsychotic naïve FEP is associated with lower FC between the anterior hippocampus and cortical regions previously implicated in schizophrenia. Preliminary analysis suggests that random forest models based on hippocampal FC may predict treatment response in FEP patients, and hence could be a useful biomarker for treatment development.

Radiomic prediction of mutation status based on MR imaging of lung cancer brain metastases

Lung cancer metastases comprise most of all brain metastases in adults and most brain metastases are diagnosed by magnetic resonance (MR) scans. The purpose of this study was to conduct an MR imaging-based radiomic analysis of brain metastatic lesions from patients with primary lung cancer to classify mutational status of the metastatic disease. We retrospectively identified lung cancer patients with brain metastases treated at our institution between 2009 and 2017 who underwent genotype testing of their primary lung cancer. Brain MR Images were used for segmentation of enhancing tumors and peritumoral edema, and for radiomic feature extraction. The most relevant radiomic features were identified and used with clinical data to train random forest classifiers to classify the mutation status. Of 110 patients in the study cohort (mean age 57.51 ± 12.32 years; M: F = 37:73), 75 had an EGFR mutation, 21 had an ALK translocation, and 15 had a KRAS mutation. One patient had both ALK translocation and EGFR mutation. Majority of radiomic features most relevant for mutation classification were textural. Model building using both radiomic features and clinical data yielded more accurate classifications than using either alone. For classification of EGFR, ALK, and KRAS mutation status, the model built with both radiomic features and clinical data resulted in area-under-the-curve (AUC) values based on cross-validation of 0.912, 0.915, and 0.985, respectively. Our study demonstrated that MR imaging-based radiomic analysis of brain metastases in patients with primary lung cancer may be used to classify mutation status. This approach may be useful for devising treatment strategies and informing prognosis.

Random Forest Classification of Alcohol Use Disorder Using EEG Source Functional Connectivity, Neuropsychological Functioning, and Impulsivity Measures

: Individuals with alcohol use disorder (AUD) manifest a variety of impairments that can be attributed to alterations in specific brain networks. The current study aims to identify features of EEG-based functional connectivity, neuropsychological performance, and impulsivity that can classify individuals with AUD (N = 30) from unaffected controls (CTL, N = 30) using random forest classification. The features included were: (i) EEG source functional connectivity (FC) of the default mode network (DMN) derived using eLORETA algorithm, (ii) neuropsychological scores from the Tower of London test (TOLT) and the visual span test (VST), and (iii) impulsivity factors from the Barratt impulsiveness scale (BIS). The random forest model achieved a classification accuracy of 80% and identified 29 FC connections (among 66 connections per frequency band), 3 neuropsychological variables from VST (total number of correctly performed trials in forward and backward sequences and average time for correct trials in forward sequence) and all four impulsivity scores (motor, non-planning, attentional, and total) as significantly contributing to classifying individuals as either AUD or CTL. Although there was a significant age difference between the groups, most of the top variables that contributed to the classification were not significantly correlated with age. The AUD group showed a predominant pattern of hyperconnectivity among 25 of 29 significant connections, indicating aberrant network functioning during resting state suggestive of neural hyperexcitability and impulsivity. Further, parahippocampal hyperconnectivity with other DMN regions was identified as a major hub region dysregulated in AUD (13 connections overall), possibly due to neural damage from chronic drinking, which may give rise to cognitive impairments, including memory deficits and blackouts. Furthermore, hypoconnectivity observed in four connections (prefrontal nodes connecting posterior right-hemispheric regions) may indicate a weaker or fractured prefrontal connectivity with other regions, which may be related to impaired higher cognitive functions. The AUD group also showed poorer memory performance on the VST task and increased impulsivity in all factors compared to controls. Features from all three domains had significant associations with one another. These results indicate that dysregulated neural connectivity across the DMN regions, especially relating to hyperconnected parahippocampal hub as well as hypoconnected prefrontal hub, may potentially represent neurophysiological biomarkers of AUD, while poor visual memory performance and heightened impulsivity may serve as cognitive-behavioral indices of AUD.

Random Forest Classification of Alcohol Use Disorder Using fMRI Functional Connectivity, Neuropsychological Functioning, and Impulsivity Measures

Individuals with alcohol use disorder (AUD) are known to manifest a variety of neurocognitive impairments that can be attributed to alterations in specific brain networks. The current study aims to identify specific features of brain connectivity, neuropsychological performance, and impulsivity traits that can classify adult males with AUD (n = 30) from healthy controls (CTL, n = 30) using the Random Forest (RF) classification method. The predictor variables were: (i) fMRI-based within-network functional connectivity (FC) of the Default Mode Network (DMN), (ii) neuropsychological scores from the Tower of London Test (TOLT), and the Visual Span Test (VST), and (iii) impulsivity factors from the Barratt Impulsiveness Scale (BIS). The RF model, with a classification accuracy of 76.67%, identified fourteen DMN connections, two neuropsychological variables (memory span and total correct scores of the forward condition of the VST), and all impulsivity factors as significantly important for classifying participants into either the AUD or CTL group. Specifically, the AUD group manifested hyperconnectivity across the bilateral anterior cingulate cortex and the prefrontal cortex as well as between the bilateral posterior cingulate cortex and the left inferior parietal lobule, while showing hypoconnectivity in long-range anterior-posterior and interhemispheric long-range connections. Individuals with AUD also showed poorer memory performance and increased impulsivity compared to CTL individuals. Furthermore, there were significant associations among FC, impulsivity, neuropsychological performance, and AUD status. These results confirm the previous findings that alterations in specific brain networks coupled with poor neuropsychological functioning and heightened impulsivity may characterize individuals with AUD, who can be efficiently identified using classification algorithms such as Random Forest.

Feature Selection and Combination of Information in the Functional Brain Connectome for Discrimination of Mild Cognitive Impairment and Analyses of Altered Brain Patterns

Mild cognitive impairment (MCI) is often considered a critical time window for predicting early conversion to Alzheimer’s disease (AD). Brain functional connectome data (i.e., functional connections, global and nodal graph metrics) based on resting-state functional magnetic resonance imaging (rs-fMRI) provides numerous information about brain networks and has been used to discriminate normal controls (NCs) from subjects with MCI. In this paper, Student’s t-tests and group-least absolute shrinkage and selection operator (group-LASSO) were used to extract functional connections with significant differences and the most discriminative network nodes, respectively. Based on group-LASSO, the middle temporal, inferior temporal, lingual, posterior cingulate, and middle frontal gyri were the most predominant brain regions for nodal observation in MCI patients. Nodal graph metrics (within-module degree, participation coefficient, and degree centrality) showed the maximum discriminative ability. To effectively combine the multipattern information, we employed the multiple kernel learning support vector machine (MKL-SVM). Combined with functional connectome information, the MKL-SVM achieved a good classification performance (area under the receiving operating characteristic curve = 0.9728). Additionally, the altered brain connectome pattern revealed that functional connectivity was generally decreased in the whole-brain network, whereas graph theory topological attributes of some special nodes in the brain network were increased in MCI patients. Our findings demonstrate that optimal feature selection and combination of all connectome features (i.e., functional connections, global and nodal graph metrics) can achieve good performance in discriminating NCs from MCI subjects. Thus, the combination of functional connections and global and nodal graph metrics of brain networks can predict the occurrence of MCI and contribute to the early clinical diagnosis of AD.

Evaluation of Brain Tortuosity Measurement for the Automatic Multimodal Classification of Subjects with Alzheimer's Disease

The 3D tortuosity determined in several brain areas is proposed as a new morphological biomarker (BM) to be considered in early detection of Alzheimer's disease (AD). It is measured using the sum of angles method and it has proven to be sensitive to anatomical changes that appear in gray and white matter and temporal and parietal lobes during mild cognitive impairment (MCI). Statistical analysis showed significant differences (p < 0.05) between tortuosity indices determined for healthy controls (HC) vs. MCI and HC vs. AD in most of the analyzed structures. Other clinically used BMs have also been incorporated in the analysis: beta-amyloid and tau protein CSF and plasma concentrations, as well as other image-extracted parameters. A classification strategy using random forest (RF) algorithms was implemented to discriminate between three samples of the studied populations, selected from the ADNI database. Classification rates considering only image-extracted parameters show an increase of 9.17%, when tortuosity is incorporated. An enhancement of 1.67% is obtained when BMs measured from several modalities are combined with tortuosity.

Artificial intelligence with multi-functional machine learning platform development for better healthcare and precision medicine

Precision medicine is one of the recent and powerful developments in medical care, which has the potential to improve the traditional symptom-driven practice of medicine, allowing earlier interventions using advanced diagnostics and tailoring better and economically personalized treatments. Identifying the best pathway to personalized and population medicine involves the ability to analyze comprehensive patient information together with broader aspects to monitor and distinguish between sick and relatively healthy people, which will lead to a better understanding of biological indicators that can signal shifts in health. While the complexities of disease at the individual level have made it difficult to utilize healthcare information in clinical decision-making, some of the existing constraints have been greatly minimized by technological advancements. To implement effective precision medicine with enhanced ability to positively impact patient outcomes and provide real-time decision support, it is important to harness the power of electronic health records by integrating disparate data sources and discovering patient-specific patterns of disease progression. Useful analytic tools, technologies, databases, and approaches are required to augment networking and interoperability of clinical, laboratory and public health systems, as well as addressing ethical and social issues related to the privacy and protection of healthcare data with effective balance. Developing multifunctional machine learning platforms for clinical data extraction, aggregation, management and analysis can support clinicians by efficiently stratifying subjects to understand specific scenarios and optimize decision-making. Implementation of artificial intelligence in healthcare is a compelling vision that has the potential in leading to the significant improvements for achieving the goals of providing real-time, better personalized and population medicine at lower costs. In this study, we focused on analyzing and discussing various published artificial intelligence and machine learning solutions, approaches and perspectives, aiming to advance academic solutions in paving the way for a new data-centric era of discovery in healthcare.

Identification of clinical and urine biomarkers for uncomplicated urinary tract infection using machine learning algorithms

Women with uncomplicated urinary tract infection (UTI) symptoms are commonly treated with empirical antibiotics, resulting in overuse of antibiotics, which promotes antimicrobial resistance. Available diagnostic tools are either not cost-effective or diagnostically sub-optimal. Here, we identified clinical and urinary immunological predictors for UTI diagnosis. We explored 17 clinical and 42 immunological potential predictors for bacterial culture among women with uncomplicated UTI symptoms using random forest or support vector machine coupled with recursive feature elimination. Urine cloudiness was the best performing clinical predictor to rule out (negative likelihood ratio [LR−] = 0.4) and rule in (LR+ = 2.6) UTI. Using a more discriminatory scale to assess cloudiness (turbidity) increased the accuracy of UTI prediction further (LR+ = 4.4). Urinary levels of MMP9, NGAL, CXCL8 and IL-1β together had a higher LR+ (6.1) and similar LR− (0.4), compared to cloudiness. Varying the bacterial count thresholds for urine culture positivity did not alter best clinical predictor selection, but did affect the number of immunological predictors required for reaching an optimal prediction. We conclude that urine cloudiness is particularly helpful in ruling out negative UTI cases. The identified urinary biomarkers could be used to develop a point of care test for UTI but require further validation.

Deep Feature Selection and Causal Analysis of Alzheimer's Disease

Deep convolutional neural networks (DCNNs) have achieved great success for image classification in medical research. Deep learning with brain imaging is the imaging method of choice for the diagnosis and prediction of Alzheimer’s disease (AD). However, it is also well known that DCNNs are “black boxes” owing to their low interpretability to humans. The lack of transparency of deep learning compromises its application to the prediction and mechanism investigation in AD. To overcome this limitation, we develop a novel general framework that integrates deep leaning, feature selection, causal inference, and genetic-imaging data analysis for predicting and understanding AD. The proposed algorithm not only improves the prediction accuracy but also identifies the brain regions underlying the development of AD and causal paths from genetic variants to AD via image mediation. The proposed algorithm is applied to the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset with diffusion tensor imaging (DTI) in 151 subjects (51 AD and 100 non-AD) who were measured at four time points of baseline, 6 months, 12 months, and 24 months. The algorithm identified brain regions underlying AD consisting of the temporal lobes (including the hippocampus) and the ventricular system.

Pharmacologically informed machine learning approach for identifying pathological states of unconsciousness via resting-state fMRI

Determining the level of consciousness in patients with disorders of consciousness (DOC) remains challenging. To address this challenge, resting-state fMRI (rs-fMRI) has been widely used for detecting the local, regional, and network activity differences between DOC patients and healthy controls. Although substantial progress has been made towards this endeavor, the identification of robust rs-fMRI-based biomarkers for level of consciousness is still lacking. Recent developments in machine learning show promise as a tool to augment the discrimination between different states of consciousness in clinical practice. Here, we investigated whether machine learning models trained to make a binary distinction between conscious wakefulness and anesthetic-induced unconsciousness would then be capable of reliably identifying pathologically induced unconsciousness. We did so by extracting rs-fMRI-based features associated with local activity, regional homogeneity, and interregional functional activity in 44 subjects during wakefulness, light sedation, and unresponsiveness (deep sedation and general anesthesia), and subsequently using those features to train three distinct candidate machine learning classifiers: support vector machine, Extra Trees, artificial neural network. First, we show that all three classifiers achieve reliable performance within-dataset (via nested cross-validation), with a mean area under the receiver operating characteristic curve (AUC) of 0.95, 0.92, and 0.94, respectively. Additionally, we observed comparable cross-dataset performance (making predictions on the DOC data) as the anesthesia-trained classifiers demonstrated a consistent ability to discriminate between unresponsive wakefulness syndrome (UWS/VS) patients and healthy controls with mean AUC's of 0.99, 0.94, 0.98, respectively. Lastly, we explored the potential of applying the aforementioned classifiers towards discriminating intermediate states of consciousness, specifically, subjects under light anesthetic sedation and patients diagnosed as having a minimally conscious state (MCS). Our findings demonstrate that machine learning classifiers trained on rs-fMRI features derived from participants under anesthesia have potential to aid the discrimination between degrees of pathological unconsciousness in clinical patients.