Analysis of neuropsychological and neuroradiological features for diagnosis of Alzheimer's disease and mild cognitive impairment.

BACKGROUND: Age-related neurodegenerative diseases are constantly increasing with prediction that in 2050 over 60 % of population will suffer from some level of cognitive impairment. A cure for the Alzheimer's disease (AD) does not exist, so early diagnosis is of a great importance. Machine learning techniques can help in early diagnosis with deep medical data processing, disease understanding, intervention analysis and knowledge discovery for achieving better medical decision making. METHODS: In this paper, we analyze the dataset consisting of 90 individuals and 482 input features. We investigate the achieved AD prediction performances using seven classifiers and five feature selection algorithms. We pay special focus on analyzing performance by utilizing only a subset of best ranked attributes to establish the minimum amount of input features that ensure acceptable performance. We also investigate the significance of neuropsychological (NP) and neuroradiological (NR) attributes for the AD diagnosis. RESULTS: The accuracy for the whole set of attributes ranged between 66.22 % and 81.00 %, and the weighted average AUROC was between 76.3 % and 95.0 %. The best results were achieved by the naive Bayes classifier and the Relief feature selection algorithm. Additionally, Support Vector Machines classifier shows the most stable results since it depends the least on the feature selection algorithm which is used. As the main result of this paper, we compare the performance of models trained with automatically selected features to models trained with hand-selected features performed by medical experts (NP and NR features). CONCLUSIONS: The results reveal that unlike the NR attributes, the NP attributes achieve a good performance that is comparable to the full set of attributes, which suggests that they possess a high predictive power for AD diagnosis.

Machine learning and physical based modeling for cardiac hypertrophy.

Background and objective: Predicting the long-term expansion and remodeling of the left ventricle in patients is challenging task but it has the potential to be clinically very useful. Methods: In our study, we present machine learning models based on random forests, gradient boosting, and neural networks, used to track cardiac hypertrophy. We collected data from multiple patients, and then the model was trained using the patient's medical history and present level of cardiac health. We also demonstrate a physical-based model, using the finite element procedure to simulate the development of cardiac hypertrophy. Results: Our models were used to forecast the evolution of hypertrophy over six years. The machine learning model and finite element model provided similar results. Conclusions: The finite element model is much slower, but it's more accurate compared to the machine learning model since it's based on physical laws guiding the hypertrophy process. On the other hand, the machine learning model is fast but the results can be less trustworthy in some cases. Both of our models, enable us to monitor the development of the disease. Because of its speed machine learning model is more likely to be used in clinical practice. Further improvements to our machine learning model could be achieved by collecting data from finite element simulations, adding them to the dataset, and retraining the model. This can result in a fast and more accurate model combining the advantages of physical-based and machine learning modeling.

A Machine Learning Approach to Developing an Accurate Prediction of Maximal Heart Rate During Exercise Testing in Apparently Healthy Adults.

PURPOSE: Maximal heart rate (HR max ) continues to be an important measure of adequate effort during an exercise test. The aim of this study was to improve the accuracy of HR max prediction using a machine learning (ML) approach. METHODS: We used a sample from the Fitness Registry of the Importance of Exercise National Database, which included 17 325 apparently healthy individuals (81% males) who performed a maximal cardiopulmonary exercise test. Two standard formulas for HR max prediction were tested: Formula1 = 220 - age (yr), root-mean-squared error (RMSE) 21.9, relative root-mean-squared error (RRMSE) 1.1; and Formula2 = 209.3 - 0.72 × age (yr), RMSE 22.7 and RRMSE 1.1. For ML model prediction, we used age, weight, height, resting HR, and systolic and diastolic blood pressure. The following ML algorithms to predict HR max were applied: lasso regression (LR), neural networks (NN), support vector machine (SVM) and random forests (RF). An evaluation was performed using cross-validation and by computing the RMSE and RRMSE, Pearson correlation, and Bland-Altman plots. The best predictive model was explained with Shapley Additive Explanations (SHAP). RESULTS: The HR max for the cohort was 162 ± 20 bpm. All ML models improved HR max prediction and reduced RMSE and RRMSE compared with Formula1 (LR: 20.2%, NN: 20.4%, SVM: 22.2%, and RF: 24.7%). The predictions of all algorithms significantly correlated with HR max ( r = 0.49, 0.51, 0.54, 0.57, respectively; P < .001). Bland-Altman analysis demonstrated lower bias and 95% CI for all ML models in comparison with standard equations. The SHAP explanation showed a high impact of all selected variables. CONCLUSIONS: Machine learning, particularly the RF model, improved prediction of HR max using readily available measures. This approach should be considered for clinical application to refine HR max prediction.

SuperFormer: Continual learning superposition method for text classification.

One of the biggest challenges in continual learning domains is the tendency of machine learning models to forget previously learned information over time. While overcoming this issue, the existing approaches often exploit large amounts of additional memory and apply model forgetting mitigation mechanisms which substantially prolong the training process. Therefore, we propose a novel SuperFormer method that alleviates model forgetting, while spending negligible additional memory and time. We tackle the continual learning challenges in a learning scenario, where we learn different tasks in a sequential order. We compare our method against several prominent continual learning methods, i.e., EWC, SI, MAS, GEM, PSP, etc. on a set of text classification tasks. We achieve the best average performance in terms of AUROC and AUPRC (0.7% and 0.9% gain on average, respectively) and the lowest training time among all the methods of comparison. On average, our method reduces the total training time by a factor of 5.4-8.5 in comparison to similarly performing methods. In terms of the additional memory, our method is on par with the most memory-efficient approaches.

Disease Progression of Hypertrophic Cardiomyopathy: Modeling Using Machine Learning.

BACKGROUND: Cardiovascular disorders in general are responsible for 30% of deaths worldwide. Among them, hypertrophic cardiomyopathy (HCM) is a genetic cardiac disease that is present in about 1 of 500 young adults and can cause sudden cardiac death (SCD). OBJECTIVE: Although the current state-of-the-art methods model the risk of SCD for patients, to the best of our knowledge, no methods are available for modeling the patient's clinical status up to 10 years ahead. In this paper, we propose a novel machine learning (ML)-based tool for predicting disease progression for patients diagnosed with HCM in terms of adverse remodeling of the heart during a 10-year period. METHODS: The method consisted of 6 predictive regression models that independently predict future values of 6 clinical characteristics: left atrial size, left atrial volume, left ventricular ejection fraction, New York Heart Association functional classification, left ventricular internal diastolic diameter, and left ventricular internal systolic diameter. We supplemented each prediction with the explanation that is generated using the Shapley additive explanation method. RESULTS: The final experiments showed that predictive error is lower on 5 of the 6 constructed models in comparison to experts (on average, by 0.34) or a consortium of experts (on average, by 0.22). The experiments revealed that semisupervised learning and the artificial data from virtual patients help improve predictive accuracies. The best-performing random forest model improved R2 from 0.3 to 0.6. CONCLUSIONS: By engaging medical experts to provide interpretation and validation of the results, we determined the models' favorable performance compared to the performance of experts for 5 of 6 targets.

CAT-Site: Predicting Protein Binding Sites Using a Convolutional Neural Network.

Identifying binding sites on the protein surface is an important part of computer-assisted drug design processes. Reliable prediction of binding sites not only assists with docking algorithms, but it can also explain the possible side-effects of a potential drug as well as its efficiency. In this work, we propose a novel workflow for predicting possible binding sites of a ligand on a protein surface. We use proteins from the PDBbind and sc-PDB databases, from which we combine available ligand information for similar proteins using all the possible ligands rather than only a special sub-selection to generalize the work of existing research. After performing protein clustering and merging of ligands of similar proteins, we use a three-dimensional convolutional neural network that takes into account the spatial structure of a protein. Lastly, we combine ligandability predictions for points on protein surfaces into joint binding sites. Analysis of our model's performance shows that its achieved sensitivity is 0.829, specificity is 0.98, and F1 score is 0.517, and that for 54% of larger and pharmacologically relevant binding sites, the distance between their real and predicted centers amounts to less than 4 Å.

Cardiopulmonary assessment of patients diagnosed with Gaucher's disease type I.

BACKGROUND: Understanding the basis of the phenotypic variation in Gaucher's disease (GD) has proven to be challenging for efficient treatment. The current study examined cardiopulmonary characteristics of patients with GD type 1. METHODS: Twenty Caucasian subjects (8/20 female) with diagnosed GD type I (GD-S) and 20 age- and sex-matched healthy controls (C), were assessed (mean age GD-S: 32.6 ± 13.1 vs. C: 36.2 ± 10.6, p > .05) before the initiation of treatment. Standard echocardiography at rest was used to assess left ventricular ejection fraction (LVEF) and pulmonary artery systolic pressure (PASP). Cardiopulmonary exercise testing (CPET) was performed on a recumbent ergometer using a ramp protocol. RESULTS: LVEF was similar in both groups (GD-S: 65.1 ± 5.2% vs. C: 65.2 ± 5.2%, p > .05), as well as PAPS (24.1 ± 4.2 mmHg vs. C: 25.5 ± 1.3 mmHg, p > .05). GD-S had lower weight (p < .05) and worse CPET responses compared to C, including peak values of heart rate, oxygen consumption, carbondioxide production (VCO2 ), end-tidal pressure of CO2 , and O2 pulse, as well as HR reserve after 3 min of recovery and the minute ventilation/VCO2  slope. CONCLUSIONS: Patients with GD type I have an abnormal CPET response compared to healthy controls likely due to the complex pathophysiologic process in GD that impacts multiple systems integral to the physiologic response to exercise.

A machine learning-based risk stratification model for ventricular tachycardia and heart failure in hypertrophic cardiomyopathy.

BACKGROUND: Machine learning (ML) and artificial intelligence are emerging as important components of precision medicine that enhance diagnosis and risk stratification. Risk stratification tools for hypertrophic cardiomyopathy (HCM) exist, but they are based on traditional statistical methods. The aim was to develop a novel machine learning risk stratification tool for the prediction of 5-year risk in HCM. The goal was to determine if its predictive accuracy is higher than the accuracy of the state-of-the-art tools. METHOD: Data from a total of 2302 patients were used. The data were comprised of demographic characteristics, genetic data, clinical investigations, medications, and disease-related events. Four classification models were applied to model the risk level, and their decisions were explained using the SHAP (SHapley Additive exPlanations) method. Unwanted cardiac events were defined as sustained ventricular tachycardia occurrence (VT), heart failure (HF), ICD activation, sudden cardiac death (SCD), cardiac death, and all-cause death. RESULTS: The proposed machine learning approach outperformed the similar existing risk-stratification models for SCD, cardiac death, and all-cause death risk-stratification: it achieved higher AUC by 17%, 9%, and 1%, respectively. The boosted trees achieved the best performing AUC of 0.82. The resulting model most accurately predicts VT, HF, and ICD with AUCs of 0.90, 0.88, and 0.87, respectively. CONCLUSIONS: The proposed risk-stratification model demonstrates high accuracy in predicting events in patients with hypertrophic cardiomyopathy. The use of a machine-learning risk stratification model may improve patient management, clinical practice, and outcomes in general.

A computational pipeline for data augmentation towards the improvement of disease classification and risk stratification models: A case study in two clinical domains.

Virtual population generation is an emerging field in data science with numerous applications in healthcare towards the augmentation of clinical research databases with significant lack of population size. However, the impact of data augmentation on the development of AI (artificial intelligence) models to address clinical unmet needs has not yet been investigated. In this work, we assess whether the aggregation of real with virtual patient data can improve the performance of the existing risk stratification and disease classification models in two rare clinical domains, namely the primary Sjögren's Syndrome (pSS) and the hypertrophic cardiomyopathy (HCM), for the first time in the literature. To do so, multivariate approaches, such as, the multivariate normal distribution (MVND), and straightforward ones, such as, the Bayesian networks, the artificial neural networks (ANNs), and the tree ensembles are compared against their performance towards the generation of high-quality virtual data. Both boosting and bagging algorithms, such as, the Gradient boosting trees (XGBoost), the AdaBoost and the Random Forests (RFs) were trained on the augmented data to evaluate the performance improvement for lymphoma classification and HCM risk stratification. Our results revealed the favorable performance of the tree ensemble generators, in both domains, yielding virtual data with goodness-of-fit 0.021 and KL-divergence 0.029 in pSS and 0.029, 0.027 in HCM, respectively. The application of the XGBoost on the augmented data revealed an increase by 10.9% in accuracy, 10.7% in sensitivity, 11.5% in specificity for lymphoma classification and 16.1% in accuracy, 16.9% in sensitivity, 13.7% in specificity in HCM risk stratification.

Extracorporeal Hemadsorption versus Glucocorticoids during Cardiopulmonary Bypass: A Prospective, Randomized, Controlled Trial.

Extracorporeal hemadsorption may reduce inflammatory reaction in cardiopulmonary bypass (CPB) surgery. Glucocorticoids have been used during open-heart surgery for alleviation of systemic inflammation after CPB. We compared intraoperative hemadsorption and methylprednisolone, with usual care, during complex cardiac surgery on CPB, for inflammatory responses, hemodynamics, and perioperative course. Seventy-six patients with prolonged CPB were recruited and randomized, with 60 included in final analysis. Allocation was into three groups: Methylprednisolone (n = 20), Cytosorb (n = 20), and Control group (usual care, n = 20). Proinflammatory (TNF-α, IL-1ß, IL-6, and IL-8) and anti-inflammatory (IL-10) cytokines which complement C5a, CD64, and CD163 expression by immune cells were analyzed within the first five postoperative days, in addition to hemodynamic and clinical outcome parameters. Methylprednisolone group, compared to Cytosorb and Control had significantly lower levels of TNF-α (until the end of surgery, p < 0.001), IL-6 (until 48 h after surgery, p < 0.001), and IL-8 (until 24 h after surgery, p < 0.016). CD64 expression on monocytes was the highest in the Cytosorb group and lasted until the 5th postoperative day (p < 0.016). IL-10 concentration (until the end of surgery) and CD163 expression on monocytes (until 48 h after surgery) were the highest in the Methylprednisolone group (p < 0.016, for all measurements between three groups). No differences between groups in the cardiac index or clinical outcome parameters were found. Methylprednisolone more effectively ameliorates inflammatory responses after CPB surgery compared to hemadsorption and usual care. Hemadsorption compared with usual care causes higher prolonged expression of CD64 on monocytes but short lasting expression of CD163 on granulocytes. Hemadsorption with CytoSorb® was safe and well tolerated. This trial is registered with clinicaltrials.gov (NCT02666703).

Machine Learning for Predicting Cognitive Diseases: Methods, Data Sources and Risk Factors.

Machine learning and data mining approaches are being successfully applied to different fields of life sciences for the past 20 years. Medicine is one of the most suitable application domains for these techniques since they help model diagnostic information based on causal and/or statistical data and therefore reveal hidden dependencies between symptoms and illnesses. In this paper we give a detailed overview of the recent machine learning research and its applications for predicting cognitive diseases, especially the Alzheimer's disease, mild cognitive impairment and the Parkinson's disease. We survey different state-of-the-art methodological approaches, data sources and public data, and provide their comparative analysis. We conclude by identifying the open problems within the field that include an early detection of the cognitive diseases and inclusion of machine learning tools into diagnostic practice and therapy planning.

Influence of dexmedetomidine and lidocaine on perioperative opioid consumption in laparoscopic intestine resection: a randomized controlled clinical trial.

OBJECTIVE: The consumption of opioid analgesics could be reduced by the use of analgesics with different mechanisms of action. We investigated whether additional treatment with dexmedetomidine or lidocaine could reduce opioid consumption. METHODS: We randomized 59 study participants into three groups and examined: (i) fentanyl consumption, (ii) consumption of piritramide, and (iii) cognitive function and neuropathic pain. The control group received continuous propofol infusion and fentanyl boluses. Continuous intravenous infusion of dexmedetomidine (0.5 µg/kg/h) was administered to the dexmedetomidine group and lidocaine (1.5 mg/kg/h) was administered to the lidocaine group. RESULTS: No reduction in fentanyl consumption was observed among the groups. However, we noted a significantly lower consumption of piritramide on the first and second postoperative day in the lidocaine group. Total consumption of piritramide was significantly lower in the lidocaine group compared with the control group. CONCLUSIONS: Lidocaine and dexmedetomidine reduced intraoperative propofol consumption, while lidocaine reduced postoperative piritramide consumption. Clinical trial registration: NCT02616523.

Assessment of machine learning reliability methods for quantifying the applicability domain of QSAR regression models.

The vastness of chemical space and the relatively small coverage by experimental data recording molecular properties require us to identify subspaces, or domains, for which we can confidently apply QSAR models. The prediction of QSAR models in these domains is reliable, and potential subsequent investigations of such compounds would find that the predictions closely match the experimental values. Standard approaches in QSAR assume that predictions are more reliable for compounds that are "similar" to those in subspaces with denser experimental data. Here, we report on a study of an alternative set of techniques recently proposed in the machine learning community. These methods quantify prediction confidence through estimation of the prediction error at the point of interest. Our study includes 20 public QSAR data sets with continuous response and assesses the quality of 10 reliability scoring methods by observing their correlation with prediction error. We show that these new alternative approaches can outperform standard reliability scores that rely only on similarity to compounds in the training set. The results also indicate that the quality of reliability scoring methods is sensitive to data set characteristics and to the regression method used in QSAR. We demonstrate that at the cost of increased computational complexity these dependencies can be leveraged by integration of scores from various reliability estimation approaches. The reliability estimation techniques described in this paper have been implemented in an open source add-on package ( https://bitbucket.org/biolab/orange-reliability ) to the Orange data mining suite.

Mining data from hemodynamic simulations for generating prediction and explanation models.

One of the most common causes of human death is stroke, which can be caused by carotid bifurcation stenosis. In our work, we aim at proposing a prototype of a medical expert system that could significantly aid medical experts to detect hemodynamic abnormalities (increased artery wall shear stress). Based on the acquired simulated data, we apply several methodologies for1) predicting magnitudes and locations of maximum wall shear stress in the artery, 2) estimating reliability of computed predictions, and 3) providing user-friendly explanation of the model's decision. The obtained results indicate that the evaluated methodologies can provide a useful tool for the given problem domain.