Complex Predictive Models Research Articles

Application of Deep Learning for Heart Attack Prediction with Explainable Artificial Intelligence

Heart disease remains a leading cause of mortality worldwide, and the timely and accurate prediction of heart attack is crucial yet challenging due to the complexity of the condition and the limitations of traditional diagnostic methods. These challenges include the need for resource-intensive diagnostics and the difficulty in interpreting complex predictive models in clinical settings. In this study, we apply and compare the performance of five well-known Deep Learning (DL) models, namely Multi-Layer Perceptron (MLP), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and a Hybrid model, to a heart attack prediction dataset. Each model was properly tuned and evaluated using accuracy, precision, recall, F1-score, and Area Under the Receiver Operating Characteristic Curve (AUC) as performance metrics. Additionally, by integrating an Explainable Artificial intelligence (XAI) technique, specifically Shapley Additive Explanations (SHAP), we enhance the interpretability of the predictions, making them actionable for healthcare professionals and thereby enhancing clinical applicability. The experimental results revealed that the Hybrid model prevailed, achieving the highest performance across all metrics. Specifically, the Hybrid model attained an accuracy of 91%, precision of 89%, recall of 90%, F1-score of 89%, and an AUC of 0.95. These results highlighted the Hybrid model’s superior ability to predict heart attacks, attributed to its efficient handling of sequential data and long-term dependencies.

How Sports Involvement and Brand Fit Influence the Effectiveness of Sports Sponsorship from the Perspective of Predictive Coding Theory: An Event-Related Potential (ERP)-Based Study.

Background/Objectives: With the rapid expansion of the global sports market, the significance of sports sponsorship has attracted growing attention. However, during the golden age of the sports industry's development in China, international sports brand giants such as Nike, Adidas, and Under Armour have rapidly captured a substantial share of the Chinese sports consumer market through their distinctive product designs and varied marketing strategies. This has resulted in a highly competitive environment for China's sports goods industry. Therefore, fostering the improved development of domestic sports brands has become a crucial issue deserving of thorough scholarly investigation. This study examines how consumers' differing levels of sports involvement and the degree of fit between the sponsoring brand and the sponsored event affect their cognitive and emotional responses to sports sponsorships. Methods: By employing Predictive Coding Theory and ERP (event-related potential) brainwave technology, this study delves into the psychological and neurobiological levels to analyze the impact of consumer sports involvement on the processing of sponsorship information. Results: The results indicate significant differences in cognitive and emotional responses between high-involvement and low-involvement consumers. Additionally, the fit between the sponsoring brand and the sponsored event also significantly affects consumers' cognitive and emotional responses. These differences stem from consumers' complex and sophisticated predictive coding models. Conclusions: This study not only provides scientific evidence for sports brands in selecting and executing sponsorship activities, but also offers new perspectives for evaluating and optimizing sponsorship effectiveness.

A multiscale time-series decomposition learning for crude oil price forecasting

Crude oil price forecasting is important for market participants and policymakers. However, accurately tracking oil prices is quite a challenging task due to the complexity of temporal oil data and the nonlinear relationships involved in the forecasting task. In this study, a multiscale time-series decomposition learning framework is proposed to deal with this issue. First, a multiscale temporal processing module is designed to capture different frequency time-series patterns in historical data at various scales. Then, a multiscale decomposition technique is applied to decompose historical crude oil data into various temporal modes, involving global shared information across multiple scales, as well as local specific information that varies at each scale. Finally, a multiscale fusion mechanism is employed to combine these information, which are further used as inputs to construct nonlinear and complex predictive models for crude oil prices. A series of experiments conducted on Shanghai crude oil market demonstrate that the proposed approach outperforms several econometric and machine learning models.

Breast tumor classification using adam and optuna model optimization based on CNN architecture

Breast cancer presents a significant challenge due to its complexity and the urgency of the intervention required to prevent metastasis and potential fatality. This article highlights the innovative application of Convolutional Neural Networks (CNN) in breast tumor classification, marking substantial progress in the field. The key to this advancement is the collaboration among medical professionals, scientists, and artificial intelligence experts, which maximizes the potential of technology. The research involved three phases of training with varying proportions of training data. The first training phase achieved the highest accuracy rate of 99.72%, with an average accuracy of 99.05% in all three phases. Metrics such as precision, recall, and F1 score were also highly satisfactory, underscoring the model's efficacy in accurately classifying breast tumors. Future research aims to develop more complex and precise predictive models by incorporating larger and more representative datasets. This progression promises to improve understanding, prevention, and management of breast cancer, offering hope for significant advances in 2024 and beyond.

Advances in Explainable, Fair, and Trustworthy AI

This special issue encapsulates the multifaceted landscape of contemporary challenges and innovations in Artificial Intelligence (AI) and Machine Learning (ML), with a particular focus on issues related to explainability, fairness, and trustworthiness. The exploration begins with the computational intricacies of understanding and explaining the behavior of binary neurons within neural networks. Simultaneously, ethical dimensions in AI are scrutinized, emphasizing the nuanced considerations required in defining autonomous ethical agents. The pursuit of fairness is exemplified through frameworks and methodologies in machine learning, addressing biases and promoting trust, particularly in predictive policing systems. Human-agent interaction dynamics are elucidated, revealing the nuanced relationship between task allocation, performance, and user satisfaction. The imperative of interpretability in complex predictive models is highlighted, emphasizing a query-driven methodology. Lastly, in the context of trauma triage, the study underscores the delicate trade-off between model accuracy and practitioner-friendly interpretability, introducing innovative strategies to address biases and trust-related metrics.

Fusing Diverse Decision Rules in 3D-Radiomics for Assisting Diagnosis of Lung Adenocarcinoma

This study aimed to develop an interpretable diagnostic model for subtyping of pulmonary adenocarcinoma, including minimally invasive adenocarcinoma (MIA), adenocarcinoma in situ (AIS), and invasive adenocarcinoma (IAC), by integrating 3D-radiomic features and clinical data. Data from multiple hospitals were collected, and 10 key features were selected from 1600 3D radiomic signatures and 11 radiological features. Diverse decision rules were extracted using ensemble learning methods (gradient boosting, random forest, and AdaBoost), fused, ranked, and selected via RuleFit and SHAP to construct a rule-based diagnostic model. The model’s performance was evaluated using AUC, precision, accuracy, recall, and F1-score and compared with other models. The rule-based diagnostic model exhibited excellent performance in the training, testing, and validation cohorts, with AUC values of 0.9621, 0.9529, and 0.8953, respectively. This model outperformed counterparts relying solely on selected features and previous research models. Specifically, the AUC values for the previous research models in the three cohorts were 0.851, 0.893, and 0.836. It is noteworthy that individual models employing GBDT, random forest, and AdaBoost demonstrated AUC values of 0.9391, 0.8681, and 0.9449 in the training cohort, 0.9093, 0.8722, and 0.9363 in the testing cohort, and 0.8440, 0.8640, and 0.8750 in the validation cohort, respectively. These results highlight the superiority of the rule-based diagnostic model in the assessment of lung adenocarcinoma subtypes, while also providing insights into the performance of individual models. Integrating diverse decision rules enhanced the accuracy and interpretability of the diagnostic model for lung adenocarcinoma subtypes. This approach bridges the gap between complex predictive models and clinical utility, offering valuable support to healthcare professionals and patients.

RETRACTED ARTICLE: Wildfire risk exploration: leveraging SHAP and TabNet for precise factor analysis

BackgroundUnderstanding the intricacies of wildfire impact across diverse geographical landscapes necessitates a nuanced comprehension of fire dynamics and areas of vulnerability, particularly in regions prone to high wildfire risks. Machine learning (ML) stands as a formidable ally in addressing the complexities associated with predicting and mapping these risks, offering advanced analytical capabilities. Nevertheless, the reliability of such ML approaches is heavily contingent on the integrity of data and the robustness of training protocols. The scientific community has raised concerns about the transparency and interpretability of ML models in the context of wildfire management, recognizing the need for these models to be both accurate and understandable. The often-opaque nature of complex ML algorithms can obscure the rationale behind their outputs, making it imperative to prioritize clarity and interpretability to ensure that model predictions are not only precise but also actionable. Furthermore, a thorough evaluation of model performance must account for multiple critical factors to ensure the utility and dependability of the results in practical wildfire suppression and management strategies.ResultsThis study unveils a sophisticated spatial deep learning framework grounded in TabNet technology, tailored specifically for delineating areas susceptible to wildfires. To elucidate the predictive interplay between the model’s outputs and the contributing variables across a spectrum of inputs, we embark on an exhaustive analysis using SHapley Additive exPlanations (SHAP). This approach affords a granular understanding of how individual features sway the model’s predictions. Furthermore, the robustness of the predictive model is rigorously validated through 5-fold cross-validation techniques, ensuring the dependability of the findings. The research meticulously investigates the spatial heterogeneity of wildfire susceptibility within the designated study locale, unearthing pivotal insights into the nuanced fabric of fire risk that is distinctly local in nature.ConclusionUtilizing SHapley Additive exPlanations (SHAP) visualizations, this research meticulously identifies key variables, quantifies their importance, and demystifies the decision-making mechanics of the model. Critical factors, including temperature, elevation, the Normalized Difference Vegetation Index (NDVI), aspect, and wind speed, are discerned to have significant sway over the predictions of wildfire susceptibility. The findings of this study accentuate the criticality of transparency in modeling, which facilitates a deeper understanding of wildfire risk factors. By shedding light on the significant predictors within the models, this work enhances our ability to interpret complex predictive models and drives forward the field of wildfire risk management, ultimately contributing to the development of more effective prevention and mitigation strategies.

Discovery of sparse, reliable omic biomarkers with Stabl

Adoption of high-content omic technologies in clinical studies, coupled with computational methods, has yielded an abundance of candidate biomarkers. However, translating such findings into bona fide clinical biomarkers remains challenging. To facilitate this process, we introduce Stabl, a general machine learning method that identifies a sparse, reliable set of biomarkers by integrating noise injection and a data-driven signal-to-noise threshold into multivariable predictive modeling. Evaluation of Stabl on synthetic datasets and five independent clinical studies demonstrates improved biomarker sparsity and reliability compared to commonly used sparsity-promoting regularization methods while maintaining predictive performance; it distills datasets containing 1,400–35,000 features down to 4–34 candidate biomarkers. Stabl extends to multi-omic integration tasks, enabling biological interpretation of complex predictive models, as it hones in on a shortlist of proteomic, metabolomic and cytometric events predicting labor onset, microbial biomarkers of pre-term birth and a pre-operative immune signature of post-surgical infections. Stabl is available at https://github.com/gregbellan/Stabl.

РАЗВИТИЕ ИССЛЕДОВАНИЙ В СФЕРЕ РЫНКА НЕДВИЖИМОСТИ

The article analyzes in detail the dynamics and complexity of the real estate market with an emphasis on forecasting methodology and its practical application in an economic context. The main theme is the integration of various forecasting models that predict trends and fluctuations in real estate prices based on a multidisciplinary approach combining economic theory, statistical data and market analysis. First, the basic principles defining the real estate economy, its differences from other sectors in terms of market conditions and consumer behavior patterns are outlined. The importance of location and consumer characteristics in determining the value of the property is emphasized. It is explained how these factors distinguish the real estate market into a separate area of research. The challenges and imperfections of the real estate market, such as lack of competitiveness and inefficiency in comparison with finance, are shown. It is argued that these characteristics require complex predictive models to improve market efficiency and investor decisionmaking. In this context, various statistical and analytical tools that have been improved over the years to improve the accuracy of forecasts in the real estate market are discussed. Theoretical judgments are supplemented by examples and empirical data confirming the effectiveness of models in various economic conditions. The author advocates continuous innovative development and adaptation of forecasting techniques used to analyze the real estate market. The need for a multidisciplinary approach is emphasized, taking into account economic indicators, opinions of market participants and geopolitical factors in forecasts. This comprehensive strategy is presented as key to navigating the complexities of global real estate markets, taking advantage of investment opportunities in an everchanging economic environment.

Consumption–Production Profile Categorization in Energy Communities

Energy Transition is changing the renewable energy participation in new distributed generation systems like the Local Energy Markets. Due to its inherent intermittent and variable nature, forecasting production and consumption load profiles will be more challenging and demand more complex predictive models. This paper analyzes the production, consumption load profile, and storage headroom% of the Cornwall Local Energy Market, using advanced statistical time series methods to optimize the opportunity market the storage units provide. These models also help the Energy Community storage reserves to meet contract conditions with the Distribution Network Operator. With this more accurate and detailed knowledge, all sites from this Local Energy Market will benefit more from their installation by optimizing their energy consumption, production, and storage. This better accuracy will make the Local Energy Market more fluid and safer, creating a flexible system that will guarantee the technical quality of the product for the whole community. The training of several SARIMAX, Exponential Smoothing, and Temporal Causal models improved the fitness of consumption, production, and headroom% time series. These models properly decomposed the time series in trend, seasonality, and stochastic dynamic components that help us to understand how the Local Energy Market consumes, produces, and stores energy. The model design used all power flows and battery energy storage system state-of-charge site characteristics at daily and hourly granularity levels. All model building follows an analytical methodology detailed step by step. A benchmark between these sequence models and the incumbent forecasting models utilized by the Energy Community shows a better performance measured with model error reduction. The best models present mean squared error reduction between 88.89% and 99.93%, while the mean absolute error reduction goes from 65.73% to 97.08%. These predictive models built at different prediction scales will help the Energy Communities better contribute to the Network Management and optimize their energy and power management performance. In conclusion, the expected outcome of these implementations is a cost-optimal management of the Local Energy Market and its contribution to the needed new Flexibility Electricity System Scheme, extending the adoption of renewable energies.

Errors of professional activity of transport system operators in conditions of high information load

Introduction. Ensuring traffic safety on the railway network involves the development of complex predictive models that take into account both the state of technology and the influence of the "human factor" in order to minimize the risk of erroneous actions and prevent accidents.
 The study aims to research the mistakes of railway transport operators (using the example of locomotive crew workers).
 Materials and methods. The authors have conducted the study on the basis of analysis of the database of erroneous actions of employees of locomotive crews (2020–2021), submitted by the Research Institute of Railway Transport (according to the agreement on joint scientific cooperation between the Izmerov Research Institute of Occupational Health and Research Institute of Railway Transport).
 Results. The database recorded 104,515 cases attributed to 291 types of erroneous actions with qualitative differences that reflected the degree of risk to the safety of train traffic. The most common categories of violations related to the maintenance of train documentation, errors in locomotive control, non-compliance with the speed limit, disconnection of locomotive safety devices, speeding, brake control violations, violations of the rules of negotiations, the passage of a forbidding signal. In relation to employees who committed violations, the most common disciplinary penalties were identified (analysis of the case in technical classes, deprivation of an employee of a bonus in the amount of 25, 50 and 100%). In a number of cases, several different disciplinary penalties were applied to one employee at the depot for the same violation.
 Limitations. Inability to get access to the protocols of psychophysiological examination of employees of locomotive crews.
 Conclusion. The obtained scientific results allow us to assert that the productivity and safety of the operator's activity correlates with the level of professional training and the mechanisms of its inclusion in the activity. The prospects for further research of the problem of ensuring the safety of locomotive crews' employees consist in a detailed study of the combined effect of factors of the working environment and the labor process in conditions of increasing labor intensity, increasing speeds and volumes of railway transportation. It is necessary to study the ratio and role of such indicators as reliability, professional suitability and readiness of the employee to perform functional duties in the event of an emergency.
 Ethics. The study was carried out on the basis of documents submitted by JSC "Russian Railways". The documents contained the number and typology of mistakes made by employees of locomotive crews in the period 2020-2021. The scientists have conducted the study in accordance with the Ethical principles of anonymity, confidentiality, respect for the rights of the subject of the study, the principles of scientific and impartiality.

Explaining machine-learning models for gamma-ray detection and identification.

As more complex predictive models are used for gamma-ray spectral analysis, methods are needed to probe and understand their predictions and behavior. Recent work has begun to bring the latest techniques from the field of Explainable Artificial Intelligence (XAI) into the applications of gamma-ray spectroscopy, including the introduction of gradient-based methods like saliency mapping and Gradient-weighted Class Activation Mapping (Grad-CAM), and black box methods like Local Interpretable Model-agnostic Explanations (LIME) and SHapley Additive exPlanations (SHAP). In addition, new sources of synthetic radiological data are becoming available, and these new data sets present opportunities to train models using more data than ever before. In this work, we use a neural network model trained on synthetic NaI(Tl) urban search data to compare some of these explanation methods and identify modifications that need to be applied to adapt the methods to gamma-ray spectral data. We find that the black box methods LIME and SHAP are especially accurate in their results, and recommend SHAP since it requires little hyperparameter tuning. We also propose and demonstrate a technique for generating counterfactual explanations using orthogonal projections of LIME and SHAP explanations.

HIPPYlib-MUQ: A Bayesian Inference Software Framework for Integration of Data with Complex Predictive Models under Uncertainty

Bayesian inference provides a systematic framework for integration of data with mathematical models to quantify the uncertainty in the solution of the inverse problem. However, the solution of Bayesian inverse problems governed by complex forward models described bypartial differential equations (PDEs)remains prohibitive with black-boxMarkov chain Monte Carlo (MCMC)methods. We present hIPPYlib-MUQ, an extensible and scalable software framework that contains implementations of state-of-the art algorithms aimed to overcome the challenges of high-dimensional, PDE-constrained Bayesian inverse problems. These algorithms accelerate MCMC sampling by exploiting the geometry and intrinsic low-dimensionality of parameter space via derivative information and low rank approximation. The software integrates two complementary open-source software packages, hIPPYlib and MUQ. hIPPYlib solves PDE-constrained inverse problems using automatically-generated adjoint-based derivatives, but it lacks full Bayesian capabilities. MUQ provides a spectrum of powerful Bayesian inversion models and algorithms, but expects forward models to come equipped with gradients and Hessians to permit large-scale solution. By combining these two complementary libraries, we created a robust, scalable, and efficient software framework that realizes the benefits of each and allows us to tackle complex large-scale Bayesian inverse problems across a broad spectrum of scientific and engineering disciplines. To illustrate the capabilities of hIPPYlib-MUQ, we present a comparison of a number of MCMC methods available in the integrated software on several high-dimensional Bayesian inverse problems. These include problems characterized by both linear and nonlinear PDEs, various noise models, and different parameter dimensions. The results demonstrate that large (∼ 50×) speedups over conventional black box and gradient-based MCMC algorithms can be obtained by exploiting Hessian information (from the log-posterior), underscoring the power of the integrated hIPPYlib-MUQ framework.

Technical Perspective on 'R2T: Instance-optimal Truncation for Differentially Private Query Evaluation with Foreign Keys

Increased use of data to inform decision making has brought with it a rising awareness of the importance of privacy, and the need for appropriate mitigations to be put in place to protect the interests of individuals whose data is being processed. From the demographic statistics that are produced by national censuses, to the complex predictive models built by "big tech" companies, data is the fuel that powers these applications. A majority of such uses rely on data that is derived from the properties and actions of individual people. This data is therefore considered sensitive, and in need of protections to prevent inappropriate use or disclosure. Some protections come from enforcing policies, access control, and contractual agreements. But in addition, we also seek technical interventions: definitions and algorithms that can be applied by computer systems in order to protect the private information while still enabling the intended use.

Editorial: Toyota Production System practices asFast‐and‐Frugalheuristics

Editorial: Toyota Production System practices as<scp>Fast‐and‐Frugal</scp>heuristics

Unsupervised Transformer-Based Anomaly Detection in ECG Signals

Anomaly detection is one of the basic issues in data processing that addresses different problems in healthcare sensory data. Technology has made it easier to collect large and highly variant time series data; however, complex predictive analysis models are required to ensure consistency and reliability. With the rise in the size and dimensionality of collected data, deep learning techniques, such as autoencoder (AE), recurrent neural networks (RNN), and long short-term memory (LSTM), have gained more attention and are recognized as state-of-the-art anomaly detection techniques. Recently, developments in transformer-based architecture have been proposed as an improved attention-based knowledge representation scheme. We present an unsupervised transformer-based method to evaluate and detect anomalies in electrocardiogram (ECG) signals. The model architecture comprises two parts: an embedding layer and a standard transformer encoder. We introduce, implement, test, and validate our model in two well-known datasets: ECG5000 and MIT-BIH Arrhythmia. Anomalies are detected based on loss function results between real and predicted ECG time series sequences. We found that the use of a transformer encoder as an alternative model for anomaly detection enables better performance in ECG time series data. The suggested model has a remarkable ability to detect anomalies in ECG signal and outperforms deep learning approaches found in the literature on both datasets. In the ECG5000 dataset, the model can detect anomalies with 99% accuracy, 99% F1-score, 99% AUC score, 98.1% recall, and 100% precision. In the MIT-BIH Arrhythmia dataset, the model achieved an accuracy of 89.5%, F1 score of 92.3%, AUC score of 93%, recall of 98.2%, and precision of 87.1%.

Less is More: Brain Functional Connectivity Empowered Generalisable Intention Classification with Task-relevant Channel Selection.

Electroencephalography (EEG) signals are gaining popularity in Brain-Computer Interface (BCI)-based rehabilitation and neural engineering applications thanks to their portability and availability. Inevitably, the sensory electrodes on the entire scalp would collect signals irrelevant to the particular BCI task, increasing the risks of overfitting in machine learning-based predictions. While this issue is being addressed by scaling up the EEG datasets and handcrafting the complex predictive models, this also leads to increased computation costs. Moreover, the model trained for one set of subjects cannot easily be adapted to other sets due to inter-subject variability, which creates even higher over-fitting risks. Meanwhile, despite previous studies using either convolutional neural networks (CNNs) or graph neural networks (GNNs) to determine spatial correlations between brain regions, they fail to capture brain functional connectivity beyond physical proximity. To this end, we propose 1) removing task-irrelevant noises instead of merely complicating models; 2) extracting subject-invariant discriminative EEG encodings, by taking functional connectivity into account. Specifically, we construct a task-adaptive graph representation of the brain network based on topological functional connectivity rather than distance-based connections. Further, non-contributory EEG channels are excluded by selecting only functional regions relevant to the corresponding intention. We empirically show that the proposed approach outperforms the state-of-the-art, with around 1% and 11% improvements over CNN-based and GNN-based models, on performing motor imagery predictions. Also, the task-adaptive channel selection demonstrates similar predictive performance with only 20% of raw EEG data, suggesting a possible shift in direction for future works other than simply scaling up the model.

ECG and Biomarker Profile in Patients with Acute Heart Failure: A Pilot Study.

Biomarkers, electrocardiogram (ECG) and Holter ECG are basic, accessible and feasible cardiac investigations. The combination of their results may lead to a more complex predictive model that may improve the clinical approach in acute heart failure (AHF). The main objective was to investigate which ECG parameters are correlated with the usual cardiac biomarkers (prohormone N-terminal proBNP, high-sensitive cardiac troponin I) in patients with acute heart failure, in a population from Romania. The relationship between certain ECG parameters and cardiac biomarkers may support future research on their combined prognostic value. In this prospective case-control study were included 49 patients with acute heart failure and 31 participants in the control group. For all patients we measured levels of prohormone N-terminal proBNP (NT-proBNP), high-sensitive cardiac troponin I (hs-cTnI) and MB isoenzyme of creatine phosphokinase (CK-MB) and evaluated the 12-lead ECG and 24 h Holter monitoring. Complete clinical and paraclinical evaluation was performed. NT-proBNP level was significantly higher in patients with AHF (p &lt; 0.001). In patients with AHF, NT-proBNP correlated with cQTi (p = 0.027), pathological Q wave (p = 0.029), complex premature ventricular contractions (PVCs) (p = 0.034) and ventricular tachycardia (p = 0.048). Hs-cTnI and CK-MB were correlated with ST-segment modification (p = 0.038; p = 0.018) and hs-cTnI alone with complex PVCs (p = 0.031). The statistical relationships found between cardiac biomarkers and ECG patterns support the added value of ECG in the diagnosis of AHF. We emphasize the importance of proper ECG analysis of more subtle parameters that can easily be missed. As a non-invasive technique, ECG can be used in the outpatient setting as a warning signal, announcing the acute decompensation of HF. In addition, the information provided by the ECG complements the biomarker results, supporting the diagnosis of AHF in cases of dyspnea of uncertain etiology. Further studies are needed to confirm long-term prognosis in a multi-marker approach.

NIMG-69. MRI T2 RELAXIVITY IN THE PREDICTION OF PROGRESSION-FREE SURVIVAL IN NEWLY-DIAGNOSED GLIOBLASTOMA: WHOLE BRAIN MEASUREMENTS AND DEEP LEARNING

Abstract Glioblastoma (GBM) is the most common and aggressive primary brain tumour; we aimed to evaluate T2 relaxivity in their follow. Fifty-one newly diagnosed GBM followed at our department after surgery and combined chemo-radiotherapy were included (&amp;gt; 180 days progression-free survival [PFS] as determined by RANO, ≥ 3 MRI examinations in the PFS interval). All data used were prior to progression. T2 relaxation rates (1/T2f) were calculated voxel-wise assuming mono-exponential decay. Whole-brain (WB) values were extracted including skewness, kurtosis, mean, median, 15 and 85 percentile histogram values and variance. We additionally segmented the 1/T2f volumes using a 5 compartment Gaussian mixture model, producing mean, variance and percent per component. To evaluate predictive potential with respect to PFS, we used a deep long short-term memory (LSTM) model in tensorflow (4 timesteps). Twenty subjects were included in the predictive model, as inclusion criteria differed (progression, ≤ 120 days between all MRIs [mean 44.3 days], ≥ 4 PFS MRIs). Two models were tested: a regression model (days to progression) and a classification model (±18 months PFS; models differed in output layer). Both models were run 10 times; mean results are presented. We found WB median 1/T2f correlated with PFS, as values decreased prior to progression. WB median 1/T2f linear regression slopes also differed in progression vs. pseudo-progression. The deep LSTM regression model achieved an R2 of 86%. The deep LSTM classification model achieved a mean macro precision of 85%, recall 84% and F1 accuracy of 84% in classifying progression/no progression at an 18 month cutoff. We found very intriguing results with WB median 1/T2f measurements in distinguishing progression from pseudo-progression, and in suggesting progression despite otherwise unremarkable imaging. A more complex predictive model, using a fully-automated segmentation method followed by deep learning, was very promising. Our results may find utility in the monitoring of GBM patients.

CMB-HUNT: Automatic detection of cerebral microbleeds using a deep neural network

Cerebral microbleeds (CMBs) are gaining increasing interest due to their importance in diagnosing cerebral small vessel diseases. However, manual inspection of CMBs is time-consuming and prone to human error. Existing automated or semi-automated solutions still have insufficient detection sensitivity and specificity. Furthermore, they frequently use more than one magnetic resonance imaging modality, but these are not always available. The majority of AI-based solutions use either numeric or image data, which may not provide sufficient information about the true nature of CMBs.This paper proposes a deep neural network with multi-type input data for automated CMB detection (CMB-HUNT) using only susceptibility-weighted imaging data (SWI). Combination of SWIs and radiomic-type numerical features allowed us to identify CMBs with high accuracy without the need for additional imaging modalities or complex predictive models. Two independent datasets were used: one with 304 patients (39 with CMBs) for training and internal system validation and one with 61 patients (21 with CMBs) for external validation.For the hold-out testing dataset, CMB-HUNT reached a sensitivity of 90.0%. As results of testing showed, CMB-HUNT outperforms existing methods in terms of the number of FPs per case, which is the lowest reported thus far (0.54 FPs/patient). The proposed system was successfully applied to the independent validation set, reaching a sensitivity of 91.5% with 1.9 false positives per patient and proving its generalization potential. The results were comparable to previous studies. Our research confirms the usefulness of deep learning solutions for CMB detection based only on one MRI modality.