Machine learning models and over-fitting considerations

A gated recurrent unit model based on ultrasound images of dynamic tongue movement for determining the severity of obstructive sleep apnea

Obstructive sleep apnea (OSA) presents as a respiratory disorder characterized by recurrent upper pharyngeal airway collapse during sleep. Dynamic tongue movement (DTM) analysis emerges as a promising avenue for elucidating the pathophysiological underpinnings of OSA, thereby facilitating its diagnosis. Recent endeavors have utilized artificial intelligence techniques to categorize OSA severity leveraging electrocardiography and blood oxygen saturation data. Nonetheless, the integration of ultrasound (US) imaging of the tongue remains largely untapped in the development of machine learning models aimed at determining the severity of OSA. This study endeavors to bridge this gap by capturing US images of DTM dynamics during wakefulness, encompassing transitions from normal breathing (NB) to the performance of the Müller maneuver (MM) in a cohort of 53 patients. Leveraging the modified optical flow method (MOFM), the trajectories of patients' DTM were tracked, facililtating the extraction of 27 parameters vital for model training. These parameters encompassed nine-point lateral movement, nine-point axial movement, and nine-point total displacement of the tongue, resulting in a dataset of 186,030 samples. The gated recurrent unit (GRU) method, renowned for its efficacy in motion tracking, was employed for model development in this study. Validation of the developed model was conducted via stratified k-fold cross-validation (SCV). The systems' overall performance in classifying OSA severity, as quantified by mean accuracy (MA), yielded a value of 43.49%. This pilot investigation marks an exploratory endeavor into the utilization of artificial intelligence for the classification of OSA severity based on US images and dynamic movement patterns. This novel model holds potential to assist clinicians in categorizing OSA severity and guiding the selection of pertinent treatment modalities tailored to the individual needs of patients afflicted with OSA.

Harnessing artificial intelligence for predicting and managing postoperative pain: a narrative literature review

PURPOSE OF REVIEW

This review examines recent research on artificial intelligence focusing on machine learning (ML) models for predicting postoperative pain outcomes. We also identify technical, ethical, and practical hurdles that demand continued investigation and research.

RECENT FINDINGS

Current ML models leverage diverse datasets, algorithmic techniques, and validation methods to identify predictive biomarkers, risk factors, and phenotypic signatures associated with increased acute and chronic postoperative pain and persistent opioid use. ML models demonstrate satisfactory performance to predict pain outcomes and their prognostic trajectories, identify modifiable risk factors and at-risk patients who benefit from targeted pain management strategies, and show promise in pain prevention applications. However, further evidence is needed to evaluate the reliability, generalizability, effectiveness, and safety of ML-driven approaches before their integration into perioperative pain management practices.

SUMMARY

Artificial intelligence (AI) has the potential to enhance perioperative pain management by providing more accurate predictive models and personalized interventions. By leveraging ML algorithms, clinicians can better identify at-risk patients and tailor treatment strategies accordingly. However, successful implementation needs to address challenges in data quality, algorithmic complexity, and ethical and practical considerations. Future research should focus on validating AI-driven interventions in clinical practice and fostering interdisciplinary collaboration to advance perioperative care.

The role of various physiological and bioelectrical parameters for estimating the weight status in infants and juveniles cohort from the Southern Cuba region: a machine learning study

OBJECTIVE

The search for other indicators to assess the weight status of individuals is important as it may provide more accurate information and assist in personalized medicine.This work is aimed to develop a machine learning predictions of weigh status derived from bioimpedance measurements and other physical parameters of healthy infant juvenile cohort from the Southern Cuba Region, Santiago de Cuba.

METHODS

The volunteers were selected between 2002 and 2008, ranging in age between 2 and 18 years old. In total, 393 female and male infant and juvenile individuals are studied. The bioimpedance parameters are obtained by measuring standard tetrapolar whole-body configuration. A classification model are performed, followed by a prediction of other bioparameters influencing the weight status.

RESULTS

The results obtained from the classification model indicate that fat-free mass, reactance, and corrected resistance primarily influence the weight status of the studied population. Specifically, the regression model demonstrates that other bioparameters derived from impedance measurements can be highly accurate in estimating weight status.

CONCLUSION

The classification and regression predictive models developed in this work are of the great importance for accessing to the weigh status with high accuracy of younger individuals at the Oncological Hospital in Santiago de Cuba, Cuba.

Hyperspectral and Fluorescence Imaging Approaches for Nondestructive Detection of Rice Chlorophyll

Estimating and monitoring chlorophyll content is a critical step in crop spectral image analysis. The quick, non-destructive assessment of chlorophyll content in rice leaves can optimize nitrogen fertilization, benefit the environment and economy, and improve rice production management and quality. In this research, spectral analysis of rice leaves is performed using hyperspectral and fluorescence spectroscopy for the detection of chlorophyll content in rice leaves. This study generated ninety experimental spectral datasets by collecting rice leaf samples from a farm in Sichuan Province, China. By implementing a feature extraction algorithm, this study compresses redundant spectral bands and subsequently constructs machine learning models to reveal latent correlations among the extracted features. The prediction capabilities of six feature extraction methods and four machine learning algorithms in two types of spectral data are examined, and an accurate method of predicting chlorophyll concentration in rice leaves was devised. The IVSO-IVISSA (Iteratively Variable Subset Optimization-Interval Variable Iterative Space Shrinkage Approach) quadratic feature combination approach, based on fluorescence spectrum data, has the best prediction performance among the CNN+LSTM (Convolutional Neural Network Long Short-Term Memory) algorithms, with corresponding RMSE-Train (Root Mean Squared Error), RMSE-Test, and RPD (Ratio of standard deviation of the validation set to standard error of prediction) indexes of 0.26, 0.29, and 2.64, respectively. We demonstrated in this study that hyperspectral and fluorescence spectroscopy, when analyzed with feature extraction and machine learning methods, provide a new avenue for rapid and non-destructive crop health monitoring, which is critical to the advancement of smart and precision agriculture.

Editorial: Using machine learning to predict significant fibrosis in metabolic dysfunction-associated steatotic liver disease-authors' reply

This article is linked to Verma et al papers. To view these articles, visit https://doi.org/10.1111/apt.17891 and https://doi.org/10.1111/apt.17902

What does it mean for a clinical AI to be just: conflicts between local fairness and being fit-for-purpose?

There have been repeated calls to ensure that clinical artificial intelligence (AI) is not discriminatory, that is, it provides its intended benefit to all members of society irrespective of the status of any protected characteristics of individuals in whose healthcare the AI might participate. There have also been repeated calls to ensure that any clinical AI is tailored to the local population in which it is being used to ensure that it is fit-for-purpose. Yet, there might be a clash between these two calls since tailoring an AI to a local population might reduce its effectiveness when the AI is used in the care of individuals who have characteristics which are not represented in the local population. Here, I explore the bioethical concept of local fairness as applied to clinical AI. I first introduce the discussion concerning fairness and inequalities in healthcare and how this problem has continued in attempts to develop AI-enhanced healthcare. I then discuss various technical aspects which might affect the implementation of local fairness. Next, I introduce some rule of law considerations into the discussion to contextualise the issue better by drawing key parallels. I then discuss some potential technical solutions which have been proposed to address the issue of local fairness. Finally, I outline which solutions I consider most likely to contribute to a fit-for-purpose and fair AI.

CSANet: a lightweight channel and spatial attention neural network for grading diabetic retinopathy with optical coherence tomography angiography

Background

Diabetic retinopathy (DR) is one of the most common eye diseases. Convolutional neural networks (CNNs) have proven to be a powerful tool for learning DR features; however, accurate DR grading remains challenging due to the small lesions in optical coherence tomography angiography (OCTA) images and the small number of samples.

Methods

In this article, we developed a novel deep-learning framework to achieve the fine-grained classification of DR; that is, the lightweight channel and spatial attention network (CSANet). Our CSANet comprises two modules: the baseline model, and the hybrid attention module (HAM) based on spatial attention and channel attention. The spatial attention module is used to mine small lesions and obtain a set of spatial position weights to address the problem of small lesions being ignored during the convolution process. The channel attention module uses a set of channel weights to focus on useful features and suppress irrelevant features.

Results

The extensive experimental results for the OCTA-DR and diabetic retinopathy analysis challenge (DRAC) 2022 data sets showed that the CSANet achieved state-of-the-art DR grading results, showing the effectiveness of the proposed model. The CSANet had an accuracy rate of 97.41% for the OCTA-DR data set and 85.71% for the DRAC 2022 data set.

Conclusions

Extensive experiments using the OCTA-DR and DRAC 2022 data sets showed that the proposed model effectively mitigated the problems of mutual confusion between DRs of different severity and small lesions being neglected in the convolution process, and thus improved the accuracy of DR classification.

The value of machine learning approaches in the diagnosis of early gastric cancer: a systematic review and meta-analysis

BACKGROUND

The application of machine learning (ML) for identifying early gastric cancer (EGC) has drawn increasing attention. However, there lacks evidence-based support for its specific diagnostic performance. Hence, this systematic review and meta-analysis was implemented to assess the performance of image-based ML in EGC diagnosis.

METHODS

We performed a comprehensive electronic search in PubMed, Embase, Cochrane Library, and Web of Science up to September 25, 2022. QUADAS-2 was selected to judge the risk of bias of included articles. We did the meta-analysis using a bivariant mixed-effect model. Sensitivity analysis and heterogeneity test were performed.

RESULTS

Twenty-one articles were enrolled. The sensitivity (SEN), specificity (SPE), and SROC of ML-based models were 0.91 (95% CI: 0.87-0.94), 0.85 (95% CI: 0.81-0.89), and 0.94 (95% CI: 0.39-1.00) in the training set and 0.90 (95% CI: 0.86-0.93), 0.90 (95% CI: 0.86-0.92), and 0.96 (95% CI: 0.19-1.00) in the validation set. The SEN, SPE, and SROC of EGC diagnosis by non-specialist clinicians were 0.64 (95% CI: 0.56-0.71), 0.84 (95% CI: 0.77-0.89), and 0.80 (95% CI: 0.29-0.97), and those by specialist clinicians were 0.80 (95% CI: 0.74-0.85), 0.88 (95% CI: 0.85-0.91), and 0.91 (95% CI: 0.37-0.99). With the assistance of ML models, the SEN of non-specialist physicians in the diagnosis of EGC was significantly improved (0.76 vs 0.64).

CONCLUSION

ML-based diagnostic models have greater performance in the identification of EGC. The diagnostic accuracy of non-specialist clinicians can be improved to the level of the specialists with the assistance of ML models. The results suggest that ML models can better assist less experienced clinicians in diagnosing EGC under endoscopy and have broad clinical application value.

Integrated web portal for non-destructive salt sensitivity detection of Camelina sativa seeds using fluorescent and visible light images coupled with machine learning algorithms

Climate change has created unprecedented stresses in the agricultural sector, driving the necessity of adapting agricultural practices and developing novel solutions to the food crisis. Camelina sativa (Camelina) is a recently emerging oilseed crop with high nutrient-density and economic potential. Camelina seeds are rich in essential fatty acids and contain potent antioxidants required to maintain a healthy diet. Camelina seeds are equally amenable to economic applications such as jet fuel, biodiesel and high-value industrial lubricants due to their favorable proportions of unsaturated fatty acids. High soil salinity is one of the major abiotic stresses threatening the yield and usability of such crops. A promising mitigation strategy is automated, non-destructive, image-based phenotyping to assess seed quality in the food manufacturing process. In this study, we evaluate the effectiveness of image-based phenotyping on fluorescent and visible light images to quantify and qualify Camelina seeds. We developed a user-friendly web portal called SeedML that can uncover key morpho-colorimetric features to accurately identify Camelina seeds coming from plants grown in high salt conditions using a phenomics platform equipped with fluorescent and visible light cameras. This portal may be used to enhance quality control, identify stress markers and observe yield trends relevant to the agricultural sector in a high throughput manner. Findings of this work may positively contribute to similar research in the context of the climate crisis, while supporting the implementation of new quality controls tools in the agri-food domain.

Using machine learning methods to analyze the association between urinary polycyclic aromatic hydrocarbons and chronic bowel disorders in American adults

The etiology of chronic bowel disorders is multifaceted, with environmental exposure to harmful substances potentially playing a significant role in their pathogenesis. However, research on the correlation between polycyclic aromatic hydrocarbons (PAHs) and chronic bowel disorders remains limited. Using data from the National Health and Nutrition Examination Survey (NHANES) conducted in 2009-2010, we investigated the relationship between 9 PAHs and chronic diarrhea and constipation in U.S. adults. We employed unsupervised methods such as clustering and Principal Component Analysis (PCA) to identify participants with similar exposure patterns. Additionally, we used supervised learning techniques, namely weighted quantile sum (WQS) and Bayesian kernel machine (BKMR) regressions, to assess the association between PAHs and the occurrence of chronic diarrhea and chronic constipation. PCA identified three principal components in the unsupervised analysis, explaining 86.5% of the total PAH variability. The first component displayed a mild association with chronic diarrhea, but no correlation with chronic constipation. Participants were divided into three clusters via K-means clustering, based on PAH concentrations. Clusters with higher PAH exposure demonstrated an increased odds ratio for chronic diarrhea, but no meaningful connection with chronic constipation. In the supervised analysis, the WQS regression underscored a positive relationship between the PAH mixture and chronic diarrhea, with three PAHs significantly impacting the mixture effect. The mixture index showed no correlation with chronic constipation. BKMR analysis illustrated a positive trend in the impact of four specific PAHs on chronic diarrhea, given other metabolites were fixed at their 50th percentiles. Our results suggest a clear association between higher PAH exposure and an increased risk of chronic diarrhea, but not chronic constipation. It also underscores the potential role of specific PAHs in contributing to the risk of chronic diarrhea.

Application of Data Fusion in Traditional Chinese Medicine: A Review

Traditional Chinese medicine is characterized by numerous chemical constituents, complex components, and unpredictable interactions among constituents. Therefore, a single analytical technique is usually unable to obtain comprehensive chemical information. Data fusion is an information processing technology that can improve the accuracy of test results by fusing data from multiple devices, which has a broad application prospect by utilizing chemometrics methods, adopting low-level, mid-level, and high-level data fusion techniques, and establishing final classification or prediction models. This paper summarizes the current status of the application of data fusion strategies based on spectroscopy, mass spectrometry, chromatography, and sensor technologies in traditional Chinese medicine (TCM) in light of the latest research progress of data fusion technology at home and abroad. It also gives an outlook on the development of data fusion technology in TCM analysis to provide references for the research and development of TCM.

Using machine learning to predict outcomes following open abdominal aortic aneurysm repair

OBJECTIVE

Prediction of outcomes following open abdominal aortic aneurysm (AAA) repair remains challenging with a lack of widely used tools to guide perioperative management. We developed machine learning (ML) algorithms that predict outcomes following open AAA repair.

METHODS

The Vascular Quality Initiative (VQI) database was used to identify patients who underwent elective open AAA repair between 2003 and 2023. Input features included 52 preoperative demographic/clinical variables. All available preoperative variables from VQI were used to maximize predictive performance. The primary outcome was in-hospital major adverse cardiovascular event (MACE; composite of myocardial infarction, stroke, or death). Secondary outcomes were individual components of the primary outcome, other in-hospital complications, and 1-year mortality and any reintervention. We split our data into training (70%) and test (30%) sets. Using 10-fold cross-validation, six ML models were trained using preoperative features (Extreme Gradient Boosting [XGBoost], random forest, Naïve Bayes classifier, support vector machine, artificial neural network, and logistic regression). The primary model evaluation metric was area under the receiver operating characteristic curve (AUROC). Model robustness was evaluated with calibration plot and Brier score. The top 10 predictive features in our final model were determined based on variable importance scores. Performance was assessed on subgroups based on age, sex, race, ethnicity, rurality, median area deprivation index, proximal clamp site, prior aortic surgery, and concomitant procedures.

RESULTS

Overall, 12,027 patients were included. The primary outcome of in-hospital MACE occurred in 630 patients (5.2%). Compared with patients without a primary outcome, those who developed in-hospital MACE were older with more comorbidities, demonstrated poorer functional status, had more complex aneurysms, and were more likely to require concomitant procedures. Our best performing prediction model for in-hospital MACE was XGBoost, achieving an AUROC of 0.93 (95% confidence interval, 0.92-0.94). Comparatively, logistic regression had an AUROC of 0.71 (95% confidence interval, 0.70-0.73). For secondary outcomes, XGBoost achieved AUROCs between 0.84 and 0.94. The calibration plot showed good agreement between predicted and observed event probabilities with a Brier score of 0.05. These findings highlight the excellent predictive performance of the XGBoost model. The top three predictive features in our algorithm for in-hospital MACE following open AAA repair were: (1) coronary artery disease; (2) American Society of Anesthesiologists classification; and (3) proximal clamp site. Model performance remained robust on all subgroup analyses.

CONCLUSIONS

Open AAA repair outcomes can be accurately predicted using preoperative data with our ML models, which perform better than logistic regression. Our automated algorithms can help guide risk-mitigation strategies for patients being considered for open AAA repair to improve outcomes.

Antiphospholipid antibodies are enriched post-acute COVID-19 but do not modulate the thrombotic risk

BACKGROUND AND OBJECTIVES

COVID-19-associated coagulopathy, shown to increase the risk for the occurrence of thromboses and microthromboses, displays phenotypic features of the antiphospholipid syndrome (APS), a prototype antibody-mediated autoimmune disease. Several groups have reported elevated levels of criteria and non-criteria antiphospholipid antibodies (aPL), assumed to cause APS, during acute or post-acute COVID-19. However, disease heterogeneity of COVID-19 is accompanied by heterogeneity in molecular signatures, including aberrant cytokine profiles and an increased occurrence of autoantibodies. Moreover, little is known about the association between autoantibodies and the clinical events. Here, we first aim to characterise the antiphospholipid antibody, anti-SARS-CoV-2 antibody, and the cytokine profiles in a diverse collective of COVID-19 patients (disease severity: asymptomatic to intensive care), using vaccinated individuals and influenza patients as comparisons. We then aim to assess whether the presence of aPL in COVID-19 is associated with an increased incidence of thrombotic events in COVID-19.

METHODS AND RESULTS

We conducted anti-SARS-CoV-2 IgG and IgA microELISA and IgG, IgA, and IgM antiphospholipid line immunoassay (LIA) against 10 criteria and non-criteria antigens in 155 plasma samples of 124 individuals, and we measured 16 cytokines and chemokines in 112 plasma samples. We additionally employed clinical and demographic parameters to conduct multivariable regression analyses within multiple paradigms. In line with recent results, we find that IgM autoantibodies against annexin V (AnV), β2-glycoprotein I (β2GPI), and prothrombin (PT) are enriched upon infection with SARS-CoV-2. There was no evidence for seroconversion from IgM to IgG or IgA. PT, β2GPI, and AnV IgM as well as cardiolipin (CL) IgG antiphospholipid levels were significantly elevated in the COVID-19 but not in the influenza or control groups. They were associated predominantly with the strength of the anti-SARS-CoV-2 antibody titres and the major correlate for thromboses was SARS-CoV-2 disease severity.

CONCLUSION

While we have recapitulated previous findings, we conclude that the presence of the aPL, most notably PT, β2GPI, AnV IgM, and CL IgG in COVID-19 are not associated with a higher incidence of thrombotic events.

Harmonization of supervised machine learning practices for efficient source attribution of Listeria monocytogenes based on genomic data

BACKGROUND

Genomic data-based machine learning tools are promising for real-time surveillance activities performing source attribution of foodborne bacteria such as Listeria monocytogenes. Given the heterogeneity of machine learning practices, our aim was to identify those influencing the source prediction performance of the usual holdout method combined with the repeated k-fold cross-validation method.

METHODS

A large collection of 1 100 L. monocytogenes genomes with known sources was built according to several genomic metrics to ensure authenticity and completeness of genomic profiles. Based on these genomic profiles (i.e. 7-locus alleles, core alleles, accessory genes, core SNPs and pan kmers), we developed a versatile workflow assessing prediction performance of different combinations of training dataset splitting (i.e. 50, 60, 70, 80 and 90%), data preprocessing (i.e. with or without near-zero variance removal), and learning models (i.e. BLR, ERT, RF, SGB, SVM and XGB). The performance metrics included accuracy, Cohen's kappa, F1-score, area under the curves from receiver operating characteristic curve, precision recall curve or precision recall gain curve, and execution time.

RESULTS

The testing average accuracies from accessory genes and pan kmers were significantly higher than accuracies from core alleles or SNPs. While the accuracies from 70 and 80% of training dataset splitting were not significantly different, those from 80% were significantly higher than the other tested proportions. The near-zero variance removal did not allow to produce results for 7-locus alleles, did not impact significantly the accuracy for core alleles, accessory genes and pan kmers, and decreased significantly accuracy for core SNPs. The SVM and XGB models did not present significant differences in accuracy between each other and reached significantly higher accuracies than BLR, SGB, ERT and RF, in this order of magnitude. However, the SVM model required more computing power than the XGB model, especially for high amount of descriptors such like core SNPs and pan kmers.

CONCLUSIONS

In addition to recommendations about machine learning practices for L. monocytogenes source attribution based on genomic data, the present study also provides a freely available workflow to solve other balanced or unbalanced multiclass phenotypes from binary and categorical genomic profiles of other microorganisms without source code modifications.

Current state of radiomics in pediatric neuro-oncology practice: a systematic review

BACKGROUND

Radiomics is the process of converting radiological images into high-dimensional data that may be used to create machine learning models capable of predicting clinical outcomes, such as disease progression, treatment response and survival. Pediatric central nervous system (CNS) tumors differ from adult CNS tumors in terms of their tissue morphology, molecular subtype and textural features. We set out to appraise the current impact of this technology in clinical pediatric neuro-oncology practice.

OBJECTIVES

The aims of the study were to assess radiomics' current impact and potential utility in pediatric neuro-oncology practice; to evaluate the accuracy of radiomics-based machine learning models and compare this to the current standard which is stereotactic brain biopsy; and finally, to identify the current limitations of radiomics applications in pediatric neuro-oncology.

MATERIALS AND METHODS

Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) standards, a systematic review of the literature was carried out with protocol number CRD42022372485 in the prospective register of systematic reviews (PROSPERO). We performed a systematic literature search via PubMed, Embase, Web of Science and Google Scholar. Studies involving CNS tumors, studies that utilized radiomics and studies involving pediatric patients (age<18 years) were included. Several parameters were collected including imaging modality, sample size, image segmentation technique, machine learning model used, tumor type, radiomics utility, model accuracy, radiomics quality score and reported limitations.

RESULTS

The study included a total of 17 articles that underwent full-text review, after excluding duplicates, conference abstracts and studies that did not meet the inclusion criteria. The most commonly used machine learning models were support vector machines (n=7) and random forests (n=6), with an area under the curve (AUC) range of 0.60-0.94. The included studies investigated several pediatric CNS tumors, with ependymoma and medulloblastoma being the most frequently studied. Radiomics was primarily used for lesion identification, molecular subtyping, survival prognostication and metastasis prediction in pediatric neuro-oncology. The low sample size of studies was a commonly reported limitation.

CONCLUSION

The current state of radiomics in pediatric neuro-oncology is promising, in terms of distinguishing between tumor types; however, its utility in response assessment requires further evaluation which, given the relatively low number of pediatric tumors, calls for multicenter collaboration.

Machine Learning in Clinical Trials: A Primer with Applications to Neurology

We reviewed foundational concepts in artificial intelligence (AI) and machine learning (ML) and discussed ways in which these methodologies may be employed to enhance progress in clinical trials and research, with particular attention to applications in the design, conduct, and interpretation of clinical trials for neurologic diseases. We discussed ways in which ML may help to accelerate the pace of subject recruitment, provide realistic simulation of medical interventions, and enhance remote trial administration via novel digital biomarkers and therapeutics. Lastly, we provide a brief overview of the technical, administrative, and regulatory challenges that must be addressed as ML achieves greater integration into clinical trial workflows.

The Application of Deep Learning to Electroencephalograms, Magnetic Resonance Imaging, and Implants for the Detection of Epileptic Seizures: A Narrative Review

Epilepsy is a neurological disorder characterized by recurrent seizures affecting millions worldwide. Medically intractable seizures in epilepsy patients are not only detrimental to the quality of life but also pose a significant threat to their safety. Outcomes of epilepsy therapy can be improved by early detection and intervention during the interictal window period. Electroencephalography is the primary diagnostic tool for epilepsy, but accurate interpretation of seizure activity is challenging and highly time-consuming. Machine learning (ML) and deep learning (DL) algorithms enable us to analyze complex EEG data, which can not only help us diagnose but also locate epileptogenic zones and predict medical and surgical treatment outcomes. DL models such as convolutional neural networks (CNNs), inspired by visual processing, can be used to classify EEG activity. By applying preprocessing techniques, signal quality can be enhanced by denoising and artifact removal. DL can also be incorporated into the analysis of magnetic resonance imaging (MRI) data, which can help in the localization of epileptogenic zones in the brain. Proper detection of these zones can help in good neurosurgical outcomes. Recent advancements in DL have facilitated the implementation of these systems in neural implants and wearable devices, allowing for real-time seizure detection. This has the potential to transform the management of drug-refractory epilepsy. This review explores the application of ML and DL techniques to Electroencephalograms (EEGs), MRI, and wearable devices for epileptic seizure detection. This review briefly explains the fundamentals of both artificial intelligence (AI) and DL, highlighting these systems' potential advantages and undeniable limitations.

Exploring new prognostic biomarkers in Mantle Cell Lymphoma: a comparison of the circSCORE and the MCL35 score

Mantle cell lymphoma (MCL) is a biologically and clinically heterogeneous disease, emphasizing the need for prognostic biomarkers. In this study we aimed at comparing the prognostic value of two RNA-based risk scores, circSCORE and MCL35, in 149 patients from the MCL2 (ISRCTN87866680) and MCL3 (NCT00514475) patient cohorts. Both risk scores provided significant stratification of high versus low risk for progression free survival (PFS) and overall survival (OS). The circSCORE retained significant prognostic value in adjusted multivariable Cox regressions for PFS, but not for OS. Furthermore, circSCORE added significant prognostic value to MIPI in the pooled cohort (MCL2 and MCL3) for PFS and OS, and for PFS in MCL3 alone, outperforming Ki67 and MCL35. We suggest a new, combined MIPI-circSCORE with improved prognostic value, and with potential for future clinical implementation, if validated in a larger, independent cohort.

Applying a nomogram based on preoperative CT to predict early recurrence of laryngeal squamous cell carcinoma after surgery

PURPOSE

To identify the value of a computed tomography (CT)-based radiomics model to predict probability of early recurrence (ER) in patients diagnosed with laryngeal squamous cell carcinoma (LSCC) after surgery.

MATERIALS AND METHOD

Pre-operative CT scans of 140 LSCC patients treated by surgery are reviewed and selected. These patients are randomly split into the training set (n = 97) and test set (n = 43). The regions of interest of each patient were delineated manually by two senior radiologists. Radiomics features are extracted from CT images acquired in non-enhanced, arterial, and venous phases. Variance threshold, one-way ANOVA, and least absolute shrinkage and selection operator algorithm are used for feature selection. Then, radiomics models are built with five algorithms namely, k-nearest neighbor (KNN), logistic regression (LR), linear support vector machine (LSVM), radial basis function SVM (RSVM), and polynomial SVM (PSVM). Clinical factors are selected using univariate and multivariate logistic regressions. Last, a radiomics nomogram incorporating the radiomics signature and clinical factors is built to predict ER and its efficiency is evaluated by receiver operating characteristic (ROC) curve and calibration curve. Decision curve analysis (DCA) is also used to evaluate clinical usefulness.

RESULTS

Four features are remarkably associated with ER in patients with LSCC. Applying to test set, the area under the ROC curves (AUCs) of KNN, LR, LSVM, RSVM, and PSVM are 0.936, 0.855, 0.845, 0.829, and 0.794, respectively. The radiomics nomogram shows better discrimination (with AUC: 0.939, 95% CI: 0.867-0.989) than the best radiomics model and the clinical model. Predicted and actual ERs in the calibration curves are in good agreement. DCA shows that the radiomics nomogram is clinically useful.

CONCLUSION

The radiomics nomogram, as a noninvasive prediction tool, exhibits favorable performance for ER prediction of LSCC patients after surgery.

Modeling, optimization and understanding of adsorption process for pollutant removal via machine learning: Recent progress and future perspectives

It is crucial to reduce the concentration of pollutants in water environment to below safe levels. Some cost-effective pollutant removal technologies have been developed, among which adsorption technology is considered as a promising solution. However, the batch experiments and adsorption isotherms widely employed at present are inefficient and time-consuming to some extent, which limits the development of adsorption technology. As a new research paradigm, machine learning (ML) is expected to innovate traditional adsorption models. This reviews summarized the general workflow of ML and commonly employed ML algorithms for pollutant adsorption. Then, the latest progress of ML for pollutant adsorption was reviewed from the perspective of all-round regulation of adsorption process, including adsorption efficiency, operating conditions and adsorption mechanism. General guidelines of ML for pollutant adsorption were presented. Finally, the existing problems and future perspectives of ML for pollutant adsorption were put forward. We highly expect that this review will promote the application of ML in pollutant adsorption and improve the interpretability of ML.

Prediction of Work-Related Risk Factors among Bus Drivers Using Machine Learning

A recent development in ergonomics research is using machine learning techniques for risk assessment and injury prevention. Bus drivers are more likely than other workers to suffer musculoskeletal diseases because of the nature of their jobs and their working conditions (WMSDs). The basic idea of this study is to forecast important work-related risk variables linked to WMSDs in bus drivers using machine learning approaches. A total of 400 full-time male bus drivers from the east and west zone depots of Bengaluru Metropolitan Transport Corporation (BMTC), which is based in Bengaluru, south India, took part in this study. In total, 92.5% of participants responded to the questionnaire. The Modified Nordic Musculoskeletal Questionnaire was used to gather data on symptoms of WMSD during the past 12 months (MNMQ). Machine learning techniques including decision tree, random forest, and naïve Bayes were used to forecast the important risk factors related to WMSDs. It was discovered that WMSDs and work-related characteristics were statistically significant. In total, 66.75% of subjects reported having WMSDs. Various classifiers were used to derive the simulation results for the frequency of pain in the musculoskeletal systems throughout the last 12 months with the important risk variables. With 100% accuracy, decision tree and random forest algorithms produce the same results. Naïve Bayes yields 93.28% accuracy. In this study, through a questionnaire survey and data analysis, several health and work-related risk factors were identified among the bus drivers. Risk factors such as involvement in physical activities, frequent posture change, exposure to vibration, egress ingress, on-duty breaks, and seat adaptability issues have the highest influence on the frequency of pain due to WMSDs among bus drivers. From this study, it is recommended that drivers get involved in physical activities, adopt a healthy lifestyle, and maintain proper posture while driving. For any transport organization/company, it is recommended to design driver cabins ergonomically to mitigate the WMSDs among bus drivers.

Artificial intelligence-based immunoprofiling serves as a potentially predictive biomarker of nivolumab treatment for advanced hepatocellular carcinoma

Immune checkpoint inhibitors (ICI) have been applied in treating advanced hepatocellular carcinoma (aHCC) patients, but few patients exhibit stable and lasting responses. Moreover, identifying aHCC patients suitable for ICI treatment is still challenged. This study aimed to evaluate whether dissecting peripheral immune cell subsets by Mann-Whitney U test and artificial intelligence (AI) algorithms could serve as predictive biomarkers of nivolumab treatment for aHCC. Disease control group carried significantly increased percentages of PD-L1+ monocytes, PD-L1+ CD8 T cells, PD-L1+ CD8 NKT cells, and decreased percentages of PD-L1+ CD8 NKT cells via Mann-Whitney U test. By recursive feature elimination method, five featured subsets (CD4 NKTreg, PD-1+ CD8 T cells, PD-1+ CD8 NKT cells, PD-L1+ CD8 T cells and PD-L1+ monocytes) were selected for AI training. The featured subsets were highly overlapping with ones identified via Mann-Whitney U test. Trained AI algorithms committed valuable AUC from 0.8417 to 0.875 to significantly separate disease control group from disease progression group, and SHAP value ranking also revealed PD-L1+ monocytes and PD-L1+ CD8 T cells exclusively and significantly contributed to this discrimination. In summary, the current study demonstrated that integrally analyzing immune cell profiling with AI algorithms could serve as predictive biomarkers of ICI treatment.

Prediction of B cell epitopes in proteins using a novel sequence similarity-based method

Prediction of B cell epitopes that can replace the antigen for antibody production and detection is of great interest for research and the biotech industry. Here, we developed a novel BLAST-based method to predict linear B cell epitopes. To that end, we generated a BLAST-formatted database upon a dataset of 62,730 known linear B cell epitope sequences and considered as a B cell epitope any peptide sequence producing ungapped BLAST hits to this database with identity ≥ 80% and length ≥ 8. We examined B cell epitope predictions by this method in tenfold cross-validations in which we considered various types of non-B cell epitopes, including 62,730 peptide sequences with verified negative B cell assays. As a result, we obtained values of accuracy, specificity and sensitivity of 72.54 ± 0.27%, 81.59 ± 0.37% and 63.49 ± 0.43%, respectively. In an independent dataset incorporating 503 B cell epitopes, this method reached accuracy, specificity and sensitivity of 74.85%, 99.20% and 50.50%, respectively, outperforming state-of-the-art methods to predict linear B cell epitopes. We implemented this BLAST-based approach to predict B cell epitopes at http://imath.med.ucm.es/bepiblast.