Research Appraisals
Evidence-based critical appraisals of the latest medical research, systematically evaluated using Oxford CEBM methodology.
Showing 29 appraisals
Rheumatic diseases clinics of North America
Machine Learning-Enhanced Autoantibody Discovery and Diagnostics in Systemic Autoimmune Rheumatic Diseases
The growing implementation of machine learning (ML) has extended into autoantibody research for the study of systemic autoimmune rheumatic diseases (SARDs). ML methods offer a promising approach for efficiently handling and identifying important signals within the big data generated by modern autoantibody technologies. The novel biomarkers identified through advanced ML techniques show promise in outperforming current clinical tools, bringing us closer to the goal of precision medicine. In this article, we will provide an overview of ML approaches and how they have been applied in autoantibody research to improve the diagnosis and characterization of SARDs.
2 Aug 2026
Read appraisal →Journal of psychiatric research
Systematic review of machine learning and deep learning models for EEG-based detection of depression.
OBJECTIVE: Depression is a leading cause of global disability, motivating the development of objective and scalable diagnostic approaches. Quantitative electroencephalography (QEEG) combined with machine learning (ML) and deep learning (DL) techniques has gained increasing attention for depression detection. This systematic review aimed to critically examine and descriptively compare the methodologies, performance metrics, and limitations of ML- and DL-based models applied to EEG data for depression detection. METHODS: A systematic review was conducted in accordance with PRISMA 2020 guidelines. Seven electronic databases (PubMed, Scopus, IEEE Xplore, ScienceDirect, Web of Science, SAGE Journals, and MDPI) were searched for peer-reviewed studies published between 2020 and 2024. Eligible studies included human participants, used EEG signals for depression detection, and applied ML or DL algorithms. Extracted information comprised algorithm type, sample size, EEG acquisition parameters, validation strategies, and reported performance metrics, which were synthesized descriptively across studies. Risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool. RESULTS: A total of 42 studies met the inclusion criteria, including 23 ML-based and 19 DL-based investigations. Reported classification accuracy ranged from approximately 76% to 100%. DL studies showed a higher mean reported accuracy than ML studies (93.92% vs. 90.78%); however, this difference was not statistically significant in the exploratory non-parametric comparison. Near-perfect performance values were frequently observed in studies with small sample sizes and subject-dependent or exclusively internal validation strategies, raising concerns regarding overfitting and limited generalizability. Studies relying on publicly available datasets tended to report more stable performance. QUADAS-2 assessment revealed recurrent risk-of-bias concerns, particularly in the domains of patient selection and index test conduct. CONCLUSIONS: Both ML and DL approaches demonstrate potential for EEG-based depression detection, but reported performance differences between them should be interpreted cautiously. Although DL studies tended to report higher accuracy values, this pattern was not statistically significant in exploratory analyses and was strongly influenced by sample size, validation strategy, and methodological design. Future research should prioritize larger and more diverse samples, subject-independent or external validation strategies, and standardized reporting frameworks to enhance methodological rigor and clinical applicability.
2 Aug 2026
Read appraisal →Scientific reports
Manual federated simulation for multiple sclerosis integrating XGBoost algorithm with SHAP explanation
Multiple sclerosis (MS) is a chronic autoimmune disorder of the central nervous system, underscoring the importance of early and accurate diagnosis. In this study investigates the predictive modelling of MS progression in patients with Clinically Isolated Syndrome (CIS), privacy-preserving for a federated and explainable Machine Learning (ML) framework. To address missing data while preserving inter-feature dependencies, Multivariate Imputation by Chained Equations (MICE) with iterative imputers was employed. Classification was performed using the Extreme Gradient Boosting (XGBoost) algorithm. Model interpretability was developed through Explainable Artificial Intelligence (XAI) techniques, specifically Shapley Additive Explanations (SHAP). To ensure data confidentiality and simulate decentralized clinical environments, an in silico federated learning framework was applied. Experimental results demonstrated strong predictive performance, achieving 96.7% accuracy and 99% ROC-AUC during training, 92.5% accuracy in validation, and 81.8% accuracy with an AUC of 88% on the test set. For the Federated Learning (FL) simulation, the model maintained competitive performance, yielding an accuracy of 76.3% and an AUC of 83.9%. The proposed approach supports early diagnosis, enhances clinical trust through interpretability, and promotes secure data collaboration, thereby contributing to more informed and transparent clinical decision-making and improved patient care.
1 Aug 2026
Read appraisal →Scientific data
A Dataset of Widefield Fundus Images From Patients With Birdshot Chorioretinitis and Matched Control
Birdshot chorioretinitis is a rare uveitis where research is limited by the lack of large, publicly available imaging datasets. To address this gap, we present Birdshot-Wide, a standardized retinal imaging dataset comprising 5,042 widefield fundus photographs from 742 eyes with birdshot chorioretinitis (BSCR) and 1,310 images from 742 age-and sex- matched healthy eyes controls. The BSCR cohort is longitudinal, with a median follow-up of 4.31 years. Each image in the BSCR cohort was manually reviewed for quality and classified by dominant lesion subtype. To demonstrate its utility, the dataset was benchmarked on a classification task, where deep learning models trained to differentiate BSCR from control images achieved an area under the receiver operating characteristic curve of up to 0.96. This dataset, with its detailed metadata, provides a valuable resource for biomarker discovery, the development of automated diagnostic tools, and disease progression modeling.
8 June 2026
Read appraisal →Biomedical physics & engineering express
Acoustic-based intraoperative assessment of femoral stem initial stability in cementless THA via sensitive frequency band identification
Intraoperative assessment of femoral stem initial stability in cementless total hip arthroplasty (THA) predominantly relies on subjective tactile and auditory feedback, which lacks standardization and increases complication risks. This study develops a quantitative framework to identify sensitive acoustic frequency bands characterizing the bone-implant fixation state. Specifically, we acquired high-fidelity acoustic signals from 118 primary cementless THA procedures. Utilizing a statistical difference analysis framework, we identified two distinct sensitive bands (640-1920 Hz and 3460-3560 Hz) to extract band power fraction features. Machine learning evaluation demonstrated that these bands significantly outperform traditional empirical partitioning, achieving a classification accuracy of 0.816±0.028, which improved to 0.834±0.042 after feature fusion. Furthermore, stratified validation confirmed high consistency across diverse bone qualities and implant designs (Summit and Corail). While robust under steady-state Gaussian noise, the system exhibited sensitivity to non-steady-state speech interference. In conclusion, the identified sensitive bands effectively reflect the mechanical evolution of the bone-implant interface, providing a physically interpretable decision-support tool for enhancing surgical safety.
27 May 2026
Read appraisal →Clinical chemistry and laboratory medicine
T-cell receptor beta constant 1 flow cytometry for diagnosing clonality in mature T-cell neoplasms: a systematic review and meta-analysis of diagnostic accuracy
INTRODUCTION: Accurate assessment of T-cell clonality is key for diagnosing mature T-cell neoplasms. TRBC1-based flow cytometry provides a rapid, robust, and cost-efficient approach. This systematic review and meta-analysis assessed the diagnostic accuracy of TRBC1 flow cytometry (TRBC1-FC) for detecting T-cell clonality in mature T-cell neoplasms. CONTENT: We systematically searched Scopus, PubMed (MEDLINE), and Google Scholar for articles on TRBC1-FC diagnostic accuracy published up to July 1, 2025. Pooled sensitivity and specificity were estimated, between-study heterogeneity was evaluated and small-study effects were examined. SUMMARY: This meta-analysis included 10 studies. The pooled sensitivity was 97.6 % (95 % CI, 95.1-99.4 %) and specificity 90.7 % (95 % CI, 76.0-99.3 %). The pooled LR+ was 10.9 (95 % CI, 4.1-28.9), LR-was 0.053 (95 % CI, 0.025-0.12), and DOR 339 (95 % CI, 64-1,788). The HSROC curve demonstrated an AUC of 0.974 (partial AUC=0.970), confirming excellent global discriminatory capacity. Between-study heterogeneity was substantial (I2=83.3 %), mainly affecting specificity and DOR, while sensitivity remained highly consistent. No evidence of a threshold effect was found. Deeks' test showed significant small-study effects (p<0.001), and sensitivity analyses identified one influential study whose exclusion markedly reduced heterogeneity. These results confirm the high diagnostic performance and robustness of TRBC1-FC for T-cell clonality assessment. OUTLOOK: TRBC1-FC demonstrates high sensitivity and low LR-, supporting its role as a rule-out test. Variability in specificity, LR+ and DOR, mainly due to small-study effects, advises caution for rule-in use. Standardized protocols and cost-effectiveness analyses are needed before broad clinical adoption.
26 May 2026
Read appraisal →ACS sensors
Diagnosis of Primary Open-Angle Glaucoma Using Spectral Profiling of Aqueous Humor-Derived Exosomes
Primary open-angle glaucoma (POAG) is one of the most common neurodegenerative diseases that cause irreversible optic nerve damage. POAG diagnosis requires multimodal assessments; however, current methods like intraocular pressure (IOP) and optical coherence tomography (OCT) imaging suffer from low sensitivity and inter-patient variability, respectively. Liquid biopsy provides objective molecular signatures, offering a robust alternative to overcome limitations of conventional clinical hallmarks. Here, we present an efficient and highly sensitive diagnostic platform operating on spectral profiles from aqueous humor (AH)-derived exosomes. Following morphological and compositional validation of exosome presence in AH samples, antibody-functionalized substrates were fabricated for selective capture. Immobilized AH-derived exosomes are then coated with gold nanoparticles to generate molecular fingerprint Raman signals. A total of 7600 spectra are acquired and used to train and evaluate the convolutional neural network (CNN) model for binary classification between healthy controls and glaucoma patients. The trained CNN model achieved an AUC of 0.96 and a prediction accuracy of 91%. The system enables rapid, individualized diagnosis, overcoming sample volume limitations via integrated immunoassay-based isolation and artificial intelligence (AI)-driven classification, highlighting the potential of ocular fluid-derived EVs in the diagnosis of neurodegenerative diseases.
23 May 2026
Read appraisal →The Cochrane database of systematic reviews
Artificial intelligence for detection and staging of Alzheimer's disease using retinal images
This is a protocol for a Cochrane Review (diagnostic). The objectives are as follows: To determine the diagnostic accuracy of AI algorithm-based retinal image reading for detecting referral-requiring Alzheimer's disease in adults who have behavioral symptoms or cognitive decline in the primary care setting. Secondary objectives Where data are available, we will investigate heterogeneity by incorporating covariates in a metaregression.
22 May 2026
Read appraisal →Scientific reports
Use of machine learning and voice for multiclass classification of Parkinson's disease, chronic obstructive pulmonary disease, and healthy controls.
Parkinson's disease (PD) and chronic obstructive pulmonary disease (COPD) are prevalent conditions with substantial impact on quality of life and health care systems. Both disorders affect voice production through different physiological mechanisms, yet neither condition has a widely adopted objective biomarker for routine clinical use. Voice analysis has emerged as a non-invasive digital biomarker candidate, but existing studies have largely focused on binary classification within a single disorder or language. This study aimed to evaluate whether an unified multiclass machine learning (ML) framework applied to sustained vowel "a" phonation can discriminate between PD, COPD, and healthy controls (HC) across linguistically distinct cohorts. Sustained vowel recordings were analyzed from Swedish speaking individuals with COPD and HC, and English-speaking individuals with PD and HC, collected under comparable mobile recording conditions. Acoustic features included baseline voice measures and Mel Frequency Cepstral Coefficients. A soft voting ML framework integrating support vector machine, random forest, CatBoost, and light gradient boosting classifiers was trained using nested cross validation with hyperparameter optimization. Data were partitioned at the participant level into a development cohort and an independent test cohort. Model performance was evaluated using accuracy, macro averaged precision, recall, F1 score, receiver operating characteristic analysis, and confusion matrices. Model interpretability was assessed using Shapley additive explanations and vowel space analysis. The final soft voting classifier achieved robust multiclass discrimination on the participant disjoint independent test set, with an overall accuracy of 0.842 and a macro averaged F1 score of 0.839. Classification performance differed across groups, with the highest performance observed for PD, intermediate performance for HC, and lower performance for COPD. Misclassifications occurred primarily between HC and COPD, while confusion between PD and COPD was minimal. Feature attribution analysis revealed class dependent relevance patterns, and vowel space analysis demonstrated subtle but consistent group level differences. These findings demonstrate the feasibility of using an explainable soft voting machine learning framework applied to sustained vowel phonation to distinguish between neurologically and respiratory driven voice impairments across linguistic contexts. The study supports voice as a promising digital biomarker modality for multiclass clinical discrimination using mobile recordings.
21 May 2026
Read appraisal →PloS one
Deep Learning outperforms physicians in myopathy and neuropathy classification based on Needle Electromyography Signal
Needle electromyography (nEMG) is a valuable tool for diagnosing patients with neuromuscular diseases. However, it is labor-intensive and is prone to diagnostic inaccuracies stemming from human biases. To address these challenges, we validated an nEMG diagnosis-aiding system with minimal preprocessing using deep learning model to classify patients into three categories: normal, myopathy, and neuropathy. Using 376 nEMG signals from 57 patients from a tertiary university hospital database through nested k-fold cross validation, deep learning model surpassed the classification performance of six electromyographers. The median patient classification accuracy, precision, sensitivity, and specificity of the deep learning model was 0.70, 0.70, 0.70, and 0.85, respectively, whereas those of the physicians were 0.55, 0.60, 0.54, and 0.78, respectively. Model interpretability and failure analysis showed that the deep learning model classifies based on relevant signal features. Despite higher accuracy of DL model, the number of unanimously misclassified cases were higher in the DL model than physicians. Our study validates deep learning is a fast, accurate, and practical application to aid physicians in diagnosing patients using nEMG signals.
20 May 2026
Read appraisal →Talanta
SERS technology helps bacterial inactivation in vivo: including detection, identification, and evaluation.
The problem of bacterial infection and antimicrobial resistance has become a major challenge threatening to human health. Detection, identification, and real-time evaluation of bacteria during the sterilization process are of great significance for interpreting the sterilization mechanism and ensuring the effectiveness of sterilization. With the advantages of high sensitivity, fast detection speed, and real-time monitoring, Surface-enhanced Raman Scattering (SERS) technology provides a revolutionary solution to in vivo bacterial detection, identification, and antibacterial effect evaluation through functional substrate optimization, microfluidic chip integration, and deep collaboration of machine learning algorithms. This review focuses on the prominent applications of SERS technology in chemical, biological, and physical sterilization processes. By analyzing the changes in bacterial SERS fingerprint patterns, precise detection and identification of sensitive/resistant bacteria have been achieved, deepening the interpretation of drug resistance mechanisms at the gene, cell, and biofilm levels. Meanwhile, SERS technology provides an ideal tool for monitoring and evaluating the biological sterilization process. In the process of physical sterilization, enhancing the substrate material not only helps to improve SERS signals, but also promotes bacterial inactivation. By monitoring real-time changes in bacterial metabolism, the sterilization process can be precisely regulated. This review summarizes the innovative application of SERS technology in accurately guiding the sterilization process in vivo, providing important support for optimizing antibacterial treatment strategies.
18 May 2026
Read appraisal →JMIR medical informatics
From Flow to Feature Using a Proof-of-Concept Spectral-Driven Machine Learning Approach Using Smart Urinary and Drainage Catheter Systems: Algorithm Development and Validation
BACKGROUND: Current urinary and drainage catheter systems collect fluids for visual inspection or manual sampling, offering limited diagnostic value while being labor-intensive and prone to error. Machine learning (ML) has the potential to automate the analysis of these fluids. However, existing methods rely on complex preprocessing steps, which hinder real-time analysis. OBJECTIVE: We aim to develop and evaluate a fully automated, real-time diagnostic approach for smart urinary and drainage catheter systems by leveraging spectral data and ML to differentiate pathological from healthy excreted fluids without the need for manual preprocessing. METHODS: This study proposes a novel, fully automated approach for smart urinary and drainage catheter systems that uses spectra and ML to extract features from excreted fluids, enabling real-time analysis directly. A total of 454 surgical drainage fluid samples (from 181 patients) and 401 urine catheter samples (from 168 patients) were analyzed using smart catheters and drains equipped with compact mini-spectrometer sensors. The collected spectral data were fed into 3 different ML models: a random forest, a partial least squares discriminant analysis regression, and a convolutional neural network (CNN). Each model aimed to extract features and differentiate between pathological and healthy urine and drainage samples based on the various biomarkers available from previously conducted laboratory analyses. RESULTS: All 3 approaches (random forest, partial least squares discriminant analysis regression, and CNN) achieved promising results, demonstrating the potential of the overall approach. In particular, the CNN models trained on the drainage biomarkers hemoglobin and bilirubin achieved the best results. Matthews correlation coefficient scores of 0.83 and 0.81 were obtained for hemoglobin and bilirubin, respectively, when differentiating between pathological and healthy samples using the extracted spectral features. CONCLUSIONS: This work demonstrates the potential of spectral-driven ML for smart urinary and drainage catheter systems. This approach offers a real-time, noninvasive method for analyzing excreted fluids, paving the way for improved diagnostics and personalized patient care. Further research will explore the optimal ML model for this application.
17 May 2026
Read appraisal →PloS one
Systematic evaluation of the DeepSeek large language model for clinical diagnostic reasoning
BACKGROUND: Artificial intelligence (AI) is undergoing an era of transformative advancement, particularly through the emergence of Transformer-based large language models (LLMs). While these systems demonstrate strong reasoning and generalization capabilities, their clinical applicability, particularly in emergency and critical care decision-making, remains underexplored.. In time-sensitive settings, diagnostic reasoning must align rigorously with evidence-based standards and ensure the relevance of timing to clinical decisions. OBJECTIVE: This study aims to provide a preliminary evaluation of the decision-support performance of the DeepSeek model in acute medical scenarios. We systematically evaluate its diagnostic reasoning, temporal consistency of recommendations, and adherence to evidence-based critical care protocols using standardized case-based assessments. METHODS: Twenty-nine representative clinical cases were extracted from the Merck Manual of Diagnosis and Therapy, a widely used medical reference providing standardized case descriptions. The model's outputs were evaluated across four decision-making dimensions: differential diagnosis, diagnostic testing, final diagnosis, and management planning. Human raters scored each response for accuracy, and multivariable linear regression was applied to assess associations between performance and case parameters (age, gender, and Rapid Emergency Medicine Score [REMS]). RESULTS: DeepSeek achieved an overall mean accuracy of 82.9% (95% CI: 80.2-85.6%) across all cases. Accuracy peaked in final diagnosis (97.7%), but declined in differential diagnosis (73.0%). Model performance showed no significant variation across demographic or severity strata. CONCLUSIONS: DeepSeek shows promising performance in structured case-based diagnostic tasks, particularly in confirmatory diagnostic reasoning. However, its early-stage reasoning and handling of ambiguous cases require enhancement. Future studies using larger and more diverse clinical datasets are needed to further evaluate the model's robustness and potential clinical applicability.
13 May 2026
Read appraisal →JMIR medical informatics
Machine Learning-Based Multidimensional Oximetry for Obstructive Sleep Apnea Screening: Development and External Validation
BACKGROUND: Obstructive sleep apnea (OSA) affects nearly one billion people globally and poses a substantial public health threat. Effective and accessible methods for OSA risk identification are urgently needed. OBJECTIVE: This study aims to develop and externally validate a machine learning model derived from multi-parameter pulse oximetry (SpO2) for OSA screening, and to evaluate its performance, interpretability, and robustness across sex and age subgroups. METHODS: Of 4156 screened participants, 2195 underwent polysomnography (internal cohort) and 446 received home sleep apnea testing (external cohort). Eight SpO2-derived parameters, including oxygen desaturation index (ODI), hypoxic burden (HB), and ST90 (percentage of sleep time with SpO2 < 90%), were used to construct models. Six machine learning algorithms were trained, with F1-score as the primary metric and area under the curve as the secondary metric. Model interpretability was assessed using Shapley additive explanations and intrinsic feature importance scores. RESULTS: Nonlinear parameter-risk relationships were observed between oximetry indices and OSA probability. The 4-parameter ODI-HB-MinSpO2-ST90 model achieved optimal performance (F1-score = 0.9516, area under the curve = 0.9879), surpassing all single-parameter models. Shapley additive explanations analysis identified ODI, HB, and MinSpO2 as key predictors. The ODI-HB-MinSpO2-MeanSpO2 configuration demonstrated superior performance in female and younger subgroups, whereas the ODI-HB-MinSpO2-ST90 model remained optimal for male and older participants. Categorical boosting outperformed other algorithms across multiple metrics and remained robust in both subgroup and external validation analyses. CONCLUSIONS: The multi-parameter oximetry model based on the categorical boosting algorithm provides a simple and accurate tool for OSA screening. Sex- and age-stratified strategies can further enhance its clinical applicability.
12 May 2026
Read appraisal →Analytical chemistry
Quantitative Analysis of Multiple Serum Tumor Biomarkers by an Interpretable Stacked Ensemble Model
Surface-enhanced Raman spectroscopy (SERS) offers exceptional sensitivity and specificity for biomolecular detection, particularly in analyzing serum tumor biomarkers. However, its application is hindered by spectral complexity and matrix interference. Although machine learning (ML) has shown promise in SERS data analysis, existing quantitative models for biomarker detection lack generalizability and interpretability. To address these challenges, we developed an interpretable stacked ensemble model (ISEM) that integrates feature selection, ensemble learning, and explainable artificial intelligence (XAI) for quantifying 12 serum tumor biomarkers. To ensure reliable quantitative analysis, we validated the acquired SERS data using uniform and reproducible Au nanoparticle substrates. Preprocessing, least absolute shrinkage, and selection operator (LASSO) feature selection were employed to establish a foundation for accurate quantification. Subsequently, 12 serum tumor biomarkers were quantified, demonstrating the superior performance of ISEM, which achieved a high predictive accuracy, with R2 values exceeding 0.9 for all biomarkers. Crucially, based on the quantitative results, we provided molecular-level interpretability for structure-spectrum correlations using Shapley additive explanations (SHAP), revealing how glycosylation reactions, matrix interference, and spectral overlap influence prediction accuracy. Furthermore, the capability of ISEM to quantify multiple biomarkers in unseen samples was confirmed through validation of its generalizability. Our study establishes an ensemble-driven, interpretable framework for quantitative biomarker analysis in a complex biological matrix, demonstrating significant potential for early cancer diagnosis and screening.
11 May 2026
Read appraisal →Pancreas
Whole Metagenomic Profiling Identifies a Gut Microbial Signature for Chronic Pancreatitis via Machine Learning
BACKGROUND: Pancreatitis significantly alters the microbial composition of the oral and intestinal compartments, causing dysbiosis that may contribute to disease mechanisms and potentially serve as a basis for diagnosis or treatment. OBJECTIVE: To determine whether the oral or gut microbial signature can classify chronic pancreatitis (CP). METHODS: Stool samples (n=707) were collected from participants in the Prospective Evaluation of Chronic Pancreatitis for Epidemiologic and Translational Studies (PROCEED). Samples were distributed among 200 healthy (HC), 310 CP, 49 acute pancreatitis (AP), and 148 recurrent acute pancreatitis (RAP). In addition, saliva samples were collected for a subset of participants (n=156). Whole genome sequencing was performed to assess microbiome composition. Machine learning algorithms were utilized to identify a signature with microbial features predictive of CP. RESULTS: Gut alpha diversity was significantly decreased in AP, RAP, and CP compared with HC, with CP exhibiting the lowest diversity. In contrast, oral microbial diversity showed no significant variation across groups. Beta diversity analysis revealed distinct gut microbiome compositions between HC and pancreatitis subtypes, with CP showing the most pronounced differences. Random forest models using gut microbial species demonstrated robust predictive performance for CP using a minimum of 10 species (Area under the curve-AUC: 0.834; accuracy: 0.774). Despite similarities in gut microbiome composition across pancreatitis subtypes, a unique gut microbial signature for CP was identified highlighting the microbiome's potential in CP diagnosis. CONCLUSION: Our study reveals a gut microbial signature predictive of CP using machine learning models in a large US multi-institutional cohort.
6 May 2026
Read appraisal →Science (New York, N.Y.)
Performance of a large language model on the reasoning tasks of a physician.
More than 65 years ago, complex clinical diagnostic reasoning cases were introduced as the gold standard for the evaluation of expert medical computing systems, a standard that has held ever since. In this study, we report the results of a physician evaluation of a large language model (LLM) on challenging clinical cases across five experiments with a baseline of hundreds of physicians. We then report a real-world study comparing human expert and artificial intelligence (AI) second opinions in randomly selected patients in the emergency room of a major tertiary academic medical center. In all experiments, the LLM outperformed physician baselines and displayed continued improvement from prior generations of AI clinical decision support. Our study suggests that LLMs have eclipsed most benchmarks of clinical reasoning, motivating the urgent need for prospective trials.
5 May 2026
Read appraisal →Computer methods in biomechanics and biomedical engineering
Multiple instance learning method based on convolutional neural network and self-attention for early cancer detection.
Early cancer detection using T-cell receptor sequencing (TCR-seq) and multiple instances learning methods has shown significant effectiveness. We introduce a multiple instance learning method based on convolutional neural networks and self-attention (MICA). First, MICA preprocesses TCR-seq using word vectors and then extracts features using convolutional neural networks. Second, MICA uses an enhanced self-attention mechanism to extract relational features of instances. Finally, MICA can extract the crucial TCR-seq. After cross-validation, MICA achieves an area under the curve (AUC) of 0.911 and 0.946 on the lung and thyroid cancer datasets, which are 7.1% and 2.1% higher than other methods, respectively.
4 May 2026
Read appraisal →Journal of neurology
Identifying diagnostic biomarkers in functional motor disorders through multimodal behavioral, neurophysiological, and imaging assessment using explainable machine learning
BACKGROUND: Functional motor disorders (FMDs) represent a frequent and disabling neurological condition. The lack of reliable diagnostic biomarkers and their heterogeneity might affect diagnosis. We identified multimodal biomarkers distinguishing FMDs from healthy controls (HCs) using machine-learning approaches. METHODS: In this multicenter cross-sectional study, consecutive adults with a clinically established FMDs diagnosis (n = 75, 74.7% female; mean age 44.20 ± 12.92) and age- and sex-matched HCs (n = 75; 58.6% female; 48.42 ± 11.67) were recruited. All participants underwent standardized behavioral, neurophysiological, and brain MRI assessment exploring motor, exteroceptive, and interoceptive domains. A Random Forest (RF) classifier combined with repeated stratified k-fold cross-validation was trained on the collected features. Predictive performance was evaluated using accuracy, sensitivity, specificity, precision, F1-score, and AUC-ROC. SHapley Additive exPlanations interpreted feature importance. RESULTS: The strongest diagnostic biomarkers were lower dual-task effect scores for postural sway area under eyes-closed motor and cognitive conditions, and gait speed during the motor dual-task, followed by increased vDMN and basal ganglia networks functional connectivity, reduced baseline ipsilateral-contralateral R2 blink reflex area, and higher DNIC-to-baseline N2P2 amplitude ratios for the lower limb. The RF classifier achieved robust performance (accuracy 85.0%, sensitivity 83.9%, specificity 86.1%, F1-score 85.7%, AUC-ROC 0.921). CONCLUSIONS: Motor, functional neuroimaging, and neurophysiological markers demonstrated diagnostic value in distinguishing FMDs from healthy controls, addressing the current lack of objective tools and supporting more confident and accurate diagnosis of these heterogeneous conditions. TRIAL REGISTRATION: Trial registration number NCT06328790. Registered on 26 March 2024.
3 May 2026
Read appraisal →Scientific reports
LASSO-HHO two-stage hybrid gene selection framework for accurate Alzheimer's disease diagnosis
Selecting informative genes is essential for building accurate and efficient diagnostic models, especially for high-dimensional gene expression data such as those related to Alzheimer's disease. These datasets typically involve thousands of genes with limited samples, increasing the risk of overfitting and computational complexity. To address this, we propose a LASSO-HHO Gene Selection (LHGS) framework. LASSO is first applied to reduce dimensionality by filtering irrelevant genes. A conditional HHO-based optimization stage is then applied only when the LASSO-selected subset does not achieve sufficient accuracy or remains relatively large. Otherwise, the LASSO-selected features are directly used without further optimization. Experimental results show that the proposed method reduces the number of selected genes by up to 99.9% while maintaining high performance. The experimental results indicate that 100% accuracy can be achieved on specific datasets. In particular, GSE48350 and GSE36980 achieved 100% accuracy using LASSO alone, whereas GSE118553 and GSE132903 required the full LHGS framework to achieve the same performance. The framework also improves computational efficiency within a consistent experimental setup by reducing the optimization search space after LASSO filtering. Overall, LHGS provides a practical and efficient solution for gene selection in high-dimensional biomedical data.
28 Apr 2026
Read appraisal →ACS sensors
Nanobiosensors and Artificial Intelligence Strategies for Glycan Profiling in Cancer Progression: A Critical Review
Glycans are complex carbohydrates that are integral to cancer progression and tumor biology. Conventional methods face challenges due to the structural diversity and low abundance of glycans, limiting the detection of cancer-associated glycan signatures. This bottleneck hampers the accurate monitoring of cancer progression and the development of effective diagnostic tools. Nanobiosensor-enabled glycan profiling overcomes conventional limitations by providing highly sensitive detection of low-abundance glycans and precise discrimination of similar forms through nanoscale electrochemical or optical signal amplification and selective biorecognition. This review evaluates emerging nanobiosensors, including optical modalities such as surface plasmon resonance, localized surface plasmon resonance, surface-enhanced Raman scattering, fluorescence, and colorimetric nanobiosensors. It also covers electrochemical systems, including impedimetric, voltammetric, and amperometric sensors that use nanomaterials to enhance the sensitivity, specificity, and reliability of glycan detection. After a critical evaluation of nanobiosensors, this review highlights their applications for glycan analysis in cancer progression and potential of miniaturized point-of-care platforms. In parallel, this review highlights emerging artificial intelligence (AI) and machine learning (ML) approaches that are being explored to analyse complex biosensor outputs. By enabling multivariate pattern recognition, dimensionality reduction, and nonlinear modeling of biosensor responses, these computational frameworks may facilitate the extraction of diagnostically relevant signatures from high-dimensional optical and electrochemical datasets. The review identifies challenges in standardization and clinical scalability of nanobiosensors and explores future directions, including AI integration to advance early cancer detection and personalized patient management.
27 Apr 2026
Read appraisal →Brain imaging and behavior
Meta-analysis and meta-regression of diagnostic test accuracy of connectome-based predictive modeling in OCD
Functional magnetic resonance imaging studies have reported disruptions in functional connectivity within brain networks, known as connectomes. Researchers have tested connectomes to see whether they serve as biomarkers for various psychiatric conditions. This meta-analysis aims to evaluate the diagnostic test accuracy of predictive models of connectomes derived from resting-state fMRI in diagnosing obsessive-compulsive disorder. A systematic review and meta-analysis were conducted on previous studies assessing the sensitivity, specificity, and accuracy of connectome-based diagnostic models in obsessive-compulsive disorder and healthy controls. Eight studies were identified, comprising 563 individuals with obsessive-compulsive disorder and 564 healthy controls. The results revealed robust diagnostic performance with a pooled sensitivity of 0.827 (95% CI: 0.779-0.867) and specificity of 0.794 (95% CI: 0.759-0.826). Connectome-based diagnostic models demonstrated excellent clinical utility, with an area under the curve of 87% (95% CI: 84%-90%), and significant predictive power as indicated by positive (4.15) and negative (0.21) likelihood ratios, as well as strong diagnostic odds ratio of 18.69 (95% CI: 11.84-29.49). The results highlight the potential of functional connectome-based predictive modeling as a robust tool for accurately diagnosing obsessive-compulsive disorder, with possibility of future implications for early diagnosis, monitoring treatment-related changes, involving in decision modeling, and understanding the biological mechanisms underlying obsessive-compulsive disorder.
20 Apr 2026
Read appraisal →Biomedical physics & engineering express
A multimodal attention fusion-based model for pathological voice detection
Pathological voice detection provides a non-invasive and cost-effective approach for the early screening of laryngeal disorders. However, most existing methods rely on single-modal acoustic features and mainly focus on binary classification, which limits their ability to achieve fine-grained discrimination among multiple pathological voice categories. To address these limitations, this paper proposes a Multimodal Fusion Network (MFNet) with attention-based fusion for pathological voice classification. The proposed framework jointly models time-domain and frequency-domain information within a unified architecture. Specifically, SincNet is employed to extract task-relevant narrowband acoustic features directly from raw speech waveforms, while the Tunable Q-factor Wavelet Transform (TQWT) and SKNet are combined to learn multi-scale spectral representations from Mel-spectrograms. In addition, an Attentional Feature Fusion (AFF) module is introduced to adaptively integrate features from the two branches, thereby enhancing the discriminative capability of pathological voice representations. Comparative experiments on three public pathological voice datasets show that the proposed method outperforms representative baseline models and demonstrates strong robustness and generalization across different datasets. Furthermore, ablation studies further validate the effectiveness of the proposed architectural design. These results indicate that MFNet provides an effective solution for pathological voice classification and has promising potential for computer-aided voice disorder screening.
18 Apr 2026
Read appraisal →BMJ open respiratory research
Parasternal electromyography in participants with mild or moderate chronic obstructive pulmonary disease in primary care: cohort study to assess technical and clinical application
BACKGROUND: Neural respiratory drive (NRD) measurement, reflecting the balance between respiratory muscle load and capacity, is quantified using surface parasternal electromyography (EMGpara). EMGpara tracks recovery from severe exacerbations of chronic obstructive pulmonary disease (COPD). Among stable COPD participants, we hypothesised the existence of a relationship between NRD, breathlessness and airway obstruction. STUDY AIMS: (1) assessing the feasibility of measuring EMGpara in COPD participants with forced expiratory volume in 1 s (FEV1) ≥50% predicted in primary care; (2) investigating relationships between NRD measures, self-reported breathlessness, airflow obstruction severity and health-related quality of life (HRQoL). METHODS: Participants with stable mild/moderate COPD, using inhaled corticosteroid (ICS) therapy, were recruited from 20 general practices. Participants were randomly allocated to continue using ICS (maintenance group) or to withdraw ICS (withdrawal group) over 6 weeks. EMGpara, spirometry, self-reported breathlessness (modified Borg dyspnoea scale), COPD Assessment Test and Chronic Respiratory Disease Questionnaire Self-Administered Standardised were measured at baseline, 3- and 6-month follow-up. Bland-Altman plots examined agreement between serial measurements. RESULTS: Forty COPD participants were recruited: age 70±9.2 years; body mass index 26±5.3 kg/m2; FEV1 1.74±0.54 L; and FEV1% pred 69.6±14.0%. High-quality EMGpara data were obtained from 35 participants at baseline and 31 participants on three occasions. High intra-rater and inter-rater agreement for EMGpara (intraclass correlation coefficient >0.9) and moderate correlation between EMGpara and FEV1% predicted (r=-0.42; p=0.01) were recorded. No correlation was observed between resting EMGpara and breathlessness or HRQoL measures across the three time points. CONCLUSIONS: EMGpara measurement is feasible in primary care. In this group of COPD patients, lung function was stable across the three time points and EMGpara was associated with the degree of airflow obstruction. In the resting stable state in mild/moderate disease, there was no association between EMGpara and participant-reported outcomes. Further work should investigate the utility of EMGpara in mild/moderate COPD participants during acute exacerbation and recovery.
12 Apr 2026
Read appraisal →PloS one
Applying machine learning and natural language processing to patient safety event reports: Identifying patterns of cardiovascular diagnostic errors
OBJECTIVE: To explore the utility of natural language processing (NLP) and machine learning (ML) techniques to identify unsafe conditions leading to cardiovascular diagnostic errors using patient safety event (PSE) reports data. METHODS: PSE reports from January 2016 to August 2021 from a multi-hospital healthcare system in the mid-Atlantic region of the United States were included in this study. To have the true cardiovascular diagnostic errors labels for PSE reports, each individual PSE report was manually reviewed to find clinician-reported narratives describing current definitions of cardiovascular diagnostic errors. The PSE reports which contained cardiovascular diagnostic errors-related narratives were labeled as one and zero otherwise. Four binary ML models were employed to identify cardiovascular diagnostic errors narratives and common features from annotated PSE reports data: (1) simple logistic regression, (2) elastic net, (3) XGBoost, and (4) deep neural networks. RESULTS: XGBoost outperformed the rest of the models in identifying cardiovascular diagnostic errors -related reports and achieved high performance metrics on the testing data (AUROC = 0.914, specificity = 0.982, PPV = 0.866, accuracy = 0.929, F-1 score = 0.738, and AUPRC = 0.783). Pacemaker emerged as a significant signal for cardiovascular diagnostic errors in our PSE reports. Our analysis demonstrated that ordering MRI for patients with pacemaker was a frequent theme among cardiovascular diagnostic errors-related PSE reports containing the word pacemaker. Order, EKG, cardiac, and chest were the five most important features in identifying cardiovascular diagnostic errors events. These words were utilized in explaining heart conditions, associated care process, and related safety incidents. CONCLUSIONS: Findings from our study demonstrates the feasibility of ML and NLP techniques in identifying cardiovascular diagnostic errors-related reports among the existing PSE data sources. However, validation in external healthcare systems is needed before broader application.
11 Apr 2026
Read appraisal →Journal of neurology
Systematic review of movement disorders mislabeled as functional: when incongruence misleads
BACKGROUND: Misdiagnosis of functional movement disorders (FMD) remains a concern for clinicians. We sought to review the phenomenology and clinical features associated with FMD misdiagnosis. METHODS: We conducted a systematic review of case reports, case series, and observational studies of adults diagnosed with FMD and subsequently reclassified as having another neurological or medical condition. PubMed was searched from inception to September 13, 2025, focusing on reported features of inconsistency, incongruence with known diseases, and low-validity features. Isolated (iMD) and mixed (mMD) movement disorders were analyzed separately. RESULTS: Forty-two included studies comprised 150 patients, of which 73 cases underwent detailed analysis. Gait disturbance (35%) and dystonia (27.5%) were the most common iMD phenomenologies (n = 40), while jerks and rigidity (24%) were the most common mMD phenomenology (n = 33). Incongruence with known diseases (50.7%), psychological factors (34.2%), and symptom variability (31.5%) were the main factors associated with misdiagnosis. Application of incongruence alone was associated with the highest rate of FMD misdiagnosis, similar to that of applying neither incongruence nor inconsistency (49% in both cases). Conversely, reliance on inconsistency alone was least associated with an FMD misdiagnosis (13.7%). Misdiagnoses of iMD were more likely in the absence of documented inconsistency (OR 3.94, 95% CI 1.32-12.82). CONCLUSIONS: FMD misdiagnoses, particularly of gait disorders and dystonia, were most commonly associated with the application of incongruence as diagnostic criterion. The lowest rate of FMD misdiagnosis was associated with ascertaining at least two positive neurological signs of inconsistency.
9 Apr 2026
Read appraisal →PloS one
Dizziness in the emergency department and risk of stroke: A systematic review and meta-analysis
BACKGROUND: Dizziness is a frequent emergency department (ED) presentation, and a subset of patients, especially those with isolated dizziness without focal neurological deficits, have stroke but are prone to misdiagnosis and adverse outcomes. Reported stroke prevalence in ED dizziness cohorts is highly heterogeneous, and comprehensive assessments of isolated dizziness remain limited. METHODS: Following PRISMA guidelines, we systematically searched PubMed, Web of Science, Embase, and the Cochrane Library for relevant studies. We included cross-sectional studies reporting stroke prevalence among all ED patients with dizziness or isolated dizziness. A random-effects model was used for meta-analysis to calculate pooled prevalence. Subgroup analyses and Egger's test were employed to explore heterogeneity and publication bias. The diagnostic accuracy of bedside diagnostic tools was also systematically reviewed. RESULTS: Twenty-nine studies involving 161,013 ED patients presenting with dizziness were included. The pooled stroke prevalence among all ED dizziness patients (n = 158,583) was 5.5% (95% CI: 4.1-7.1). Among patients with isolated dizziness (n = 2,559), the pooled prevalence was 13.9% (95% CI: 8.2-20.9), substantially higher than in the overall dizziness cohort. Subgroup analyses indicated diagnostic methods and hospital level as major contributors to heterogeneity. Summary analysis of bedside diagnostic tools showed that HINTS and STANDING examinations have high diagnostic accuracy overall, while the TriAGe+ score can be applied flexibly for screening or confirmation based on different cut-off points. CONCLUSIONS: ED patients with dizziness carry a meaningful, setting-dependent stroke risk. Standardized bedside exams (HINTS, STANDING) improve early triage, and selectively deploying neuroimaging helps prevent missed strokes when risk is high.
9 Apr 2026
Read appraisal →Journal of clinical microbiology
Identification of Aspergillus at section and species levels by artificial intelligence-based microscopic morphology image recognition
Rapid and accurate identification of Aspergillus species in clinical microbiology laboratories is crucial for aspergillosis diagnosis and antifungal therapy. However, traditional methods still face challenges in distinguishing phylogenetically related species due to their morphological similarities. This study presents FungalNet, a deep learning model integrating ResNet-50 architecture with Focal Loss algorithm, specifically designed to enhance feature extraction for Aspergillus identification. A total of 12,000 high-resolution images were obtained from lactophenol cotton blue-stained slide preparations under a 100× oil immersion objective, among which 311 images were excluded through a novel quality control approach combining fivefold cross-validation and expert manual review. The performance of four deep learning models (FungalNet and three established models) for identifying Aspergillus species and sections was evaluated using the remaining 11,689 qualified images. FungalNet demonstrated superior classification performance, achieving overall accuracies of 98.45% and 97.85% at the section and species levels, respectively. These results indicate that FungalNet shows significant promise for rapid and accurate identification of Aspergillus species. With further optimization and multicenter validation, this tool could potentially be integrated into routine diagnostic workflows to enhance the efficiency and reliability of fungal identification in clinical settings.IMPORTANCEThis study integrates microscopic morphology identification with deep learning to address the challenge of accurate Aspergillus species identification. Twelve clinically isolated Aspergillus species belonging to eight different sections were included. From touch-tape slide preparations with lactophenol cotton blue staining under a 100× oil immersion objective, 11,689 qualified images were collected and analyzed using FungalNet (our proposed model) along with three established models (GoogLeNet, ResNet-50, and Xception). The results showed that FungalNet demonstrated superior performance in Aspergillus identification, achieving the highest classification accuracy at both section (98.45%) and species (97.85%) levels. Given its rapid turnaround time and cost-effectiveness, this AI-based image analysis approach shows promising potential for the rapid and accurate identification of Aspergillus species in clinical microbiology laboratories.
9 Apr 2026
Read appraisal →Clinical chemistry
Towards Clinical Integration of Deep Learning-Based Classification of Urinary Sediment Particles from Digital Microscopy Images: A Prospective Study
BACKGROUND: Urinalysis is a standard clinical test that includes the microscopic examination of urinary sediment to identify formed elements. Manual evaluation by laboratory technicians is time-intensive and subject to human error. Automated analysis using digital microscopy images presents a potential alternative. This study evaluates the integration of a deep learning approach to automatically classify urinary sediment images in the clinical laboratory, including independent prospective validation of its performance. METHODS: An annotated data set comprising 13 classes of urinary sediment elements was created from a database of Sysmex UD-10 digital microscope images. An EfficientNet-based model was trained and tested across three experimental scenarios to evaluate the effects of data collection strategies on performance. Uncertainty calibration was examined. The model's robustness and interpretability were examined using gradient-weighted class activation mapping (Grad-CAM) to visualize influential image regions and t-distributed stochastic neighbor embedding (t-SNE) to analyze learned feature embeddings. Lastly, a graphical user interface was developed for a prospective evaluation in the laboratory. RESULTS: The model achieved approximately 97% overall accuracy on the test set. Experiments revealed sensitivity to data set variability, suggesting that performance may improve by integrating additional training examples. Confidence scores aligned with accuracy, and interpretability analyses showed that the model focused on relevant image regions and learned embeddings demonstrated clear class separation. In the prospective evaluation, top 1 and top 3 accuracies decreased to approximately 78% and 92%, respectively. CONCLUSIONS: Our results indicate that a lightweight deep learning model can achieve high performance in classifying urine particles. Analysis of discrepancies between retrospective and prospective evaluations provides important insights toward reliable clinical application.
4 Apr 2026
Read appraisal →