Cell. 2025 Apr 4:S0092-8674(25)00289-2. doi: 10.1016/j.cell.2025.03.016. Online ahead of print.

ABSTRACT

Diverse microbes utilize redox shuttles to exchange electrons with their environment through mediated extracellular electron transfer (EET), supporting anaerobic survival. Although mediated EET has been leveraged for bioelectrocatalysis for decades, fundamental questions remain about how these redox shuttles are reduced within cells and their role in cellular bioenergetics. Here, we integrate genome editing, electrochemistry, and systems biology to investigate the mechanism and bioenergetics of mediated EET in Escherichia coli, elusive for over two decades. In the absence of alternative electron sinks, the redox cycling of 2-hydroxy-1,4-naphthoquinone (HNQ) via the cytoplasmic nitroreductases NfsB and NfsA enables E. coli respiration on an extracellular electrode. E. coli also exhibits rapid genetic adaptation in the outer membrane porin OmpC, enhancing HNQ-mediated EET levels coupled to growth. This work demonstrates that E. coli can grow independently of classic electron transport chains and fermentation, unveiling a potentially widespread new type of anaerobic energy metabolism.

PMID:40215961 | DOI:10.1016/j.cell.2025.03.016

EBioMedicine. 2025 Apr 10;115:105688. doi: 10.1016/j.ebiom.2025.105688. Online ahead of print.

ABSTRACT

BACKGROUND: Endometriosis is a chronic inflammatory condition characterised by pain and infertility. We conducted a prospective study to elucidate the pathophysiological mechanisms underlying endometriosis development.

METHODS: We examined the association between 1305 proteins measured by SomaScan proteomics and risk of endometriosis diagnosis in prospectively collected plasma from 200 laparoscopically-confirmed endometriosis cases and 200 risk-set sampling matched controls within the Nurses' Health Study II (NHSII) cohort. Using conditional logistic regression, we calculated odds ratios (OR) and 95% confidence intervals (CI) per one standard deviation increase in protein levels and area under the curve (AUC) to assess the multi-protein model in discriminating cases from controls. Analytical validation for three proteins was performed using immunoassays. Ingenuity Pathway Analysis and STRING analyses identified biological pathways and protein interactions.

FINDINGS: Blood samples from cases were collected up to 9 years before diagnosis (median = 4 years). Among 61 individual proteins nominally significantly associated with risk of endometriosis diagnosis compared to controls, endometriosis cases had higher plasma levels of S100A9 (OR = 1.52, 95%CI = 1.19-1.94), ICAM2 (OR = 1.47, 95%CI = 1.17-1.85), HIST1H3A (OR = 1.42, 95%CI = 1.31-1.78), TOP1 (OR = 1.95, 95%CI = 1.24-3.06), CD5L (OR = 1.23, 95%CI = 1.00-1.51) and lower levels of IGFBP1 (OR = 0.70, 95%CI = 0.52-0.94). We further evaluated three of the proteins in an independent set of 103 matched case-control pairs within the NHSII cohort. Pathway analyses revealed upregulation of multiple immune-related pathways in blood samples collected years before endometriosis diagnosis.

INTERPRETATION: In this prospective analysis using aptamer-based proteomics, we identified multiple proteins and biological pathways related to innate immune response upregulated years before endometriosis surgical diagnosis, suggesting the role of immune dysregulation in endometriosis development.

FUNDING: This study was supported by the Department of Defence, the 2017 Boston Center for Endometriosis Trainee Award. Investigators were supported by Aspira Women's Health and NIH which were not directly related to this project.

PMID:40215752 | DOI:10.1016/j.ebiom.2025.105688

JCO Glob Oncol. 2025 Apr;11:e2400608. doi: 10.1200/GO-24-00608. Epub 2025 Apr 11.

ABSTRACT

Latin America's cultural diversity, political instability, and socioeconomic inequality contribute to fragmented health systems, hindering high-quality care and health equity. Financial limitations in health care exacerbate these issues, affecting care quality and research capacity. In cancer research, such barriers result in inconsistent data, limiting the creation of evidence-based strategies and slowing progress in understanding disease and treatment outcomes. Collaborative academic groups have become essential in overcoming these obstacles. By promoting multidisciplinary cooperation, these groups enhance research, covering areas such as epidemiology, molecular studies, and pharmacogenomics, while also bridging gaps in health care education and infrastructure. Partnerships with various actors provide critical funding and training, supporting sustainable research infrastructure. This review underscores the role of academic collaborations in advancing cancer research and health care delivery in Latin America, fostering innovation and equity. Developing and sustaining these networks will be key to addressing ongoing public health challenges in the region.

PMID:40215435 | DOI:10.1200/GO-24-00608

PLoS One. 2025 Apr 11;20(4):e0319042. doi: 10.1371/journal.pone.0319042. eCollection 2025.

ABSTRACT

BACKGROUND AND OBJECTIVES: Pharmacogenomics (PGx) leverages genomic information to tailor drug therapies, enhancing precision medicine. Despite global advancements, its implementation in Lebanon, Qatar, and Saudi Arabia faces unique challenges in clinical integration. This study aimed to investigate PGx attitudes, knowledge implementation, associated challenges, forecast future educational needs, and compare findings across the three countries.

METHODS: This cross-sectional study utilized an anonymous, self-administered online survey distributed to healthcare professionals, academics, and clinicians in Lebanon, Qatar, and Saudi Arabia. The survey comprised 18 questions to assess participants' familiarity with PGx, current implementation practices, perceived obstacles, potential integration strategies, and future educational needs.

RESULTS: The survey yielded 337 responses from healthcare professionals across the three countries. Data revealed significant variations in PGx familiarity and educational involvement. Qatar and Saudi Arabia participants were more familiar with PGx compared to Lebanon (83%, 75%, and 67%, respectively). Participation in PGx-related talks was most prevalent in Saudi Arabia (96%), followed by Qatar (53%) and Lebanon (35%). Key challenges identified included test cost and reimbursement, insufficient physician knowledge, and lack of infrastructure. Lebanon reported the highest concern for test costs (16%), compared to the lowest in Saudi Arabia (5%). Despite these challenges, a strong consensus emerged on PGx's potential to improve patient outcomes, with over 86% of respondents in all three countries expressing this belief. Educational interest areas varied by country, with strong interest in PGx for cancer chemotherapy in Saudi Arabia and Lebanon and for diabetes mellitus in Qatar.

CONCLUSION: This study highlights the significant influence of varied educational backgrounds and infrastructural limitations on PGx implementation across Lebanon, Qatar, and Saudi Arabia. The findings emphasize the need for targeted strategies in each country to address these distinct barriers. Integrating PGx education into healthcare training programs and clinical workflows could unlock PGx's potential to optimize patient care.

PMID:40215419 | DOI:10.1371/journal.pone.0319042

J Investig Med High Impact Case Rep. 2025 Jan-Dec;13:23247096251334248. doi: 10.1177/23247096251334248. Epub 2025 Apr 11.

ABSTRACT

Cystic fibrosis (CF) is a genetic disorder typically diagnosed in early childhood, caused by mutations in the cystic fibrosis transmembrane conductance regulator gene, leading to thick mucus accumulation in the lungs, pancreas, and other organs. While most diagnoses occur in childhood, a growing number of cases are being identified in adulthood, presenting unique challenges for recognition and management. This case highlights a 37-year-old patient diagnosed with CF after presenting with chronic respiratory symptoms, and weight loss. Late diagnosis of CF remains rare but can delay appropriate treatment, potentially impacting long-term outcomes.

PMID:40215399 | DOI:10.1177/23247096251334248

Rhinology. 2025 Apr 11. doi: 10.4193/Rhin24.545. Online ahead of print.

ABSTRACT

BACKGROUND: Chronic rhinosinusitis (CRS) is a highly prevalent upper airway disease. Its pathogenesis remains poorly understood, especially non-eosinophilic CRS. Currently, no validated mouse model exists to study disease mechanisms, indicating an important research gap. We aimed at establishing a reproducible mouse model of non-eosinophilic CRS to allow further research on its pathophysiology.

METHODOLOGY: Mice were infected with relevant bacteria for sinus disease via surgical insertion of a nasal tampon in their nasal cavity. Inflammatory features in sinus mucosa were evaluated after 4, 8 and 12 weeks on decalcified skulls by histology and immunohistochemistry and by cytospins and enzyme-linked immunoassay on nasal lavage.

RESULTS: S. aureus-inoculated mice showed better survival than S. pneumoniae- and P. aeruginosa- inoculated mice. S. aureus and, to lesser extent, P. aeruginosa were still detectable in the nasal lavage up to 12 weeks. Mice with S. aureus and P. aeruginosa-induced CRS showed significant hypertrophia of the epithelium, neutrophilic infiltration and fibrosis in the sinus mucosa, with increased non-Type 2 cytokines in the nasal lavage.

CONCLUSIONS: S. aureus and P. aeruginosa are more potent inducers of neutrophilic inflammation than S. pneumoniae in mice. This model allows us to further study non-eosinophilic chronic rhinosinusitis pathophysiology in vivo.

PMID:40215396 | DOI:10.4193/Rhin24.545

PLoS One. 2025 Apr 11;20(4):e0319202. doi: 10.1371/journal.pone.0319202. eCollection 2025.

ABSTRACT

In industrial production, obtaining sufficient bearing fault signals is often extremely difficult, leading to a significant degradation in the performance of traditional deep learning-based fault diagnosis models. Many recent studies have shown that data augmentation using generative adversarial networks (GAN) can effectively alleviate this problem. However, the quality of generated samples is closely related to the performance of fault diagnosis models. For this reason, this paper proposes a new GAN-based small-sample bearing fault diagnosis method. Specifically, this study proposes a continuous wavelet convolution strategy (CWCL) instead of the traditional convolution operation in GAN, which can additionally capture the signal's frequency domain features. Meanwhile, this study designed a new multi-size kernel attention mechanism (MSKAM), which can extract the features of bearing vibration signals from different scales and adaptively select the features that are more important for the generation task to improve the accuracy and authenticity of the generated signals. In addition, the structural similarity index (SSIM) is adopted to quantitatively evaluate the quality of the generated signal by calculating the similarity between the generated signal and the real signal in both the time and frequency domains. Finally, we conducted extensive experiments on the CWRU and MFPT datasets and made a comprehensive comparison with existing small-sample bearing fault diagnosis methods, which verified the effectiveness of the proposed approach.

PMID:40215467 | DOI:10.1371/journal.pone.0319202

PLoS One. 2025 Apr 11;20(4):e0321116. doi: 10.1371/journal.pone.0321116. eCollection 2025.

ABSTRACT

Estimating fishing grounds is an important task in the fishing industry. This study modeled the fisher's decision-making process based on sea surface temperature patterns as a pattern recognition task. We used a deep learning-based keypoint detector to estimate fishing ground locations from these patterns. However, training the model required catch data for annotation, the amount of which was limited. To address this, we proposed a training strategy that combines weak supervision and meta-learning to estimate fishing grounds. Weak supervision involves using partially annotated or noisy data, where the labels are incomplete or imprecise. In our case, catch data cover only a subset of fishing grounds, and trajectory data, which are readily available and larger in volume than catch data, provide imprecise representations of fishing grounds. Meta-learning helps the model adapt to the noise by refining its learning rate during training. Our approach involved pre-training with trajectory data and fine-tuning with catch data, with a meta-learner further mitigating label noise during pre-training. Experimental results showed that our method improved the F1-score by 64% compared to the baseline using only catch data, demonstrating the effectiveness of pre-training and meta-learning.

PMID:40215460 | DOI:10.1371/journal.pone.0321116

PLoS One. 2025 Apr 11;20(4):e0320764. doi: 10.1371/journal.pone.0320764. eCollection 2025.

ABSTRACT

Cucumbers play a significant role as a greenhouse crop globally. In numerous countries, they are fundamental to dietary practices, contributing significantly to the nutritional patterns of various populations. Due to unfavorable environmental conditions, they are highly vulnerable to various diseases. Therefore the accurate detection of cucumber diseases is essential for maintaining crop quality and ensuring food security. Traditional methods, reliant on human inspection, are prone to errors, especially in the early stages of disease progression. Based on a VGG19 architecture, this paper uses an innovative transfer learning approach for detecting and classifying cucumber diseases, showing the applicability of artificial intelligence in this area. The model effectively distinguishes between healthy and diseased cucumber images, including Anthracnose, Bacterial Wilt, Belly Rot, Downy Mildew, Fresh Cucumber, Fresh Leaf, Pythium Fruit Rot, and Gummy Stem Blight. Using this novel approach, a balanced accuracy of 97.66% on unseen test data is achieved, compared to a balanced accuracy of 93.87% obtained with the conventional transfer learning approach, where fine-tuning is employed. This result sets a new benchmark within the dataset, highlighting the potential of deep learning techniques in agricultural disease detection. By enabling early disease diagnosis and informed agricultural management, this research contributes to enhancing crop productivity and sustainability.

PMID:40215456 | DOI:10.1371/journal.pone.0320764

Proc Natl Acad Sci U S A. 2025 Apr 15;122(15):e2423072122. doi: 10.1073/pnas.2423072122. Epub 2025 Apr 11.

ABSTRACT

Differentially private gradient descent (DP-GD) is a popular algorithm to train deep learning models with provable guarantees on the privacy of the training data. In the last decade, the problem of understanding its performance cost with respect to standard GD has received remarkable attention from the research community, which has led to upper bounds on the excess population risk [Formula: see text] in different learning settings. However, such bounds typically degrade with overparameterization, i.e., as the number of parameters [Formula: see text] gets larger than the number of training samples [Formula: see text]-a regime which is ubiquitous in current deep-learning practice. As a result, the lack of theoretical insights leaves practitioners without clear guidance, leading some to reduce the effective number of trainable parameters to improve performance, while others use larger models to achieve better results through scale. In this work, we show that in the popular random features model with quadratic loss, for any sufficiently large [Formula: see text], privacy can be obtained for free, i.e., [Formula: see text], not only when the privacy parameter [Formula: see text] has constant order but also in the strongly private setting [Formula: see text]. This challenges the common wisdom that overparameterization inherently hinders performance in private learning.

PMID:40215275 | DOI:10.1073/pnas.2423072122

IEEE Trans Image Process. 2025 Apr 11;PP. doi: 10.1109/TIP.2025.3558089. Online ahead of print.

ABSTRACT

Video Anomaly Detection (VAD) remains a fundamental yet formidable task in the video understanding community, with promising applications in areas such as information forensics and public safety protection. Due to the rarity and diversity of anomalies, existing methods only use easily collected regular events to model the inherent normality of normal spatial-temporal patterns in an unsupervised manner. Although such methods have made significant progress benefiting from the development of deep learning, they attempt to model the statistical dependency between observable videos and semantic labels, which is a crude description of normality and lacks a systematic exploration of its underlying causal relationships. Previous studies have shown that existing unsupervised VAD models are incapable of label-independent data offsets (e.g., scene changes) in real-world scenarios and may fail to respond to light anomalies due to the overgeneralization of deep neural networks. Inspired by causality learning, we argue that there exist causal factors that can adequately generalize the prototypical patterns of regular events and present significant deviations when anomalous instances occur. In this regard, we propose Causal Representation Consistency Learning (CRCL) to implicitly mine potential scene-robust causal variable in unsupervised video normality learning. Specifically, building on the structural causal models, we propose scene-debiasing learning and causality-inspired normality learning to strip away entangled scene bias in deep representations and learn causal video normality, respectively. Extensive experiments on benchmarks validate the superiority of our method over conventional deep representation learning. Moreover, ablation studies and extension validation show that the CRCL can cope with label-independent biases in multi-scene settings and maintain stable performance with only limited training data available.

PMID:40215152 | DOI:10.1109/TIP.2025.3558089

IEEE Trans Image Process. 2025 Apr 11;PP. doi: 10.1109/TIP.2025.3558420. Online ahead of print.

ABSTRACT

In this paper, we propose a new approach to train deep learning models using game theory concepts including Generative Adversarial Networks (GANs) and Adversarial Training (AT) where we deploy a double-oracle framework using best response oracles. GAN is essentially a two-player zero-sum game between the generator and the discriminator. The same concept can be applied to AT with attacker and classifier as players. Training these models is challenging as a pure Nash equilibrium may not exist and even finding the mixed Nash equilibrium is difficult as training algorithms for both GAN and AT have a large-scale strategy space. Extending our preliminary model DO-GAN, we propose the methods to apply the double oracle framework concept to Adversarial Neural Architecture Search (NAS for GAN) and Adversarial Training (NAS for AT) algorithms. We first generalize the players' strategies as the trained models of generator and discriminator from the best response oracles. We then compute the meta-strategies using a linear program. For scalability of the framework where multiple network models of best responses are stored in the memory, we prune the weakly-dominated players' strategies to keep the oracles from becoming intractable. Finally, we conduct experiments on MNIST, CIFAR-10 and TinyImageNet for DONAS-GAN. We also evaluate the robustness under FGSM and PGD attacks on CIFAR-10, SVHN and TinyImageNet for DONAS-AT. We show that all our variants have significant improvements in both subjective qualitative evaluation and quantitative metrics, compared with their respective base architectures.

PMID:40215149 | DOI:10.1109/TIP.2025.3558420

Phys Rev Lett. 2025 Mar 28;134(12):128401. doi: 10.1103/PhysRevLett.134.128401.

ABSTRACT

The spatial organization of DNA involves DNA loop extrusion and the formation of protein-DNA condensates. While the significance of each process is increasingly recognized, their interplay remains unexplored. Using molecular dynamics simulation and theory we investigate this interplay. Our findings reveal that loop extrusion can enhance the dynamics of condensation and promotes coalescence and ripening of condensates. Further, the DNA loop enables condensate formation under DNA tension and position condensates. The concurrent presence of loop extrusion and condensate formation results in the formation of distinct domains similar to TADs, an outcome not achieved by either process alone.

PMID:40215530 | DOI:10.1103/PhysRevLett.134.128401

Cells. 2025 Apr 2;14(7):533. doi: 10.3390/cells14070533.

ABSTRACT

Failure in the proliferation, recruitment, mobilization, and/or differentiation of oligodendrocyte precursor cells (OPCs) impedes remyelination in central nervous system (CNS) demyelinating diseases. Our group has recently achieved the generation of functional oligodendroglia through direct lineage conversion by expressing Sox10, Olig2, and Zfp536 genes in adult rat adipose tissue-derived stromal cells. The present study aimed to determine whether various repurposed drugs or molecules could enhance the myelinating capacities of these induced OPCs (iOPCs). We report that kainate, benztropine, miconazole, clobetasol, and baclofen promote in vitro iOPCs migration, differentiation, and ensheathing abilities through mechanisms similar to those observed in rat neural stem cell-derived OPCs. This research supports the potential use of iOPCs as they provide an alternative and reliable cell source for testing the effects of in vitro promyelinating repurposed drugs and for assessing the molecular and cellular mechanisms involved in therapeutic strategies for demyelinating diseases.

PMID:40214487 | DOI:10.3390/cells14070533

Clin Transl Oncol. 2025 Apr 11. doi: 10.1007/s12094-025-03908-y. Online ahead of print.

ABSTRACT

PURPOSE: CYP2D6 is a key enzyme involved in converting tamoxifen into its active metabolites. However, polymorphisms in CYP2D6 lead to variable enzymatic capacities. We aimed to examine the impact of CYP2D6 polymorphisms on tamoxifen-derived side effects in breast cancer patients.

METHODS: Eighty-six patients with hormone receptor-positive breast cancer who received tamoxifen were classified as poor (PM), intermediate (IM), normal (NM), or ultrarapid (UM) metabolizers according to Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines. All patients received 20 mg/day tamoxifen for 5 years, except PM, who were dose-escalated (20 mg/day for 4 months, 40 mg/day for 4 months, 60 mg/day for 4 months, then back to 20 mg/day). Adverse events-osteoarticular pain, hot flashes, asthenia, and uterine changes-were analyzed by Kaplan-Meier and Cox regression. A propensity score-matched (PSM) subgroup was also examined.

RESULTS: Rapid metabolizers (RM: NM + UM) consistently showed fewer uterine changes compared to slow metabolizers (SM: PM + IM) in both the entire cohort (HR 0.20, p = 0.001) and the PSM subgroup (HR 0.07, p = 0.011). Excluding PM and UM, comparison of IM vs. NM showed similar differences (complete group: HR 0.20, p = 0.002; PSM subgroup: HR 0.23, p = 0.068). Other side effects (joint pain, hot flashes, asthenia) were not significantly associated with CYP2D6 phenotype.

CONCLUSION: Uterine alterations in breast cancer patients treated with tamoxifen appear linked to decreased CYP2D6 activity, although we observed no association between CYP2D6 and other toxicities. These findings suggest closer monitoring for uterine toxicity in individuals with impaired CYP2D6 metabolism.

PMID:40214721 | DOI:10.1007/s12094-025-03908-y

Diabetes Care. 2025 Apr 11:dc250160. doi: 10.2337/dc25-0160. Online ahead of print.

ABSTRACT

OBJECTIVE: Bariatric surgery is an effective treatment option for individuals with obesity and type 2 diabetes (T2D). However, whether outcomes in subtypes of individuals at risk for T2D and/or comorbidities (Tübingen Clusters) differ, is unknown. Of these, cluster 5 (C5) and cluster 6 (C6) are high-risk clusters for developing T2D and/or comorbidities, while cluster 4 (C4) is a low-risk cluster. We investigated bariatric surgery outcomes, hypothesizing that high-risk clusters benefit most due to great potential for metabolic improvement.

RESEARCH DESIGN AND METHODS: We allocated participants without T2D but at risk for T2D, defined by elevated BMI, to the Tübingen Clusters. Participants had normal glucose regulation or prediabetes according to American Diabetes Association criteria. Two cohorts underwent bariatric surgery: a discovery (Lille, France) and a replication cohort (Rome, Italy). A control cohort (Tübingen, Germany) received behavioral modification counseling. Main outcomes included alteration of glucose regulation parameters and prediabetes remission.

RESULTS: In the discovery cohort, 15.0% of participants (n = 121) were allocated to C4, 22.3% (n = 180) to C5, and 62.4% (n = 503) to C6. Relative body weight loss was similar among all clusters; however, reduction of insulin resistance and improvement of β-cell function were strongest in C5. Prediabetes remission rate was lowest in low-risk C4 and highest in high-risk C5. Individuals from high-risk clusters changed to low-risk clusters in both bariatric surgery cohorts but not in the control cohort.

CONCLUSIONS: Participants in C5 had the highest benefit from bariatric surgery in terms of improvement in insulin resistance, β-cell function, and prediabetes remission. This novel classification might help identify individuals who will benefit specifically from bariatric surgery.

PMID:40214701 | DOI:10.2337/dc25-0160

Intern Med J. 2025 Apr 11. doi: 10.1111/imj.16576. Online ahead of print.

ABSTRACT

BACKGROUND: Implementing pharmacogenomic-guided management in cancer patients equitably and effectively in a large population presents challenges. DPYD genotyping determines clinically significant variants of patients at increased risk of developing grade3-5 fluoropyrimidine (FP) toxicity. FP chemotherapies are prescribed for ~16,000 Australians with a 10%-40% grade3-4 toxicity incidence and 1% mortality. Variant carriers can have FP dosing adjusted to improve treatment tolerance without compromising anticancer effect. This strategy has not been formally adopted within Australia, despite widespread international standardisation.

AIM: This pilot study determined genotyping turnaround-times (TAT) for 4 DPYD variants (c.1905+1G>A, c.1679T>G, c.2846A>T and c.1236G>A/Haplotype B3) in Australian patients. Secondary objectives were identification of FP toxicities of DPYD variant carriers, and analysis of healthcare stakeholder perspectives, including enablers/barriers to implementation.

METHODS: Genotyping was determined by Real-Time Polymerase Chain Reaction. Qualitative data were determined through semi-structured questionnaire.

RESULTS: 104 patients recruited over 24 months had a mean TAT of 7.2 days, 5.2 business days (range 1-30). Grade3-4 toxicity occurred in 9/16 DPYD variant carriers, including 2 ICU admissions and 1 death. Themes from 30 questionnaire respondents suggest that clinical environment and resources were fundamental barriers, and motivation to improve patient care was the predominant enabler of change.

CONCLUSION: DPYD genotyping is feasible for improving precision-oncology for patients requiring FP chemotherapies. A TAT of 7 days is acceptable by both stakeholder respondents and national oncology clinician groups. This pilot study, although small, informs a large national project evaluating prospective DPYD genotyping and its impact on FP tolerability, patient safety and cost-effectiveness in Australia.

PMID:40214188 | DOI:10.1111/imj.16576

Orthod Craniofac Res. 2025 Apr 11. doi: 10.1111/ocr.12918. Online ahead of print.

ABSTRACT

Craniosynostosis is a condition characterised by the premature fusion of cranial sutures, which can lead to significant neurodevelopmental and aesthetic issues if not diagnosed and treated early. This study aimed to systematically review and conduct a meta-analysis of studies utilising machine learning (ML) models to diagnose craniosynostosis in photographs or radiographs from humans, evaluating their accuracy through sensitivity, specificity and diagnostic odds ratio. A comprehensive search was conducted on PubMed, Web of Science and Scopus until October 2024 regarding the following PECO question: 'Should ML models (E) be used to diagnose craniosynostosis in photographs or radiographs from humans (P) compared to a reference standard (C) based on their sensitivity, specificity, and diagnostic odds ratio (O)?'. Studies employing ML to diagnose craniofacial deformities on photographs and radiographs of human subjects were included. Using Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2), the risk of bias was assessed. A bivariate random-effect meta-analysis was conducted to pool the diagnostic odds ratio, sensitivity and specificity of the included studies. The GRADE approach was used to evaluate the overall strength of the clinical recommendation and estimated meta-evidence. An initial search yielded 685 articles. After screening, 47 articles were selected for a full-text review. Eventually, 28 studies were selected for the systematic review, and 17 were included in the meta-analysis. The results, with an overall moderate certainty, indicated an AUC of 0.99 (95% CI: 0.98-1.00), an overall sensitivity of 97% (95% CI: 94%-98%) and an overall specificity of 97% (95% CI: 94%-99%). The estimated pooled diagnostic odds ratio was 1131 (95% CI: 290-4419). The present study showed that the ML approaches possess high efficiency and applicability in the diagnosis of craniosynostosis in photographs or radiographs from humans. These findings affirm that ML models should be considered viable diagnostic tools for craniosynostosis.

PMID:40215002 | DOI:10.1111/ocr.12918

Jpn J Radiol. 2025 Apr 11. doi: 10.1007/s11604-025-01780-y. Online ahead of print.

ABSTRACT

BACKGROUND: Esophageal squamous cell carcinoma (ESCC) poses a significant global health challenge with a particularly grim prognosis. Accurate prediction of lymph node metastasis (LNM) in ESCC is crucial for optimizing treatment strategies and improving patient outcomes. This study leverages the power of deep learning, specifically Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks, to analyze arterial phase enhanced CT images and predict LNM in ESCC patients.

METHODS: A retrospective study included 441 ESCC patients who underwent radical esophagectomy and regional lymphadenectomy. CT imaging was performed using contrast-enhanced CT scanners. Tumor region segmentation was conducted to determine the region of interest (ROI), where local tumor 3D volumes were extracted as input for the model. The novel deep learning model, LymphoReso-Net, combined CNN and LSTM networks to process and learn from medical imaging data. The model outputs a binary prediction for LNM. GRAD-CAM was integrated to enhance model interpretability. Performance was evaluated using fivefold cross-validation with metrics including accuracy, sensitivity, specificity, and AUC-ROC. The gold standard for LNM confirmation was pathologically confirmed LNM shortly after the CT.

RESULTS: LymphoReso-Net demonstrated promising performance with an average accuracy of 0.789, an AUC of 0.836, a sensitivity of 0.784, and a specificity of 0.797. GRAD-CAM provided visual explanations of the model's decision-making, aiding in identifying critical regions associated with LNM prediction.

CONCLUSION: This study introduces a novel deep learning framework, LymphoReso-Net, for predicting LNM in ESCC patients. The model's accuracy and interpretability offer valuable insights into lymphatic spread patterns, enabling more informed therapeutic decisions.

PMID:40214915 | DOI:10.1007/s11604-025-01780-y

Sleep Breath. 2025 Apr 11;29(2):155. doi: 10.1007/s11325-025-03325-z.

ABSTRACT

PURPOSE: Accurately identifying sleep states (REM, NREM, and Wake) and brief awakenings (arousals) is essential for diagnosing sleep disorders. Polysomnography (PSG) is the gold standard for such assessments but is costly and requires overnight monitoring in a lab. Home sleep testing (HST) offers a more accessible alternative, relying primarily on breathing measurements but lacks electroencephalography, limiting its ability to evaluate sleep and arousals directly. This study evaluates a deep learning algorithm which determines sleep states and arousals from breathing signals.

METHODS: A novel deep learning algorithm was developed to classify sleep states and detect arousals from respiratory inductance plethysmography signals. Sleep states were predicted for 30-s intervals (one sleep epoch), while arousal probabilities were calculated at 1-s resolution. Validation was conducted on a clinical dataset of 1,299 adults with suspected sleep disorders. Performance was assessed at the epoch level for sensitivity and specificity, with agreement analyses for arousal index (ArI) and total sleep time (TST).

RESULTS: The algorithm achieved sensitivity and specificity of 77.9% and 96.2% for Wake, 93.9% and 80.4% for NREM, 80.5% and 98.2% for REM, and 66.1% and 86.7% for arousals. Bland-Altman analysis showed ArI limits of agreement ranging from - 32 to 24 events/hour (bias: - 4.4) and TST limits from - 47 to 64 min (bias: 8.0). Intraclass correlation was 0.74 for ArI and 0.91 for TST.

CONCLUSION: The algorithm identifies sleep states and arousals from breathing signals with agreement comparable to established variability in manual scoring. These results highlight its potential to advance HST capabilities, enabling more accessible, cost-effective and reliable sleep diagnostics.

PMID:40214714 | DOI:10.1007/s11325-025-03325-z