Biomater Res. 2025 May 28;29:0211. doi: 10.34133/bmr.0211. eCollection 2025.

ABSTRACT

While fluorescent labeling has been the standard for visualizing fibers within fibrillar scaffold models of the extracellular matrix (ECM), the use of fluorescent dyes can compromise cell viability and photobleach prematurely. The intricate fibrillar composition of ECM is crucial for its viscoelastic properties, which regulate intracellular signaling and provide structural support for cells. Naturally derived biomaterials such as fibrin and collagen replicate these fibrillar structures, but longitudinal confocal imaging of fibers using fluorescent dyes may impact cell function and photobleach the sample long before termination of the experiment. An alternative technique is reflection confocal microscopy (RCM) that provides high-resolution images of fibers. However, RCM is sensitive to fiber orientation relative to the optical axis, and consequently, many fibers are not detected. We aim to recover these fibers. Here, we propose a deep learning tool for predicting fluorescently labeled optical sections from unlabeled image stacks. Specifically, our model is conditioned to reproduce fluorescent labeling using RCM images at 3 laser wavelengths and a single laser transmission image. The model is implemented using a fully convolutional image-to-image mapping architecture with a hybrid loss function that includes both low-dimensional statistical and high-dimensional structural components. Upon convergence, the proposed method accurately recovers 3-dimensional fibrous architecture without substantial differences in fiber length or fiber count. However, the predicted fibers were slightly wider than original fluorescent labels (0.213 ± 0.009 μm). The model can be implemented on any commercial laser scanning microscope, providing wide use in the study of ECM biology.

PMID:40438124 | PMC:PMC12117218 | DOI:10.34133/bmr.0211

Adv Mater. 2025 May 29:e2502322. doi: 10.1002/adma.202502322. Online ahead of print.

ABSTRACT

Nanozymes hold great potential in protecting human health. However, constructing new and efficient nanozymes is a significant challenge. Developing atomic-level nanozymes is a promising approach. Despite their potential, atomic-level high-entropy nanozymes have not been reported due to thermodynamic instability. Therefore, developing atomic-level high-entropy nanozymes are of great significance. What's more, further exploring their biomedical applications can open up new horizons for nanozymology. Here, the atomic-level high-entropy nanozyme system capable of remarkable endogenous targeted catalysis and enhancing tumor photothermal therapy is successfully constructed. The system is prepared by reduction-diffusion and grafting methods. The RuRhPtIrMo sub-nanometer high-entropy nanozyme (snHEAzyme) with about 8-10 atoms thickness is first prepared. Then, they are grafted by targeting agent DSPE-PEG2000-cRGD and imaging agent Cy7 to obtain the snHEAzyme@DSPE-PEG2000-cRGD@Cy7 nanozyme system. The synthesized snHEAzyme@DSPE-PEG2000-cRGD@Cy7 system exhibits excellent peroxidase-like activity and high absorbance in the near-infrared (NIR) range. Under NIR irradiation, the nanozyme shows efficient photothermal conversion and reactive oxygen species generation effects. In vitro and in vivo experiments demonstrated that the snHEAzyme@DSPE-PEG2000-cRGD@Cy7 system can be effectively targeted to penetrate tumor cell membranes and treat tumors. This work offers a new perspective on snHEAzyme fabrication and its biomedical applications.

PMID:40437935 | DOI:10.1002/adma.202502322

Ethiop J Health Sci. 2024 Nov;34(6):459-468. doi: 10.4314/ejhs.v34i6.5.

ABSTRACT

BACKGROUND: Chronic Liver Disease (CLD) is a long-term condition marked by a gradual decline in liver function. Patients with CLD often experience multimorbidity and polypharmacy, which can adversely affect their health outcomes. The objective of the current study is to identify and resolve the drug-related problems associated with chronic liver disease.

METHODS: This prospective observational study involved 150 patients with CLD over a six-month period. Eligible participants included individuals over 18 years old, diagnosed with CLD based on the Child-Pugh score, and currently receiving treatment. Drug-related problems (DRPs) were identified using the Pharmaceutical Care Network Europe (PCNE) classification version 9.1. Data analysis was conducted using Chi-square and Fisher's exact tests with SPSS software version 29.

RESULTS: A total of 212 DRPs were identified and resolved. The most frequent type of DRP was related to treatment efficacy, with 96 instances (45.29%). Within this category, the subcategory 'effect of drug treatment not optimal' was the most common, accounting for 45 patients (21.23%). Drug interactions were identified as the leading cause of DRPs, comprising 65 cases (30.66%). Most issues were addressed at the prescriber level, with 48.11% of interventions accepted by physicians.

CONCLUSION: This study provides valuable insights into identifying and managing DRPs that can negatively impact treatment outcomes in CLD patients. The findings can assist healthcare professionals in prioritizing strategies to enhance clinical results.

PMID:40438433 | PMC:PMC12110266 | DOI:10.4314/ejhs.v34i6.5

Chin Med J (Engl). 2025 May 29. doi: 10.1097/CM9.0000000000003672. Online ahead of print.

ABSTRACT

BACKGROUND: Chronic kidney disease (CKD)-associated anemia (CKD-anemia) is associated with poor survival, and hemoglobin targets are often not achieved with current therapies. Phase 3 trials have demonstrated the treatment efficacy of roxadustat for CKD-anemia. This phase four study aims to evaluate the long-term (52-week) safety and effectiveness of roxadustat in a broad real-world patient population with CKD-anemia with and without dialysis in China.

METHODS: This Phase 4 multicenter, open-label, prospective study, conducted from 24 November 2020 to 11 November 2022, evaluated the long-term safety and effectiveness of roxadustat for CKD-anemia in China. Patients aged ≥18 years with CKD-anemia with or without dialysis were included. The initial oral dose was 70-120 mg (weight-based followed by dose adjustment) over 52 weeks. The primary endpoint was safety based on adverse events (AEs). The secondary endpoints were hemoglobin changes from baseline and the proportion of patients who achieved mean hemoglobin ≥100 g/L. Effectiveness evaluable populations 1 (EE1) and EE2 included roxadustat-naïve and previously roxadustat-treated patients, respectively. The safety analysis set (SAF) included all patients who received ≥1 dose.

RESULTS: The EE1, EE2, and SAF populations included 1804, 193, and 2021 patients, respectively. In the SAF, the mean age was 50 ± 14 years, and 1087 patients (53.8%) were male. Mean baseline hemoglobin was 96.9 ± 14.0 g/L in EE1 and 100.3 ± 12.9 g/L in EE2. In EE1, the mean (95% confidence interval) hemoglobin changes from baseline over weeks 24-36 and 36-52 were 14.2 (13.5-14.9) g/L and 14.3 (13.5-15.0) g/L, respectively. Over weeks 24-36 and 36-52, 83.3% and 86.1% of patients in EE1 and 82.7% and 84.7% in EE2 achieved mean hemoglobin ≥100 g/L, respectively. In the SAF, 1643 patients (81.3%) experienced treatment-emergent AEs (TEAEs). Overall, 219 patients (10.8%) experienced drug-related TEAEs. Thirty-eight patients (1.9%) died of TEAEs (unrelated to the study drug). Vascular access thrombosis was uncommon.

CONCLUSIONS: Roxadustat (52 weeks) increased hemoglobin and maintained the treatment target in Chinese patients with CKD-anemia with acceptable safety, supporting its use in real-world settings.

REGISTRATION: Chinese Clinical Trial Registry (www.chictr.org.cn) ChiCTR2100046322 CDE (www.chinadrugtrials.org.cn) CTR20201568.

PMID:40437668 | DOI:10.1097/CM9.0000000000003672

Medicina (Kaunas). 2025 May 12;61(5):881. doi: 10.3390/medicina61050881.

ABSTRACT

Background and Objectives: Poretti-Boltshauser syndrome (PBS) is a rare, autosomal recessive disorder caused by pathogenic variants in the LAMA1 gene, resulting in laminin dysfunction. This manifests as a cerebellar malformation with cysts, and patients present with developmental delay and ataxia; however, ocular features are not well-characterised. We aimed to summarise the ocular phenotypes of PBS based on cases reported in the literature. Materials and Methods: A literature search was conducted on Medline, Embase, and PubMed on PBS and its ocular associations. Genetically confirmed PBS cases were reviewed, and genotype-phenotype correlations were investigated. Results: Comprehensive reporting of genotypes and associated systemic and ocular phenotypes was available in 51 patients with PBS, who had 52 distinct variants in LAMA1. Most patients carried homozygous variants. The most common genotype was a c.2935delA homozygous mutation, followed by the c.768+1G>A; c.6701delC compound heterozygous mutation. High myopia was the most common ocular phenotype (n = 39), followed by strabismus (n = 27) and ocular motor apraxia (n = 26). A wide range of other ocular manifestations, including retinal dystrophy, retinal neovascularisation, retinal detachment, strabismus, nystagmus, optic disc and iris hypoplasia, were reported. Patients with the same genotype exhibited variable expressivity. Conclusions: PBS has a broad ocular phenotypic spectrum, and characterisation of this variability is important for making an accurate diagnosis and informing genetic counselling.

PMID:40428839 | PMC:PMC12113114 | DOI:10.3390/medicina61050881

Future Oncol. 2025 Jun;21(13):1639-1645. doi: 10.1080/14796694.2025.2497257. Epub 2025 May 28.

ABSTRACT

BACKGROUND: Alleviating the toxic and adverse reactions associated with chemotherapy is crucial for improving patient outcomes. This study aimed to assess the impacts of positive emotion, engagement, relationships, meaning, and accomplishment (PERMA) model-based psychological interventions and predictive chemotherapy reaction nursing on patients with malignant tumors following chemotherapy.

RESEARCH DESIGN AND METHODS: The control group (n = 43) received conventional care, while the observation group (n = 43) received psychological intervention based on PERMA model alongside predictive nursing care. Chemotherapy-induced toxicity and side effects, fatigue levels, coping mode, psychological status, and quality of life were assessed.

RESULTS: Compared to the control group, the observation group exhibited a lower incidence of gastrointestinal adverse reactions, myelosuppression, alopecia, and oral ulcers (p < 0.05), reduced behavioral, cognitive, somatic, and emotional fatigue (p < 0.001), lower scores in avoidance and yielding coping styles (p < 0.001), higher scores in confrontation coding mode (p = 0.056), improved quality of life, and better outcomes in anxiety, depression, and overall psychological state of patients (p < 0.001).

CONCLUSION: PERMA model-based psychological interventions and predictive chemotherapy reaction nursing interventions effectively reduce the incidence of chemotherapy-induced toxicity, alleviate fatigue, enhance quality of life, and improve psychological well-being in cancer patients.

PMID:40432478 | DOI:10.1080/14796694.2025.2497257

Oncoimmunology. 2025 Dec;14(1):2512109. doi: 10.1080/2162402X.2025.2512109. Epub 2025 May 28.

ABSTRACT

Nonmutated mitochondrial DNA (mtDNA) from T lymphocytes can be incorporated into cancer cells bearing mutated mtDNA to repair their bioenergetic deficiency. However, a recent paper by Ikeda et al. indicates that mutated mtDNA from malignant cells can also be transferred into tumor-infiltrating T lymphocytes to subvert their function in cancer immunosurveillance.

PMID:40434021 | PMC:PMC12123971 | DOI:10.1080/2162402X.2025.2512109

Children (Basel). 2025 Apr 28;12(5):567. doi: 10.3390/children12050567.

ABSTRACT

INTRODUCTION: Methadone is an opioid-sparing opioid and it is increasingly used in children undergoing surgery due to its beneficial effects on postoperative pain scores, decreased opioid requirements, and fewer adverse effects compared to other opioids. Intraoperative methadone is not well studied in pediatric cardiac surgery. We hypothesized that intraoperative methadone-based analgesia would provide comparable effectiveness in pain management to non-methadone-based analgesia, including caudal morphine, following pediatric cardiac surgery.

METHODS: We conducted a retrospective cohort study of 287 children undergoing cardiac surgery using single institutional electronic health records with Society of Thoracic Surgeons database outcomes. Patients were administered intravenous opioids plus caudal morphine (≤6 years) or intravenous opioids in the non-methadone group versus intravenous methadone (two 0.1 mg/kg doses given intraoperatively) with or without additional intraoperative opioids. The primary outcome was postoperative opioid use in morphine milligram equivalents (MME)/kg.

RESULTS: This study included 287 pediatric cardiac surgical patients with a mean age of 3.8 years, 59% male, and 72% White. Among 287 patients, 67 (23%) received intraoperative methadone. Unadjusted analysis showed the methadone group had lower postoperative opioid use on the day of surgery (median = 0.3 vs. 0.5 MME/kg, p = 0.005). Adjusted analyses showed there were no significant differences in postoperative opioid use, average pain, maximum pain, antiemetic use, reintubation, and use of naloxone between methadone and non-methadone groups. Hospital length of stay was 2.62 times longer (95% CI: [1.55, 4.41] p < 0.001) in the methadone group vs non-methadone group, but this was only shown in the younger children (≤6 years), who also had higher max pain scores in the methadone group. All outcomes were similar between analgesia groups in older children (>6 years).

CONCLUSIONS: Intraoperative methadone-based analgesia had comparable effectiveness in postoperative opioid use, pain, and antiemetic use compared to non-methadone-based intraoperative pain management for pediatric cardiac surgery. Large prospective studies of perioperative methadone are needed to examine methadone's analgesic benefits in children undergoing cardiac surgery.

PMID:40426746 | PMC:PMC12109820 | DOI:10.3390/children12050567

Orphanet J Rare Dis. 2025 May 28;20(1):258. doi: 10.1186/s13023-025-03754-9.

ABSTRACT

INTRODUCTION: In Thailand, obtaining medicines for rare diseases presents significant challenges, with limited evidence highlighting these issues.

OBJECTIVES: To evaluate the accessibility of medicines and the extent of health insurance coverage for treatments of rare diseases in Thailand.

METHOD: This study utilized a thorough review of current health policies, drug registration database, and insurance coverage conditions. Additionally, procurement data from the Ministry of Finance was analyzed to verify the acquisition of medicines intended for the treatment of rare diseases.

RESULTS: A review of the availability and procurement of medicines for rare diseases in Thailand revealed considerable limitations in both registration and accessibility. According to the International Rare Diseases Research Consortium, only 46.80% of their recommended medicines were registered in Thailand, and of these, just 22.93% were included in the national essential medicines list. Additionally, a review of the state's pharmaceutical procurement dataset over the past 5 years showed that merely 31.70% of these registered drugs had been purchased from suppliers for use in hospitals.

CONCLUSION: To address these issues, the study recommended accelerating the approval process for rare disease medicines, expanding health insurance coverage, establishing financial support for patients, and creating a specific pricing policy for orphan drugs. Collaborative efforts among stakeholders were emphasized as crucial for improving access to essential medicines and enhancing treatment outcomes for patients with rare diseases in Thailand.

PMID:40437602 | DOI:10.1186/s13023-025-03754-9

BMC Pharmacol Toxicol. 2025 May 28;26(1):112. doi: 10.1186/s40360-025-00951-x.

ABSTRACT

BACKGROUND: Hepatic injury is a common pathological process for a wide spectrum of liver diseases. Quercetin has been found to counteract this process by scavenging free radicals, but its therapeutic effect is limited due to poor water-solubility. Thus, the question of how to deliver quercetin to a target organ effectively with minimal side effects has remained a clinical challenge. Our previous research findings indicate that when quercetin is delivered in the form of liposomal nanoparticles, its targeting efficiency to the liver is significantly enhanced. Although quercetin liposomal nanoparticles have been shown to improve the therapeutic effect on liver damage compared to traditional quercetin treatment, the optimal dosage of liposomal quercetin still warrants further exploration. The aim of this study was therefore to ascertain whether there are differences in the therapeutic effects on liver damage at different dosages of quercetin liposomes and to determine the optimal dosage.

METHODS: 62 rats modeled with liver injury were enrolled and distributed into 4 groups, where they were treated with quercetin liposome nanoparticles, blank liposome nanoparticles, simple quercetin, and normal saline accordingly. Serum samples were measured for liver function indicators, and tissue samples were analyzed by pathohistological examination. Statistical analysis was performed to quantify the difference between the experimental and control groups.

RESULTS: Both liver function and histopathological examinations demonstrated enhanced therapeutic effects as the concentration of quercetin liposome drugs increased. Moreover, compared to traditional quercetin treatments, liposomal quercetin nanoparticles of varying concentrations uniformly provide better liver protection, with the highest dose group showing the best therapeutic effect. In addition, low concentration carrier liposome nanoparticles also showed a certain protective effect on the liver damage in rats.

CONCLUSION: Liposomal quercetin nanoparticles exhibit superior efficacy in liver protection and repair compared to pure quercetin, with the highest dose group showing the best therapeutic effect.

PMID:40437639 | DOI:10.1186/s40360-025-00951-x

Sci Rep. 2025 May 28;15(1):18722. doi: 10.1038/s41598-025-02773-7.

ABSTRACT

The COVID-19 pandemic has initiated a global health emergency, with an exigent need for an effective cure. Progressively, drug repurposing is emerging as a promising solution for saving time, cost, and labor. However, the number of drug candidates that have been identified for the treatment of COVID-19 is still insufficient, so more effective and thorough drug exploration strategies are required. In this study, we joined the molecular docking with machine learning approaches to find some prospective therapeutic candidates for COVID-19 treatment. We screened the 5903 approved drugs for their inhibition by targeting the replicating enzyme 3CLpro of SARS-CoV-2. Molecular docking is used to calculate the binding affinities of these drugs towards 3CLpro. We employed several machine learning approaches for QSAR modeling to explore some potential drugs with high binding affinities. Our outcomes demonstrated that the Decision Tree Regression (DTR) model, with the best scores of R² and RMSE, is the most suitable model to explore the potential drugs. We shortlisted six favorable drugs with their respective Zinc IDs (3873365, 85432544, 203757351, 85536956, 8214470, and 261494640) within the range of -15 kcal/mol to -13 kcal/mol. We further examined the physiochemical and pharmacokinetic properties of these most potent drugs. Our study provides an efficient framework to explore the potential drugs against COVID-19 and establishes the impending combination of molecular docking with machine learning approaches to accelerate the identification of potential therapeutic candidates. Our verdicts contribute to the larger goal of finding effective cures for COVID-19, which is an acute global health challenge. The outcomes of our study provide valuable insights into potential therapeutic candidates for COVID-19 treatment.

PMID:40436944 | DOI:10.1038/s41598-025-02773-7

Hum Genomics. 2025 May 28;19(1):62. doi: 10.1186/s40246-025-00765-2.

ABSTRACT

BACKGROUND: Variations in pharmacogenes that regulate drug absorption, distribution, metabolism, and excretion (ADME) contribute to approximately 20-30% of interindividual differences in drug response. While many common variants are successfully utilized in clinical settings to predict individual drug responses, a significant portion of the genetic basis underlying this variability remains unidentified. This includes rare variants, which are estimated to account for 4-6% of drug response variability.

RESULTS: To comprehensively elucidate the functional consequences and molecular mechanisms of rare variants, we conducted in vitro enzyme expression studies combined with in silico structure-function analyses. We selected 11 rare variants in the CYP2C19 and CYP2D6 genes identified among participants within the Estonian Biobank. Variant cDNAs were heterologously expressed in HEK-293 cells, and detailed enzyme activity analyses were performed. The experimental results were further validated against average scores from five optimized in silico prediction models: LRT, Mutation Assessor, PROVEAN, VEST3, and CADD. To explore structure-activity relationships, we performed in silico docking of substrates into available 3D enzyme structures. Our findings reveal that most of the rare genetic variants caused significant functional alterations, including: (i) Likely impairments in substrate transport to the active site due to narrowing of access channels; (ii) Changes in catalytic rates; and (iii) Potential effects on substrate extrusion rates from the active site. The in silico prediction tools accurately anticipated the functional impact of 6 out of the 11 variants (54%).

CONCLUSIONS: Evaluating the functionality of rare variants will become increasingly essential as rapid and cost-effective whole-genome sequencing technologies continue to advance. Our results highlight the need for further refinement of in silico prediction models, particularly those leveraging 3D crystal enzyme structures, to enhance the accuracy of functional predictions for rare genetic variants.

PMID:40437642 | DOI:10.1186/s40246-025-00765-2

BMC Pediatr. 2025 May 28;25(1):430. doi: 10.1186/s12887-025-05789-0.

ABSTRACT

OBJECTIVE: Cystic fibrosis (CF) is a fatal hereditary disorder that leads to respiratory infections and gastrointestinal inflammation with possible association with intestinal dysbiosis. The present study was conducted with the aim of investigating the effects of probiotic consumption in improving pulmonary, gastrointestinal, and growth symptoms in patients with CF.

MATERIALS AND METHODS: In this double-blind randomized clinical trial, 110 CF patients were examined. Patients were divided into two equal groups of 55 subjects. Patients in the probiotic group consumed Lactobacillus reuteri at the rate of 108 CFU/d for one month, and the control group received a placebo. Then, pulmonary, gastrointestinal, and growth-related outcomes as well as quality of life were assessed after one month of intervention as well as at three-month follow-up.

RESULTS: The results of our study showed that in both intervention and control groups, weight increases significantly after 12 weeks (P = 0.01). However, no remarkable difference was reported between the two groups after 12 weeks (P = 0.09). In addition, no significant changes were observed between the two groups after 4 and 12 weeks regarding BMI and FEV1. Based on the findings, the score of the CFQ questionnaire in the intervention group increased significantly in the 4th and 12th week. No significant differences were observed between the two groups in terms of factors related to lung function or exacerbations after 12 weeks.The only notable effect reported was related to pain attacks in the probiotic group compared to the placebo group after 4 weeks (P = 0.02).

CONCLUSION: In general, treatment with probiotics improved the quality of life in patients with CF. However, no significant effect was observed on pulmonary, gastrointestinal, and growth-related outcomes.

TRIAL REGISTRATION: This study was retrospectively registered IRCT registration number: IRCT20240105060622N1 (Registration date: 2024-08-16).

PMID:40437397 | DOI:10.1186/s12887-025-05789-0

J Cardiothorac Surg. 2025 May 28;20(1):246. doi: 10.1186/s13019-025-03488-6.

ABSTRACT

AIM: To assess the predictive performance, risk stratification capabilities, and auxiliary diagnostic utility of radiomics, deep learning, and fusion models in identifying visceral pleural invasion (VPI) in lung adenocarcinoma.

MATERIALS AND METHODS: A total of 449 patients (female:male, 263:186; 59.8 ± 10.5 years) diagnosed with clinical IA stage lung adenocarcinoma (LAC) from two distinct hospitals were enrolled in the study and divided into a training cohort (n = 289) and an external test cohort (n = 160). The fusion models were constructed from the feature level and the decision level respectively. A comprehensive analysis was conducted to assess the prediction ability and prognostic value of radiomics, deep learning, and fusion models. The diagnostic performance of radiologists of varying seniority with and without the assistance of the optimal model was compared.

RESULTS: The late fusion model demonstrated superior diagnostic performance (AUC = 0.812) compared to clinical (AUC = 0.650), radiomics (AUC = 0.710), deep learning (AUC = 0.770), and the early fusion models (AUC = 0.586) in the external test cohort. The multivariate Cox regression analysis showed that the VPI status predicted by the late fusion model were independently associated with patient disease-free survival (DFS) (p = 0.044). Furthermore, model assistance significantly improved radiologist performance, particularly for junior radiologists; the AUC increased by 0.133 (p < 0.001) reaching levels comparable to the senior radiologist without model assistance (AUC: 0.745 vs. 0.730, p = 0.790).

CONCLUSIONS: The proposed decision-level (late fusion) model significantly reducing the risk of overfitting and demonstrating excellent robustness in multicenter external validation, which can predict VPI status in LAC, aid in prognostic stratification, and assist radiologists in achieving higher diagnostic performance.

PMID:40437608 | DOI:10.1186/s13019-025-03488-6

Acta Neuropathol Commun. 2025 May 28;13(1):120. doi: 10.1186/s40478-025-02030-y.

ABSTRACT

Postmortem neuropathological examination, while the gold standard for diagnosing neurodegenerative diseases, often relies on limited regional sampling that may miss critical areas affected by Alzheimer's disease and related disorders. Ultra-high resolution postmortem MRI can help identify regions that fall outside the diagnostic sampling criteria for additional histopathologic evaluation. However, there are no standardized guidelines for integrating histology and MRI in a traditional brain bank. We developed a comprehensive protocol for whole hemisphere postmortem 7T MRI-guided histopathological sampling with whole-slide digital imaging and histopathological analysis, providing a reliable pipeline for high-volume brain banking in heterogeneous brain tissue. Our method uses patient-specific 3D printed molds built from postmortem MRI, allowing standardized tissue processing with a permanent spatial reference frame. To facilitate pathology-MRI association studies, we created a semi-automated MRI to histology registration pipeline and developed a quantitative pathology scoring system using weakly supervised deep learning. We validated this protocol on a cohort of 29 brains with diagnosis on the AD spectrum that revealed correlations between cortical thickness and phosphorylated tau accumulation. This pipeline has broad applicability across neuropathological research and brain banking, facilitating large-scale studies that integrate histology with neuroimaging. The innovations presented here provide a scalable and reproducible approach to studying postmortem brain pathology, with implications for advancing diagnostic and therapeutic strategies for Alzheimer's disease and related disorders.

PMID:40437594 | DOI:10.1186/s40478-025-02030-y

BMC Med. 2025 May 28;23(1):308. doi: 10.1186/s12916-025-04150-7.

ABSTRACT

BACKGROUND: There is a lack of studies exploring the performance of Transformers-based language models in common risks assessment among psychiatric inpatients. We aim to develop a scalable risk assessment model using multidimensional textualized data and test the stability, robustness, and benefit of this approach.

METHODS: In this real-world cohort study, a deep learning language model was developed and validated using first hospitalized cases diagnosed with schizophrenia, bipolar disorder, and depressive disorder between January 2016 and March 2023 in three hospitals. The algorithm was externally validated on an independent testing cohort comprising 1180 patients. A total of 140 features, including first medical records (FMR), laboratory examinations, medical orders, and psychological scales, were assessed for analysis. The outcomes were short- and long-term impulsivity (STI and LTI), risk of suicide (STSS and LTSS), and need of physical restraint (STPR and LTPR) assessed by qualified nurses or clinicians. Analysis was carried out between August 2024 and June 2024. Models with different architectures and input settings were compared with each other. The area under the receiver operating characteristic curve (AUROC) was used to assess the primary performance of models. The clinical utility was determined by the net benefit under Youden's threshold.

RESULTS: Of 7451 patients included in this study, 2982 (47.6%) were male, and the median (interquartile range) age was 42 (28-57) years. The overall incidence of outcomes was 635 (8.5%), 728 (10.5%), 659 (8.8%), 803 (10.8%), 588 (7.9%), and 728 (9.8%) for STPR, LTPR, STSS, LTSS, STI, and LTI, respectively. The multitask semi-structured Transformers-based language (SSTL) model showed more promising AUROCs (STPR: 0.915; LTPR: 0.844; STSS: 0.867; LTSS: 0.879; STI: 0.899; LTI: 0.894) in the prediction of these outcomes than single-tasked or multimodal language models and traditional structured data models. Combining FMR with other data from electronic health records led to significant improvements in the performance and clinical utility of SSTL models based on demographic, diagnosis, laboratory tests, treatment, and psychological scales.

CONCLUSIONS: The SSTL model shows potential advantages in prognostic evaluation. FMR is a strong predictor for common risks prediction and may benefit other tasks in psychiatry with minimum requirements for data and data processing.

PMID:40437564 | DOI:10.1186/s12916-025-04150-7

BMC Biol. 2025 May 28;23(1):142. doi: 10.1186/s12915-025-02243-6.

ABSTRACT

BACKGROUND: Neuropeptides (NPs) are critical signaling molecules involved in various physiological and behavioral processes, including development, metabolism, and memory. They function within both the nervous and endocrine systems and have emerged as promising therapeutic targets for a range of diseases. Despite their significance, the accurate identification of NPs remains a challenge, necessitating the development of more effective computational approaches.

RESULTS: In this study, we introduce NeuroScale, a multi-channel neural network model leveraging evolutionary scale modeling (ESM) for the precise prediction of NPs. By integrating the GoogLeNet framework, NeuroScale effectively captures multi-scale NP features, enabling robust and accurate classification. Extensive benchmarking demonstrates its superior performance, consistently achieving an area under the receiver operating characteristic curve (AUC) exceeding 0.97. Additionally, we systematically analyzed the impact of protein sequence similarity thresholds and multi-scale sequence lengths on model performance, further validating NeuroScale's robustness and generalizability.

CONCLUSIONS: NeuroScale represents a significant advancement in neuropeptide prediction, offering both high accuracy and adaptability to diverse sequence characteristics. Its ability to generalize across different sequence similarity thresholds and lengths underscores its potential as a reliable tool for neuropeptide discovery and peptide-based drug development. By providing a scalable and efficient deep learning framework, NeuroScale paves the way for future research in neuropeptide function, disease mechanisms, and therapeutic applications.

PMID:40437538 | DOI:10.1186/s12915-025-02243-6

BMC Oral Health. 2025 May 28;25(1):828. doi: 10.1186/s12903-025-06196-8.

ABSTRACT

OBJECTIVES: Ponticulus posticus (PP) is a bony structure in the cervical spine, often difficult to identify in radiographic images, and its detection is important for both orthodontic diagnosis and clinical decision-making related to craniovertebral pathologies. The purpose of this study is to develop a deep learning-based approach for detecting the PP in lateral cephalometric radiographs using the YOLOv8-seg model.

METHODS: This retrospective study analyzed a dataset of 1000 anonymized lateral cephalometric radiographs, focusing on the segmentation and detection of the PP. Images were resized to 640 × 640 pixels and labeled by two experienced dentomaxillofacial radiologists. The YOLOv8-seg model, designed for segmentation tasks, was trained over 100 epochs with a batch size of sixteen, using pre-trained weights from the COCO dataset. Model performance was evaluated using precision, recall, mean average precision (mAP), and F1 score metrics.

RESULTS: The YOLOv8s-seg model demonstrated high accuracy in detecting the PP, with a precision of 62.81%, recall of 88.7%, mAP50 of 75.27%, mAP95 of 62.28%, and an F1 score of 73.54%. Even in cases where the boundaries of the C1 cervical vertebra were not clearly distinguishable, the model performed effectively, suggesting its reliability in clinical applications.

CONCLUSIONS: The proposed YOLOv8-seg model shows promising potential for improving the accuracy and efficiency of PP detection in lateral cephalometric radiographs. By integrating AI into the diagnostic process, orthodontic practices can benefit from more precise and reliable identification of small but clinically significant anatomical structures, ultimately enhancing patient care and diagnostic accuracy.

PMID:40437474 | DOI:10.1186/s12903-025-06196-8

BMC Cancer. 2025 May 28;25(1):963. doi: 10.1186/s12885-025-14353-z.

ABSTRACT

BACKGROUND: Cervical cancer remains a significant global health issue, with accurate differentiation between low-grade (LSIL) and high-grade squamous intraepithelial lesions (HSIL) crucial for effective screening and management. Current methods, such as Pap smears and HPV testing, often fall short in sensitivity and specificity. Deep learning models hold the potential to enhance the accuracy of cervical cancer screening but require thorough evaluation to ascertain their practical utility.

METHODS: This study compares the performance of two advanced deep learning models, SEResNet101 and SE-VGG19, in classifying cervical lesions using a dataset of 3,305 high-quality colposcopy images. We assessed the models based on their accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC).

RESULTS: The SEResNet101 model demonstrated superior performance over SE-VGG19 across all evaluated metrics. Specifically, SEResNet101 achieved a sensitivity of 95%, a specificity of 97%, and an AUC of 0.98, compared to 89% sensitivity, 93% specificity, and an AUC of 0.94 for SE-VGG19. These findings suggest that SEResNet101 could significantly reduce both over- and under-treatment rates by enhancing diagnostic precision.

CONCLUSION: Our results indicate that SEResNet101 offers a promising enhancement over existing screening methods, integrating advanced deep learning algorithms to significantly improve the precision of cervical lesion classification. This study advocates for the inclusion of SEResNet101 in clinical workflows to enhance cervical cancer screening protocols, thereby improving patient outcomes. Future work should focus on multicentric trials to validate these findings and facilitate widespread clinical adoption.

PMID:40437403 | DOI:10.1186/s12885-025-14353-z

Phys Eng Sci Med. 2025 May 28. doi: 10.1007/s13246-025-01566-6. Online ahead of print.

ABSTRACT

Ultrasound fetal imaging is beneficial to support prenatal development because it is affordable and non-intrusive. Nevertheless, fetal plane classification (FPC) remains challenging and time-consuming for obstetricians since it depends on nuanced clinical aspects, which increases the difficulty in identifying relevant features of the fetal anatomy. Thus, to assist with its accurate feature extraction, a lightweight artificial intelligence architecture leveraging convolutional neural networks and attention mechanisms is proposed to classify the largest benchmark ultrasound dataset. The approach fine-tunes from lightweight EfficientNet feature extraction backbones pre-trained on the ImageNet1k. to classify key fetal planes such as the brain, femur, thorax, cervix, and abdomen. Our methodology incorporates the attention mechanism to refine features and 3-layer perceptrons for classification, achieving superior performance with the highest Top-1 accuracy of 96.25%, Top-2 accuracy of 99.80% and F1-Score of 0.9576. Importantly, the model has 40x fewer trainable parameters than existing benchmark ensemble or transformer pipelines, facilitating easy deployment on edge devices to help clinical practitioners with real-time FPC. The findings are also interpreted using GradCAM to carry out clinical correlation to aid doctors with diagnostics and improve treatment plans for expectant mothers.

PMID:40437331 | DOI:10.1007/s13246-025-01566-6