Neural Netw. 2025 Jan 22;185:107185. doi: 10.1016/j.neunet.2025.107185. Online ahead of print.

ABSTRACT

Model-based diagnosis (MBD) is a critical problem in artificial intelligence. Recent advancements have made it possible to address this challenge using methods like deep learning. However, current approaches that use deep learning for MBD often struggle with accuracy and computation time due to the limited diagnostic information provided by a single observation. To address this challenge, we introduce two novel algorithms, Discret2DiMO (Discret2Di with Multiple Observations) and Discret2DiMO-DC (Discret2Di with Multiple Observations and Dictionary Cache), which enhance MBD by integrating multiple observations with deep learning techniques. Experimental evaluations on a simulated three-tank model demonstrate that Discret2DiMO significantly improves diagnostic accuracy, achieving up to a 685.06% increase and an average improvement of 59.18% over Discret2Di across all test cases. To address computational overhead, Discret2DiMO-DC additionally implements a caching mechanism that eliminates redundant computations during diagnosis. Remarkably, Discret2DiMO-DC achieves comparable accuracy while reducing computation time by an average of 95.74% compared to Discret2DiMO and 89.42% compared to Discret2Di, with computation times reduced by two orders of magnitude. These results indicate that our proposed algorithms significantly enhance diagnostic accuracy and efficiency in MBD compared with the state-of-the-art algorithm, highlighting the potential of integrating multiple observations with deep learning for more accurate and efficient diagnostics in complex systems.

PMID:39862533 | DOI:10.1016/j.neunet.2025.107185

Comput Methods Programs Biomed. 2025 Jan 15;261:108585. doi: 10.1016/j.cmpb.2025.108585. Online ahead of print.

ABSTRACT

BACKGROUND AND OBJECTIVES: Hypertensive Retinopathy (HR) is a retinal manifestation resulting from persistently elevated blood pressure. Severity grading of HR is essential for patient risk stratification, effective management, progression monitoring, timely intervention, and minimizing the risk of vision impairment. Computer-aided diagnosis and artificial intelligence (AI) systems play vital roles in the diagnosis and grading of HR. Over the years, very limited research has been conducted for the grading of HR. Nevertheless, there are no publicly available datasets for HR grading. Moreover, one of the key challenges observed is high-class imbalance.

METHODS: To address these issues, in this paper, we develop "HRSG: Expert-Annotated Hypertensive Retinopathy Severity Grading" dataset, classifying HR severity into four distinct classes: normal, mild, moderate, and severe. Further, to enhance the grading performance on limited datasets, this paper introduces a novel hybrid architecture that combines the strengths of pretrained ResNet-50 via transfer learning, and a modified Vision Transformer (ViT) architecture enhanced with a combination of global self-attention and locality self-attention mechanisms. The locality self-attention addresses the common issue of a lack of inductive bias in ViT architecture. This architecture effectively captures both local and global contextual information, resulting in a robust and resilient classification model. To overcome class imbalance, Decouple Representation and Classifier (DRC) - based training approach is proposed. This method improves the model's ability to learn effective features while preserving the original dataset's distribution, leading to better diagnostic accuracy.

RESULTS: Performance evaluation results show the competence of the proposed method in accurately grading the severity of HR. The proposed method achieved an average accuracy of 0.9688, sensitivity of 0.9435, specificity of 0.9766, F1-score of 0.9442, and precision of 0.9474. The comparative results indicate that the proposed method outperforms existing HR methods, state-of-the-art CNN models, and baseline pretrained ViT models. Additionally, we compared our method with a CNNViT model, which combines a shallow CNN architecture with 3 convolution blocks consisting of a convolution layer, a batch normalization layer, a max pooling layer, and lightweight ViT architecture, due to limited datasets. In comparison with the CNNViT, the proposed method achieved superior performance, demonstrating its effectiveness.

CONCLUSION: The experimental results demonstrate the efficacy of the proposed method in accurately grading HR severity.

PMID:39862474 | DOI:10.1016/j.cmpb.2025.108585

Ophthalmology. 2025 Jan 23:S0161-6420(25)00070-3. doi: 10.1016/j.ophtha.2025.01.016. Online ahead of print.

ABSTRACT

PURPOSE: Glucagon-like peptide-1 receptor agonists (GLP-1RAs) have risen exponentially in usage and have been shown to exert neuroprotective and anti-inflammatory effects across multiple organ systems. This study investigates whether GLP-1RAs influence the risk for age-related ocular diseases.

DESIGN: Retrospective cohort study.

SUBJECTS AND PARTICIPANTS: This study utilized an electronic health records platform of patients in the United States. Patients older than 60 years of age with at least five years of ophthalmology follow-up and medication prescription documentation were included. Patients were categorized into five medication groups: GLP-1RAs, metformin, insulin, statins, or aspirin users. Cohorts were propensity-matched on demographics and chronic health conditions using a greedy matching algorithm.

MAIN OUTCOME MEASURES: Outcomes of cataract, ocular hypertension, primary open angle glaucoma, non-exudative AMD, and exudative AMD were compared five years following initial medication prescription. We then examined earlier timepoints within the five-year period. Significance was defined as p<0.05 and HR threshold > 1.1 or < 0.9 to improve signal to noise ratio.

RESULTS: Of the 9,669 patients taking GLP-1RAs, 84.4 percent were diabetic with an average BMI of 36.2. Propensity matched cohorts demonstrated GLP-1RAs were associated with reduced hazard of non-exudative AMD compared to metformin (HR 0.68, 95%CI: 0.56-0.84), insulin (HR 0.72, 95%CI: 0.58-0.89), and statins (HR 0.7, 95%CI: 0.57-0.87). These findings were validated compared to aspirin and in an independent older cohort of patients. This significant reduction appeared after three years compared to metformin (HR 0.69, 95%CI: 0.52-0.91), insulin (HR 0.66, 95%CI: 0.5-0.87), and statins (HR 0.67, 95%CI: 0.51-0.88). Time course results were validated using independent cohorts of propensity matched patients taking medications for three years. Notably, GLP-1RAs also significantly reduced the risk of exudative AMD (HR 0.7, 95%CI: 0.58-0.84) and POAG (HR 0.58, 95% CI 0.45-0.76) compared to insulin after three years. Usage of GLP-1RAs showed no persistent significant impact on the risk of cataract formation nor ocular hypertension after five years compared other medications.

CONCLUSIONS: This study suggests GLP-1RAs may reduce the risk of multiple age-related ocular diseases and suggests the need for future prospective studies to validate these findings.

PMID:39863057 | DOI:10.1016/j.ophtha.2025.01.016

Biochim Biophys Acta Mol Basis Dis. 2025 Jan 23:167688. doi: 10.1016/j.bbadis.2025.167688. Online ahead of print.

ABSTRACT

Cancer, a heterogeneous disease, presents significant challenges for drug development due to its complex etiology. Drug repurposing, particularly through network medicine approaches, offers a promising avenue for cancer treatment by analyzing how drugs influence cellular networks on a systemic scale. The advent of large-scale proteomics data provides new opportunities to elucidate regulatory mechanisms specific to cancer subtypes. Herein, we present NetSDR, a Network-based Subtype-specific Drug Repurposing framework for prioritizing repurposed drugs specific to certain cancer subtypes, guided by subtype-specific proteomic signatures and network perturbations. First, by integrating cancer subtype information into a network-based method, we developed a pipeline to recognize subtype-specific functional modules. Next, we conducted drug response analysis for each module to identify the "therapeutic module" and then used deep learning to construct weighted drug response network for the particular subtype. Finally, we employed a perturbation response scanning-based drug repurposing method, which incorporates dynamic information, to facilitate the prioritization of candidate drugs. Applying the framework to gastric cancer, we attested the significance of the extracellular matrix module in treatment strategies and discovered a promising potential drug target, LAMB2, as well as a series of possible repurposed drugs. This study demonstrates a systems biology framework for precise drug repurposing in cancer and other complex diseases.

PMID:39862994 | DOI:10.1016/j.bbadis.2025.167688

Naunyn Schmiedebergs Arch Pharmacol. 2025 Jan 25. doi: 10.1007/s00210-025-03809-5. Online ahead of print.

ABSTRACT

As a promising candidate for tackling drug-resistant cancers, triptolide, a diterpenoid derived from the Chinese medicinal plant Tripterygium wilfordii, has been developed. This review summarizes potential antitumor activities, including the suppression of RNA polymerase II, the suppression of heat shock proteins (HSP70 and HSP90), and the blockade of NF-kB signalling. Triptolide is the first known compound to target cancer cells specifically but spare normal cells, and it has success in treating cancers that are difficult to treat, including pancreatic, breast, and lung cancers. It acts against the tolerance mechanisms, including efflux pump upregulation, epithelial-mesenchymal transition, and cancer stem cells. Triptolide modulates important cascades, including PI3K/AKT/mTOR, enhancing the efficacy of conventional therapies. Nonetheless, its clinical application is constrained by toxicity and bioavailability challenges. Emerging drug delivery systems, such as nanoparticles and micellar formulations, are being developed to address these limitations. It has strong interactions with key anticancer targets, like PARP, as determined in preclinical and computational studies consistent with its mechanism of action. Early-phase clinical trials of Minnelide, a water-soluble derivative of triptolide, are promising, but additional work is necessary to optimize dosing, delivery, and safety. This comprehensive analysis demonstrates that triptolide may constitute a repurposed precision medicine tool to overcome tolerance in cancer therapy.

PMID:39862263 | DOI:10.1007/s00210-025-03809-5

BMJ Open. 2025 Jan 25;15(1):e086527. doi: 10.1136/bmjopen-2024-086527.

ABSTRACT

OBJECTIVES: This study sheds light on the available global definitions, classifications, and criteria used for rare diseases (RDs), ultrarare diseases (URDs), orphan drugs (ODs) and ultraorphan drugs (UODs) and provides insights into the rationale behind these definitions.

DESIGN: A systematic literature review was conducted to identify existing definitions and the criteria used to define RDs, ODs and their subtypes.

DATA SOURCES: Searches were performed in the PubMed/Medline, Embase, Scopus and Web of Science (Science and Social Sciences Citation Index) databases covering articles published from 1985 to 2021.

ELIGIBILITY CRITERIA FOR SELECTING STUDIES: English-language studies on the general human population were included if they provided definitions or criteria for RDs, ODs and/or their subtypes without restrictions on publication year, country or jurisdiction.

DATA EXTRACTION AND SYNTHESIS: Two independent reviewers conducted the search, screening and data extraction. Narrative synthesis, content analysis and descriptive analyses were conducted to extract and categorise definitions and criteria from these sources. Study quality was assessed using the Joanna Briggs Institute (JBI) critical appraisal tools.

RESULTS: Online searches identified 2712 published articles. Only 93 articles met the inclusion criteria, with 209 distinct definitions extracted. Specifically, 93 of these articles pertained to 119 RDs, 11 URDs, 67 ODs and 12 UODs. These definitions varied in their reliance on prevalence based and other contextual criteria.

CONCLUSION: Prevalence-based criteria alone pose challenges, as disease frequencies differ by country. Establishing country-specific definitions can enhance understanding, support intercountry evaluations, improve healthcare efficiency and access to ODs, and strengthen equity and equality in healthcare. Such efforts would also promote research and development and support better outcomes for patients with complex and rare conditions.

PROSPERO REGISTRATION NUMBER: CRD42021252701.

PMID:39863413 | DOI:10.1136/bmjopen-2024-086527

J Biomed Inform. 2025 Jan 23:104761. doi: 10.1016/j.jbi.2024.104761. Online ahead of print.

ABSTRACT

OBJECTIVE: Electronic health record (EHR) systems contain a wealth of clinical data stored as both codified data and free-text narrative notes (NLP). The complexity of EHR presents challenges in feature representation, information extraction, and uncertainty quantification. To address these challenges, we proposed an efficient Aggregated naRrative Codified Health (ARCH) records analysis to generate a large-scale knowledge graph (KG) for a comprehensive set of EHR codified and narrative features.

METHODS: Using data from 12.5 million Veterans Affairs patients, ARCH first derives embedding vectors and generates similarities along with associated p-values to measure the strength of relatedness between clinical features with statistical certainty quantification. Next, ARCH performs a sparse embedding regression to remove indirect linkage between features to build a sparse KG. Finally, ARCH was validated on various clinical tasks, including detecting known relationships between entity pairs, predicting drug side effects, disease phenotyping, as well as sub-typing Alzheimer's disease patients.

RESULTS: ARCH produces high-quality clinical embeddings and KG for over 60,000 codified and narrative EHR concepts. The KG and embeddings are visualized in the R-shiny powered web-API3. ARCH achieved high accuracy in detecting EHR concept relationships, with AUCs of 0.926 (codified) and 0.861 (NLP) for similar EHR concepts, and 0.810 (codified) and 0.843 (NLP) for related pairs. It detected drug side effects with a 0.723 AUC, which improved to 0.826 after fine-tuning. Using both codified and NLP features, the detection power increased significantly. Compared to other methods, ARCH has superior accuracy and enhances weakly supervised phenotyping algorithms' performance. Notably, it successfully categorized Alzheimer's patients into two subgroups with varying mortality rates.

CONCLUSION: The proposed ARCH algorithm generates large-scale high-quality semantic representations and knowledge graph for both codified and NLP EHR features, useful for a wide range of predictive modeling tasks.

PMID:39863245 | DOI:10.1016/j.jbi.2024.104761

Cortex. 2024 Dec 27;184:172-193. doi: 10.1016/j.cortex.2024.11.019. Online ahead of print.

ABSTRACT

Retrieving words quickly and correctly is an important language competence. Semantic contexts, such as prior naming of categorically related objects, can induce conceptual priming but also lexical-semantic interference, the latter likely due to enhanced competition during lexical selection. In the continuous naming (CN) paradigm, such semantic interference is evident in a linear increase in naming latency with each additional member of a category out of a seemingly random sequence of pictures being named (cumulative semantic interference/CSI effect). Extensively studied in neurotypical participants, CSI studies in people with aphasia (PWA) are rare, although some lesions regularly and persistently impair word retrieval. In the present study, 20 PWA with lesions in the extended left hemispheric language network and 20 matched controls underwent a CN paradigm, naming photographs of closely related objects from 24 categories (e.g., birds) with 5 members each. The experiment was conducted web-based (Stark et al., 2022) on three days (day 1, 2, and 8). The main results are: (i) Mild-moderate aphasia does not preclude web-based testing. (ii) The CSI effect in naming latencies (∼21 ms per ordinal position) did not differ significantly between groups but was more variable in the PWA; the effect was stable across days. (iii) Overall response times decreased between day 1 and day 2, but remained stable on day 8. (iv) In PWA, increased error-rates paralleled the latency-based CSI effect, suggesting stronger interference in this group. (v) Exploratory analyses suggest that lesions in a large area, including frontal, inferior parietal, pre- and post-central opercular cortices, are linked to a larger CSI effect. At a more lenient statistical threshold, lesions in occipital and supramarginal cortices were associated with increased overall naming latencies. These results offer an initial step toward identifying the neuronal underpinnings of semantic context effects in PWA. We conclude that web-based assessment is feasible in PWA and yields a stable CSI effect over repetitive testing. While not directly clinically applicable, the findings could serve as a foundation for exploring training-interventions targeting lexical activation, interference resolution, or word selection.

PMID:39862560 | DOI:10.1016/j.cortex.2024.11.019

Dig Liver Dis. 2025 Jan 24:S1590-8658(24)01151-4. doi: 10.1016/j.dld.2024.12.026. Online ahead of print.

ABSTRACT

BACKGROUND: Azathioprine (AZA) is part of the standard treatment for autoimmune hepatitis (AIH). The first step in the complex bioconversion of AZA to active metabolites is mediated by glutathione transferases (GSTs).

AIMS: Elucidate the association between GSTM1 and GSTT1 copy number variation (CNV), genetic variation in GSTA2, GSTP1, and inosine-triphosphate-pyrophosphatase, and the response to AZA in AIH.

METHODS: Genotyping was performed in AIH patients (n = 131) on AZA, and in a Swedish background population (n = 283). Thiopurine metabolites in blood erythrocytes were determined by high performance liquid chromatography.

RESULTS: GSTM1 and GSTT1 CNV were associated with treatment response to AZA. Gene deletion of GSTM1-but not of GSTT1-was associated with the liver transaminase levels. None of the studied genetic variants were associated with the thiopurine metabolite concentrations, suggesting non-enzymatic mechanisms of GSTM1 and GSTT1 in the context of AZA efficacy in AIH. The prevalence of GSTM1 and GSTT1 CNV genotypes was similar in AIH and in the background population.

CONCLUSION: This study shows the effects of GSTM1 and GSTT1 CNV on AZA efficacy in AIH, not previously described. It also elaborates on the impact of the definition of treatment response, on the importance of the various GSTs studied. Furthermore, the GSTM1 and GSTT1 CNV frequencies previously reported in European populations were confirmed.

PMID:39863504 | DOI:10.1016/j.dld.2024.12.026

J Mol Diagn. 2025 Jan 23:S1525-1578(25)00018-2. doi: 10.1016/j.jmoldx.2024.12.011. Online ahead of print.

ABSTRACT

This article examines the frequency distribution of Tier 1 pharmacogenetic variants of the Association for Molecular Pathology Pharmacogenomics Working Group Recommendations in two large (>1.000 individuals) cohorts of the admixed Brazilian population, and in patients from the Brazilian Public Health System enrolled in pharmacogenetic trials. Three Tier 1 variants, all in DPYD, were consistently absent, which may justify their non-inclusion in genotyping panels for Brazilians; 13 variants had frequency < 1.0% and the remaining 21 variants ranged in frequency from 1.2% (NUDT15*3) to 76.4% (CYP3A5*3). The frequency of some CYP2C9, CYP2D6, CYP3A4 and VKORC1 variants differed significantly across the three major "race/Color" categories of the Brazilian Census (White, Brown and Black), as a consequence of different proportions of individual European and African ancestry. However, it is recommended that selection of variants for inclusion in pharmacogenetic testing panels and implementation of pharmacogenetic-informed dosing guidelines for Brazilians should not be determined by race/Color categories. Native Americans (0.4% of the Brazilian population), virtually absent from the study cohorts, display wide inter-ethnic diversity in frequency of some Tier 1 variants (e.g. NUDT15*3 and TPMT*3A) and/or differ markedly from non-Indigenous people in frequency of some variant alleles (e.g. CYP2C19*17). Collectively, the data support the notion that population diversity must be taken into account on the design and implementation of pharmacogenetic testing panels.

PMID:39863018 | DOI:10.1016/j.jmoldx.2024.12.011

Cancer Lett. 2025 Jan 23:217494. doi: 10.1016/j.canlet.2025.217494. Online ahead of print.

ABSTRACT

Lung cancer, particularly non-small cell lung cancer (NSCLC), remains a leading cause of cancer-related mortality. Resistance to platinum-based chemotherapy, such as cisplatin, significantly limits treatment efficacy. Circular RNAs (circRNAs) have emerged as key regulators of cancer progression and chemotherapy resistance due to their stable structure, which protects them from degradation. In this study, we focus on circKIAA0182, a circRNA identified as highly expressed in cisplatin-resistant NSCLC cells through profiling. We explore its role in cell proliferation, migration, invasion, apoptosis, and cisplatin resistance. Our findings show that circKIAA0182 promotes cisplatin resistance and tumor progression in NSCLC, in vitro and in vivo. Furthermore, we discovered that circKIAA0182 may interact with the RNA-binding protein YBX1, potentially mediating its oncogenic and cisplatin-resistant functions. The biological role of circKIAA0182 presents a promising target for developing therapeutic strategies to overcome NSCLC progression and cisplatin resistance.

PMID:39862920 | DOI:10.1016/j.canlet.2025.217494

Lancet. 2025 Jan 25;405(10475):314-328. doi: 10.1016/S0140-6736(24)02560-1.

ABSTRACT

BACKGROUND: In the UK, booster COVID-19 vaccinations have been recommended biannually to people considered immune vulnerable. We investigated, at a population level, whether the absence of detectable anti-SARS-CoV-2 spike protein IgG antibody (anti-S Ab) following three or more vaccinations in immunosuppressed individuals was associated with greater risks of infection and severity of infection.

METHODS: In this prospective cohort study using UK national disease registers, we recruited participants with solid organ transplants (SOTs), rare autoimmune rheumatic diseases (RAIRDs), and lymphoid malignancies. All participants were tested for anti-S Ab using a lateral flow immunoassay, completed a questionnaire on sociodemographic and clinical characteristics, and were followed up for 6 months using linked data from the National Health Service in England. SARS-CoV-2 infection was primarily defined using UK Health Security Agency data and supplemented with hospitalisation and therapeutics data, and hospitalisation due to SARS-CoV-2 was defined as an admission within 14 days of a positive test.

FINDINGS: Between Dec 7, 2021, and June 26, 2022, we recruited 21 575 participants. Anti-S Ab was detected in 6519 (77·0%) of 8466 participants with SOTs, 5594 (85·9%) of 6516 with RAIRDs, and 5227 (79·3%) of 6593 with lymphoid malignancies. COVID-19 infection was recorded in 3907 (18·5%) participants, with 556 requiring a COVID-19-related hospital admission and 17 dying within 28 days of infection. Rates of infection varied by sociodemographic and clinical characteristics but, in adjusted analysis, having detectable anti-S Ab was independently associated with a reduced incidence of infection, with incident rate ratios (IRRs) of 0·69 (95% CI 0·65-0·73) in the SOT cohort, 0·57 (0·49-0·67) in the RAIRD cohort, and 0·62 (0·54-0·71) in the lymphoid malignancy cohort. In adjusted analysis, having detectable anti-S Ab was also associated with a reduced incidence of hospitalisation, with IRRs of 0·40 (0·35-0·46) in the SOT cohort, 0·32 (0·22-0·46) in the RAIRD cohort, and 0·41 (0·29-0·58) in the lymphoid malignancy cohort.

INTERPRETATION: All people with immunosuppression require ongoing access to COVID-19 protection strategies. Assessment of anti-S Ab responses, which can be performed at scale, can identify people with immunosuppression who remain most at risk, providing a mechanism to further individualise protection approaches.

FUNDING: UK Research and Innovation, Kidney Research UK, Blood Cancer UK, Vasculitis UK, and Cystic Fibrosis Trust.

PMID:39863371 | DOI:10.1016/S0140-6736(24)02560-1

J Biol Chem. 2025 Jan 23:108220. doi: 10.1016/j.jbc.2025.108220. Online ahead of print.

ABSTRACT

Carboxyl-terminus of Hsp70-Interacting Protein (CHIP) is an E3 ubiquitin ligase that marks misfolded substrates for degradation. Hyper-activation of CHIP has been implicated in multiple diseases, including cystic fibrosis and cancer, suggesting that it may be a potential drug target. However, there are few tools available for exploring this possibility. Moreover, the best ways of inhibiting CHIP's function are not obvious, as this complex protein is composed of a tetratricopeptide repeat (TPR) domain, a U-box domain, and a coiled-coil domain that mediates homodimerization. To probe the structure and function of CHIP, we report an antibody panning campaign that yielded six recombinant Fabs with affinity for CHIP. Interestingly, these antibodies varied in their binding site(s) and impact on CHIP function, such as inhibiting TPR interactions, autoubiquitination, and/or substrate ubiquitination. Of particular interest, antibody 2F1 nearly eliminated substrate binding (IC50 = 2.7 μM) and limited ubiquitination and autoubiquitination. Cryo-electron microscopy of the 2F1:CHIP complex revealed a 2:1 binding mode (Fab:CHIP dimer), with 2F1 bound to the U-box domain and simultaneously displacing the TPR domain. Together, these studies provide insight into ways of inhibiting CHIP's activity and provide a series of new probes for exploring the function of this important E3 ubiquitin ligase.

PMID:39863102 | DOI:10.1016/j.jbc.2025.108220

Int J Pharm. 2025 Jan 23:125241. doi: 10.1016/j.ijpharm.2025.125241. Online ahead of print.

ABSTRACT

Cystic fibrosis (CF) is characterized by abnormal mucus hydration due to a defective CF Transmembrane Regulator (CFTR) protein, leading to the production of difficult-to-clear mucus. This causes airflow obstruction, recurrent infections, and respiratory complications. Chronic lung infections are the leading cause of death for CF patients and inhaled tobramycin is the first-in-line antibiotic treatment against these infections, mainly caused by Pseudomonas aeruginosa in adult patients. KuDa-tob, a nanoformulation of tobramycin (tob) as the active ingredient and dextran single chain nanoparticles, a drug carrier platform (KuDa) as an excipient, has been developed. The neutralization of the positive charges of the drug by KuDa nanoparticles facilitates its diffusion through the mucus and biofilm, reaching the bacteria. The polar interactions existing between tobramycin and KuDa have been thoroughly characterized by electrophoresis (ζ-potential) and diffusion experiments (diffusion ordered spectroscopy and Taylor dispersion analysis) demonstrating that up to 40 wt% tobramycin could be loaded into the KuDa-tob nanoformulation. The drug product was developed following Quality by Design (QbD) principles. Critical quality attributes (CQAs), critical process parameters (CPPs) and critical material attributes (CMAs) have been defined to obtain a robust production process that was then scaled-up to 40 g, allowing the production of KuDa-tob for further preclinical evaluation. Finally, the final pharmaceutical form of KuDa-tob was defined based on stability studies, and nebulization assays showed that the aerosols generated by reconstituted KuDa-tob were in the ideal range size for lung deposition (Median Mass Aerodynamic Diameter - MMAD - 2.2 μm).

PMID:39863028 | DOI:10.1016/j.ijpharm.2025.125241

Comput Biol Med. 2025 Jan 24;186:109703. doi: 10.1016/j.compbiomed.2025.109703. Online ahead of print.

ABSTRACT

- Brain tumors (BT), both benign and malignant, pose a substantial impact on human health and need precise and early detection for successful treatment. Analysing magnetic resonance imaging (MRI) image is a common method for BT diagnosis and segmentation, yet misdiagnoses yield effective medical responses, impacting patient survival rates. Recent technological advancements have popularized deep learning-based medical image analysis, leveraging transfer learning to reuse pre-trained models for various applications. BT segmentation with MRI remains challenging despite advancements in image acquisition techniques. Accurate detection and segmentation are essential for proper diagnosis and treatment planning. This study aims to enhance BT detection and segmentation accuracy and effectiveness of categorization through the implementation of an advanced stacking ensemble learning (SEL) approach. This study explores the efficiency of SEL architecture in augmenting the precision of BT segmentation. SEL, a prominent approach within the machine learning paradigm, combines the predictions of base-level models and improves the overall performance of predictions in order to reduce the errors and biases of each model. The proposed approach involves designing a stacked DenseNet201 as the meta-model called SEL-DenseNet201, complemented by six diverse base models such as mobile network version 3 (MobileNet-v3), 3-dimensional convolutional neural network (3D-CNN), visual geometry group network with 16 and 19 layers (VGG-16 and VGG-19), residual network with 50 layers (ResNet50), and Alex network (AlexNet). The strengths of the base models are calculated to capture distinct aspects of the BT MRI, aiming for enhanced segmentation performance. The proposed SEL-DenseNet201 is trained using BT MRI datasets. The augmentation techniques are applied to MRI scans to balance and enhance the model performance through the application of image enhancement and segmentation techniques. The proposed SEL-DenseNet201 achieves impressive results with an accuracy of 99.65 % and a dice coefficient of 97.43 %. These outcomes underscore the superiority of the proposed model over existing approaches. This study holds the potential to be an initial screening approach for early BT detection, with a high success rate.

PMID:39862469 | DOI:10.1016/j.compbiomed.2025.109703

Comput Biol Med. 2025 Jan 24;186:109706. doi: 10.1016/j.compbiomed.2025.109706. Online ahead of print.

ABSTRACT

The fusion index is a critical metric for quantitatively assessing the transformation of in vitro muscle cells into myotubes in the biological and medical fields. Traditional methods for calculating this index manually involve the labor-intensive counting of numerous muscle cell nuclei in images, which necessitates determining whether each nucleus is located inside or outside the myotubes, leading to significant inter-observer variation. To address these challenges, this study proposes a three-stage process that integrates the strengths of pattern recognition and deep-learning to automatically calculate the fusion index. The experimental results demonstrate that the proposed process achieves significantly higher performance in cell nuclei detection and classification, with an F1-score of 0.953, whereas traditional object detection methods achieve less than 0.5. In addition, the fusion index obtained using the proposed method is closely aligned with the human-assessed values, showing minimal discrepancy and strong agreement with human evaluations. This process is incorporated into the development of "SEPO-FI" as public software, automating cell detection and classification to enable effective fusion index calculation and broaden access to this methodology within the scientific community.

PMID:39862466 | DOI:10.1016/j.compbiomed.2025.109706

NPJ Digit Med. 2025 Jan 26;8(1):56. doi: 10.1038/s41746-025-01454-z.

ABSTRACT

Deep-learning models have shown promise in differentiating between benign and malignant lesions. Previous studies have primarily focused on specific anatomical regions, overlooking tumors occurring throughout the body with highly heterogeneous whole-body backgrounds. Using neurofibromatosis type 1 (NF1) as an example, this study developed highly accurate MRI-based deep-learning models for the early automated screening of malignant peripheral nerve sheath tumors (MPNSTs) against complex whole-body background. In a Chinese seven-center cohort, data from 347 subjects were analyzed. Our one-step model incorporated normal tissue/organ labels to provide contextual information, offering a solution for tumors with complex backgrounds. To address privacy concerns, we utilized a lightweight deep neural network suitable for hospital deployment. The final model achieved an accuracy of 85.71% for MPNST diagnosis in the validation cohort and 84.75% accuracy in the independent test set, outperforming another classic two-step model. This success suggests potential for AI models in screening other whole-body primary/metastatic tumors.

PMID:39863790 | DOI:10.1038/s41746-025-01454-z

Sci Rep. 2025 Jan 25;15(1):3273. doi: 10.1038/s41598-025-87803-0.

ABSTRACT

We aimed to build a robust classifier for the MGMT methylation status of glioblastoma in multiparametric MRI. We focused on multi-habitat deep image descriptors as our basic focus. A subset of the BRATS 2021 MGMT methylation dataset containing both MGMT class labels and segmentation masks was used. A comprehensive mask fusion approach was developed to select relevant image crops of diseased tissue. These fusion masks, which were guided by multiple sequences, helped collect information from the regions that seem disease-free to radiologists in standard MRI sequences while harboring pathology. Integrating the information in different MRI sequences and leveraging the high entropic capacity of deep neural networks, we built a 3D ROI-based custom CNN classifier for the automatic prediction of MGMT methylation status of glioblastoma in multi-parametric MRI. Single sequence-based classifiers reached intermediate predictive performance with 0.65, 0.71, 0.77, and 0.82 accuracy for T1W, T2W, T1 contrast-enhanced, and FLAIR sequences, respectively. The multiparametric classifier using T1 contrast-enhanced and FLAIR images reached 0.88 accuracy. The accuracy of the four-input model that used all sequences was 0.81. The best model reached 0.90 ROC AUC value. Integrating human knowledge in the form of relevant target selection was a useful approach in MGMT methylation status prediction in MRI. Exploration of means to integrate radiology knowledge into the models and achieve human-machine collaboration may help to develop better models. MGMT methylation status of glioblastoma is an important prognostic marker and is also important for treatment decisions. The preoperative non-invasive predictive ability and the explanation tools of the developed model may help clinicians to better understand imaging phenotypes of MGMT methylation status of glial tumors.

PMID:39863759 | DOI:10.1038/s41598-025-87803-0

Sci Rep. 2025 Jan 25;15(1):3226. doi: 10.1038/s41598-025-87966-w.

ABSTRACT

Exploring the potential of advanced artificial intelligence technology in predicting microsatellite instability (MSI) and Ki-67 expression of endometrial cancer (EC) is highly significant. This study aimed to develop a novel hybrid radiomics approach integrating multiparametric magnetic resonance imaging (MRI), deep learning, and multichannel image analysis for predicting MSI and Ki-67 status. A retrospective study included 156 EC patients who were subsequently categorized into MSI and Ki-67 groups. The hybrid radiomics model (HMRadSum) was developed by extracting quantitative imaging features and deep learning features from multiparametric MRI using emerging attention mechanism. Tumor markers were subsequently predicted utilizing an XGBoost classifier. Model performance and interpretability were evaluated using standard classification metrics, Gradient-weighted Class Activation Mapping (Grad-CAM), and SHapley Additive exPlanations (SHAP) techniques. For the MSI prediction task, the HMRadSum model achieved area-under-curve (AUC) value of 0.945 (95% CI 0.862-1.000) and accuracy of 0.889. For the Ki-67 prediction task, the AUC and accuracy of HMRadSum model was 0.888 (95% CI 0.743-1.000) and 0.810. This hybrid radiomics model effectively extracted features associated with EC gene expression, providing potential clinical implications for personalized diagnosis, treatment, and treatment strategy optimization.

PMID:39863695 | DOI:10.1038/s41598-025-87966-w

Sci Rep. 2025 Jan 25;15(1):3204. doi: 10.1038/s41598-025-87450-5.

ABSTRACT

In order to solve the limitations of flipped classroom in personalized teaching and interactive effect improvement, this paper designs a new model of flipped classroom in colleges and universities based on Virtual Reality (VR) by combining the algorithm of Contrastive Language-Image Pre-Training (CLIP). Through cross-modal data fusion, the model deeply combines students' operation behavior with teaching content, and improves teaching effect through intelligent feedback mechanism. The test data shows that the similarity between video and image modes reaches 0.89, which indicates that different modal information can be effectively integrated to ensure the semantic consistency and intuitive understanding of teaching content. The minimum Kullback-Leibler (KL) divergence is 0.12, which ensures the stability of data distribution and avoids information loss. The accuracy of automatically generating feedback reaches 93.72%, which significantly improves the efficiency of personalized learning guidance. In the adaptability test of virtual scene, the frequency of scene adjustment is 2.5 times/minute, and the consistency score is stable above 8.6, ensuring the consistency of teaching goals under complex interaction. This paper aims to enhance personalized learning experience, improve teaching efficiency and autonomous learning effect through VR technology and intelligent feedback, and promote the innovation of interactive teaching mode.

PMID:39863690 | DOI:10.1038/s41598-025-87450-5