Sci Rep. 2025 Apr 14;15(1):12801. doi: 10.1038/s41598-025-97719-4.

ABSTRACT

Current artificial intelligence (AI) trends are revolutionizing medical image processing, greatly improving cervical cancer diagnosis. Machine learning (ML) algorithms can discover patterns and anomalies in medical images, whereas deep learning (DL) methods, specifically convolutional neural networks (CNNs), are extremely accurate at identifying malignant lesions. Deep models that have been pre-trained and tailored through transfer learning and fine-tuning become faster and more effective, even when data is scarce. This paper implements a state-of-the-art Hybrid Learning Network that combines the Progressive Resizing approach and Principal Component Analysis (PCA) for enhanced cervical cancer diagnostics of whole slide images (WSI) slides. ResNet-152 and VGG-16, two fine-tuned DL models, are employed together with transfer learning to train on augmented and progressively resized training data with dimensions of 224 × 224, 512 × 512, and 1024 × 1024 pixels for enhanced feature extraction. Principal component analysis (PCA) is subsequently employed to process the combined features extracted from two DL models and reduce the dimensional space of the feature set. Furthermore, two ML methods, Support Vector Machine (SVM) and Random Forest (RF) models, are trained on this reduced feature set, and their predictions are integrated using a majority voting approach for evaluating the final classification results, thereby enhancing overall accuracy and reliability. The accuracy of the suggested framework on SIPaKMeD data is 99.29% for two-class classification and 98.47% for five-class classification. Furthermore, it achieves 100% accuracy for four-class categorization on the LBC dataset.

PMID:40229435 | DOI:10.1038/s41598-025-97719-4

Sci Rep. 2025 Apr 14;15(1):12811. doi: 10.1038/s41598-025-96459-9.

ABSTRACT

Diffusion tensor imaging (DTI) is essential for assessing brain microstructure but requires long acquisition times, limiting clinical use. Recent deep learning (DL) approaches, such as SuperDTI or deepDTI, improve DTI metrics but demand large, high-quality datasets for training. We propose a self-supervised deep learning with fine-tuning (SSDLFT) framework to reduce training data requirements. SSDLFT involves self-supervised pretraining, which denoises data without clean labels, followed by fine-tuning with limited high-quality data. Experiments using Human Connectome Project data show that SSDLFT outperforms traditional methods and other DL approaches in qualitative and quantitative assessments of DWI reconstructions and tensor metrics. SSDLFT's ability to maintain high performance with fewer training subjects and DWIs presents a significant advancement, enhancing DTI's practical applications in clinical and research settings.

PMID:40229411 | DOI:10.1038/s41598-025-96459-9

Nat Biotechnol. 2025 Apr;43(4):491-500. doi: 10.1038/s41587-025-02626-8.

NO ABSTRACT

PMID:40229366 | DOI:10.1038/s41587-025-02626-8

Sci Rep. 2025 Apr 14;15(1):12812. doi: 10.1038/s41598-025-97297-5.

ABSTRACT

Acute Lymphoblastic Leukemia (ALL) is a life-threatening malignancy characterized by its aggressive progression and detrimental effects on the hematopoietic system. Early and accurate diagnosis is paramount to optimizing therapeutic interventions and improving clinical outcomes. This study introduces a novel diagnostic framework that synergizes the EfficientNet-B7 architecture with Explainable Artificial Intelligence (XAI) methodologies to address challenges in performance, computational efficiency, and explainability. The proposed model achieves improved diagnostic performance, with accuracies exceeding 96% on the Taleqani Hospital dataset and 95.50% on the C-NMC-19 and Multi-Cancer datasets. Rigorous evaluation across multiple metrics-including Area Under the Curve (AUC), mean Average Precision (mAP), Accuracy, Precision, Recall, and F1-score-demonstrates the model's robustness and establishes its superiority over state-of-the-art architectures namely VGG-19, InceptionResNetV2, ResNet50, DenseNet50 and AlexNet . Furthermore, the framework significantly reduces computational overhead, achieving up to 40% faster inference times, thereby enhancing its clinical applicability. To address the opacity inherent in Deep learning (DL) models, the framework integrates advanced XAI techniques, including Gradient-weighted Class Activation Mapping (Grad-CAM), Class Activation Mapping (CAM), Local Interpretable Model-Agnostic Explanations (LIME), and Integrated Gradients (IG), providing transparent and explainable insights into model predictions. This fusion of high diagnostic precision, computational efficiency, and explainability positions the proposed framework as a transformative tool for ALL diagnosis, bridging the gap between cutting-edge AI technologies and practical clinical deployment.

PMID:40229347 | DOI:10.1038/s41598-025-97297-5

Sci Rep. 2025 Apr 14;15(1):12829. doi: 10.1038/s41598-025-96587-2.

ABSTRACT

In light of the challenges currently facing the inheritance of blue calico, including the reduction in the number of inheritors and the contraction of the market, this paper puts forth a stylistic transfer method based on an enhanced cycle consistency generative adversarial network. This approach is designed to facilitate the creation of novel designs for traditional blue calico patterns. To address the shortcomings of existing style transfer models, including the generation of blurry details, poor texture and color effects, and excessive model parameters, we propose the incorporation of the Ghost convolution module and the SRM attention module into the generator network structure. This approach aims to reduce the model parameter quantity and computational cost while enhancing the feature extraction ability of the network. The experimental results demonstrate that the method proposed in this paper not only effectively enhances the content details, texture, and color effects of the generated images, but also successfully combines traditional blue calico with modern daily necessities, thereby enhancing its appeal to young people. This research provides novel insights into the digital protection and innovative development of traditional culture, and illustrates the extensive potential applications of deep learning technology in the field of cultural heritage.

PMID:40229316 | DOI:10.1038/s41598-025-96587-2

Mol Syst Biol. 2025 Apr 14. doi: 10.1038/s44320-025-00101-9. Online ahead of print.

NO ABSTRACT

PMID:40229558 | DOI:10.1038/s44320-025-00101-9

Commun Biol. 2025 Apr 15;8(1):609. doi: 10.1038/s42003-025-07810-9.

ABSTRACT

Early life gut microbiota is being increasingly recognized as a major contributor to short and/or long-term human health and diseases. However, little is known about these early-life events in the human microbiome of the lower respiratory tract. This study aims to investigate fungal and bacterial colonization in the lower airways over the first year of life by analyzing lung tissue from autopsied infants. The fungal and bacterial communities of lung tissue samples from 53 autopsied infants were characterized by Next-Generation Sequencing (NGS), based on universal PCR amplification of the ITS region and the 16S rRNA gene, respectively. Our study highlights a high degree of inter-individual variability in both fungal and bacterial communities inhabiting the infant lung. The lower respiratory tract microbiota is mainly composed of transient microorganisms that likely travel from the upper respiratory tract and do not establish permanent residence. However, it could also contain some genera identified as long-term inhabitants of the lung, which could potentially play a role in lung physiology or disease. At 3-4 months of age, important dynamic changes to the microbial community were observed, which might correspond to a transitional time period in the maturation of the lung microbiome. This timeframe represents a susceptibility period for the colonization of pathogens such as Pneumocystis. The asymptomatic colonization of Pneumocystis was associated with changes in the fungal and bacterial communities. These findings suggest that the period of 2-4 months of age is a "critical window" early in life. Pneumocystis jirovecii could be a potential pivotal ecological driver contributing to shifts in microbial equilibrium during the early-life lower airway microbiome assembly, and to the future health of children.

PMID:40229539 | DOI:10.1038/s42003-025-07810-9

NPJ Vaccines. 2025 Apr 14;10(1):69. doi: 10.1038/s41541-025-01103-2.

ABSTRACT

Correlates of protection (CoPs) for SARS-CoV-2 have yet to be sufficiently defined. This study uses the machine learning platform, SIMON, to accurately predict the immunological parameters that reduced clinical pathology or viral load following SARS-CoV-2 challenge in a cohort of 90 non-human primates. We found that anti-SARS-CoV-2 spike antibody and neutralising antibody titres were the best predictors of clinical protection and low viral load in the lung. Since antibodies to SARS-CoV-2 spike showed the greatest association with clinical protection and reduced viral load, we next used SIMON to investigate the immunological features that predict high antibody titres. It was found that a pre-immunisation response to seasonal beta-HCoVs and a high frequency of peripheral intermediate and non-classical monocytes predicted low SARS-CoV-2 spike IgG titres. In contrast, an elevated T cell response as measured by IFNγ ELISpot predicted high IgG titres. Additional predictors of clinical protection and low SARS-CoV-2 burden included a high abundance of peripheral T cells. In contrast, increased numbers of intermediate monocytes predicted clinical pathology and high viral burden in the throat. We also conclude that an immunisation strategy that minimises pathology post-challenge did not necessarily mediate viral control. This would be an important finding to take forward into the development of future vaccines aimed at limiting the transmission of SARS-CoV-2. These results contribute to SARS-CoV-2 CoP definition and shed light on the factors influencing the success of SARS-CoV-2 vaccination.

PMID:40229322 | DOI:10.1038/s41541-025-01103-2

Life Sci Alliance. 2025 Apr 14;8(7):e202503222. doi: 10.26508/lsa.202503222. Print 2025 Jul.

ABSTRACT

STAT5B is a vital transcription factor for lymphocytes. Here, the function of two STAT5B mutations from human T-cell leukemias: one substituting tyrosine 665 with phenylalanine (STAT5BY665F) and the other with histidine (STAT5BY665H), was interrogated. In silico modeling predicted divergent energetic effects on homodimerization with a range of pathogenicity. In primary T cells in vitro, STAT5BY665F showed gain-of-function, whereas STAT5BY665H demonstrated loss-of-function. Introducing the mutation into the mouse genome illustrated that the gain-of-function Stat5b Y665F mutation resulted in accumulation of CD8+ effector and memory and CD4+ regulatory T cells, altering CD8+/CD4+ ratios. In contrast, STAT5BY665H "knock-in" mice showed diminished CD8+ effector and memory and CD4+ regulatory T cells. In contrast to WT STAT5B, the STAT5BY665F variant displayed greater STAT5 phosphorylation, DNA binding, and transcriptional activity after cytokine activation, whereas the STAT5BY665H variant resembled a null. The work exemplifies how joining in silico and in vivo studies of single nucleotides deepens our understanding of disease-associated variants, revealing structural determinants of altered function, defining mechanistic roles, and, specifically here, identifying a gain-of-function variant that does not directly induce hematopoietic malignancy.

PMID:40228864 | DOI:10.26508/lsa.202503222

BMJ Open. 2025 Apr 14;15(4):e102029. doi: 10.1136/bmjopen-2025-102029.

ABSTRACT

INTRODUCTION: Myocardial protection against ischaemia-reperfusion injury is a key determinant of heart function and outcome following cardiac surgery in children. However, myocardial injury still occurs routinely following aortic cross-clamping, as demonstrated by the ubiquitous rise in circulating troponin. del Nido cardioplegia was designed to protect the immature myocardium and is widely used in the USA but has not previously been available in the UK, where St. Thomas' blood cardioplegia is most common. The del Nido versus St. Thomas' blood cardioplegia in the young (DESTINY) trial will evaluate whether one solution is better than the other at improving myocardial protection by reducing myocardial injury, shortening ischaemic time and improving clinical outcomes.

METHODS AND ANALYSIS: The DESTINY trial is a multicentre, patient-blinded and assessor-blinded, parallel-group, individually randomised controlled trial recruiting up to 220 children undergoing surgery for congenital heart disease. Participants will be randomised in a 1:1 ratio to either del Nido cardioplegia or St. Thomas' blood cardioplegia, with follow-up until 30 days following surgery. The primary outcome is area under the time-concentration curve for plasma high-sensitivity troponin I in the first 24 hours after aortic cross-clamp release. Secondary outcome measures include the incidence of low cardiac output syndrome and Vasoactive-Inotropic Score in the first 48 hours, total aortic cross-clamp time, duration of mechanical ventilation and lengths of stay in the paediatric intensive care unit and the hospital.

ETHICS AND DISSEMINATION: The trial was approved by the West Midlands-Coventry and Warwickshire National Health Service Research Ethics Committee (21/WM/0149) on 30 June 2021. Findings will be disseminated to the academic community through peer-reviewed publications and presentation at national and international meetings. Parents will be informed of the results through a newsletter in conjunction with a national charity.

TRIAL REGISTRATION NUMBER: ISRCTN13638147; Pre-results.

PMID:40228861 | DOI:10.1136/bmjopen-2025-102029

J Biotechnol. 2025 Apr 12:S0168-1656(25)00093-8. doi: 10.1016/j.jbiotec.2025.04.008. Online ahead of print.

ABSTRACT

Beta-ketoadipate (βKA) is an intermediate of the βKA pathway involved in the degradation of aromatic compounds in several bacteria and fungi. Beta-ketoadipate also represents a promising chemical for the manufacturing of performance-advantaged nylons. We established a strategy for the in planta synthesis of βKA via manipulation of the shikimate pathway and the expression of bacterial enzymes from the βKA pathway. Using Nicotiana benthamiana as a transient expression system, we demonstrated the efficient conversion of protocatechuate (PCA) to βKA when plastid-targeted bacterial-derived PCA 3,4-dioxygenase (PcaHG) and 3-carboxy-cis,cis-muconate cycloisomerase (PcaB) were co-expressed with 3-deoxy-D-arabinoheptulosonate 7-phosphate synthase (AroG) and 3-dehydroshikimate dehydratase (QsuB). This metabolic pathway was reconstituted in Arabidopsis by introducing a construct (pAtβKA) with stacked pcaG, pcaH, and pcaB genes into a PCA-overproducing genetic background that expresses AroG and QsuB (referred as QsuB-2). The resulting QsuB-2 x pAtβKA stable lines displayed βKA titers as high as 0.25% on a dry weight basis in stems, along with a drastic reduction in lignin content and improvement of biomass saccharification efficiency compared to wild-type controls, and without any significant reduction in biomass yields. Using biomass sorghum as a potential crop for large-scale βKA production, techno-economic analysis indicated that βKA accumulated at titers of 0.25% and 4% on a dry weight basis could be competitively priced in the range of $2.04-34.49/kg and $0.47-2.12/kg, respectively, depending on the selling price of the residual biomass recovered after βKA extraction. This study lays the foundation for a more environmentally-friendly synthesis of βKA using plants as production hosts.

PMID:40228630 | DOI:10.1016/j.jbiotec.2025.04.008

Hum Exp Toxicol. 2025 Jan-Dec;44:9603271251332914. doi: 10.1177/09603271251332914. Epub 2025 Apr 14.

ABSTRACT

IntroductionNephrotoxicity and hepatotoxicity are substantial side effects triggered in individuals injected with 5-fluorouracil (5-FU), an anticancer drug. This study aimed to investigate the impact of the natural antioxidant and anti-inflammatory phenolic compound; protocatechuic acid (PCA) on 5-FU-provoked renal and hepatic injury in rats.MethodsRats were allocated to 4 groups: control, 5-FU, 5-FU + PCA (50 mg/kg), and 5-FU + PCA (100 mg/kg). Rats were intraperitoneally injected 5-FU (75 mg/kg; once a week for 21 days. Protocatechuic acid (50 and 100 mg/kg/day; orally) was administered for 3 weeks.ResultsRats co-treated with PCA had lower serum kidney and liver function markers than those receiving 5-FU alone. Furthermore, co-treatment with PCA successfully modulated kidney and liver contents of TNF-α, NF-κB p65, active caspase-1, IL-1β, p-p38 MAPK, SOD, GSH, Nrf-2, HO-1 and MDA. Moreover, PCA improved histopathological alterations of both kidney and liver tissues.ConclusionPCA exerts its hepatoprotective and nephroprotective effects against 5-FU-triggered toxicity through modulation of oxidative stress and inflammatory pathways, particularly via Nrf-2 activation and NF-κB inhibition.

PMID:40228806 | DOI:10.1177/09603271251332914

Vox Sang. 2025 Apr 14. doi: 10.1111/vox.70032. Online ahead of print.

ABSTRACT

BACKGROUND AND OBJECTIVES: Granulocyte-colony stimulating factor (G-CSF) is commonly used for peripheral blood stem cell harvesting (PBSCH). Although its well-documented adverse effects include thrombocytopaenia and bone pain, drug-induced liver injury (DILI) is rare.

MATERIALS AND METHODS: We present the case of a 40-year-old male donor who developed DILI 4 days after G-CSF administration for PBSCH.

RESULTS: Laboratory results indicated elevated hepatobiliary enzymes, with aspartate aminotransferase (AST) peaking at 171 U/L (×6 the upper limit of normal [ULN]) on Day 7 of G-CSF administration, alanine aminotransferase (ALT) at 244 U/L (×6 ULN) on Day 11, alkaline phosphatase (ALP) at 371 U/L (×3 ULN) on Day 5 and gamma-glutamyl transpeptidase (γ-GTP) 93 U/L (×1.5 ULN) on Day 9. The hepatobiliary dysfunction became evident after G-CSF administration had ended, despite other parameters-including white blood cell and platelet counts-remaining within acceptable ranges. DILI was confirmed by positive drug lymphocyte stimulation test results. The donor's liver function normalized within 1 month of supportive treatment, and the recipient achieved successful engraftment without G-CSF administration.

CONCLUSION: As G-CSF allergy screening is not mandatory in current Japanese protocols, DILI due to G-CSF could present a risk during PBSCH. This case emphasizes the importance of vigilant monitoring and comprehensive risk assessment to ensure the safety of healthy donors.

PMID:40228804 | DOI:10.1111/vox.70032

ACS Appl Bio Mater. 2025 Apr 14. doi: 10.1021/acsabm.5c00362. Online ahead of print.

ABSTRACT

Fruit spoilage caused by oxidation and microbial infection exacerbates resource wastage. Although starch films including chitosan possessed admirable biocompatibility owing to great biodegradability compared with conventional plastics, deficient antibacterial and antioxidant capacity restricted food shelf life. Herein, an environmentally friendly antibacterial film (CS/G-CDs) was constructed by carbon dots derived from Cirsii Herba (CDs), which was formed through high affinity resulting from hydrogen bonding between chitosan molecules and hydroxyl originating from CDs. The prepared CDs presented homogeneous and monodisperse spherical structures with an ultrasmall size, providing favorable conditions for uniform film formation. Encouragingly, the antioxidant capacity of CS/G-CDs increased 5.00-fold, followed by an antibacterial rate of up to 97.0%. Dramatically, CS/G-CDs revealed glorious UV shielding efficacy (99.9% for UVB and 98.2% for UVA), and its preservation time for blueberries was remarkably extended 8 days longer than that of the chitosan film. Overall, Chinese herb-derived antibacterial films exhibited magnified antibacterial/antioxidant properties and great biocompatibility, which provided a promising strategy for sustainable development of packaging materials.

PMID:40227972 | DOI:10.1021/acsabm.5c00362

Am Soc Clin Oncol Educ Book. 2025 Jun;45(3):e100051. doi: 10.1200/EDBK-25-100051. Epub 2025 Apr 14.

ABSTRACT

The field of rare cancer research is rapidly transforming, marked by significant progress in clinical trials and treatment strategies. Rare cancers, as defined by the National Cancer Institute, occur in fewer than 150 cases per million people each year, yet they collectively represent a significant portion of all cancer diagnoses. Because of their infrequency, these cancers pose distinct challenges for clinical trials, including limited patient populations, geographical dispersion, and a general lack of awareness of treatment options. Economic limitations further complicate drug development, making initiatives such as the Orphan Drug Act essential for incentivizing research. The advent of next-generation sequencing (NGS) and precision medicine has been instrumental in identifying actionable genetic alterations in parallel with an explosion in the development of genomically targeted therapies, immunotherapies, and antibody drug conjugates. Advances in clinical NGS, precision medicine, and tumor-agnostic therapies have become central to the progress in rare cancer research. The development and approval of tumor-agnostic drugs, such as BRAF, NTRK, and RET inhibitors, and immunotherapy for mismatch repair deficient/microsatellite instability-high status cancers highlight the potential of personalized treatments across diverse cancer types and across the age spectrum. Collaborative trials from cooperative groups including SWOG DART, ASCO TAPUR, NCI-MATCH, pediatric COG-match, DRUP, IMPRESS, and innovative registrational basket and platform trials (eg, VE-Basket, ROAR, LIBRETTO-001, ARROW), along with patient advocacy group-run trials like TRACK, are enhancing access to clinical trials. In addition, artificial intelligence has the potential to improve the trial matching process. An integrated approach, combining these innovations in collaboration with multiple stakeholders, is crucial for advancing rare cancer research, offering hope for better patient outcomes and quality of life.

PMID:40228175 | DOI:10.1200/EDBK-25-100051

Crit Rev Eukaryot Gene Expr. 2025;35(4):55-64. doi: 10.1615/CritRevEukaryotGeneExpr.2025057731.

ABSTRACT

BACKGROUND: Cystic fibrosis (CF) is common genetic disorder in Europe and North America but rarer in Asian populations.

OBJECTIVE: To explore the clinical manifestations and gene mutations of cystic fibrosis.

METHODS: This case series study enrolled children with CF diagnosed in the pediatric respiratory department of Shandong Provincial Hospital affiliated to Shandong First Medical University between June 2016 and August 2022.

RESULTS: Seven children, including 6 girls and 1 boy, were enrolled. All 7 patients had recurrent wet cough and (chronic) pneumonia. Six patients suffered from chronic sinusitis, 4 patients had recurrent wheezing; 2 patients had chronic diarrhea, malnutrition and growth lag; 2 patients were complicated by allergic bronchopulmonary aspergillosis; and 1 patient had pancreatic insufficiency. Bronchiectasis, thickening of bronchial wall and mucous impaction, were seen in the chest CT of 7 children. Six patients showed a large amount of viscous sputum adhered to the bronchial wall by bronchoscopy. Infection of Pseudomonas aeruginosa was found in 6 cases, Staphylococcus aureus in 2 cases, and Aspergillus fumigatus in 2 cases by bronchoalveolar lavage fluid or sputum culture. Sweat sodium chloride test was performed in 3 cases, and the result showed that Cl-> 60 mmol/L. CFTR gene mutations were found in 7 cases, which were rare mutations of Caucasians, including 2 cases with new mutation sites (c.325T>G and 326A>G).

CONCLUSIONS: The major clinical presentations of CF could be chronic and recurrent upper and lower respiratory tract infections, malnutrition, and digestive tract diseases. The rare and even new mutations of Caucasians on CFTR gene may occur in Chinese children.

PMID:40228226 | DOI:10.1615/CritRevEukaryotGeneExpr.2025057731

PLoS One. 2025 Apr 14;20(4):e0320297. doi: 10.1371/journal.pone.0320297. eCollection 2025.

ABSTRACT

Cardiac auscultation requires the mechanical vibrations occurring on the body's surface, which carries a range of sound frequencies. These sounds are generated by the movement and pulsation of different cardiac structures as they facilitate blood circulation. Subsequently, these sounds are identified as phonocardiogram (PCG). In this research, deep learning models, namely gated recurrent neural Network (GRU), Bidirectional-GRU, and Bi-directional long-term memory (BILSTM) are applied separately to segment four specific regions within the PCG signal, namely S1 (lub sound), the systolic region, S2 (dub sound), and the diastolic region. These models are applied to three well-known datasets: PhysioNet/Computing in Cardiology Challenge 2016, Massachusetts Institute of Technology (MITHSDB), and CirCor DigiScope Phonocardiogram.The PCG signal underwent a series of pre-processing steps, including digital filtering and empirical mode decomposition, after then deep learning algorithms were applied to achieve the highest level of segmentation accuracy. Remarkably, the proposed approach achieved an accuracy of 97.2% for the PhysioNet dataset and 96.98% for the MITHSDB dataset. Notably, this paper represents the first investigation into the segmentation process of the CirCor DigiScop dataset, achieving an accuracy of 92.5%. This study compared the performance of various deep learning models using the aforementioned datasets, demonstrating its efficiency, accuracy, and reliability as a software tool in healthcare settings.

PMID:40228205 | DOI:10.1371/journal.pone.0320297

PLoS One. 2025 Apr 14;20(4):e0320360. doi: 10.1371/journal.pone.0320360. eCollection 2025.

ABSTRACT

Alzheimer's disease (AD) is a neurodegenerative disorder that requires early diagnosis for effective management. However, issues with currently available diagnostic biomarkers preclude early diagnosis, necessitating the development of alternative biomarkers and methods, such as blood-based diagnostics. We propose c-Triadem (constrained triple-input Alzheimer's disease model), a novel deep neural network to identify potential blood-based biomarkers for AD and predict mild cognitive impairment (MCI) and AD with high accuracy. The model utilizes genotyping data, gene expression data, and clinical information to predict the disease status of participants, i.e., cognitively normal (CN), MCI, or AD. The nodes of the neural network represent genes and their related pathways, and the edges represent known relationships among the genes and pathways. Simulated data validation further highlights the robustness of key features identified by SHapley Additive exPlanations (SHAP). We trained the model with blood genotyping data, microarray, and clinical features from the Alzheimer's Neuroimaging Disease Initiative (ADNI). We demonstrate that our model's performance is superior to previous models with an AUC of 97% and accuracy of 89%. We then identified the most influential genes and clinical features for prediction using SHapley Additive exPlanations (SHAP). Our SHAP analysis shows that CASP9, LCK, and SDC3 SNPs and PINK1, ATG5, and ubiquitin (UBB, UBC) expression have a higher impact on model performance. Our model has facilitated the identification of potential blood-based genetic markers of DNA damage response and mitophagy in affected regions of the brain. The model can be used for detection and biomarker identification in other related dementias.

PMID:40228177 | DOI:10.1371/journal.pone.0320360

Environ Health Perspect. 2025 Apr 14. doi: 10.1289/EHP15400. Online ahead of print.

NO ABSTRACT

PMID:40228076 | DOI:10.1289/EHP15400

Magn Reson Med. 2025 Apr 14. doi: 10.1002/mrm.30532. Online ahead of print.

ABSTRACT

PURPOSE: The aim of this study was to develop a saturation transfer MR fingerprinting (ST-MRF) technique using a biophysics model-driven deep learning approach.

METHODS: A deep learning-based quantitative saturation transfer framework was proposed to estimate water, magnetization transfer contrast, and amide proton transfer (APT) parameters plus B0 field inhomogeneity. This framework incorporated a Bloch-McConnell simulator during neural network training and enforced consistency between synthesized MRF signals and experimentally acquired ST-MRF signals. Ground-truth numerical phantoms were used to assess the accuracy of estimated tissue parameters, and in vivo tissue parameters were validated using synthetic MRI analysis.

RESULTS: The proposed ST-MRF reconstruction network achieved a normalized root mean square error (nRMSE) of 9.3% when tested against numerical phantoms with a signal-to-noise ratio of 46 dB, which outperformed conventional Bloch-McConnell fitting (nRMSE of 15.3%) and dictionary-matching approaches (nRMSE of 19.5%). Synthetic MRI analysis indicated excellent similarity (RMSE = 3.2%) between acquired and synthesized ST-MRF images, demonstrating high in vivo reconstruction accuracy. In healthy human brains, the APT pool size ratios for gray and white matter were 0.16 ± 0.02% and 0.13 ± 0.02%, respectively, and the exchange rates for gray and white matter were 101 ± 25 Hz and 131 ± 27 Hz, respectively. The reconstruction network processed the eight tissue parameter maps in approximately 27 s for ST-MRF data sized at 256 × 256 × 9 × 103.

CONCLUSION: This study highlights the feasibility of the deep learning-based ST-MRF imaging for rapid and accurate quantification of free bulk water, magnetization transfer contrast, APT parameters, and B0 field inhomogeneity.

PMID:40228056 | DOI:10.1002/mrm.30532