Sci Rep. 2025 Apr 6;15(1):11754. doi: 10.1038/s41598-025-95871-5.

ABSTRACT

To investigate the potential of employing artificial intelligence (AI) -driven breast ultrasound analysis models for the classification of glandular tissue components (GTC) in dense breast tissue. A total of 1,848 healthy women with mammograms classified as dense breast were enrolled in this prospective study. Residual Network (ResNet) 101 classification model and ResNet with fully Convolutional Networks (ResNet + FCN) segmentation model were trained. The better effective model was selected to appraise the classification performance of 3 breast radiologists and 3 non-breast radiologists. The evaluation metrics included sensitivity, specificity, and positive predictive value (PPV). The ResNet101 model demonstrated superior performance compared to the ResNet + FCN model. It significantly enhanced the classification sensitivity of all radiologists by 0.060, 0.021, 0.170, 0.009, 0.052, and 0.047, respectively. For P1 to P4 glandular, the PPVs of all radiologists increased by 0.154, 0.178, 0.027, and 0.109 with Ai-assisted. Notably, the non-breast radiologists experienced a particularly substantial rise in PPV (p < 0.01). This study trained ResNet 101 deep learning model is a reliable and accurate system for assisting different experienced radiologists differentiate dense breast glandular tissue components in ultrasound images.

PMID:40189689 | DOI:10.1038/s41598-025-95871-5

Sci Rep. 2025 Apr 6;15(1):11750. doi: 10.1038/s41598-025-95985-w.

ABSTRACT

Cassava is a tuberous edible plant native to the American tropics and is essential for its versatile applications including cassava flour, bread, tapioca, and laundry starch. Cassava leaf diseases reduce crop yields, elevate production costs, and disrupt market stability. This places significant burdens on farmers and economies while highlighting the need for effective management strategies. Traditional methods of manual disease diagnosis are costly, labor-intensive, and time-consuming. This research aims to address the challenge of accurate disease classification by overcoming the limitations of existing methods, which encounter difficulties with the complexity and variability of leaf disease symptoms. To the best of our knowledge, this is the first study to propose a novel dual-track feature aggregation architecture that integrates the Residual Inception Positional Encoding Attention (RIPEA) Network with EfficientNet for the classification of cassava leaf diseases. The proposed model employs a dual-track feature aggregation architecture which integrates the RIPEA Network with EfficientNet. The RIPEA track extracts significant features by leveraging residual connections for preserving gradients and uses multi-scale feature fusion for combining fine-grained details with broader patterns. It also incorporates Coordinate and Mixed Attention mechanisms which focus on cross-channel and long-range dependencies. The extracted features from both tracks are aggregated for classification. Furthermore, it incorporates an image augmentation method and a cosine decay learning rate schedule to improve model training. This improves the ability of the model to accurately differentiate between Cassava Bacterial Blight (CBB), Brown Streak Disease (CBSD), Green Mottle (CGM), Mosaic Disease (CMD), and healthy leaves, addressing both local textures and global structures. Additionally, to enhance the interpretability of the model, we apply Grad-CAM to provide visual explanations for the model's decision-making process, helping to understand which regions of the leaf images contribute to the classification results. The proposed network achieved a classification accuracy of 93.06%.

PMID:40189680 | DOI:10.1038/s41598-025-95985-w

Sci Rep. 2025 Apr 6;15(1):11757. doi: 10.1038/s41598-025-96385-w.

ABSTRACT

Coherent beam combination has emerged as a promising strategy for overcoming the power limitations of individual fibre lasers. This approach relies on maintaining precise phase difference between the constituent beamlets, which are typically established using phase retrieval algorithms. However, phase locking is often studied under the assumption that the power levels of the beamlets remain stable, an idealisation that does not hold always in practical applications. Over the operational lifetime of fibre lasers, power degradation inevitably occurs, introducing additional challenges to phase retrieval. To address this, we propose a deep learning algorithm for single-step simultaneous phase and amplitude identification, directly from a single camera observation of the intensity distribution of the combined beam. By leveraging its ability to detect and interpret subtle variations in intensity interference patterns, the deep learning approach can accurately disentangle phase and power contributions, even in the presence of significant power fluctuations. Using a spatial light modulator, we systematically investigate the impact of power-level fluctuations on phase retrieval within a simulated coherent beam combination system. Furthermore, we explore the scalability of this deep learning approach by evaluating its ability to achieve the required phase and amplitude precision as the number of beamlets increases.

PMID:40189661 | DOI:10.1038/s41598-025-96385-w

Sci Rep. 2025 Apr 6;15(1):11747. doi: 10.1038/s41598-025-96053-z.

ABSTRACT

The utility of CT-derived parameters for hepatic steatosis assessment has primarily focused on non-alcoholic fatty liver disease. This study aimed to evaluate their applicability in chronic hepatitis B (CHB) through a retrospective analysis of 243 CHB patients. Using deep-learning-based 3D organ segmentation on abdominal CT scans at 100 kVp, the mean volumetric CT attenuation of the liver and spleen was automatically measured on pre-contrast (liver (L)_pre and spleen (S)_pre) and post-contrast (L_post and S_post) portal venous phase images. To identify mild, moderate, and severe steatosis (S1, S2, and S3 based on the controlled attenuation parameter), L_pre showed areas under the receiver operating characteristic curve (AUROCs) of 0.695, 0.779, and 0.795, significantly higher than L-S_pre (0.633, 0.691, and 0.732; Ps = 0.02, 0.003, and 0.03). Post-contrast parameters demonstrated slightly lower AUROCs than their pre-contrast counterparts (Ps = 0.15-0.81). Concomitant hepatic fibrosis influenced diagnostic performance, with CT parameters performing better in patients without severe fibrosis than those with (F3-4 on transient elastography), though statistical significance was only observed for L-S_post in severe steatosis (P = 0.037). In conclusion, CT attenuation-based parameters extracted through automated 3D analysis show promise as a tool for assessing hepatic steatosis in patients with CHB.

PMID:40189652 | DOI:10.1038/s41598-025-96053-z

Sci Rep. 2025 Apr 6;15(1):11744. doi: 10.1038/s41598-025-95819-9.

ABSTRACT

Neural style transfer (NST) has opened new possibilities for digital art by enabling the blending of distinct artistic styles with content from various images. However, traditional NST methods often need help balancing style fidelity and content preservation, and many models need more computational efficiency, limiting their applicability for real-time applications. This study aims to enhance the efficiency and quality of NST by proposing a refined model that addresses key challenges in content retention, style fidelity, and computational performance. Specifically, the research explores techniques to improve the visual coherence of style transfer, ensuring consistency and accessibility for practical use. The proposed model integrates Adaptive Instance Normalization (AdaIN) and Gram matrix-based style representation within a convolutional neural network (CNN) architecture. The model is evaluated using quantitative metrics such as content loss, style loss, Structural Similarity Index (SSIM), and processing time, along with a qualitative assessment of content and style consistency across various image pairs. The proposed model significantly improves content and style balance, with content and style loss values reduced by 15% compared to baseline models. The optimal configuration yields an SSIM score of 0.88 for medium style intensity, maintaining structural integrity while achieving stylistic effects. Additionally, the model's processing time is reduced by 76%, making it suitable for near-real-time applications. Style fidelity scores remain high across various artistic styles, with minimal loss in content retention. The refined NST model balances style and content effectively, enhancing visual quality and computational efficiency. These advancements make NST more accessible for real-time artistic applications, providing a versatile digital art, design, and multimedia production tool.

PMID:40189651 | DOI:10.1038/s41598-025-95819-9

Sci Rep. 2025 Apr 6;15(1):11778. doi: 10.1038/s41598-025-96523-4.

ABSTRACT

Agriculture is an essential foundation that supports numerous economies, and the longevity of the coffee business is of paramount significance. Controlling and safeguarding coffee farms from harmful pests, including the Coffee Berry Borer, Mealybugs, Scales, and Leaf Miners, which may drastically affect crop productivity and quality. Standard methods for detecting pest diseases sometimes need specialized knowledge or thorough analysis, leading to a substantial commitment of time and effort. To address this challenge, researchers have explored the use of computer vision and deep learning techniques for the automated detection of plant pest diseases. This paper presents a novel strategy for the early detection of coffee crop killers using Hybrid Vision Graph Neural Networks (HV-GNN) in coffee plantations. The model was trained and validated using a curated dataset of 2850 labelled coffee plant images, which included diverse insect infestations. The HV-GNN design allows the model to recognize individual pests within images and capture the complex relationships between them, potentially leading to improved detection accuracy. HV-GNN proficiently detect pests by analyzing their visual characteristics and elucidating the interconnections among pests in images. Experimental findings indicate that HV-GNN attain a detection accuracy of 93.6625%, exceeding that of leading models. The increased accuracy underscores the feasibility of practical implementation, enabling proactive pest control to protect coffee farms and improve agricultural output.

PMID:40189644 | DOI:10.1038/s41598-025-96523-4

Sci Rep. 2025 Apr 6;15(1):11772. doi: 10.1038/s41598-025-91225-3.

ABSTRACT

Image observation method is a key method for shale reservoir evaluation. At the micro scale, scanning electron microscope images can be used to accurately understand the structural characteristics of shale. Most of the current research is to artificially identify the microstructure of shale. This approach has subjective limitations and makes it difficult to process images in batches on a large scale. We take the scanning electron microscope image as the research object, and the shale deep learning theory as the research method to realize the intelligent identification of microscopic substances in the shale scanning electron microscope image. The results show that the improved deep learning model performs better than other deep learning models. The maximum values of Precision, Recall, mAP50 and mAP50-95 reached 0.94442, 0.91695, 0.9579 and 0.71547, respectively. The functions of the optimized Yolov8 model were integrated with SEM technology. In engineering practice, it can assist researchers to quickly locate object substances and obtain high-quality SEM images, effectively improving the efficiency and accuracy of reservoir evaluation. In addition, this technology has great potential for development, and it is expected to play an important role in expanding to a variety of fields such as medicine and materials science by changing the test object.

PMID:40189619 | DOI:10.1038/s41598-025-91225-3

Radiat Oncol. 2025 Apr 5;20(1):50. doi: 10.1186/s13014-025-02622-x.

ABSTRACT

BACKGROUND: Brain metastases are common complications in patients with cancer and significantly affect prognosis and treatment strategies. The accurate segmentation of brain metastases is crucial for effective radiation therapy planning. However, in resource-limited areas, the unavailability of MRI imaging is a significant challenge that necessitates the development of reliable segmentation models for computed tomography images (CT).

PURPOSE: This study aimed to develop and evaluate a Diffusion-CSPAM-U-Net model for the segmentation of brain metastases on CT images and thereby provide a robust tool for radiation oncologists in regions where magnetic resonance imaging (MRI) is not accessible.

METHODS: The proposed Diffusion-CSPAM-U-Net model integrates diffusion models with channel-spatial-positional attention mechanisms to enhance the segmentation performance. The model was trained and validated on a dataset consisting of CT images from two centers (n = 205) and (n = 45). Performance metrics, including the Dice similarity coefficient (DSC), intersection over union (IoU), accuracy, sensitivity, and specificity, were calculated. Additionally, this study compared models proposed for brain metastases of different sizes with those proposed in other studies.

RESULTS: The diffusion-CSPAM-U-Net model achieved promising results on the external validation set. Overall average DSC of 79.3% ± 13.3%, IoU of 69.2% ± 13.3%, accuracy of 95.5% ± 11.8%, sensitivity of 80.3% ± 12.1%, specificity of 93.8% ± 14.0%, and HD of 5.606 ± 0.990 mm were measured. These results demonstrate favorable improvements over existing models.

CONCLUSIONS: The diffusion-CSPAM-U-Net model showed promising results in segmenting brain metastases in CT images, particularly in terms of sensitivity and accuracy. The proposed diffusion-CSPAM-U-Net model provides an effective tool for radiation oncologists for the segmentation of brain metastases in CT images.

PMID:40188354 | DOI:10.1186/s13014-025-02622-x

NPJ Digit Med. 2025 Apr 5;8(1):188. doi: 10.1038/s41746-025-01582-6.

ABSTRACT

Preeclampsia (PE), a severe hypertensive disorder during pregnancy, significantly contributes to maternal and neonatal mortality. Existing prediction biomarkers are often invasive and expensive, hindering their widespread application. This study introduces PROMPT (Preeclampsia Risk factor + Ophthalmic data + Mean arterial pressure Prediction Test), an AI-driven model leveraging retinal photography for PE prediction, registered at ChiCTR (ChiCTR2100049850) in August 2021. Analyzing 1812 pregnancies before 14 gestational weeks, we extracted retinal parameters using a deep learning system. The PROMPT achieved an AUC of 0.87 (0.83-0.90) for PE prediction and 0.91 (0.85-0.97) for preterm PE prediction using machine learning, significantly outperforming the baseline model (p < 0.001). It also improved detection of severe adverse pregnancy outcomes from 35% to 41%. Economically, PROMPT was estimated to avert 1809 PE cases and saved over $50 million per 100,000 screenings. These results position PROMPT as a non-invasive and cost-effective tool for prenatal care, especially valuable in low- and middle-income countries.

PMID:40188283 | DOI:10.1038/s41746-025-01582-6

Sci Rep. 2025 Apr 5;15(1):11712. doi: 10.1038/s41598-025-95291-5.

ABSTRACT

Autism spectrum disorder (ASD) presents significant challenges in diagnosis and intervention due to its diverse clinical manifestations and underlying biological complexity. This study explored machine learning approaches to enhance ASD screening accuracy and identify meaningful subtypes using clinical assessments from AGRE database integrated with molecular data from GSE15402. Analysis of ADI-R scores from a large cohort of 2794 individuals demonstrated that deep learning models could achieve exceptional screening accuracy of 95.23% (CI 94.32-95.99%). Notably, comparable performance was maintained using a streamlined set of just 27 ADI-R sub-items, suggesting potential for more efficient diagnostic tools. Clustering analyses revealed three distinct subgroups identifiable through both clinical symptoms and gene expression patterns. When ASD were grouped based on clinical features, stronger associations emerged between symptoms and underlying molecular profiles compared to grouping based on gene expression alone. These findings suggest that starting with detailed clinical observations may be more effective for identifying biologically meaningful ASD subtypes than beginning with molecular data. This integrated approach combining clinical and molecular data through machine learning offers promising directions for developing more precise screening methods and personalized intervention strategies for individuals with ASD.

PMID:40188264 | DOI:10.1038/s41598-025-95291-5

Sci Rep. 2025 Apr 5;15(1):11687. doi: 10.1038/s41598-025-96575-6.

ABSTRACT

Parkinson's disease (PD) is a neurodegenerative disorder affecting motor control, leading to symptoms such as tremors and stiffness. Early diagnosis is essential for effective treatment, but traditional methods are often time-consuming and expensive. This study leverages Artificial Intelligence (AI) and Machine Learning (ML) techniques, using voice analysis to detect early signs of PD. We applied a hybrid model combining Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Multiple Kernel Learning (MKL), and Multilayer Perceptron (MLP) to a dataset of 81 voice recordings. Acoustic features such as Mel-Frequency Cepstral Coefficients (MFCCs), jitter, and shimmer were analyzed. The model achieved 91.11% accuracy, 92.50% recall, 89.84% precision, 91.13% F1 score, and an area-under-the-curve (AUC) of 0.9125. SHapley Additive exPlanations (SHAP) provided data explainability, identifying key features driving the PD diagnosis, thus enhancing AI interpretability and trustability. Furthermore, a probability-based scoring system was developed to enable PD patients and clinicians to track disease progression. This AI-driven approach offers a non-invasive, cost-effective, and rapid tool for early PD detection, facilitating personalized treatment through vocal biomarkers.

PMID:40188263 | DOI:10.1038/s41598-025-96575-6

Abdom Radiol (NY). 2025 Apr 5. doi: 10.1007/s00261-025-04934-8. Online ahead of print.

ABSTRACT

PURPOSE: To develop deep learning models for automated detection and segmentation of uterine fibroids using multi-orientation MRI.

METHODS: Pre-treatment sagittal and axial T2-weighted MRI scans acquired from patients diagnosed with uterine fibroids were collected. The proposed segmentation models were constructed based on the three-dimensional nnU-Net framework. Fibroid detection efficacy was assessed, with subgroup analyses by size and location. The segmentation performance was evaluated using Dice similarity coefficients (DSCs), 95% Hausdorff distance (HD95), and average surface distance (ASD).

RESULTS: The internal dataset comprised 299 patients who were divided into the training set (n = 239) and the internal test set (n = 60). The external dataset comprised 45 patients. The sagittal T2WI model and the axial T2WI model demonstrated recalls of 74.4%/76.4% and precision of 98.9%/97.9% for fibroid detection in the internal test set. The models achieved recalls of 93.7%/95.3% for fibroids ≥ 4 cm. The recalls for International Federation of Gynecology and Obstetrics (FIGO) type 2-5, FIGO types 0\1\2(submucous), fibroids FIGO types 5\6\7(subserous) were 100%/100%, 73.3%/78.6%, and 80.3%/81.9%, respectively. The proposed models demonstrated good performance in segmentation of the uterine fibroids with mean DSCs of 0.789 and 0.804, HD95s of 9.996 and 10.855 mm, and ASDs of 2.035 and 2.115 mm in the internal test set, and with mean DSCs of 0.834 and 0.818, HD95s of 9.971 and 11.874 mm, and ASDs of 2.031 and 2.273 mm in the external test set.

CONCLUSION: The proposed deep learning models showed promise as reliable methods for automating the detection and segmentation of the uterine fibroids, particularly those of clinical relevance.

PMID:40188260 | DOI:10.1007/s00261-025-04934-8

Nat Commun. 2025 Apr 6;16(1):3274. doi: 10.1038/s41467-025-58521-y.

ABSTRACT

The UniProt database is a valuable resource for biocatalyst discovery, yet predicting enzymatic functions remains challenging, especially for low-similarity sequences. Identifying superior enzymes with enhanced catalytic properties is even harder. To overcome these challenges, we develop ESM-Ezy, an enzyme mining strategy leveraging the ESM-1b protein language model and similarity calculations in semantic space. Using ESM-Ezy, we identify novel multicopper oxidases (MCOs) with superior catalytic properties, achieving a 44% success rate in outperforming query enzymes (QEs) in at least one property, including catalytic efficiency, heat and organic solvent tolerance, and pH stability. Notably, 51% of the MCOs excel in environmental remediation applications, and some exhibited unique structural motifs and unique active centers enhancing their functions. Beyond MCOs, 40% of L-asparaginases identified show higher specific activity and catalytic efficiency than QEs. ESM-Ezy thus provides a promising approach for discovering high-performance biocatalysts with low sequence similarity, accelerating enzyme discovery for industrial applications.

PMID:40188191 | DOI:10.1038/s41467-025-58521-y

J Transl Med. 2025 Apr 5;23(1):405. doi: 10.1186/s12967-025-06414-5.

ABSTRACT

BACKGROUND: Identification of seizures is essential for the treatment of epilepsy. Current machine-learning and deep-learning models often perform well on public datasets when classifying generalized seizures with prominent features. However, their performance was less effective in detecting brief, localized seizures. These seizure-like patterns can be masked by fixed brain rhythms.

METHODS: Our study proposes a supervised multilayer hybrid model called GEM-CRAP (gradient-enhanced modulation with CNN-RES, attention-like, and pre-policy networks), with three parallel feature extraction channels: a CNN-RES module, an amplitude-aware channel with attention-like mechanisms, and an LSTM-based pre-policy layer integrated into the recurrent neural network. The model was trained on the Xuanwu Hospital and HUP iEEG dataset, including intracranial, cortical, and stereotactic EEG data from 83 patients, covering over 8500 labeled electrode channels for hybrid classification (wakefulness and sleep). A post-SVM network was used for secondary training on channels with classification accuracy below 80%. We introduced an average channel deviation rate metric to assess seizure detection accuracy.

RESULTS: For public datasets, the model achieved over 97% accuracy for intracranial and cortical EEG sequences in patients, and over 95% for mixed sequences, with deviations below 5%. In the Xuanwu Hospital dataset, it maintained over 94% accuracy for wakefulness seizures and around 90% during sleep. SVM secondary training improved average channel accuracy by over 10%. Additionally, a strong positive correlation was found between channel accuracy distribution and the temporal distribution of seizure states.

CONCLUSIONS: GEM-CRAP enhances focal epilepsy detection through adaptive adjustments and attention mechanisms, achieving higher precision and robustness in complex signal environments. Beyond improving seizure interval detection, it excels in identifying and analyzing specific epileptic waveforms, such as high-frequency oscillations. This advancement may pave the way for more precise epilepsy diagnostics and provide a suitable artificial intelligence algorithm for closed-loop neurostimulation.

PMID:40188070 | DOI:10.1186/s12967-025-06414-5

Chin J Traumatol. 2025 Mar 15:S1008-1275(25)00029-X. doi: 10.1016/j.cjtee.2024.10.006. Online ahead of print.

ABSTRACT

PURPOSE: Distal radius fracture is a common trauma fracture and timely preoperative diagnosis is crucial for the patient's recovery. With the rise of deep-learning applications in the medical field, utilizing deep-learning for diagnosing distal radius fractures has become a significant topic. However, previous research has suffered from low detection accuracy and poor identification of occult fractures. This study aims to design an improved deep-learning model to assist surgeons in diagnosing distal radius fractures more quickly and accurately.

METHODS: This study, inspired by the comprehensive analysis of anteroposterior and lateral X-ray images by surgeons in diagnosing distal radius fractures, designs a dual-channel feature fusion network for detecting distal radius fractures. Based on the Faster region-based convolutional neural network framework, an additional Residual Network 50, which is integrated with the Deformable and Separable Attention mechanism, was introduced to extract semantic information from lateral X-ray images of the distal radius. The features extracted from the 2 channels were then combined via feature fusion, thus enriching the network's feature information. The focal loss function was also employed to address the sample imbalance problem during the training process.The selection of cases in this study was based on distal radius X-ray images retrieved from the hospital's imaging database, which met the following criteria: inclusion criteria comprised clear anteroposterior and lateral X-ray images, which were diagnosed as distal radius fractures by experienced radiologists. The exclusion criteria encompassed poor image quality, the presence of severe multiple or complex fractures, as well as non-adult or special populations (e.g., pregnant women). All cases meeting the inclusion criteria were labeled as distal radius fracture cases for model training and evaluation. To assess the model's performance, this study employed several metrics, including accuracy, precision, recall, area under the precision-recall curve, and intersection over union.

RESULTS: The proposed dual-channel feature fusion network achieved an average precision (AP)50 of 98.5%, an AP75 of 78.4%, an accuracy of 96.5%, and a recall of 94.7%. When compared to traditional models, such as Faster region-based convolutional neural network, which achieved an AP50 of 94.1%, an AP75 of 70.6%, a precision of 91.1%, and a recall of 92.3%, our method shows notable improvements in all key metrics. Similarly, when compared to other classic object detection networks like You Only Look Once version 4 (AP50=95.2%, AP75=72.2 %, precision=91.2%, recall=92.4%) and You Only Look Once version 5s (AP50=95.1%, AP75=73.8%, precision=93.7%, recall=92.8%), the dual-channel feature fusion network outperforms them in precision, recall, and AP scores. These results highlight the superior accuracy and reliability of the proposed method, particularly in identifying both apparent and occult distal radius fractures, demonstrating its effectiveness in clinical applications where precise detection of subtle fractures is critical.

CONCLUSION: This study found that combining anteroposterior and lateral X-ray images of the distal radius as input for deep-learning algorithms can more accurately and efficiently identify distal radius fractures, providing a reference for research on distal radius fractures.

PMID:40187904 | DOI:10.1016/j.cjtee.2024.10.006

Anal Chim Acta. 2025 May 15;1351:343920. doi: 10.1016/j.aca.2025.343920. Epub 2025 Mar 8.

ABSTRACT

BACKGROUND: Pharmaceutical residual such as antibiotics and disinfectants in aquaculture wastewater have significant potential risks for environment and human health. Surface enhanced Raman spectroscopy (SERS) has been widely used for the detection of pharmaceuticals due to its high sensitivity, low cost, and rapidity. However, it is remain a challenge for high-sensitivity SERS detection and accurate identification of complex pollutants.

RESULTS: Hence, in this work, we developed an aluminum foil (AlF) based SERS detection substrate and established a multilayer perceptron (MLP) deep learning model for the rapid identification of antibiotic components in a mixture. The detection method demonstrated exceptional performance, achieving a high SERS enhancement factor of 4.2 × 105 and excellent sensitivity for trace amounts of fleroxacin (2.7 × 10-8 mol/L), levofloxacin (1.95 × 10-8 mol/L), and pefloxacin (6.9 × 10-8 mol/L),sulfadiazine, methylene blue, and malachite green at a concentration of 1 × 10-8 mol/L can all be detected, the concentrations of the six target compounds and their Raman intensities exhibit a good linear relationship. Moreover, the AlF SERS substrate can be prepared rapidly without adding organic reagents, and it exhibited good reproducibility, with RSD<9.6 %. Additionally, the algorithm model can accurately identify the contaminants mixture of sulfadiazine, methylene blue, and malachite green with a recognition accuracy of 97.8 %, an F1-score of 98.2 %, and a 5-fold cross validation score of 97.4 %, the interpretation analysis using Shapley Additive Explanations (SHAP) reveals that MLP model can specifically concentrate on the distribution of characteristic peaks.

SIGNIFICANCE: The experimental results indicated that the MLP model demonstrated strong performance and good robustness in complex matrices. This research provides a promising detection and identification method for the antibiotics and disinfectants in actual aquaculture wastewater treatment.

PMID:40187885 | DOI:10.1016/j.aca.2025.343920

Anal Chim Acta. 2025 May 15;1351:343900. doi: 10.1016/j.aca.2025.343900. Epub 2025 Mar 5.

ABSTRACT

Osteoarthritis (OA) is a major cause of disability worldwide, with symptoms like joint pain, limited functionality, and decreased quality of life, potentially leading to deformity and irreversible damage. Chemical changes in joint tissues precede imaging alterations, making early diagnosis challenging for conventional methods like X-rays. Although Raman imaging provides detailed chemical information, it is time-consuming. This paper aims to achieve rapid osteoarthritis diagnosis and grading using a self-developed Raman imaging system combined with deep learning denoising and acceleration algorithms. Our self-developed aberration-corrected line-scanning confocal Raman imaging device acquires a line of Raman spectra (hundreds of points) per scan using a galvanometer or displacement stage, achieving spatial and spectral resolutions of 2 μm and 0.2 nm, respectively. Deep learning algorithms enhance the imaging speed by over 4 times through effective spectrum denoising and signal-to-noise ratio (SNR) improvement. By leveraging the denoising capabilities of deep learning, we are able to acquire high-quality Raman spectral data with a reduced integration time, thereby accelerating the imaging process. Experiments on the tibial plateau of osteoarthritis patients compared three excitation wavelengths (532, 671, and 785 nm), with 671 nm chosen for optimal SNR and minimal fluorescence. Machine learning algorithms achieved a 98 % accuracy in distinguishing articular from calcified cartilage and a 97 % accuracy in differentiating osteoarthritis grades I to IV. Our fast Raman imaging system, combining an aberration-corrected line-scanning confocal Raman imager with deep learning denoising, offers improved imaging speed and enhanced spectral and spatial resolutions. It enables rapid, label-free detection of osteoarthritis severity and can identify early compositional changes before clinical imaging, allowing precise grading and tailored treatment, thus advancing orthopedic diagnostics and improving patient outcomes.

PMID:40187878 | DOI:10.1016/j.aca.2025.343900

Int J Comput Assist Radiol Surg. 2025 Apr 5. doi: 10.1007/s11548-025-03348-7. Online ahead of print.

ABSTRACT

PURPOSE: Metal artifacts remain a persistent issue in intraoperative CBCT imaging. Particularly in orthopedic and trauma applications, these artifacts obstruct clinically relevant areas around the implant, reducing the modality's clinical value. Metal artifact avoidance (MAA) methods have shown potential to improve image quality through trajectory adjustments, but often fail in clinical practice due to their focus on irrelevant objects and high computational demands. To address these limitations, we introduce the novel parametric metal artifact avoidance (P-MAA) method.

METHODS: The P-MAA method first detects keypoints in two scout views using a deep learning model. These keypoints are used to model each clinically relevant object as an ellipsoid, capturing its position, extent, and orientation. We hypothesize that fine details of object shapes are less critical for artifact reduction. Based on these ellipsoidal representations, we devise a computationally efficient metric for scoring view trajectories, enabling fast, CPU-based optimization. A detection model for object localization was trained using both simulated and real data and validated on real clinical cases. The scoring method was benchmarked against a raytracing-based approach.

RESULTS: The trained detection model achieved a mean average recall of 0.78, demonstrating generalizability to unseen clinical cases. The ellipsoid-based scoring method closely approximated results using raytracing and was effective in complex clinical scenarios. Additionally, the ellipsoid method provided a 33-fold increase in speed, without the need for GPU acceleration.

CONCLUSION: The P-MAA approach provides a feasible solution for metal artifact avoidance in CBCT imaging, enabling fast trajectory optimization while focusing on clinically relevant objects. This method represents a significant step toward practical intraoperative implementation of MAA techniques.

PMID:40186717 | DOI:10.1007/s11548-025-03348-7

Pediatr Radiol. 2025 Apr 5. doi: 10.1007/s00247-025-06225-2. Online ahead of print.

ABSTRACT

BACKGROUND: Hepatocyte cytokeratin 7 (CK7) is a reliable marker for evaluating the severity of cholestasis in chronic cholestatic cholangiopathies. However, there is currently no noninvasive test available to assess the status of hepatocyte CK7 in pancreaticobiliary maljunction patients.

OBJECTIVE: We aimed to develop a deep learning radiomics nomogram using magnetic resonance images (MRIs) to preoperatively identify the hepatocyte CK7 status and assess cholestasis progression in patients with pancreaticobiliary maljunction.

MATERIALS AND METHODS: In total, 180 pancreaticobiliary maljunction patients were retrospectively enrolled and were randomly divided into a training cohort (n = 144) and a validation cohort (n = 36). CK7 status was determined through immunohistochemical analysis. Pyradiomics and pretrained ResNet50 were used to extract radiomics and deep learning features, respectively. To construct the radiomics and deep learning signature, feature selection methods including the minimum redundancy-maximum relevance and least absolute shrinkage and selection operator were employed. The integrated deep learning radiomics nomogram model was constructed by combining the imaging signatures and valuable clinical feature.

RESULTS: The deep learning signature exhibited superior predictive performance compared with the radiomics signature, as evidenced by the higher area under the curve (AUC) values in validation cohort (0.92 vs. 0.81). Further, the deep learning radiomics nomogram, which incorporated the radiomics signature, deep learning signature, and Komi classification, demonstrated excellent predictive ability for CK7 expression, with AUC value of 0.95 in the validation cohort.

CONCLUSION: The proposed deep learning radiomics nomogram exhibits promising performance in accurately identifying hepatic CK7 expression, thus facilitating prediction of cholestasis progression and perhaps earlier initiation of treatment in pancreaticobiliary maljunction children.

PMID:40186654 | DOI:10.1007/s00247-025-06225-2

Pediatr Radiol. 2025 Apr 5. doi: 10.1007/s00247-025-06230-5. Online ahead of print.

ABSTRACT

BACKGROUND: Radial k-space sampling is widely employed in paediatric magnetic resonance imaging (MRI) to mitigate motion and aliasing artefacts. Artificial intelligence (AI)-based image reconstruction has been developed to enhance image quality and accelerate acquisition time.

OBJECTIVE: To assess image quality of deep learning (DL)-based denoising image reconstruction of body MRI in children.

MATERIALS AND METHODS: Children who underwent thoraco-abdominal MRI employing radial k-space filling technique (PROPELLER) with conventional and DL-based image reconstruction between April 2022 and January 2023 were eligible for this retrospective study. Only cases with previous MRI including comparable PROPELLER sequences with conventional image reconstruction were selected. Image quality was compared between DL-reconstructed axial T1-weighted and T2-weighted images and conventionally reconstructed images from the same PROPELLER acquisition. Quantitative image quality was assessed by signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of the liver and spleen. Qualitative image quality was evaluated by three observers using a 4-point Likert scale and included presence of noise, motion artefact, depiction of peripheral lung vessels and subsegmental bronchi at the lung bases, sharpness of abdominal organ borders, and visibility of liver and spleen vessels. Image quality was compared with the Wilcoxon signed-rank test. Scan time length was compared to prior MRI obtained with conventional image reconstruction.

RESULTS: In 21 children (median age 7 years, range 1.5 years to 15.8 years) included, the SNR and CNR of the liver and spleen on T1-weighted and T2-weighted images were significantly higher with DL-reconstruction (P<0.001) than with conventional reconstruction. The DL-reconstructed images showed higher overall image quality, with improved delineation of the peripheral vessels and the subsegmental bronchi in the lung bases, sharper abdominal organ margins and increased visibility of the peripheral vessels in the liver and spleen. Not respiratory-gated DL-reconstructed T1-weighted images demonstrated more pronounced respiratory motion artefacts in comparison to conventional reconstruction (P=0.015), while there was no difference for the respiratory-gated T2-weighted images. The median scan time per slice was reduced from 6.3 s (interquartile range, 4.2 - 7.0 s) to 4.8 s (interquartile range, 4.4 - 4.9 s) for the T1-weighted images and from 5.6 s (interquartile range, 5.4 - 5.9 s) to 4.2 s (interquartile range, 3.9 - 4.8 s) for the T2-weighted images.

CONCLUSION: DL-based denoising image reconstruction of paediatric body MRI sequences employing radial k-space sampling allowed for improved overall image quality at shorter scan times. Respiratory motion artefacts were more pronounced on ungated T1-weighted images.

PMID:40186652 | DOI:10.1007/s00247-025-06230-5