Pub Date : 2025-08-01Epub Date: 2025-07-28DOI: 10.1016/j.slast.2025.100337
Komel Tariq, Nosheen Fatima Rana, Sabah Javaid, Muneeba Khadim
Implant-associated infections remain a significant challenge in orthopaedic and dental implants because they frequently result in implant failure, extended hospital stays, reoperations, and increased healthcare costs. Studies have shown that the cost of managing orthopaedic implant infections can range from USD 30,000 to over USD 100,000 per case, depending on severity and required surgical interventions. One of the primary pathogens responsible for these infections is Staphylococcus aureus, known for its potential to make biofilms on the surfaces of implants. To address this problem, this study investigates the formation of calcium phosphate-based biomimetic coatings substituted with calcium-doped ZnO nanoparticles on titanium discs to strengthen the antibacterial properties and enhance tissue integration. The SEM analysis of discs revealed uniform and dense coating layers with negligible surface defects, indicating a strong adhesive coating on titanium discs. The biomimetic-coated titanium implants with Ca-doped ZnO NPs were then evaluated for antibacterial activity using a closed system in an in vitro biofilm model. In case of 14 days treated disc, a significant increase in the antibacterial properties was observed against (Staphylococcus aureus, p < 0.0001). These findings suggest that calcium phosphate-based biomimetic coatings, doped with calcium-doped ZnO NPs show great potential for reducing the risk for implant-associated infections and improving the success rate of implants in clinical settings.
{"title":"Titanium surface functionalization with calcium-doped ZnO nanoparticles for hard tissue implant applications","authors":"Komel Tariq, Nosheen Fatima Rana, Sabah Javaid, Muneeba Khadim","doi":"10.1016/j.slast.2025.100337","DOIUrl":"10.1016/j.slast.2025.100337","url":null,"abstract":"<div><div>Implant-associated infections remain a significant challenge in orthopaedic and dental implants because they frequently result in implant failure, extended hospital stays, reoperations, and increased healthcare costs. Studies have shown that the cost of managing orthopaedic implant infections can range from USD 30,000 to over USD 100,000 per case, depending on severity and required surgical interventions. One of the primary pathogens responsible for these infections is <em>Staphylococcus aureus,</em> known for its potential to make biofilms on the surfaces of implants. To address this problem, this study investigates the formation of calcium phosphate-based biomimetic coatings substituted with calcium-doped ZnO nanoparticles on titanium discs to strengthen the antibacterial properties and enhance tissue integration. The SEM analysis of discs revealed uniform and dense coating layers with negligible surface defects, indicating a strong adhesive coating on titanium discs. The biomimetic-coated titanium implants with Ca-doped ZnO NPs were then evaluated for antibacterial activity using a closed system in an <em>in vitro</em> biofilm model. In case of 14 days treated disc, a significant increase in the antibacterial properties was observed against (<em>Staphylococcus aureus, p</em> < 0.0001)<em>.</em> These findings suggest that calcium phosphate-based biomimetic coatings, doped with calcium-doped ZnO NPs show great potential for reducing the risk for implant-associated infections and improving the success rate of implants in clinical settings.</div></div>","PeriodicalId":54248,"journal":{"name":"SLAS Technology","volume":"33 ","pages":"Article 100337"},"PeriodicalIF":3.7,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144738961","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-01Epub Date: 2025-06-19DOI: 10.1016/j.slast.2025.100320
Shufang Xiao, Meimei Lin
Premature rupture of membranes is one of the more common symptoms of pregnant women before labor, which can lead to an increased rate of preterm birth and a higher mortality rate of the fetus born from it. The current research on premature rupture of membranes (PROM) is mainly based on multivariate regression analysis, and variables are selected for multivariate regression analysis after univariate analysis. This method may omit some independent variables, resulting in one-sided analysis results. In this context, this study uses Bayesian method and Logistic regression analysis to construct a new variable analysis model to analyze the clinical characteristics and risk factors of PROM infection. First, through Bayesian Logistic regression, the clinical features of PROM infection mainly include fever, increased white blood cells and C-reactive protein, and increased fetal heart rate. The analysis of risk factors showed that pathogen infection, maternal pregnancy number, and scarred uterus were all risk factors for PROM infection. Finally, in order to explain the effect of the analysis model used in this paper, a nonparametric test, AUC value and ROC curve were used to compare the effect of Bayesian Logistic regression and Logistic regression. The results showed that the statistic value of Bayesian logistic regression was 0.177 higher than that of logistic regression, and the AUC value was 0.014 higher. That is, the performance of the Bayesian logistic regression model is better. The method used in the experiment is feasible, and the experimental results are in line with expectations.
{"title":"Clinical characteristics and risk factors of premature rupture of membranes infection in pregnant and lying-in women","authors":"Shufang Xiao, Meimei Lin","doi":"10.1016/j.slast.2025.100320","DOIUrl":"10.1016/j.slast.2025.100320","url":null,"abstract":"<div><div>Premature rupture of membranes is one of the more common symptoms of pregnant women before labor, which can lead to an increased rate of preterm birth and a higher mortality rate of the fetus born from it. The current research on premature rupture of membranes (PROM) is mainly based on multivariate regression analysis, and variables are selected for multivariate regression analysis after univariate analysis. This method may omit some independent variables, resulting in one-sided analysis results. In this context, this study uses Bayesian method and Logistic regression analysis to construct a new variable analysis model to analyze the clinical characteristics and risk factors of PROM infection. First, through Bayesian Logistic regression, the clinical features of PROM infection mainly include fever, increased white blood cells and C-reactive protein, and increased fetal heart rate. The analysis of risk factors showed that pathogen infection, maternal pregnancy number, and scarred uterus were all risk factors for PROM infection. Finally, in order to explain the effect of the analysis model used in this paper, a nonparametric test, AUC value and ROC curve were used to compare the effect of Bayesian Logistic regression and Logistic regression. The results showed that the statistic value of Bayesian logistic regression was 0.177 higher than that of logistic regression, and the AUC value was 0.014 higher. That is, the performance of the Bayesian logistic regression model is better. The method used in the experiment is feasible, and the experimental results are in line with expectations.</div></div>","PeriodicalId":54248,"journal":{"name":"SLAS Technology","volume":"33 ","pages":"Article 100320"},"PeriodicalIF":2.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144340585","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-01Epub Date: 2025-07-04DOI: 10.1016/j.slast.2025.100328
Ru Liu , Wenxi Shen
This project aims to improve the accuracy of fitness and physical pressure ratings, focusing on basketball, by integrating artificial intelligence (AI) into data collection and training. Athletes and fitness fanatics can benefit greatly from the data collected using complex AI algorithms to determine stress levels. This study employs the Intelligent Physiological Monitoring Framework for Exercise and Fitness Pressure Measurement (IPM-EFPM) to perform automated stress tests that employ AI to enhance the precision of exercise and fitness pressure measurements. Basketball training programs can benefit from this framework's utilization of state-of-the-art technology, meticulous monitoring of exercise-induced stress, and continuous validation and improvement. The IPM-EFPM system gathers data from wearable sensors, uses real-time location systems, and employs artificial intelligence's Long Short-Term Memory (LSTM) and machine learning algorithms to uncover new insights in healthcare and sports. To accurately record fitness strain, physical activity, exercise-induced stress, and sports like basketball, this system employs cutting-edge artificial intelligence technologies, such as wearable sensors and current gathering data methods. Placement of sensors, real-time data collecting, data preprocessing and integrating, evaluation of stress by artificial intelligence algorithms, discovery and application of new information, validation and improvement are all parts of an iterative method that has been fine-tuned for use in sports and fitness settings by the IPM-EFPM. Examining the intricate relationship between AI, physical activity, and psychological stress is the main objective of this research. This could have real-world uses tailored to the sports world, particularly for basketball players.
{"title":"Data acquisition of exercise and fitness pressure measurement based on artificial intelligence technology","authors":"Ru Liu , Wenxi Shen","doi":"10.1016/j.slast.2025.100328","DOIUrl":"10.1016/j.slast.2025.100328","url":null,"abstract":"<div><div>This project aims to improve the accuracy of fitness and physical pressure ratings, focusing on basketball, by integrating artificial intelligence (AI) into data collection and training. Athletes and fitness fanatics can benefit greatly from the data collected using complex AI algorithms to determine stress levels. This study employs the Intelligent Physiological Monitoring Framework for Exercise and Fitness Pressure Measurement (IPM-EFPM) to perform automated stress tests that employ AI to enhance the precision of exercise and fitness pressure measurements. Basketball training programs can benefit from this framework's utilization of state-of-the-art technology, meticulous monitoring of exercise-induced stress, and continuous validation and improvement. The IPM-EFPM system gathers data from wearable sensors, uses real-time location systems, and employs artificial intelligence's Long Short-Term Memory (LSTM) and machine learning algorithms to uncover new insights in healthcare and sports. To accurately record fitness strain, physical activity, exercise-induced stress, and sports like basketball, this system employs cutting-edge artificial intelligence technologies, such as wearable sensors and current gathering data methods. Placement of sensors, real-time data collecting, data preprocessing and integrating, evaluation of stress by artificial intelligence algorithms, discovery and application of new information, validation and improvement are all parts of an iterative method that has been fine-tuned for use in sports and fitness settings by the IPM-EFPM. Examining the intricate relationship between AI, physical activity, and psychological stress is the main objective of this research. This could have real-world uses tailored to the sports world, particularly for basketball players.</div></div>","PeriodicalId":54248,"journal":{"name":"SLAS Technology","volume":"33 ","pages":"Article 100328"},"PeriodicalIF":2.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144576956","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-01Epub Date: 2025-06-28DOI: 10.1016/j.slast.2025.100324
Amal Al-Rasheed , Sheikh Muhammad Saqib , Muhammad Zubair Asghar , Tehseen Mazhar , Asim Seedahmed Ali Osman , Mohammad Shahid , Muhammad Iqbal , Muhammad Amir Khan
Early diagnosis and thorough management techniques are crucial for people with chronic kidney disease (CKD), a crippling and potentially fatal condition. Research has focused a lot on machine learning and deep learning systems for the detection of kidney diseases. Deep learning platforms like hidden layers, activation functions, optimizers, and epochs are also necessary for the automatic detection of these diseases. The proposed model achieved 99 % accuracy, with a precision, recall, and F1 score of 0.99, indicating highly reliable performance. Additionally, the model demonstrated strong agreement and robustness, as reflected in metrics such as the ROC AUC score of 0.9821 and Matthews Correlation Coefficient of 0.9727. The experiment used a publicly accessible dataset with 24 independent fields and independent values as chronic or not-chronic classes, building dense-layered deep neural networks based on an optimized architecture. The outcomes demonstrated that, when compared to the other models, the proposed model was the most accurate.
{"title":"Classifying kidney disease using a dense layers deep learning model","authors":"Amal Al-Rasheed , Sheikh Muhammad Saqib , Muhammad Zubair Asghar , Tehseen Mazhar , Asim Seedahmed Ali Osman , Mohammad Shahid , Muhammad Iqbal , Muhammad Amir Khan","doi":"10.1016/j.slast.2025.100324","DOIUrl":"10.1016/j.slast.2025.100324","url":null,"abstract":"<div><div>Early diagnosis and thorough management techniques are crucial for people with chronic kidney disease (CKD), a crippling and potentially fatal condition. Research has focused a lot on machine learning and deep learning systems for the detection of kidney diseases. Deep learning platforms like hidden layers, activation functions, optimizers, and epochs are also necessary for the automatic detection of these diseases. The proposed model achieved 99 % accuracy, with a precision, recall, and F1 score of 0.99, indicating highly reliable performance. Additionally, the model demonstrated strong agreement and robustness, as reflected in metrics such as the ROC AUC score of 0.9821 and Matthews Correlation Coefficient of 0.9727. The experiment used a publicly accessible dataset with 24 independent fields and independent values as chronic or not-chronic classes, building dense-layered deep neural networks based on an optimized architecture. The outcomes demonstrated that, when compared to the other models, the proposed model was the most accurate.</div></div>","PeriodicalId":54248,"journal":{"name":"SLAS Technology","volume":"33 ","pages":"Article 100324"},"PeriodicalIF":2.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144531106","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-01Epub Date: 2025-07-23DOI: 10.1016/j.slast.2025.100336
Muhammad Waqar , Zeshan Aslam Khan , Shanzey Tariq Khawaja , Naveed Ishtiaq Chaudhary , Saadia Khan , Khalid Mehmood Cheema , Muhammad Farhan Khan , Syed Sohail Ahmed , Muhammad Asif Zahoor Raja
Deep learning (DL) has had an incredible influence on many different scientific areas over the past couple of decades. Particularly in the field of healthcare, DL strategies were able to outclass other existing methodologies in image processing. The rapid expansion of the monkeypox endemic to over 40 nations apart from Africa has prompted serious worries in the realm of public health. Given that monkeypox can have symptoms that are akin to both chickenpox and measles, early detection can be difficult. Fortunately, due to the developments in artificial intelligence approaches, it can be implemented to promptly and accurately identify monkeypox disease using visual data information. Many DL driven techniques have already been exploited in the literature for skin related issues, which have provided accurate results to some extent. These models were dependent on extensive computational and time resources due to which the real-time applicability is difficult. Rather of building and training CNNs from scratch, this study uses transfer learning (TL) technique to fine-tune pre-trained networks, particularly exploiting various versions of ConvNeXt, by substituting last layer with additional task specific ones. A number of pre-processing and data augmentation methods have also been assessed and adjusted with regard to computing time and performance. The proposed study performs the binary and multi class monkeypox disease classification task. Promising accurate results of 99.9 % on the benchmark MSLD (binary class) dataset and 94 % on the MSLD v2.0 (multi-class) dataset is obtained by fine-tuned TL-based ConvNeXtSmall and ConvNeXtBase architecture with Adafactor optimization technique, demonstrating the practicality of the suggested framework as a substitute for the current ones. The proposed model is assessed through both standard train-test split and k-fold cross validation techniques. Furthermore, performance of models is also assessed on several other metrics including recall, F1 score, precision and multiple statistical tests incorporated with explainable AI methods for better interpretability of results. The concerns regarding the real-time applicability are tackled by utilizing the less time consuming and computationally efficient networks through the exploitation of transfer learning capabilities. Moreover, the explainable findings of the proposed study will be highly valuable for the healthcare professionals to understand the decisive behavior of the model and make informed clinical decisions.
{"title":"Explainable clinical diagnosis through unexploited yet optimized fine-tuned ConvNeXt Models for accurate monkeypox disease classification","authors":"Muhammad Waqar , Zeshan Aslam Khan , Shanzey Tariq Khawaja , Naveed Ishtiaq Chaudhary , Saadia Khan , Khalid Mehmood Cheema , Muhammad Farhan Khan , Syed Sohail Ahmed , Muhammad Asif Zahoor Raja","doi":"10.1016/j.slast.2025.100336","DOIUrl":"10.1016/j.slast.2025.100336","url":null,"abstract":"<div><div>Deep learning (DL) has had an incredible influence on many different scientific areas over the past couple of decades. Particularly in the field of healthcare, DL strategies were able to outclass other existing methodologies in image processing. The rapid expansion of the monkeypox endemic to over 40 nations apart from Africa has prompted serious worries in the realm of public health. Given that monkeypox can have symptoms that are akin to both chickenpox and measles, early detection can be difficult. Fortunately, due to the developments in artificial intelligence approaches, it can be implemented to promptly and accurately identify monkeypox disease using visual data information. Many DL driven techniques have already been exploited in the literature for skin related issues, which have provided accurate results to some extent. These models were dependent on extensive computational and time resources due to which the real-time applicability is difficult. Rather of building and training CNNs from scratch, this study uses transfer learning (TL) technique to fine-tune pre-trained networks, particularly exploiting various versions of ConvNeXt, by substituting last layer with additional task specific ones. A number of pre-processing and data augmentation methods have also been assessed and adjusted with regard to computing time and performance. The proposed study performs the binary and multi class monkeypox disease classification task. Promising accurate results of 99.9 % on the benchmark MSLD (binary class) dataset and 94 % on the MSLD v2.0 (multi-class) dataset is obtained by fine-tuned TL-based ConvNeXtSmall and ConvNeXtBase architecture with Adafactor optimization technique, demonstrating the practicality of the suggested framework as a substitute for the current ones. The proposed model is assessed through both standard train-test split and k-fold cross validation techniques. Furthermore, performance of models is also assessed on several other metrics including recall, F1 score, precision and multiple statistical tests incorporated with explainable AI methods for better interpretability of results. The concerns regarding the real-time applicability are tackled by utilizing the less time consuming and computationally efficient networks through the exploitation of transfer learning capabilities. Moreover, the explainable findings of the proposed study will be highly valuable for the healthcare professionals to understand the decisive behavior of the model and make informed clinical decisions.</div></div>","PeriodicalId":54248,"journal":{"name":"SLAS Technology","volume":"33 ","pages":"Article 100336"},"PeriodicalIF":3.7,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144719154","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-01Epub Date: 2025-06-11DOI: 10.1016/j.slast.2025.100307
Ronghai Cheng , Adil Muneer , Maria Hercher , Bekim Bajrami , Reza Nemati
Covalent drug discovery has garnered renewed interest due to its potential to target proteins previously considered "undruggable." Intact protein mass spectrometry (MS) is a critical technique for providing direct evidence of covalent drug modifications to protein targets. However, its application for screening covalent libraries has been hindered by low assay throughput, complex sample preparation, and high protein consumption associated with traditional liquid chromatography-MS (LC-MS) or solid-phase extraction-MS (SPE-MS) platforms. The recent integration of acoustic ejection (AE) with electrospray ionization (ESI) source of high-resolution time-of-flight (TOF) mass spectrometers—specifically, the SCIEX Echo® MS+ with the ZenoTOF 7600—has enabled the direct introduction of intact proteins without desalting at nanoliter volumes from 384 or 1536 well plates into the electrospray ionization (ESI) source of the mass spectrometer, achieving analysis rates of 1–2 seconds per sample. This advancement offers significant potential for covalent library screening and kinetic studies of identified hits due to ultrafast sample introduction and minimal sample consumption. To fully automate this pipeline, the SCIEX Echo® MS+ with ZenoTOF 7600 mass spectrometer was integrated with our internal automation system (HighRes Biosolutions) and the data analysis workflow was automated. Using Bruton’s tyrosine kinase (BTK) as a model, we demonstrated that this integrated pipeline could accelerate covalent drug discovery through covalent library screens, off-target reactivity assessment via GSH reactivity assays, and potency evaluation through kinact/Ki measurements.
{"title":"Acoustic ejection mass spectrometry: An integrated pipeline for ultra-high throughput screening, reactivity profiling, and potency analysis of covalent BTK inhibitors","authors":"Ronghai Cheng , Adil Muneer , Maria Hercher , Bekim Bajrami , Reza Nemati","doi":"10.1016/j.slast.2025.100307","DOIUrl":"10.1016/j.slast.2025.100307","url":null,"abstract":"<div><div>Covalent drug discovery has garnered renewed interest due to its potential to target proteins previously considered \"undruggable.\" Intact protein mass spectrometry (MS) is a critical technique for providing direct evidence of covalent drug modifications to protein targets. However, its application for screening covalent libraries has been hindered by low assay throughput, complex sample preparation, and high protein consumption associated with traditional liquid chromatography-MS (LC-MS) or solid-phase extraction-MS (SPE-MS) platforms. The recent integration of acoustic ejection (AE) with electrospray ionization (ESI) source of high-resolution time-of-flight (TOF) mass spectrometers—specifically, the SCIEX Echo® MS+ with the ZenoTOF 7600—has enabled the direct introduction of intact proteins without desalting at nanoliter volumes from 384 or 1536 well plates into the electrospray ionization (ESI) source of the mass spectrometer, achieving analysis rates of 1–2 seconds per sample. This advancement offers significant potential for covalent library screening and kinetic studies of identified hits due to ultrafast sample introduction and minimal sample consumption. To fully automate this pipeline, the SCIEX Echo® MS+ with ZenoTOF 7600 mass spectrometer was integrated with our internal automation system (HighRes Biosolutions) and the data analysis workflow was automated. Using Bruton’s tyrosine kinase (BTK) as a model, we demonstrated that this integrated pipeline could accelerate covalent drug discovery through covalent library screens, off-target reactivity assessment via GSH reactivity assays, and potency evaluation through k<sub>inact</sub>/K<sub>i</sub> measurements.</div></div>","PeriodicalId":54248,"journal":{"name":"SLAS Technology","volume":"33 ","pages":"Article 100307"},"PeriodicalIF":2.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144254755","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-01Epub Date: 2025-06-04DOI: 10.1016/j.slast.2025.100311
Jamien Lim , Tal Murthy
{"title":"Literature highlights column: From the literature life sciences discovery and technology highlights","authors":"Jamien Lim , Tal Murthy","doi":"10.1016/j.slast.2025.100311","DOIUrl":"10.1016/j.slast.2025.100311","url":null,"abstract":"","PeriodicalId":54248,"journal":{"name":"SLAS Technology","volume":"33 ","pages":"Article 100311"},"PeriodicalIF":3.7,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144250861","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Diabetic retinopathy (DR) remains a key contributor to eye impairment worldwide, requiring the development of efficient and accurate deep learning models for automated diagnosis. This study presents FastEffNet, a novel framework that leverages transformer-based knowledge distillation (KD) to enhance DR severity classification while reducing computational complexity. The proposed approach employs FastViT-MA26 as the teacher model and EfficientNet-B0 as the student model, striking the ideal mix between accuracy and computational efficiency. APTOS blindness detection dataset comprising 3662 images across five severity classes is collected, pre-processed, normalized, split and augmented to address class imbalance. The teacher model undergoes training and validation before transferring its knowledge to the student model, enabling the latter to approximate the teacher’s performance while maintaining a lightweight architecture. To comprehensively assess the efficacy of the proposed framework, additional student models—including HGNet, ResNet50, MobileNetV3, and DeiT—are analysed for comparative assessment. Model interpretability is enhanced through Grad-CAM++ visualizations, which highlight critical retinal regions influencing DR severity classification. Several measures are used to evaluate performance, including accuracy, precision, recall, F1-score, Cohen’s Kappa Score (CKS), Weighted Kappa Score (WKS), and Matthews Correlation Coefficient (MCC), ensuring a robust assessment. Among all student models, EfficientNet-B0 achieves the highest classification accuracy of 95.39 %, 95.43 % precision, recall of 95.39 %, F1-score of 95.37 %, CKS of 0.94, WKS of 0.97, MCC of 0.94, AUC of 0.99, and a KD loss of 0.17, with a computational cost of 0.38 G FLOPs. These results demonstrate its effectiveness as an optimized lightweight model for DR detection. The findings emphasize the potential of KD-based lightweight models in attaining high diagnostic accuracy while reducing computational complexity, paving the way for scalable and cost-effective DR screening solutions.
糖尿病视网膜病变（DR）仍然是全球眼部损伤的主要原因，需要开发高效、准确的深度学习模型来进行自动诊断。本研究提出了一种新的框架fastffnet，它利用基于变压器的知识蒸馏（KD）来增强灾难严重性分类，同时降低计算复杂度。所提出的方法采用FastViT-MA26作为教师模型，采用EfficientNet-B0作为学生模型，在准确性和计算效率之间实现了理想的结合。APTOS盲检测数据集包含5个严重级别的3662张图像，通过预处理、归一化、分割和增强来解决类别不平衡问题。教师模型在将其知识传递给学生模型之前要经过培训和验证，使后者能够在保持轻量级体系结构的同时近似教师的表现。为了全面评估拟议框架的有效性，对其他学生模型（包括HGNet、ResNet50、MobileNetV3和deit）进行了分析，以进行比较评估。通过Grad-CAM++可视化增强了模型的可解释性，突出了影响DR严重程度分类的关键视网膜区域。用于评估性能的几个指标，包括准确性、精密度、召回率、f1分数、科恩Kappa分数（CKS）、加权Kappa分数（WKS）和马修斯相关系数（MCC），以确保可靠的评估。在所有学生模型中，effentnet - b0的分类准确率最高，为95.39%，准确率为95.43%，召回率为95.39%，f1评分为95.37%，CKS为0.94，WKS为0.97，MCC为0.94，AUC为0.99，KD损失为0.17，计算成本为0.38 G FLOPs。这些结果证明了它作为一种优化的轻量级DR检测模型的有效性。研究结果强调了基于kd的轻量级模型在实现高诊断准确性的同时降低计算复杂性的潜力，为可扩展和具有成本效益的DR筛选解决方案铺平了道路。
{"title":"Leveraging FastViT based knowledge distillation with EfficientNet-B0 for diabetic retinopathy severity classification","authors":"Jyotirmayee Rautaray , Ali B.M. Ali , Meenakshi Kandpal , Pranati Mishra , Rzgar Farooq Rashid , Farzona Alimova , Mohamed Kallel , Nadia Batool","doi":"10.1016/j.slast.2025.100325","DOIUrl":"10.1016/j.slast.2025.100325","url":null,"abstract":"<div><div>Diabetic retinopathy (DR) remains a key contributor to eye impairment worldwide, requiring the development of efficient and accurate deep learning models for automated diagnosis. This study presents FastEffNet, a novel framework that leverages transformer-based knowledge distillation (KD) to enhance DR severity classification while reducing computational complexity. The proposed approach employs FastViT-MA26 as the teacher model and EfficientNet-B0 as the student model, striking the ideal mix between accuracy and computational efficiency. APTOS blindness detection dataset comprising 3662 images across five severity classes is collected, pre-processed, normalized, split and augmented to address class imbalance. The teacher model undergoes training and validation before transferring its knowledge to the student model, enabling the latter to approximate the teacher’s performance while maintaining a lightweight architecture. To comprehensively assess the efficacy of the proposed framework, additional student models—including HGNet, ResNet50, MobileNetV3, and DeiT—are analysed for comparative assessment. Model interpretability is enhanced through Grad-CAM++ visualizations, which highlight critical retinal regions influencing DR severity classification. Several measures are used to evaluate performance, including accuracy, precision, recall, F1-score, Cohen’s Kappa Score (CKS), Weighted Kappa Score (WKS), and Matthews Correlation Coefficient (MCC), ensuring a robust assessment. Among all student models, EfficientNet-B0 achieves the highest classification accuracy of 95.39 %, 95.43 % precision, recall of 95.39 %, F1-score of 95.37 %, CKS of 0.94, WKS of 0.97, MCC of 0.94, AUC of 0.99, and a KD loss of 0.17, with a computational cost of 0.38 G FLOPs. These results demonstrate its effectiveness as an optimized lightweight model for DR detection. The findings emphasize the potential of KD-based lightweight models in attaining high diagnostic accuracy while reducing computational complexity, paving the way for scalable and cost-effective DR screening solutions.</div></div>","PeriodicalId":54248,"journal":{"name":"SLAS Technology","volume":"33 ","pages":"Article 100325"},"PeriodicalIF":2.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144523640","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-01Epub Date: 2025-06-19DOI: 10.1016/j.slast.2025.100313
Xiao Li, Min Han
The study aimed to explore the potential mechanism of action of extracellular miRNA-188–3p derived from CAFs in cervical cancer. In this study, CAFs were isolated from patients with cervical cancer, and exosomes were extracted by ultrafast centrifugation method to detect the expression level of miRNA-188–3p in exosomes. Subsequently, the exosomes were co-cultured with cervical cancer cells, and the temperature changes of the cells were monitored by medical thermal image analysis technology to evaluate the metabolic activity of the cells. Western blot and qPCR were used to detect protein and mRNA expression levels related to iron metabolism in order to investigate the role of miRNA-188–3p in iron metabolism of cervical cancer cells. The results showed that the expression level of miRNA-188–3p in exosomes derived from CAFs was significantly higher than that of exosomes derived from normal fibroblasts. Medical thermal image analysis showed that cervical cancer cells treated with miRNA-188–3p showed higher metabolic activity, manifested by increased temperature. The results of cell proliferation test, scratch test and Transwell invasion test all showed that miRNA-188–3p promoted the proliferation, migration and invasion of cervical cancer cells. Further molecular mechanism studies showed that miRNA-188–3p regulates iron homeostasis in cervical cancer cells by targeting genes related to iron metabolism, thereby promoting cell proliferation and invasion.
{"title":"Exosomal miRNA-188–3p derived from cancer-associated fibroblasts promotes ferroptosis in cervical cancer: Medical biothermal image analysis","authors":"Xiao Li, Min Han","doi":"10.1016/j.slast.2025.100313","DOIUrl":"10.1016/j.slast.2025.100313","url":null,"abstract":"<div><div>The study aimed to explore the potential mechanism of action of extracellular miRNA-188–3p derived from CAFs in cervical cancer. In this study, CAFs were isolated from patients with cervical cancer, and exosomes were extracted by ultrafast centrifugation method to detect the expression level of miRNA-188–3p in exosomes. Subsequently, the exosomes were co-cultured with cervical cancer cells, and the temperature changes of the cells were monitored by medical thermal image analysis technology to evaluate the metabolic activity of the cells. Western blot and qPCR were used to detect protein and mRNA expression levels related to iron metabolism in order to investigate the role of miRNA-188–3p in iron metabolism of cervical cancer cells. The results showed that the expression level of miRNA-188–3p in exosomes derived from CAFs was significantly higher than that of exosomes derived from normal fibroblasts. Medical thermal image analysis showed that cervical cancer cells treated with miRNA-188–3p showed higher metabolic activity, manifested by increased temperature. The results of cell proliferation test, scratch test and Transwell invasion test all showed that miRNA-188–3p promoted the proliferation, migration and invasion of cervical cancer cells. Further molecular mechanism studies showed that miRNA-188–3p regulates iron homeostasis in cervical cancer cells by targeting genes related to iron metabolism, thereby promoting cell proliferation and invasion.</div></div>","PeriodicalId":54248,"journal":{"name":"SLAS Technology","volume":"33 ","pages":"Article 100313"},"PeriodicalIF":2.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144340588","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号