Journal of Radiation Research and Applied Sciences最新文献

{"title":"Controlling thermal energy through radiation, heat source/sink and thermophoretic diffusion for stagnation point-Williamson rheological model having suction/injection, inclined magnetic field and porous medium","authors":"Muhammad Yousaf ,&nbsp;Salman Zeb ,&nbsp;Rai Sajjad Saif ,&nbsp;Sania Nazeer ,&nbsp;Subhan Ali ,&nbsp;Muhammad Imran Malik ,&nbsp;Saeed Afzal ,&nbsp;Taseer Muhammad","doi":"10.1016/j.jrras.2025.101854","DOIUrl":"10.1016/j.jrras.2025.101854","url":null,"abstract":"<div><div>Extensive research into next-generation thermal energy technologies has been prompted by the growing global demand for sustainable and high-performance energy systems. The thermophysical characteristics of the working fluid, particularly its capacity for heat transmission, have a major role in the efficiency of sensible heat storage systems. However, the efficiency of thermal transmission in conventional heat transfer fluids is sometimes limited. In the present work, we investigated the stagnation point flow of a Williamson fluid under the effects of the suction/injection, inclined magnetic field, and thermophoretic diffusion along a nonlinear stretchable surface in a porous space, and with the presence of thermal radiation, heat source/sink and chemical reaction influences. The governing partial differential equations (PDEs) of the Williamson fluid flow model are transformed into nonlinear ordinary differential equations (ODEs) via similarity transformations. We obtained the results numerically solving the non-linear ODEs which characterize the behavior of flow profiles and of the physical quantities against the governing parameters. The fluid velocity increases against the velocity ratio parameter while it declines for the angle of inclination and suction/injection parameter. Heat source/sink, thermal radiation, and angle of inclination parameters increase the temperature field. The concentration profile declines for the chemical reaction parameter and Schmidt number, while it increases for the thermophoretic parameter.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 4","pages":"Article 101854"},"PeriodicalIF":2.5,"publicationDate":"2025-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144860341","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Automated multi-class classification of thyroid nodules in ultrasound imaging using transformer-based segmentation and hybrid feature learning","authors":"Mingshuang Fang ,&nbsp;Qingfeng Ma ,&nbsp;Binxiong Xu","doi":"10.1016/j.jrras.2025.101782","DOIUrl":"10.1016/j.jrras.2025.101782","url":null,"abstract":"<div><h3>Objective</h3><div>To develop and validate an end-to-end machine learning pipeline for automated multi-class classification of thyroid nodules in ultrasound imaging, using transformer-based segmentation and hybrid radiomic-deep feature integration to enhance clinical accuracy and reproducibility.</div></div><div><h3>Materials and methods</h3><div>In this multi-center study, 2654 ultrasound cases from five hospitals were used for model development, and 873 additional cases from an independent center were used for external validation. Thyroid nodules were segmented using four architectures: UNETR, nnU-Net, Swin-UNet, and UNet. From the segmented regions, handcrafted radiomic features and deep features extracted via Vision Transformer encoder layers were obtained. Features were filtered using ICC ≥0.75, followed by variance and correlation-based refinement. Three feature selection methods—Lasso, PCA, and Mutual Information—were evaluated. Classification was performed using XGBoost, Random Forest, and TabTransformer across six TI-RADS categories. Five-fold stratified cross-validation and external testing ensured robustness. Segmentation was assessed using Dice, Jaccard, and Hausdorff metrics; classification performance was evaluated via accuracy, AUC, and recall.</div></div><div><h3>Results</h3><div>UNETR achieved the highest segmentation performance and enabled the most accurate classification. The best outcome was observed with radiomic features selected via Lasso and classified with XGBoost (external accuracy: 93.0 %, AUC: 93.6 %, recall: 92.0 %). Deep features showed comparable results (accuracy: 92.8 %). Q-value analysis confirmed statistical superiority of the best-performing models. Differences across segmentation models significantly impacted classification performance, highlighting the importance of boundary quality. All models demonstrated strong generalizability and minimal overfitting.</div></div><div><h3>Conclusions</h3><div>The study demonstrates the feasibility and clinical value of a fully automated pipeline for TI-RADS-based thyroid nodule classification. The proposed framework is generalizable, interpretable, and suitable for integration into real-time diagnostic systems.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 4","pages":"Article 101782"},"PeriodicalIF":2.5,"publicationDate":"2025-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144860342","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Cardiovascular Magnetic Resonance imaging analysis using neural networks","authors":"Linyou Wang ,&nbsp;Lingfei Wang ,&nbsp;Ping Wu ,&nbsp;Li Ding","doi":"10.1016/j.jrras.2025.101874","DOIUrl":"10.1016/j.jrras.2025.101874","url":null,"abstract":"<div><div>Cardiovascular Magnetic Resonance (CMR) provides high-resolution images for diagnosing heart diseases, but interpreting these images is time-consuming and requires expert skill. Recently, deep neural networks have shown potential to automate and standardize CMR analysis, performing tasks such as image sorting, cardiac chamber segmentation, and tissue characterization with speed and precision. This review summarizes the theoretical foundations of neural networks applied to CMR, evaluates state-of-the-art deep learning architectures (e.g., convolutional and recurrent networks, U-Net variants) for various CMR tasks, and highlights key performance metrics and validation strategies. We incorporate a comprehensive survey of recent literature and discuss critical challenges such as the curse of dimensionality, limited annotated datasets, and model generalizability across imaging centers. We also address issues of bias and fairness in algorithms, the impact of data variability and annotation quality, and failure modes in deployment. In addition, we examine regulatory and practical considerations for clinical integration—covering FDA/CE approvals, patient safety, economic factors, and workflow integration—and outline future directions including hybrid physics-informed approaches and deployment science. The review is focused on a core clinical application (automated CMR image segmentation and classification) to provide depth of analysis. Our findings indicate that while neural networks can markedly improve the efficiency and reproducibility of CMR analysis, careful attention to technical and clinical validation is required to translate these advances into safe, effective real-world tools.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 4","pages":"Article 101874"},"PeriodicalIF":2.5,"publicationDate":"2025-08-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144852403","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Third‐Grade fluid flow over a stretching sheet under Lorentz force and thermal radiation: Parametric continuation algorithm","authors":"Fisal Asiri ,&nbsp;Mohd Aamir Mumtaz ,&nbsp;Naveed sarwar Abbasi ,&nbsp;Barun Haldar ,&nbsp;Temur Eshchanov ,&nbsp;Sabirov Sardor","doi":"10.1016/j.jrras.2025.101880","DOIUrl":"10.1016/j.jrras.2025.101880","url":null,"abstract":"<div><div>In the current analysis, the energy and mass transmission subject to Lorentz force and thermal radiation through the third‐grade fluid across a stretching inclined sheet is considered. The thermal radiation, heat source, and magnetic impact are also applied to the fluid flow. The flow equations are reformulated into the non-dimensional form of ODEs using the similarity transformations. The lowest order of ODEs is solved through the PCM (parametric continuation method) by using Matlab software. For the validity of the results, the outcomes are compared to published studies. The relative percent error between the present calculation and the published is 0.00537 % at Pr = 10 (Prandtl number), which reveals that the present results are accurate. The nature of the flow constraints on the physical interest quantities, skin friction, velocity, and energy fields are displayed via Figures. For the validity of the results, for particular cases, the results are compared to the published study. It has been observed that the velocity field enhances for the intensifying values of the third-grade fluid parameter. The effect of the magnetic term and the permeability factor reduces the fluid velocity. The radiation effect and the heat source elevate the temperature field. The energy transfer rate enhances up to 15.46340 % by raising the buoyancy impact from 0.3 to 0.9, whereas the flow rate declines up to −12.4772 %.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 4","pages":"Article 101880"},"PeriodicalIF":2.5,"publicationDate":"2025-08-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144852368","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Establishment of diagnostic reference levels based on administered radiopharmaceutical activity for adult and pediatric nuclear medicine procedures in Taif city, Saudi Arabia","authors":"Hamid Osman ,&nbsp;Raghad Alharthi ,&nbsp;Yara Alghamdi ,&nbsp;Rana Aljuaid ,&nbsp;Ghalaa Alosaimi ,&nbsp;Saleh Althobaiti ,&nbsp;Sahal Alotaibi ,&nbsp;Abdelmoneim Sulieman ,&nbsp;Kawthar Moh. Sharif Abdulrhman ,&nbsp;Mayeen Uddin Khandaker","doi":"10.1016/j.jrras.2025.101882","DOIUrl":"10.1016/j.jrras.2025.101882","url":null,"abstract":"<div><h3>Background</h3><div>Nuclear medicine procedures are widely used in diagnostic imaging due to their ability to provide functional information about organs and tissues. However, these procedures involve exposure to ionizing radiation, which raises concerns about patient safety, particularly for vulnerable populations like children. Optimizing radiation doses has become essential to minimize unnecessary exposure while maintaining image quality.</div></div><div><h3>Objective</h3><div>This study aims to assess patient radiation doses (administered activities), propose local diagnostic reference level (DRLs) to both adult and paediatric patients in Taif, Saudi Arabia and estimate effective radiation doses (ED), associated with commonly performed nuclear medicine procedures.</div></div><div><h3>Methodology</h3><div>Data were collected from 277 adult and 60 pediatric patients who underwent nuclear medicine procedures at three NM centers in Taif City. Patient demographics (age, gender, weight, and height when available) and procedure-specific information were obtained from electronic medical records (EMR) and radiological information system (RIS). Administered radiopharmaceutical activities were the primary dose metric for DRL establishment which then converted to effective doses (mSv) using standard conversion factors from ICRP Publication 128. Descriptive analysis was conducted to calculate mean and median doses and propose local DRLs.</div></div><div><h3>Results</h3><div>The study found that adult procedures such as bone (HDP and MDP) and thyroid scans had mean effective doses of 3.54 mSv, 3.71 mSv, and 3.46 mSv, respectively, all within international standards. Paediatric renal MAG3 procedures showed wider variation depending on age, ranging from 0.35 mSv (1-year-old) to 2.5 mSv (10-year-old), while the DRL for aforementioned procedure was 740.770 and 185 MBq respectively, for adult and for paediatric renal MAG3 showed 40–185 MBq depending on age<strong>. Conclusion</strong>: Most procedures were within global safety benchmarks. However, elevated paediatric doses highlight the need for local DRLs and ongoing dose optimization. Implementing region-specific DRLs will improve protocol consistency, enhance safety, and support better-quality nuclear medicine services.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 4","pages":"Article 101882"},"PeriodicalIF":2.5,"publicationDate":"2025-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144841770","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MiR-409-3p/Rab10/BRCA1 axis inhibits malignant transformation of endometrial carcinoma","authors":"Yibing Fan ,&nbsp;Jiayu Wen ,&nbsp;Li Shi,&nbsp;Yingqing Dai","doi":"10.1016/j.jrras.2025.101877","DOIUrl":"10.1016/j.jrras.2025.101877","url":null,"abstract":"<div><h3>Background</h3><div>Current research suggests that miR-409-3p plays the role of a tumor suppressor by regulating key oncogenes, but its mechanism in endometrial carcinoma (EC) remains unclear.</div></div><div><h3>Objectives</h3><div>This study evaluated how miR-409-3p functions in inhibiting the malignant transformation of EC by targeting ras-related protein Rab-1 (<em>Rab10)</em> and exploring its impact on breast cancer susceptibility gene 1 (BRCA1) expression.</div></div><div><h3>Materials and methods</h3><div>Experiments were performed in RL95-2 and HEC-1B EC cell lines and normal endometrial stromal cells (ESCs). miR-409-3p expression was manipulated using lentiviral vectors. CCK-8 assays assessed cell proliferation, Transwell assays measured invasion, real-time PCR and Western blot analyzed gene expression. Flow cytometry evaluated cell cycle changes. Malondialdehyde (MDA) and superoxide dismutase (SOD) assays were used to assess oxidative stress levels.</div></div><div><h3>Results</h3><div>MiR-409-3p upregulation reduced <em>Rab10</em> expression by 2.5-fold (P &lt; 0.001) and increased BRCA1 levels by 1.8-fold (P &lt; 0.01), inhibiting EC cell expansion and invasion (P &lt; 0.05). Further, BRCA1 upregulation was reversed by <em>Rab10</em> overexpression (P &lt; 0.05). BRCA1 upregulation induced cell cycle arrest and alleviated oxidative stress, suggesting a protective role (P &lt; 0.05). Knockdown of BRCA1 reversed the anticancer impact of miR-409-3p/<em>Rab10</em> signaling (P &lt; 0.05), confirming its critical role in the pathway.</div></div><div><h3>Conclusion</h3><div>This study is the first to demonstrate that miR-409-3p suppresses EC malignancy via Rab10-mediated BRCA1 upregulation, revealing a novel therapeutic axis for EC. These findings highlight miR-409-3p as a potential therapeutic target and implications for future research and clinical applications in gynecological cancer treatment.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 4","pages":"Article 101877"},"PeriodicalIF":2.5,"publicationDate":"2025-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144852366","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Roles of ZnO and La2O3 in enhancing the mechanical, optical, and radiation shielding properties of B2O3-Bi2O3-Na2O transparent glasses","authors":"M.I. Sayyed ,&nbsp;Awatif Alshamari ,&nbsp;M.H.A. Mhareb","doi":"10.1016/j.jrras.2025.101859","DOIUrl":"10.1016/j.jrras.2025.101859","url":null,"abstract":"<div><div>This work aims to propose three glass systems composed of B₂O₃-Bi₂O₃-Na₂O-ZnO-La₂O₃ for potential radiation shielding purposes. These glasses gave the following codes: 14Zn1La, 16Zn2La, and 18Zn3La. The effect of changing ZnO and La₂O₃ contents on the mechanical, optical, and gamma radiation shielding properties of the glasses was investigated, as the components were selected for their high density. Makishima and Mackenzie's principles determined the mechanical qualities, while the optical properties were determined using the Mott and Davis relation and absorption spectra. There was a slight variation in the mechanical and optical characteristics with the addition of La₂O₃ and ZnO instead of B₂O₃. For 14Zn1La and 18Zn3La, there was a 3.302–3.332 eV increase in energy. Meanwhile, for 14Zn1La and 18Zn3La, the packing density values were in the range of 0.625 to 0.615. The Phy-X program was used to determine the shielding properties, which showed promising results with the replacement of B₂O₃ by La₂O₃ and ZnO. Including ZnO and La₂O₃ improves the glasses' radiation shielding effectiveness, as indicated by the effective atomic number values, which follow the pattern 14Zn1La &lt;16Zn2La &lt;18Zn3La. Ultimately, it can be said that adding La₂O₃ and ZnO improves the shielding qualities of existing glasses while decreasing the stability of the glass.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 4","pages":"Article 101859"},"PeriodicalIF":2.5,"publicationDate":"2025-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144841768","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Intrapartum ultrasound monitoring in second-stage Labor: Impact on delivery outcomes","authors":"Xiaoying Wang,&nbsp;Fengmei Xiao,&nbsp;Jun Li","doi":"10.1016/j.jrras.2025.101846","DOIUrl":"10.1016/j.jrras.2025.101846","url":null,"abstract":"<div><h3>Objective</h3><div>This study compared the impact of dynamic ultrasound monitoring versus digital examination on second-stage labor (SSL) delivery outcomes, with the primary outcome being cesarean delivery rate and secondary outcomes including perineal lacerations, postpartum hemorrhage, and neonatal complications, providing scientific evidence to optimize obstetric decision-making in the delivery room.</div></div><div><h3>Methods</h3><div>A semi-randomized trial included 60 pregnant women with slow second stage of labor (SSL) at hospital (June 2023–December 2024), divided into an ultrasound group (observation group, n = 30) and a digital examination group (control group, n = 30). The ultrasound group used fetal position, angle of progression (AOP), and head-perineum distance (HPD) to guide delivery, while the control group used traditional vaginal exams. Outcomes (delivery mode, hemorrhage, lacerations, infections, maternal/neonatal complications, injuries, and asphyxia) were analyzed. Logistic regression assessed AOP and HPD in predicting successful assisted vaginal delivery.</div></div><div><h3>Results</h3><div>The observation group had a significantly lower cesarean section (CS) rate (23.3 % <em>vs.</em> 40.0 %, <em>P</em> &lt; 0.05), fewer perineal lacerations (16.7 % <em>vs.</em> 30.0 %, <em>P</em> &lt; 0.05), and less postpartum blood loss (210.5 ± 50.3 <em>vs.</em> 280.2 ± 65.8 mL, <em>P</em> &lt; 0.05). Neonatal outcomes (infections, injuries, asphyxia) were similar (<em>P</em> &gt; 0.05). The observation group had higher assisted vaginal delivery success (86.7 % <em>vs.</em> 66.7 %, <em>P</em> &lt; 0.05) and lower CS rates (13.3 % <em>vs.</em> 33.3 %). Multivariate analysis showed AOP &gt;125° (OR = 6.32, <em>P</em> = 0.014) and HPD ≤2.5 cm (OR = 8.17, <em>P</em> = 0.003) independently predicted successful delivery. AOP (AUC = 0.89, cutoff = 125°) and HPD (AUC = 0.92, cutoff = 2.7 cm) were strong predictors.</div></div><div><h3>Conclusion</h3><div>AOP and HPD are crucial ultrasound markers for predicting successful vaginal assisted delivery in SSL. Combining these parameters (AOP &gt;125° and HPD ≤2.7 cm) improves decision-making accuracy. Dynamic intrapartum ultrasound monitoring, compared to digital examination, optimizes delivery decisions, reduces unnecessary CS and perineal trauma, and improves outcomes. This evidence-based approach holds strong clinical value and should be integrated into practice.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 4","pages":"Article 101846"},"PeriodicalIF":2.5,"publicationDate":"2025-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144841767","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Radiological risks, mineralogical features and Petrological Aspects of some Neoproterozoic Basement Rocks","authors":"Gharam A. Alharshan ,&nbsp;Reda Elsaman ,&nbsp;A. Tolba ,&nbsp;A. Betiha ,&nbsp;Asmae Mimouni ,&nbsp;MAM Uosif","doi":"10.1016/j.jrras.2025.101813","DOIUrl":"10.1016/j.jrras.2025.101813","url":null,"abstract":"<div><div>In continuation of our program to create a radiation map of various environments and geological regions, the Wadi Mubarak region was chosen for this purpose, which is located in North Marsa Allam and South El Quser bounded by latitudes 25° 28.386′ N &amp; 25° 20.493′ N and longitudes 34° 37.084′ E &amp; 34° 36.387′ E, covering an area of about 440 km². The geological context of the study region is primarily informed by field geology. The predominant rock formations in this area consist of exposed ophiolite and related rocks, metavolcanics, gabbro, tonalite, Hammam granodiorite, and monzogranite. Numerous valleys filled with Quaternary sediments intersect the outcrops of various rock units. The region is traversed by Wadi Mubarak, which runs approximately NE-SW, and Wadi Kadabora, which flows in a N-S direction. According to the geological and sampling maps, the rock units are arranged chronologically from old to youngest. An analysis of radionuclides (⁴⁰K, ²³²Th, and ²²⁶Ra) was conducted on 36 samples of Egyptian Neoproterozoic rocks collected from Wadi Mubarak, utilizing gamma spectrometry with a NaI(Tl) 3” × 3” detector. The measured radioactivity concentrations in the rocks are as follows (in Bq kg⁻¹): ²²⁶Ra: 8 to 34, ²³²Th: 5 to 35 and ⁴⁰K: 134 to 374 Furthermore, an assessment of the radiological risks associated with the examined samples indicates that these measurements are below the global average. Consequently, it can be concluded that the living conditions for the local population are secure.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 4","pages":"Article 101813"},"PeriodicalIF":2.5,"publicationDate":"2025-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144829532","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"High-precision brain tumor classification from MRI images using an advanced hybrid deep learning method with minimal radiation exposure","authors":"Rahim Khan ,&nbsp;Sher Taj ,&nbsp;Zahid Ullah Khan ,&nbsp;Sajid Ullah Khan ,&nbsp;Javed Khan ,&nbsp;Tahir Arshad ,&nbsp;Sarra Ayouni","doi":"10.1016/j.jrras.2025.101858","DOIUrl":"10.1016/j.jrras.2025.101858","url":null,"abstract":"<div><h3>Background</h3><div>Accurate identification of brain tumors is critical to improving patient outcomes and minimizing unnecessary radiation exposure from imaging procedures. While Magnetic Resonance Imaging (MRI) is the gold standard for brain tumor detection, manual interpretation remains time-consuming, error-prone, and subject to inter-observer variability.</div></div><div><h3>Objective</h3><div>This study aims to develop a high-precision, automated MRI-based brain tumor classification model using a hybrid deep learning architecture to reduce diagnostic errors and support radiation exposure minimization strategies.</div></div><div><h3>Methods</h3><div>A novel hybrid deep learning model was developed by integrating the CE-EEN-B0 and ResGANet architectures. The model incorporates advanced feature selection and ensemble-based learning techniques to enhance classification performance across diverse datasets. The feature vectors extracted from MRI images were benchmarked against state-of-the-art (SOTA) deep learning classifiers, including InceptionV3, Vision Transformer, MobileNet, VGG-SCNet, DenseNet121, and ResNet50.</div></div><div><h3>Results</h3><div>The proposed hybrid model achieved an accuracy of 99.11 %, with a precision, recall, and F1-Score of 99.6 %. It also attained a specificity of 99.75 %, an error rate of just 0.01, and a Cohen's Kappa score of 99.10, outperforming all benchmark models.</div></div><div><h3>Conclusion</h3><div>The hybrid CE-EEN-B0-ResGANet model demonstrates high reliability and performance in MRI-based brain tumor classification. Its strong diagnostic metrics support its potential for clinical deployment as an effective, automated tool for aiding radiologists and minimizing unnecessary imaging interventions.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 4","pages":"Article 101858"},"PeriodicalIF":2.5,"publicationDate":"2025-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144829531","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}