Journal of Radiation Research and Applied Sciences最新文献

{"title":"Depth determination of simulated biological tissue using X-ray radiography and feature extraction techniques: Evaluation with Bi-LSTM neural network","authors":"Javad Tayebi ,&nbsp;Mohammadreza Rezaie ,&nbsp;Saeedeh Khezripour","doi":"10.1016/j.jrras.2025.101406","DOIUrl":"10.1016/j.jrras.2025.101406","url":null,"abstract":"<div><h3>Purpose</h3><div>Accurate determination of tissue depth in medical imaging and disease diagnosis is crucial, especially for complex biological structures. Traditional methods often lack the necessary precision for effective diagnosis and treatment planning. This study investigates the determination of heart-like tissue depths using X-ray outputs and advanced feature extraction techniques.</div></div><div><h3>Methods</h3><div>Simulated tissues at depths of 5, 10, 15, and 20 cm were analyzed using the Monte Carlo N-Particle Transport Code (MCNPX), with radiographic images captured at 70 keV. Features such as wavelet transform, Gray Level Co-occurrence Matrix (GLCM), Gray Level Run Length Matrix (GLRM), and fractal analysis were extracted. A Bidirectional Long Short-Term Memory (Bi-LSTM) network was used to predict tissue depth, comparing the performance of optimizers including Adam, RMSprop, and SGD.</div></div><div><h3>Results</h3><div>The results showed that the Stochastic Gradient Descent (SGD) optimizer achieved superior prediction accuracy compared to other optimizers. Statistical performance metrics indicated that SGD outperformed its counterparts, showcasing enhanced precision and reliability in predictive modeling of tissue depths. Mean RMSE: 0.21155, Mean MAE: 0.18522, Mean MBE: 0.03400, Mean MRE: 0.01422, Mean MAPE: 0.01422, Mean SMAPE: 0.01429.</div></div><div><h3>Discussion</h3><div>The findings demonstrate the high accuracy of the Bi-LSTM model in predicting tissue depths from radiographic images. This study represents a significant advancement in medical diagnostics, providing an innovative solution to longstanding challenges. By integrating advanced imaging techniques with machine learning algorithms and leveraging MCNPX for precise simulation, more accurate and reliable diagnostic tools can be developed, ultimately improving patient outcomes. Future research could explore clinical applications of this approach and further refine the models for greater accuracy. Accurate depth determination is crucial not only in medical applications, such as optimizing radiation doses in radiotherapy, but also in various industrial contexts, such as non-destructive testing and evaluation.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 2","pages":"Article 101406"},"PeriodicalIF":1.7,"publicationDate":"2025-03-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143579065","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Magnetohydrodynamic flow of carbon nanotubes blood based hybrid nanofluids with the impact of thermal radiation over a permeable surface","authors":"Ali Rehman ,&nbsp;Mustafa Inc ,&nbsp;Dean Chou","doi":"10.1016/j.jrras.2025.101408","DOIUrl":"10.1016/j.jrras.2025.101408","url":null,"abstract":"<div><div>This study looks at the magnetohydrodynamic (MHD) flow properties of blood-based hybrid nanofluids (HNFs) containing carbon nanotubes (CNTs), such as SWCNTs and MWCNTs with multiple walls. Thermal radiation is present during the study. The study is mostly about how these fluids behave on a permeable surface in steady-state laminar flow. To make things easier, boundary layer (BL) approximations are used to simplify and solve the equations for momentum and energy. We transform these equations into a system of nonlinear ODEs via similarity transformations (STs) and solve them semi-numerically. This study looks at how surface permeability, magnetic field (MF) strength, and thermal radiation affect the flow and heat transfer properties of fluids. It does this by looking closely at key parameters like the permeability parameter, the radiation parameter, the power law index, the CSP, the nanoparticle volume fraction (VF), the heat generation, the Eckert number (EN), and the MF strength. The results, which are shown in the form of graphs and a table with the NN and skin friction (SF) coefficients, give us important information about how blood-based (BB) HNFs with CNTs behave in MHD conditions. The moment of HNF particles decreases as the magnetic parameter (MP), CSP, and nanoparticle volume friction all go up. However, this has the opposite effect on the temperature profile as the EN, radiation parameter, heat generation parameter, and nanoparticle volume friction all go up. This study shows how important permeability and thermal radiation are in changing these dynamics. It also helps to create better ways to control temperature in engineering and biomedical settings.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 2","pages":"Article 101408"},"PeriodicalIF":1.7,"publicationDate":"2025-03-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143579116","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Thermally radiative flow of non-Newtonian Rabinowitsch fluid through a permeable artery with multiple stenoses of varying shapes","authors":"S. Khaliq ,&nbsp;M. Younas ,&nbsp;Z. Abbas ,&nbsp;M.A. Aljohani ,&nbsp;Khadijah M. Abualnaja","doi":"10.1016/j.jrras.2025.101400","DOIUrl":"10.1016/j.jrras.2025.101400","url":null,"abstract":"<div><div>Cardiovascular diseases remain one of the leading causes of mortality worldwide, with arterial stenosis playing a crucial role in impairing blood circulation. The presence of non-uniform stenoses in permeable arteries can significantly alter hemodynamics, impacting oxygen transport and increasing the risk of severe complications. Understanding the intricate interplay between thermal radiation, viscous dissipation, and flow constraints is essential for developing effective diagnostic and therapeutic strategies. This study investigates the influence of thermal radiation and viscous dissipation on the Rabinowitsch fluid model to account for the shear-thinning nature of blood in non-uniform inclined stenosed permeable arteries under slip constraints. The research aims to bridge the gap in existing models by incorporating multiple stenosis geometries bell-shaped, W-shaped, and elliptical to provide a more comprehensive analysis of arterial flow behavior. The governing equations are formulated in dimensionless form under mild stenosis assumptions and analytically solved. Computational results reveal that the Brinkman number significantly affects fluid temperature, enhancing dissipation effects across different arterial regions. However, the presence of slip conditions and Darcy number alter this trend, particularly near the arterial walls. The stenosis height exhibits minimal influence on flow dynamics, while the slip parameter plays a dominant role in modifying velocity distributions. The findings of this study have significant biomedical implications, particularly in optimizing stent designs, improving blood flow simulations in diseased arteries, and enhancing medical interventions such as targeted drug delivery and hemodynamic regulation. The results also provide a theoretical foundation for future computational fluid dynamics (CFD) and experimental studies in arterial mechanics, aiding in early disease diagnosis and prevention strategies.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 2","pages":"Article 101400"},"PeriodicalIF":1.7,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143549256","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Assessment of radiobiological impacts of threshold doses of alpha particles on A549 human lung cancer cells following direct irradiation exposure","authors":"Asaad H. Ismail ,&nbsp;Runas Y. Sula ,&nbsp;Ali H. Alomari ,&nbsp;Hoshanc S. Rashid ,&nbsp;Talal A. Alnaemi ,&nbsp;Ibrahim Y. Alshaikhi ,&nbsp;Saeed M. Al-Qahtani ,&nbsp;Simona Mancini ,&nbsp;Michele Guida","doi":"10.1016/j.jrras.2025.101394","DOIUrl":"10.1016/j.jrras.2025.101394","url":null,"abstract":"<div><h3>Purpose</h3><div>Investigation of the radiobiological effects of low-doses of alpha particles on the inhibition of human lung cancer cell growth rate.</div></div><div><h3>Materials and methods</h3><div>To examine whether alpha particles inhibit lung cancer cell growth, human adenocarcinoma A549 cells were irradiated for 10–180 s. The irradiated processes were done during cell culture, using of micro alpha irradiation collimators. The density of accumulated alpha particles is assessed using CR-39NTDs. The MTT cell viability assay was performed on 96-well plates with cell lines seeded at 1 × 10⁴ cells/well to test the cytotoxic effect. The media was removed after 72 h to determine the cytotoxic effect.</div></div><div><h3>Results</h3><div>Growth Rate Inhibition (GRI%) has been investigated with different irradiation times t_IR ranges (very short: 10–50 s, short: 60–180 s), obtaining the following results: GRI% &lt; 10% for the first range, GRI% increased to over 25% at 60 s, while, for the second range, GRI % increased with a power law relationship. Also, a reduced serum-stimulated growth in the lung cancer cell lines has been observed.</div></div><div><h3>Conclusions</h3><div>This study investigated that the low doses (i.e. low densities) of accumulated alpha particles) in human lung cancer cells have a significant impact on the Growth Rate Inhibition (GRI%), depending on the irradiation time, showing that there it occurs a critical time of 60 s at which significant changes could start to take place.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 2","pages":"Article 101394"},"PeriodicalIF":1.7,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143563723","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Applications to medical and failure time data: Using a new extension of the extended exponential model","authors":"Ibrahim E. Ragab ,&nbsp;Mohamed Kayid ,&nbsp;Oluwafemi Samson Balogun ,&nbsp;Tamer S. Helal","doi":"10.1016/j.jrras.2025.101373","DOIUrl":"10.1016/j.jrras.2025.101373","url":null,"abstract":"<div><div>This article introduces a novel extension of the extended exponential model, referred to as the Kavya-Manoharan extended exponential model. The new suggested model is very flexible model because its probability density function and hazard rate function can be right skewed, reversed J-shapes and increasing. Various statistical properties of the Kavya-Manoharan extended exponential model are calculated. Various uncertainty metrics are calculated and analyzed both theoretically and quantitatively. Furthermore, the Kavya-Manoharan extended exponential model utilizes the maximum likelihood estimation technique to determine its two parameters. A comprehensive numerical analysis is conducted to assess the effectiveness of the maximum likelihood estimation technique. The feasibility and importance of the Kavya-Manoharan extended exponential model may be demonstrated by analyzing three actual datasets about medical and failure times data. The Kavya-Manoharan extended exponential model is evaluated against many established statistical models such as; generalized Dinesh-Umesh-Sanjay-exponential, extended exponential, generalized Lindley, Kavya-Manoharan Burr X, generalized power Weibull, and the Gumbel distributions using diverse criteria. The numerical results indicated that the Kavya-Manoharan extended exponential model was more suitable for the data than the other competing models.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 2","pages":"Article 101373"},"PeriodicalIF":1.7,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143549257","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Irreversibility analysis and thermal performance of quadratic radiation and Darcy-Forchheimer flow over non-isothermal needle with velocity slip: Effects of aggregation and non-aggregation dynamics","authors":"Khadija Rafique ,&nbsp;Zafar Mahmood ,&nbsp;Mushtaq Ahmad Ansari ,&nbsp;Abhinav Kumar ,&nbsp;Umar Khan","doi":"10.1016/j.jrras.2025.101395","DOIUrl":"10.1016/j.jrras.2025.101395","url":null,"abstract":"<div><div>The importance of nanofluid flow across thin geometries in thermal engineering, biological applications, and energy systems has led to much research into this topic. Classical boundary layer theories first guided studies on heat transport across needles; subsequently, magnetohydrodynamic (MHD), porous media, radiation effects, and convective heat transfer processes were included. Although recent developments have brought nanoparticles to improve thermal conductivity, their behavior especially under aggregation and non-aggregation conditions remains a subject of much investigation. This study quantitatively analyses the flow, heat transfer, and entropy generation properties of nanofluid flow over a slender needle, accounting for the influences of an inclined magnetic field, quadratic thermal radiation, Darcy-Forchheimer porous media, heat generation/absorption, viscous dissipation, Joule heating, velocity slip, convective heating, and nanoparticle aggregation. By using the bvp4c solver inside MATLAB, the controlling nonlinear boundary layer equations may be resolved. Findings show that the velocity profile grows with velocity slip but shrinks with increasing magnetic field strength, porous medium resistance, Darcy-Forchheimer drag, and inclined angle owing to increased resistive forces. Thermal energy dissipation causes the temperature profile to rise with increasing Biot number (Bi) and Eckert (Ec) numbers. Stronger convective effects in non-aggregation scenarios provide greater entropy creation. Skin friction diminishes with rising magnetic field intensity, slip, porous resistance, and needle thickness; however, it escalates with elevated nanoparticle volume fractions. The Nusselt number increases with the radiation, and temperature ratio, whereas it decreases with the Eckert number and nanoparticle volume percentage. Heat absorption exceeds heat production by 16.51%–28.61% in the non-aggregation model and by 17.30%–29.03% in the aggregation model when Bi grows from 0.2 to 0.8. Heat absorption situations often provide greater heat transfer rates than heat production. The research shows that nanofluid properties, resistance of porous media, and MHD forces all interact intricately to determine thermal and flow parameters. The results provide the groundwork for future technical optimizations of cooling and thermal management systems based on nanofluids.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 2","pages":"Article 101395"},"PeriodicalIF":1.7,"publicationDate":"2025-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143549254","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Robust estimator for estimation of population mean under PPS sampling: Application to radiation data","authors":"Ahmed R. El-Saeed ,&nbsp;Sohaib Ahmad ,&nbsp;Badr Aloraini","doi":"10.1016/j.jrras.2025.101384","DOIUrl":"10.1016/j.jrras.2025.101384","url":null,"abstract":"<div><div>When the population features are same, choosing the units with the help of simple random sampling (SRS). However, in situations while there is a significant variation in unit dimensions, the probability proportional to size (PPS) sampling approach might be implemented. The aim of this work was to bring a new estimator for mean estimation using PPS sampling. The efficiency of the estimators was demonstrated using radiation data, also apply simulation analysis. When comparing the proposed estimator to existing counterparts, the numerical result shows that recommended estimators performs better for estimating population means. Visual representations of the data further prove the validity of the proposed estimator. As a result, we are in favor of utilizing the proposed estimator and believe it can improve outcomes when estimating the populations mean using a PPS sampling method.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 2","pages":"Article 101384"},"PeriodicalIF":1.7,"publicationDate":"2025-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143549255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enhanced prediction of radiation-induced skin toxicity in breast cancer patients using a hybrid dosiomics-clinical model","authors":"Samira Soltani ,&nbsp;Ali Akbar Aliasgharzadeh ,&nbsp;Pedram Fadavi ,&nbsp;Zahra Bagherpour ,&nbsp;Habib Moradi ,&nbsp;Mojtaba Safari ,&nbsp;Manijeh Beigi","doi":"10.1016/j.jrras.2025.101382","DOIUrl":"10.1016/j.jrras.2025.101382","url":null,"abstract":"<div><h3>Objectives</h3><div>This study aims to develop a predictive model for radiation-induced skin toxicity (RIST) in breast cancer patients using dosiomics features extracted from the dose distribution map within the clinical target volume (CTV).</div></div><div><h3>Materials and methods</h3><div>This study included breast cancer patients treated with 3D conformal radiation therapy (3D-CRT). Patients were categorized into low-grade (G0-G1) and high-grade (G2-G3) toxicity groups. Dosiomics features of CTV, clinical data of medical records, and dosimetric parameters of dose maps were extracted. Three predictive models were developed: a dosiomics model using CTV-based features, a hybrid dosiomics-clinical model (HDO), and a hybrid dose-volume histogram-clinical model (HDV). Machine learning algorithms (support vector machines and random forests) were used to build the models and their performances were assessed using the area under the receiver operating characteristic curve (AUC).</div></div><div><h3>Results</h3><div>Thirty-two patients (42%) experienced high-grade RIST (CTCAE grade ≥2) following breast radiation therapy (RT). The HDO model demonstrated superior predictive performance, attaining an AUC of 0.78, significantly higher than the HDV and single predictive models. In the dosiomics-based features group, the major axis length from the shape class is one of the most relevant features for skin toxicity (grade ≥2). In the clinical parameters group, chemo regimen, receptor state, and hormonal treatment showed significant correlation with skin toxicity (p-value&lt;0.05). In the DVH factors group V105 cc, V 110%, V107%, and Breast CTV revealed a significant correlation with skin toxicity.</div></div><div><h3>Conclusion</h3><div>The developed predictive model utilizing dosiomics features demonstrated superior performance compared to dose volume histogram (DVH) based methods, with an AUC of 0.78, leading to early prediction of skin toxicity among breast cancer patients who had received RT. Moreover, our results suggest that the integration of dosiomics features with clinical parameters significantly improves the predictive power of models.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 2","pages":"Article 101382"},"PeriodicalIF":1.7,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143534301","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"GDP growth drivers in Saudi Arabia based on machine learning algorithms","authors":"Mohamed F. Abd El-Aal ,&nbsp;Mansour Shrahili ,&nbsp;Mohamed Kayid ,&nbsp;Shahid Mohammad","doi":"10.1016/j.jrras.2025.101380","DOIUrl":"10.1016/j.jrras.2025.101380","url":null,"abstract":"<div><div>This study utilizes machine-learning algorithms to investigate the economic sectors that most significantly influence Saudi Arabia's economic growth rate, focusing on agriculture, industry, and services. The analysis shows that the random forest algorithm offers the highest predictive accuracy in identifying the key sectors driving economic growth. The research findings show that the service and industrial sectors account for 39.3% and 37.7% of Saudi Arabia's GDP growth, respectively. These results show that this country is moving significantly toward diversifying its economy as it depends more and more on non-oil sectors for growth. Even while the agricultural industry presently makes up a lower 23% of the total GDP, its comparison small share does not limit its potential for expansion. The paper emphasizes how agricultural developments, such as enhanced technologies and more efficient methods, could increase economic impact. The agricultural sector has the potential to play a significant role in boosting future economic growth, which would further help Saudi Arabia's objectives for sustainable growth and diversification.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 2","pages":"Article 101380"},"PeriodicalIF":1.7,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143548639","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-modal data integration of dosiomics, radiomics, deep features, and clinical data for radiation-induced lung damage prediction in breast cancer patients","authors":"Yan Li ,&nbsp;Jun Jiang ,&nbsp;Xuyi Li ,&nbsp;Mei Zhang","doi":"10.1016/j.jrras.2025.101389","DOIUrl":"10.1016/j.jrras.2025.101389","url":null,"abstract":"<div><h3>Objective</h3><div>Radiation-induced lung damage (RILD) is a critical complication in breast cancer patients undergoing radiotherapy. This study proposes a multi-modal predictive framework integrating dosiomics, radiomics, deep learning-based features, and clinical data to enhance early detection and risk stratification of Grade ≥2 RILD, ultimately supporting personalized radiotherapy planning.</div></div><div><h3>Materials and methods</h3><div>A dataset of 450 breast cancer patients receiving radiotherapy was analyzed, incorporating high-resolution CT scans, 3D spatial dose distributions, and comprehensive clinical parameters such as age, BMI, tumor laterality, chemotherapy regimens, and comorbidities. Imaging data were standardized through voxel resampling and intensity normalization, and features were extracted from both radiomics (215 features) and dosiomics. Mutual Information (MI)-based feature selection was applied to enhance model performance, while a 3D autoencoder with attention mechanisms was utilized to capture spatial and structural patterns linked to RILD. Five-fold cross-validation was performed to ensure robustness.</div></div><div><h3>Results</h3><div>The Intraclass Correlation Coefficient (ICC) analysis identified the most reproducible radiomics features, leading to significant feature reduction while maintaining predictive stability. Multi-modal data integration significantly improved classification performance, with the Voting Classifier achieving 95.89% accuracy and 96.98% sensitivity when using MI-based feature selection. Deep features demonstrated superior predictive power compared to standalone dosimetric data. The 3D autoencoder model with attention mechanisms further enhanced predictive accuracy, achieving 95% accuracy, 0.96 AUC, and 0.93 sensitivity.</div></div><div><h3>Conclusion</h3><div>The proposed multi-modal AI-driven approach effectively predicts Grade ≥2 RILD, addressing limitations of traditional dose-volume metrics. The integration of radiomics, dosiomics, deep learning, and clinical data enhances model accuracy and interpretability, paving the way for personalized risk assessment and optimized radiotherapy planning. Future research should focus on external validation and real-time clinical implementation to further refine predictive capabilities.</div></div>","PeriodicalId":16920,"journal":{"name":"Journal of Radiation Research and Applied Sciences","volume":"18 2","pages":"Article 101389"},"PeriodicalIF":1.7,"publicationDate":"2025-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143534302","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}