Search Results (810)

Cone beam computed tomography (CBCT) allows for rapid and accessible acquisition of three-dimensional images with a lower radiation dose compared to conventional computed tomography (CT) scans. However, the quality of CBCT images is limited by a variety of artifacts. This systematic review attempts to explore different artificial intelligence-based solutions for enhancing the quality of CBCT scans and reducing different types of artifacts in these three-dimensional images. PubMed, Web of Science, Scopus, Embase, Cochrane, and Google Scholar were searched up to March 2025. Risk of bias of included studies was assessed using the QUADAS-II tool. Extracted data included bibliographic information, aim, imaging modality, anatomical site of interest, artificial intelligence modeling approach and details, data and dataset details, qualitative and quantitative performance metrics, and main findings. A total of 27 papers from 2018 to 2025 were included. These studies focused on five areas: metal artifact reduction, scatter correction, image reconstruction improvement, motion artifact reduction, and noise reduction. Artificial intelligence models mainly used U-Net variants, though hybrid and transformer-based models were also explored. The thoracic region was the most analyzed, and the structural similarity index measure and peak signal-to-noise-ratio were common performance metrics. Data availability was limited, with only 26% of studies providing public access and 15% sharing model source codes. Artificial intelligence-driven approaches have demonstrated promising results for CBCT artifact reduction. This review highlights a wide variability in performance assessments and that most studies have not received diagnostic validation, limiting conclusions on the true clinical impact of these artificial intelligence-based improvements. Full article

Unlike high-dose scans, low-dose cardiac CT perfusion imaging reduces patient radiation exposure and thereby the risk of potential health effects. However, it introduces significant image noise, degrading diagnostic quality and limiting clinical assessment. Denoising is thus a critical preprocessing step to enhance image quality without compromising anatomical or perfusion details. Traditionally used reconstruction-domain methods, such as Iterative Reconstruction and Compressed Sensing, are often limited by algorithmic complexity, dependence on raw sinogram data, and restricted adaptability. Conversely, image-domain methods offer more adaptable denoising options. Recently, learning-based approaches have further expanded this flexibility and demonstrated state-of-the-art performance across various denoising tasks. In this work, we present a deep learning-based denoising method specifically tuned for low-dose cardiac CT perfusion imaging. Our model is trained to reduce noise while preserving structural integrity and temporal contrast dynamics, which are critical for downstream analysis. Unlike many existing methods, our approach is optimized for perfusion data, where temporal consistency is essential. Residual cardiac motion remains a separate challenge, which we aim to address in our future work. Experimental results show significant improvements in quantitative image quality, using both reference-based and no-reference metrics, such as MSE/PSNR/SSIM and NIQE/FID/KID, as well as improved accuracy of perfusion measurements. Full article

Purpose: We conducted the largest multinational review to date evaluating outcomes following radiotherapy for non-small cell lung carcinoma (NSCLC) patients with oligoprogressive disease (OPD) on immune checkpoint inhibitors (ICIs). Methods: Patients with NSCLC irradiated to ≤5 progressive lesions while receiving ICIs between 2010 and 2023 were identified. We evaluated predictors of local control (LC), progression-free survival (PFS), and overall survival (OS). Patient demographics, disease characteristics, and survival were analyzed using the Wilcoxon test, Kaplan-Meier methods, and uni-/multivariate Cox models. Results: Out of 1178 treated patients, 103 eligible ones were included. The median OPD lesion was 1; the most common site was the lung (n = 33). The median LC of irradiated OPD lesions was not reached. Median PFS and OS were 6.90 (5.75–12.91) and 23.46 (17.54–37.16) months, respectively. Patient demographics, tumor pathological factors, number of OPD lesions, cumulative tumor volume, radiation modality, and OPD response to prior ICIs before radiation were not associated with these three outcomes. However, LC was associated with intermediate/high radiation doses (p = 0.005) and local response to radiation (p = 0.007). Improved PFS was associated with visceral OPD sites following radiation (p = 0.01). A favorable OS was associated with intermediate/high radiation doses (p = 0.01), local response to radiation (p = 0.006), and duration of last ICI before OPD (p = 0.03). Conclusions: Promising outcomes were observed with ICI and radiation for visceral OPD at intermediate/high doses. Prolonged ICI use before OPD and local response to radiotherapy improved survival. These data can contribute towards guidance of multidisciplinary clinical decision-making for managing OPD in NSCLC patients receiving ICIs. Full article

Mediastinal radiotherapy plays a central role in the treatment of several malignancies, particularly Hodgkin lymphoma and breast cancer. However, exposure to thoracic radiation is associated with long-term cardiovascular complications, among which valvular heart disease (VHD) is increasingly recognized. Radiation-induced VHD typically presents after a latency period of 10–20 years and is characterized by progressive valve fibrosis, thickening, and calcification, most commonly affecting the left-sided valves. Management of radiation-induced VHD generally follows standard guidelines but remains challenging due to extensive calcification and coexisting radiation-related cardiac or pulmonary injury. A history of thoracic radiotherapy is associated with increased perioperative risk and may negatively impact surgical outcomes, which often alters the risk–benefit balance and favors less invasive therapeutic approaches. Advances in the transcatheter approach have expanded treatment options for this high-risk population; however, data on long-term outcomes remain limited. Evolving dose-reduction techniques, such as deep-inspiration breath-hold, intensity-modulated radiotherapy, and proton therapy, together with predictive dosimetric models, aim to minimize future cardiac toxicity. Given the delayed onset and progressive nature of radiation-associated VHD, structured long-term surveillance is essential to enable early detection and timely intervention in cancer survivors at risk. Full article

A Monte Carlo-based model of the CyberKnife M6 6 MV Flattening Filter-Free (FFF) beam was developed to produce the data that can be used to train artificial neural networks. The data include the energy spectra of the beam, its average energy, the spatial distributions of the beam, and the distributions of the photon propagation directions for two selected radiation fields—a large one with a diameter of 60 mm, and a small one with a diameter of 15 mm. The GEANT4 code was used to develop the beam model. The developed model was verified by comparing the depth-dose distributions along the beam axis and the profiles obtained in both simulations and measurements. The data included in this paper, intended for training neural networks, will be made available via Google Drive. Full article

This paper focuses on FinFET transistors. The degradation characteristics of FinFET devices after total ionizing dose (TID) radiation in low-temperature environments were investigated by means of a combination of experiments and TCAD simulations. By analyzing the electronic properties of radiation-induced defects in FinFET transistors under low-temperature conditions, the formation and evolution mechanisms of these defects are studied. A physical model for the low-temperature total dose effects of FinFET transistors is established, providing support for the radiation hardening and space applications of FinFET devices. Full article

Background/Objectives: The breast is a highly radiosensitive organ that is directly exposed to ionizing radiation during chest computed tomography (CT) examinations. Excessive radiation exposure increases the risk of radiation-induced malignancies, highlighting the importance of accurate and patient-specific dose estimation. This study aims to estimate the effective radiation dose absorbed by the breast during chest CT examinations using a machine learning (ML)-based personalized prediction approach. Methods: In this retrospective study, a total of 653 female patients who underwent both mammography and chest CT between 2020 and 2024 were included. A structured database was created incorporating demographic and anatomical parameters, including body weight, height, body mass index (BMI), and breast thickness (mm) obtained from mammography, along with dose length product (DLP) values from chest CT scans. Five regression-based ML algorithms—CatBoost, Gradient Boosting, Extra Trees, AdaBoost, and Random Forest—were implemented to predict breast radiation dose. Model performance was evaluated using Mean Squared Error (MSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and the Coefficient of Determination (R²). Results: Among the evaluated models, the CatBoost algorithm optimized with Particle Swarm Optimization (CatBoostPSO) achieved the best overall predictive performance, yielding the lowest MSE (0.3795), MAE (0.3846), and MAPE (4.37%), along with the highest R² value (0.9875). CatBoost and Gradient Boosting models demonstrated predictions most closely aligned with ground truth values, indicating that ensemble-based and dynamically optimized models are particularly effective for breast dose estimation. Conclusions: The proposed machine learning framework enables rapid, accurate, and clinically applicable estimation of breast radiation dose during chest CT examinations. This patient-specific approach has strong potential to support personalized radiation dose monitoring and optimization strategies, contributing to improved radiation safety in clinical practice. Full article

Low-dose computed tomography (LDCT) has become a widely adopted protocol to reduce radiation exposure during clinical imaging. However, dose reduction inevitably amplifies noise and artifacts, compromising image quality and diagnostic confidence. To address this challenge, this study introduces De-TransGAN, a transformer-augmented Generative Adversarial Network specifically designed for high-fidelity LDCT image denoising. Unlike conventional CNN-based denoising models, De-TransGAN combines convolutional layers with transformer blocks to jointly capture local texture details and long-range anatomical dependencies. To further guide the network toward diagnostically critical structures, we embed channel–spatial attention modules based on the Convolutional Block Attention Module (CBAM). On the discriminator side, a hybrid design integrating PatchGAN and vision transformer (ViT) components enhances both fine-grained texture discrimination and global structural consistency. Training stability is achieved using the Wasserstein GAN with Gradient Penalty (WGAN-GP), while a composite objective function—L1 loss, SSIM loss, and VGG perceptual loss—ensures pixel-level fidelity, structural similarity, and perceptual realism. De-TransGAN was trained on the TCIA LDCT and Projection Data dataset and validated on two additional benchmarks: the AAPM Mayo Clinic Low Dose CT Grand Challenge dataset and a private clinical chest LDCT dataset comprising 524 scans (used for qualitative assessment only, as no NDCT ground truth is available). Across these datasets, the proposed method consistently outperformed state-of-the-art CNN- and transformer-based denoising models. On the LDCT and Projection dataset head images, it achieved a PSNR of 44.9217 dB, SSIM of 0.9801, and RMSE of 1.001, while qualitative evaluation on the private dataset confirmed strong generalization with clear noise suppression and preservation of fine anatomical details. These findings establish De-TransGAN as a clinically viable approach for LDCT denoising, enabling radiation reduction without compromising diagnostic quality. Full article

Computed tomography (CT) is a major source of low-dose ionizing radiation exposure in medical imaging. Risk assessment at this dose level is difficult and relies on the hypothetical linear no-threshold model. To address the response to such low doses in patients undergoing CT scans, we examined radiation-induced alterations at the transcriptomic and DNA damage levels in peripheral blood cells. Peripheral whole blood of 60 patients was collected before and after CT. Post-CT samples were obtained 4–6 h after scan (n = 28, in vivo incubation) or alternatively immediately after the CT scan, followed by ex vivo incubation (n = 32). The gene expression of known radiation-responsive genes (n = 9) was quantified using qRT-PCR. DNA double-strand breaks (DSB) were assessed in 12 patients through microscopic γ-H2AX + 53BP1 DSB focus staining. The mean dose–length product (DLP) across all scans was 561.9 ± 384.6 mGy·cm. Significant differences in the median differential gene expression (DGE) were detected between in vivo and ex vivo incubation conditions, implicating that ex vivo incubation masked the true effect in low-dose settings. The median DGE of in vivo-incubated samples showed a significant upregulation of EDA2R, MIR34AHG, PHLDA3, DDB2, FDXR, and AEN (p ranging from <0.001 to 0.041). In vivo, we observed a linear dose-dependent upregulation for several genes and an explained variance of 0.66 and 0.56 for AEN and FDXR, respectively. DSB focus analysis revealed a slight, non-significant increase in the average DSB damage post-exposure, at a mean DLP of 321.0 mGy·cm. Our findings demonstrate that transcriptional biomarkers are sensitive indicators of low-dose radiation exposure in medical imaging and could prove themselves as clinically applicable biodosimetry tools. Furthermore, the results underscore the need for dose optimization. Full article

Redox imbalance and epigenetic dysregulation, which both contribute to tumor initiation, survival, and resistance to therapy, are intimately linked to the progression of cancer. Reactive oxygen species (ROS) have two contrasting effects: at moderate concentrations, they promote angiogenesis and oncogenic signaling, whereas at high concentrations, they trigger apoptosis. Oxidative stress alters histone modifications, DNA methylation, and non-coding RNA (ncRNA) expression, reshaping the epigenetic landscape and supporting malignant phenotypes. Plant-derived antioxidants, including flavonoids, polyphenols, alkaloids, and terpenoids, act as dual modulators of cancer biology. They scavenge or regulate reactive oxygen species (ROS), restore redox balance, activate tumor suppressor pathways, inhibit oncogenic mechanisms, and reverse abnormal epigenetic marks. Compounds such as resveratrol, curcumin, epigallocatechin gallate (EGCG), quercetin, and sulforaphane modulate DNA methyltransferases (DNMTs), histone deacetylases (HDACs), and non-coding RNA networks, and can enhance chemotherapy and radiation therapy. Despite promising mechanisms, challenges remain in translational efficacy, optimal dosing, and bioavailability. This review emphasizes the potential of plant-derived antioxidants as precision oncology adjuncts and highlights the need for biomarker-guided strategies, nano-delivery systems, and clinical validation to fully realize their therapeutic benefits. Plant-derived antioxidants mitigate ROS-induced oncogenic signaling, as evidenced by in vitro and clinical models. Full article

Amygdalin, or vitamin B17, has attracted attention due to its commercial promotion as an anticancer and immune-boosting agent, despite documented concerns regarding its potential toxicity. To address this controversy, the present study demonstrates the protective effects of vitamin B17 against the harmful effects of ultraviolet radiation (UVR), a major risk factor for skin cancer, using the model organism Drosophila melanogaster. Our results showed that vitamin B17 supplementation effectively mitigated the adverse effects of UVR. Flies fed B17-supplemented food prior to UVR exposure displayed markedly higher adult emergence rates, improved climbing ability and shortened developmental time compared to UV-exposed flies on standard food. At the cellular level, B17 supplementation reduced Caspase-3 activation, preserved the structural integrity of compound eyes and mitochondria. Furthermore, biochemical analysis revealed that vitamin B17 reduced levels of oxidative stress markers, such as malondialdehyde, while simultaneously enhancing the activity of antioxidant enzymes, such as superoxide dismutase and catalase. Overall, these results demonstrate that vitamin B17 protects against UV-induced adverse effects in adult flies, highlighting its potential as a modulator of environmental stressors. However, caution is warranted given its known toxicity profile, which warrants further studies to determine appropriate doses and potential toxicity to other organisms. Full article

Background/Objectives: Recurrence of glioblastoma (GBM) mostly occurs in close vicinity to the resection cavity. Therefore, our group has previously designed an implant to locally apply repetitive photodynamic therapy to mitigate tumor recurrence. The penetration depths of different wavelengths in brain tissue were exhaustively studied before. However, the PDT-induced biological effects of 5-ALA-based PDT against GBM cells at different depths have not been evaluated yet. Methods: Therefore, a synthetic brain substitute material of 1–10 mm thickness and with optical properties comparable to the white or gray matter of pig brain was developed. Tumor cell viability was assessed in spheroids from six GBM cell lines using disks of varying thickness prepared from pig brain substitute material to mimic in vivo radiation attenuation. Results: Using an artificial brain tissue optical model based on material science, we have established a relationship between the PDT-induced effect of our PDT implant and the distance of migrating GBM cells from the resection cavity wall. Conclusions: This model may be helpful to aid optimization of the irradiation doses and fractionation required to attain the maximal therapeutic effect by long-term PDT applications. Full article

Exposure to ionizing and non-ionizing radiation from environmental and clinical settings can significantly threaten genomic stability, especially when combined. This ex vivo study investigates the potential combined effects of gamma radiation and ultraviolet B (UVB) exposure on human peripheral blood mononuclear cells (PBMCs) from healthy donors by exposing whole blood and isolated PBMCs to 1 Gy of gamma rays, to an absolute dose of approximately 100 J/m² of UVB, or to their combination. Combined exposure resulted in significantly elevated γH2AX foci formation and chromosomal aberrations relative to individual stressors, with the most pronounced effects observed in isolated PBMCs. Notably, lymphocytes from some donors failed to proliferate after UVB or co-exposure. Based on our results, a predictive biophysical model derived from dicentric yield was developed to estimate the gamma-ray equivalent dose from co-exposure, indicating up to ~9% increase in lifetime cancer risk. Although this proof-of-concept study included only a small number of donors and focused on two endpoints (γH2AX and dicentric assays), it provides a controlled framework for investigating mechanisms of radiation-induced genomic instability. The results emphasize the importance of accounting for mixed radiation exposures in genotoxic risk assessment and radiation protection. Full article

Machine learning models such as XGBoost show strong potential for predicting pasture quality metrics like crude protein (CP) content in tamani grass (Panicum maximum). However, their ‘black box’ nature hinders practical adoption. To address this limitation, this study applied SHapley Additive exPlanations (SHAP) to interpret an XGBoost model and uncover how management practices (grazing interval, nitrogen fertilization, and pre- and post-grazing heights) and environmental factors (precipitation, temperature, and solar radiation) jointly influence CP predictions. Data were divided into 80% for training/validation and 20% for testing. Model performance was assessed with stratified 5-fold cross-validation, and hyperparameters were tuned via grid search. The XGBoost model yielded a Pearson correlation coefficient (r) of 0.78, a mean absolute error (MAE) of 1.45, and a coefficient of determination (R²) of 0.57. The results showed that precipitation in the range of 100–180 mm increased the predicted CP content. Application of 240 kg N ha⁻¹ year⁻¹ positively affected predicted CP, whereas a lower dose of 80 kg N ha⁻¹ year⁻¹ had a negative impact, reducing predicted levels of CP. These findings highlight the importance of integrated management strategies that combine grazing height, nitrogen fertilization, and grazing intervals to optimize crude protein levels in tamani grass pastures. Full article

Purpose: To improve the prediction accuracy of radiation dermatitis (RD) in breast cancer patients undergoing volumetric modulated arc therapy (VMAT), we developed a hybrid artificial intelligence (AI) model that integrates deep learning radiomics (DLR), handcrafted radiomics (HCR), clinical features, and dose–volume histogram (DVH) parameters, aiming to enhance the early identification of high-risk individuals and support personalized prevention strategies. Methods: A retrospective cohort of 156 breast cancer patients treated with VMAT at Kaohsiung Veterans General Hospital (2018–2023) was analyzed; 148 patients were eligible after exclusions, with RD graded according to the RTOG criteria. Clinical variables and 12 DVH indices were collected, while HCR features were extracted via PyRadiomics. DLR features were derived from a pretrained VGG16 network across four input designs: original CT images (DLR_Original), a 5 mm subcutaneous region (DLR_Skin5mm), a planning target volume with a 100% prescription dose (DLR_PTV100%), and a subcutaneous region receiving ≥ 5 Gy (DLR_V5Gy). The features were preselected via ANOVA (p < 0.05), followed by Boruta–SHAP refinement across 11 feature sets. Predictive models were built via logistic regression, random forest, gradient boosting decision tree, and stacking ensemble (SE) methods. Explainability was assessed via SHapley Additive exPlanations (SHAPs) and gradient-weighted class activation mapping (Grad-CAM). Results: Among the 148 patients, 49 (33%) developed Grade ≥ 2 RD. The DLR models outperformed the HCR models (AUC = 0.72 vs. 0.66). The best performance was achieved with DLR_V5Gy + clinical + DVH features, yielding an AUC = 0.76, recall = 0.68, and F1 score = 0.60. SE consistently surpassed single classifiers. SHAP identified convolutional DLR features as the strongest predictors, whereas Grad-CAM focused attention on subcutaneous high-dose regions, which was consistent with the clinical RD distribution. Conclusions: The proposed hybrid AI framework, which integrates DLR, clinical, and DVH features, provides accurate and explainable predictions of Grade ≥ 2 RD after VMAT in breast cancer patients. By combining ensemble learning with XAI methods, the model offers reliable high-risk stratification and potential clinical utility for personalized treatment planning. Full article