Background Immune checkpoint blockade provides durable benefit in a subset of patients with advanced non-small cell lung cancer (NSCLC), but response rates remain limited after progression on standard systemic therapy. Stereotactic body radiotherapy (SBRT) may enhance tumor-antigen release and systemic immune activation, thereby improving the efficacy of PD-1 blockade. This multicenter, single-arm, phase II study evaluated the efficacy and safety of SBRT plus nivolumab in previously treated advanced NSCLC. Methods Eighty-three patients who had experienced disease progression after at least one line of standard systemic therapy were enrolled across nine centers. Patients received SBRT to the primary lung lesion or, in selected cases without a visible primary tumor, to an involved nodal lesion, followed by nivolumab. Patients with symptomatic bone metastases could receive prior bone radiotherapy. The primary endpoint was objective response rate (ORR) according to RECIST version 1.1. Secondary endpoints included duration of response, progression-free survival (PFS), overall survival (OS), and safety. Bone radiotherapy and biomarker analyses were exploratory. Results The ORR was 39.7% (95% confidence interval [CI], 29.4-51.1%), including complete responses in 10.8%, and the disease control rate was 61.4%. The median duration of response was 27.0 months. Median PFS and OS were 11.1 months and 25.9 months, respectively. Among the 36 patients with bone metastases, the exploratory ORR was 57.2% in those who received bone radiotherapy and 13.6% in those who did not. Elevated baseline alkaline phosphatase and urea levels were associated with poorer survival in exploratory analyses. Treatment-related adverse events occurred in 63.9% of patients, with grade 3-4 events in 4.8% and no treatment-related deaths. Conclusions SBRT plus nivolumab demonstrated encouraging and durable antitumor activity with manageable toxicity in previously treated advanced NSCLC. The findings concerning bone radiotherapy and serum biomarkers are exploratory and hypothesis-generating and require validation in prospective randomized studies.

To evaluate whether a statistical process control (SPC) method can improve the safety and stability of radiotherapy planning for patients with nasopharyngeal carcinoma. A total of 200 patients were retrospectively enrolled and divided into 2 sequential phases. The first phase (100 patients) was used to establish baseline statistical parameters including centerlines and control limits. The second phase (100 patients) used these parameters to guide plan optimization. Dose indicators of organs at risk, target coverage, process stability, and process capability index were analyzed and compared between the 2 phases. After applying SPC, radiation doses to the larynx and parotid glands were significantly reduced, while target volume coverage was adequately maintained. All plans remained within stable statistical control limits, and process capability indices were improved, indicating enhanced process stability and reliability. SPC provides an effective, data-driven, and easy-to-implement tool for improving nasopharyngeal carcinoma radiotherapy planning. It reduces unnecessary radiation exposure to normal tissues, enhances process stability, and promotes consistent high-quality treatment. This method is clinically feasible and suitable for widespread application in routine radiotherapy quality assurance.

Radiotherapy (RT) remains a cornerstone of cancer management but is fundamentally constrained by normal tissue toxicity, intrinsic and acquired radioresistance, and hypoxia- and microenvironment-driven dose-response plateaus. Engineered nanomaterials offer a versatile toolbox to reshape this therapeutic landscape by coupling enhanced energy deposition with microenvironmental and biological reprogramming. This review summarizes recent advances in nanomaterial-enabled radiosensitization from four interlocking dimensions: (i) high-atomic-number platforms that amplify local dose via photoelectric and related interactions and thereby increase microscopic energy deposition; (ii) chemical radiosensitization through modulation of reactive oxygen species (ROS), including Fenton/Fenton-like catalysis and depletion of glutathione (GSH)/thioredoxin antioxidant networks; (iii) tumor microenvironment (TME) remodeling strategies that alleviate hypoxia, buffer acidity, rewire redox and metabolic states, relieve immune suppression, and normalize vasculature and extracellular matrix (ECM) to broaden the effective therapeutic window; and (iv) biological radiosensitization targeting DNA damage response (DDR), cell-cycle redistribution, and multiple programmed cell-death pathways such as apoptosis and ferroptosis. We further discuss nano-delivery architectures-passive EPR-based systems, ligand-directed and biomimetic carriers, and stimuli-responsive (pH, hypoxia, redox, or irradiation triggered) formulations-that co-load radiosensitizers, chemotherapeutics, and molecularly targeted agents, as well as theranostic platforms integrating computed tomography (CT)/magnetic resonance imaging (MRI)/optical contrast for image-guided, dose-adapted treatment. Emerging multimodal regimens, including radiotherapy combined with photothermal, sonodynamic, chemotherapy, and immunotherapy on a single nano-platform, are highlighted for their capacity to achieve genuine "1 + 1>2" synergy. Finally, we outline key translational challenges-industrial-scale, standardized manufacturing; long-term safety and clearance; inter-patient and spatiotemporal heterogeneity; quantitative linkage between imaging signals, dose, and biological effect; and the integration of biomarkers and artificial intelligence into personalized nano-radiotherapy-and propose future directions to accelerate clinical implementation.

BACKGROUND:With improving survival for patients receiving whole-brain radiotherapy (WBRT), mitigating long-term toxicities like cataract and dry eye syndrome has become increasingly critical. The lens and lacrimal gland are highly radiosensitive and lie in close proximity to the target volume, posing a persistent challenge for achieving sharp dose gradients. While modern techniques like multileaf collimator (MLC) shaping offer some protection, the potential of classic geometric optimization principles, such as anterior beam shift, remains underexplored in contemporary treatment planning workflows. PURPOSE:This study aimed to systematically translate, quantitatively validate, and integrate the classic geometric principle of anterior isocenter shift into the modern treatment planning workflow for the dual protection of the lens and lacrimal gland in WBRT, utilizing 3D-conformal radiotherapy (3D-CRT) and field-in-field (FIF) techniques. METHODS:For 40 patients, conventional and isocenter-optimized plans (involving an anterior shift of the isocenter within the PTV) were generated for both 3D-CRT and FIF. We compared dosimetric parameters for the planning target volume (PTV), lenses, lacrimal glands, and other organs at risk. Plan quality, normal tissue complication probability (NTCP) for cataract and dry eye syndrome, and clinical risk stratification were evaluated RESULTS: Isocenter optimization significantly reduced the median lens Dmax by 20% and PRV_lens D0.03 cm3 by 23% (p < 0.001). Lacrimal gland Dmean, V6Gy, and V10Gy were also significantly reduced. The strategy physically improved the dose gradient, narrowing the penumbra width by 31.25% and increasing the dose fall-off rate by 15%, while reducing low-dose irradiation volumes outside the PTV. These dosimetric benefits translated into meaningful reductions in projected NTCP for both complications. The 3D-FIF plans maintained target coverage and homogeneity, mitigating the heterogeneity increase observed in 3D-CRT plans. CONCLUSION:Anterior isocenter optimization is a practical, and hardware-free technique that seamlessly synergizes with modern MLC-based planning to provide significant, concurrent sparing of the lens and lacrimal gland in WBRT. As a readily implementable modification within existing planning systems, this strategy can be adopted immediately to enhance treatment safety without requiring additional resources.

Pancreatic cancer, a highly lethal malignancy with a poor prognosis, has primarily relied on gemcitabine monotherapy, yet its clinical efficacy is severely restricted by the dense extracellular matrix (ECM), which impedes drug penetration. In this study, a self-healing CMC-OHA hydrogel derived from oxidized hyaluronic acid (OHA) and carboxymethyl chitosan (CMC) was synthesized to fabricate an injectable hydrogel therapeutic platform (Gel@Col/Z-Gem). This platform incorporates type I collagenase (Col) and gemcitabine (Gem)-loaded metal organic framework ZIF-67 (Z-Gem) to enhance drug permeation in pancreatic tumors. In three-dimensional (3D) tumor spheroid models, Nile red fluorescence was detected deep within the core of tumor spheroids in both the Col+Z-Gem and Gel@Col/Z-Gem groups, whereas groups without collagenase exhibited fluorescence limited only to the tumor periphery. Furthermore, in the Panc02 mouse subcutaneous tumor model, tumors treated with Gel@Col/Z-Gem exhibited significantly greater inhibition compared to those in the control group. These findings suggest that the Gel@Col/Z-Gem hydrogel effectively disrupts the tumor stroma by enabling controlled temporal and spatial release of the encapsulated agents, thereby enhancing the anti-tumor efficacy of gemcitabine. Collectively, this hydrogel system represents a promising strategy for improving the clinical treatment of pancreatic cancer.

Objective The objective of this study was to develop a predictive model combining radiomic characteristics and clinical features to forecast overall survival in cervical cancer patients treated with intensity-modulated radiotherapy and concurrent chemotherapy. Methods In this retrospective observational study, 159 patients were divided into a training group (n = 95) and a validation group (n = 64). Radiomic characteristics were extracted from contrast-enhanced computed tomography scans. The least absolute shrinkage and selection operator regression analysis was used to filter the extracted radiomic characteristics and reduce the dimensionality of the data. A radiomic score was calculated from the selected features, and multivariate Cox regression models were established to analyze overall survival. A nomogram combining radiomic score and clinical features was developed, and its reliability was assessed using the area under the receiver operating characteristic curve. Results Four radiomic characteristics and two clinical features were extracted for overall survival analysis. A nomogram combining these factors was developed and validated, showing good performance with a high C-index. Patients were categorized as low-risk or high-risk for overall survival based on a cut-off value. Conclusions Our model combining computed tomography–extracted radiomic characteristics and clinical features shows good potential for evaluating overall survival in cervical cancer patients treated with intensity-modulated radiotherapy and concurrent chemotherapy.

This study constructed a predictive model for occurrence of radiation esophagitis during breast-cancer radiotherapy. 308 breast-cancer patients were analyzed. Lasso regression identified crucial variables that were further integrated into a radiation esophagitis risk score, which was used to segregate patients into high- and low-risk groups. A nomogram model was designed for clinical applicability. Training and validations were performed to assess robustness and generalizability of proposed models, employing C-index, AUCs, calibration curves, and decision curves. SHAP algorithm was used for model interpretation, offering insights into the major contributory factors. Seven significant variables were identified by Lasso regression. C-indexes of nomograms of individual clinical variables and risk score were 0.795 and 0.784, respectively, exhibiting strong predictive ability. In internal validation, AUCs for risk score, nomogram, and logistic models were 0.784, 0.795, and 0.812, respectively. Calibration curves showed a close fit between predicted and observed outcomes across models. Decision curve analysis indicated logistic model's superior clinical utility when the risk threshold was above 0.2. SHAP interpretation emphasized radiation dose, pruritus, molecular type, and hepatic dysfunction as top contributory factors for radiation esophagitis. Models based on interpretable machine learning offer an intuitive tool to assess risk of radiation esophagitis in breast-cancer radiotherapy.

Unsupervised brain tumor segmentation can aid brain tumor diagnosis and treatment without the high cost of manual annotations. Existing methods typically use a reconstruction-based strategy, where an image self-reconstruction network is trained with normal data and applied to images with brain tumors. The reconstruction error map is then used to indicate the tumor regions and is thresholded to obtain tumor segmentation. However, optimal threshold selection is challenging without annotations in the unsupervised case, which limits the accuracy and applicability of these reconstruction-based methods. To address the problem, in this work we propose the Bi-Level Optimization Guided by Radiological Reports (BLOGRR) framework for unsupervised brain tumor segmentation. BLOGRR extends the reconstruction-based strategy with an additional threshold estimation network. Instead of selecting an empirical fixed threshold, it determines an adaptive threshold for every sample. Specifically, we develop an iterative bi-level optimization procedure, where lower and upper loops jointly update the reconstruction network and threshold estimation network. As no manual annotation is available, BLOGRR resorts to radiological reports, which provide key descriptions of image anomalies in the form of natural language, for learning the threshold determination. The reports are processed with brain anatomical knowledge to indicate potential tumor regions. Two loss functions are developed for the two loops to optimize the reconstruction network and threshold estimation network. Experimental results on a public dataset and an in-house dataset indicate that BLOGRR outperforms existing unsupervised methods with noticeable improvements. Code is available at https://github.com/Beliefzp/BLOGRR.

Background:This study was performed to investigate the relationship of the pretreatment neutrophil count and neutrophil-to-lymphocyte ratio (NLR) with the prognosis of nasopharyngeal carcinoma (NPC), as well as to establish an NLR-related nomogram to predict survival in patients with NPC. Methods:In total, 747 patients with NPC were enrolled between January 2005 and January 2015 at our hospital. Kaplan-Meier survival analysis was used to evaluate overall survival (OS), progression-free survival (PFS), and distant metastasis-free survival (DMFS), with comparisons made using the log-rank test. Univariate and multivariate Cox regression analyses were conducted to identify independent risk factors for OS, PFS, and DMFS. The optimal NLR cut-off value was determined using receiver operating characteristic curve analysis. A nomogram model was then constructed and validated using R software (Version 3.6.0). Results:Among the 747 patients, N stage (P = 0.01, 0.042, 0.017) and NLR (P = 0.037) were identified as independent predictors of DMFS. Independent predictors of OS were sex (P = 0.024), age (P = 0.019), N stage (P = 0.006, 0.031, 0.002), American Joint Committee on Cancer (AJCC) stage (P = 0.003), adjuvant chemotherapy (P = 0.016), and NLR (P = 0.036). N stage (P = 0.001, 0.0221, 0.003), AJCC stage (P = 0.001), and NLR (P = 0.035) were also associated with PFS. The prognostic model showed good agreement with actual outcomes. Compared with the TNM staging system, the nomogram demonstrated superior accuracy and stability. Conclusions:In patients with NPC, an elevated pretreatment NLR was associated with poorer OS, PFS, and DMFS. The NLR-based nomogram provided more accurate survival prediction than clinical staging and may serve as a valuable tool in guiding prognosis and treatment planning.

Background: Kidney cancer remains a significant challenge in oncology, with accurate prognostic assessment being crucial for postoperative management. While radiomics has shown promise in cancer prognosis, there is limited research on comprehensive models that effectively integrate radiomic features with clinical parameters for kidney cancer survival prediction. Objective: This study aimed to develop and validate a comprehensive computed tomography (CT) radiomics-based nomogram for predicting overall survival in postoperative patients with kidney cancer by integrating radiomic features with clinical parameters. Methods: Radiomic features were extracted from regions of interest in CT images of 207 postoperative patients with kidney cancer. The eigenvalue data of all radiomic features were processed using z score standardization and the R software package GLMNet. We integrated survival time, survival status, and radiomic features and screened these features using the least absolute shrinkage and selection operator-Cox regression method. We conducted 10-fold cross-validation to obtain an optimal model of 5 radiomic features. Multivariate Cox regression hazard models were established to analyze patients' overall survival. The predictive ability of the nomogram (receiver operating characteristic curve and calibration curve) was evaluated using bootstrap resampling validation. Patients were divided into high-and low-risk groups based on the radiomic risk score cutoff value, and the Kaplan-Meier method was conducted to identify established models' forecasting ability. Five radiomic features were screened for predictive model construction. Results: This retrospective analysis was conducted from April 2024 to July 2024 using data from The Cancer Imaging Archive public database. The final cohort included 207 patients (3 excluded from the initial 210) who underwent nephrectomy for kidney cancer. The median follow-up time was 33 (IQR 11-47) months. The receiver operating characteristic curve and area under the curve showed that the predictive model performed well. The calibration curve of nomogram and radiomic features in the cohort study set indicated an overall net benefit. Kaplan-Meier curves indicated that overall survival time was dramatically shorter in the high-risk group. Conclusions: Our radiomics nomogram successfully integrates CT-derived radiomic features with clinical variables for kidney cancer survival prediction, demonstrating good prognostic capability and offering a noninvasive, quantitative tool for personalized postoperative management and clinical decision-making.

OBJECTIVE This study explored using statistical process control for quality control of cervical cancer interstitial brachytherapy treatment plans. MATERIALS AND METHODS For retrospective analysis, interstitial brachytherapy treatment plans were divided into first (n = 300) and second phases (n = 200). The first phase was chronologically divided 2:1 into training and validation sets. The Dn2cm3 (D2cm3 divided by the high-risk clinical target volume D90) of the organs at risk (the bladder, rectum, and sigmoid colon) were analyzed to draw individual control charts. Process capability analysis charts were drawn, and the statistical process capability was evaluated using the process capability index Cpk. The centerline of the organ at risk dose in the first-phase plan's training set was used as the optimization parameter for the second-phase dataset plan. RESULTS The Dn2cm3 centerlines for the bladder, rectum, and sigmoid colon were 0.6980, 0.5440, and 0.4910 in the training set and 0.6845, 0.4528, and 0.4144 in the second phase, respectively. The first-phase δ values were 0.0099, 0.0530, and 0.0268, respectively. The process capability analysis for the first and second phases showed that all indicators had a Cpk >1. CONCLUSION For all organs at risk, the Dn2cm3 centerlines were lower in the second phase than in the first phase, indicating that quality control of cervical cancer interstitial brachytherapy treatment plans continuously improved through statistical process control. This method is simple and practical and warrants promotion for application in radiotherapy treatment plan quality control.

Brain metastases (BM) are common complications of advanced cancer, posing significant diagnostic and therapeutic challenges for clinicians. Therefore, the ability to accurately detect, segment, and classify brain metastases is crucial. This review focuses on the application of artificial intelligence (AI) in brain metastasis imaging analysis, including classical machine learning and deep learning techniques. It also discusses the role of AI in brain metastasis detection and segmentation, the differential diagnosis of brain metastases from primary brain tumors such as glioblastoma, the identification of the source of brain metastases, and the differentiation between radiation necrosis and recurrent tumors after radiotherapy. Additionally, the advantages and limitations of various AI methods are discussed, with a focus on recent advancements and future research directions. AI-driven imaging analysis holds promise for improving the accuracy and efficiency of brain metastasis diagnosis, thereby enhancing treatment plans and patient prognosis.

Cervical cancer (CC) was a significant gynecological malignancy. Brachytherapy (BT) has found extensive application in cervical cancer radiotherapy, attributed to its remarkable features of high-precision positioning and highly conformal dose distribution. However, the field lacks a comprehensive bibliometric analysis. This research provides a comprehensive bibliometric analysis of cervical cancer brachytherapy trends and key topics, projecting future research directions. A search was executed within the Web of Science Core Collection for publications on cervical cancer brachytherapy until 2025/6/23. Analytical tools were utilized to conduct in-depth bibliometric and visual analyses of the relevant online publications. The analyses covered multiple aspects, including countries/regions, institutions, authors, journals, and keywords. A total of 2924 articles were analyzed, showing an upward trend. The USA was the most productive country, with the Medical University of Vienna leading in publications. Brachytherapy had the highest number of publications. Tanderup K was the most prolific author, while Pötter R was the most frequently co-cited author. The latest high-frequency keywords included “cervical cancer,” “interstitial brachytherapy,” and “image-guided brachytherapy,” among others. Through keyword co-occurrence-based cluster analysis, 10 distinct clusters were generated, effectively highlighting the research hotspots and frontiers in cervical cancer brachytherapy. With medical imaging informatics advancing, research on cervical cancer brachytherapy has become increasingly profound. Recently, areas such as image-guided-adaptive-brachytherapy, effective biological dose, and radionecrosis have attracted significant research attention. Future research is expected to focus on developing and enhancing artificial intelligence (AI) tools to optimize brachytherapy treatment planning, aiming to benefit a larger number of cervical cancer patients.

Brain metastases are common complications in patients with cancer and significantly affect prognosis and treatment strategies. The accurate segmentation of brain metastases is crucial for effective radiation therapy planning. However, in resource-limited areas, the unavailability of MRI imaging is a significant challenge that necessitates the development of reliable segmentation models for computed tomography images (CT). This study aimed to develop and evaluate a Diffusion-CSPAM-U-Net model for the segmentation of brain metastases on CT images and thereby provide a robust tool for radiation oncologists in regions where magnetic resonance imaging (MRI) is not accessible. The proposed Diffusion-CSPAM-U-Net model integrates diffusion models with channel-spatial-positional attention mechanisms to enhance the segmentation performance. The model was trained and validated on a dataset consisting of CT images from two centers (n = 205) and (n = 45). Performance metrics, including the Dice similarity coefficient (DSC), intersection over union (IoU), accuracy, sensitivity, and specificity, were calculated. Additionally, this study compared models proposed for brain metastases of different sizes with those proposed in other studies. The diffusion-CSPAM-U-Net model achieved promising results on the external validation set. Overall average DSC of 79.3

Background:Non-small cell lung cancer (NSCLC) represents a significant portion of lung cancer cases globally, with kirsten rats arcomaviral oncogene homolog (KRAS) mutations being a critical factor in its pathogenesis. Predicting KRAS mutation status is crucial for guiding targeted therapies and improving patient outcomes. This study aimed to develop and validate a differential evolution optimized artificial neural network (DE-ANN) model that integrates positron emission tomography/computed tomography (PET/CT) radiomics and genomics data for predicting KRAS mutation status in NSCLC patients, showcasing the potential of multi-omics integration in precision oncology. Methods:The study utilized PET/CT radiomics features and genomics data from public databases using least absolute shrinkage and selection operator (LASSO) regression and support vector machine-recursive feature elimination (SVM-RFE) to identify key predictive features. The DE-ANN model was optimized using differential evolution algorithms and validated internally using Bootstrap resampling to assess its predictive performance. Results:The DE-ANN model demonstrated superior predictive accuracy with an area under the curve (AUC) of 0.909 [95% confidence interval (CI): 0.882-0.937], outperforming traditional artificial neural network (ANN) models (AUC =0.819, 95% CI: 0.778-0.860). Key features identified included significant radiomics signatures and gene markers, with the model showing enhanced convergence rates and robust internal validation outcomes. The model's calibration and decision curve analyses further confirmed its clinical applicability and potential for improving personalized treatment strategies in NSCLC. Conclusions:The DE-ANN model represents a significant advancement in the predictive modeling of KRAS mutation status in NSCLC, leveraging the synergy between radiomics and genomic data. Its high predictive accuracy and methodological robustness highlight the model's potential as a tool in precision oncology, warranting further external validation and exploration in other cancer types.

To analyze the effect of jaw width in jaw tracking mode on the dose of radiotherapy partial arc VMAT (P-VMAT) for patients undergoing left breast-conserving surgery and to explore the best jaw width as the initial inverse optimization parameter. Twenty patients who underwent left breast-conserving surgery were randomly selected. Six groups of P-VMAT plans were designed (named Plan0, Plan0.3, Plan0.6, Plan0.9, Plan-0.3, and Plan-0.6). The width of the jaw of each plan was changed in 0.3 cm steps along the X direction (from - 0.6 to 0.9 cm) according to the beginning of the half beam (Plan0). The PTV coverage, conformity index (CI), homogeneity index (HI), monitor units (MU) and organs at risk (OARs) dose were evaluated by repeated measurement data analysis of variance between plan0 and the other plans. Additionally, the correlations between CI, HI, MU and OARs to change in jaw width were analyzed using Spearman's bivariate correlation analysis. The PTV dose distributions of Plan-0.3 and Plan-0.6, which have smaller jaw widths than those of Plan0, did not meet the clinical requirements. CI, HI and MU were correlated with jaw width (r = 0.554, -0.501, -0.641, p < 0.05, respectively). The V5, V10, V20, V40, Dmean and Dmax of the heart were correlated with jaw width (r = 0.288, 0.284, 0.191, -0.27, 0.186, -0.245, p < 0.05, respectively). The V2.5, V5, V10, V20, V40 and Dmean of the left lung (Lung-L) were correlated with jaw width (0.298, 0.421, 0.516, 0.391, -0.241, 0.356, p < 0.05, respectively). Among all the plans to ensure PTV target coverage, Plan0 had the lowest clinical indicators for the heart and Lung-L (p < 0.05, respectively). The internal boundary of the jaw set as 0 cm (Plan0) represents the optimal jaw width for the initial optimization of the plan design. This method is the simplest and most effective for radiotherapy treatment planning for breast-conserving surgery for breast cancer as well as allows ideal dose distribution.

Background:Brain metastases (BM) are the most common type of intracranial tumor and the leading cause of mortality in patients with systemic cancer. In recent years, stereotactic radiosurgery (SRS) has been widely used in the radiotherapy of BM due to its advantages of high positional accuracy and highly conformal dose distributions. However, this area lacks a bibliometric analysis. This study aims to provide an overview of recent trends and key topics related to SRS for BM treatment over the past decade and to anticipate future directions through bibliometric methods. Methods:We conducted a search in the Web of Science for publications on SRS in BM treatment from 2013 to 2023. VOSviewer, CiteSpace, and the R package "bibliometrix" were utilized to perform a bibliometric and visual analysis of online publications in this field, focusing on countries/regions, institutions, authors, journals, and keywords. Results:A total of 2,085 articles were identified in this study, with a steady increase observed in annual publications. The United States (USA) was the most productive country and the core of international cooperation; Mayo Clinic was the institution with the most publications and citations; Journal of Neuro-Oncology published the most papers; the most published author was Sahgal A, and Brown PD was the most co-cited author. The latest high-frequency keywords were immunotherapy, survival, prognosis, recurrence, leptomeningeal metastases, and so on. Keyword cooccurrence was used for cluster analysis, resulting in 7 clusters that highlight the emerging frontiers of SRS in BM treatment. Conclusions:As medical imaging informatics technology continues to advance, research into SRS for BM treatment has become increasingly in-depth. The immunoadjuvant therapy, biological effective dose, and radionecrosis have emerged as hot topics in recent years. Future work is ongoing to develop and improve artificial intelligence (AI) tools to assist in SRS treatment planning, thus benefiting more BM patients.

BackgroundKidney tumors, common in the urinary system, have widely varying survival rates post-surgery. Current prognostic methods rely on invasive biopsies, highlighting the need for non-invasive, accurate prediction models to assist in clinical decision-making.PurposeThis study aimed to construct a K-means clustering algorithm enhanced by Transformer-based feature transformation to predict the overall survival rate of patients after kidney tumor resection and provide an interpretability analysis of the model to assist in clinical decision-making.MethodsThis study was based on a publicly available C4KC-KiTS-2019 dataset from the TCIA database, including preoperative computed tomography (CT) images and survival time data of 210 patients. Initially, the radiomics features of the kidney tumor area were extracted using the 3D slicer software. Feature selection was then conducted using ICC, mRMR algorithms, and LASSO regression to calculate radiomics scores. Subsequently, the selected features were input into a pre-trained Transformer model for feature transformation to obtain a higher-dimensional feature set. Then, K-means clustering was performed using this feature set, and the model was evaluated using receiver operating characteristic (ROC) and Kaplan-Meier curves. Finally, the SHAP interpretability algorithm was used for the feature importance analysis of the K-means clustering results.ResultsEleven important features were selected from 851 radiomics features. The K-means clustering model after Transformer feature transformation showed AUCs of 0.889, 0.841, and 0.926 for predicting 1-, 3-, and 5-year overall survival rates, respectively, thereby outperforming both the K-means model with original feature inputs and the radiomics score method. A clustering analysis revealed survival prognosis differences among different patient groups, and a SHAP analysis provided insights into the features that had the most significant impacts on the model predictions.ConclusionsThe K-means clustering algorithm enhanced by the Transformer feature transformation proposed in this study demonstrates promising accuracy and interpretability in predicting the overall survival rate after kidney tumor resection. This method provides a valuable tool for clinical decision-making and contributes to improved management and treatment strategies for patients with kidney tumors.

Purpose:To evaluate the safety and efficacy of computed tomography (CT)-guided iridium-192 (192Ir) high-dose-rate (HDR) interstitial brachytherapy (ISBT) for pleural and chest wall malignant tumours. Material and methods:This single-centre retrospective cohort study involved 21 patients with pleural/chest wall malignant tumours treated between January 2024 and January 2025. All patients underwent HDR ISBT (30 Gy in a single fraction). Treatment included CT-guided needle implantation, three-dimensional dose optimisation (Oncentra system), and adherence to Radiation Therapy Oncology Group dose constraints for organs at risk (OARs). Efficacy endpoints included objective response (Response Evaluation Criteria in Solid Tumours v1.1), pain relief (Numerical Rating Scale), and dosimetric comparison with virtual stereotactic body radiotherapy. Safety was assessed using the Radiation Therapy Oncology Group/European Organisation for Research and Treatment of Cancer toxicity criteria. Results:The median follow-up duration was 7.48 months. The objective response rate (complete response + partial response) was 76.19%, with 28.57% achieving a complete response and 47.62% achieving a partial response. Pain relief was achieved in 87.5% of patients with pretreatment pain, with numerical rating scale scores decreasing from moderate to severe (median, 6) to mild (median, 3) at 1 month. No ≥ grade II complications (e.g. bronchopleural fistula, pneumothorax) occurred; only four patients experienced minor subcutaneous haemorrhage/emphysema. Dosimetric analysis showed a significantly higher target mean dose with ISBT than with stereotactic body radiotherapy (p < 0.001), while OARs doses (e.g. lungs, heart, spinal cord) remained within Radiation Therapy Oncology Group limits. Rib and chest wall dose exceedances were rare and clinically insignificant. Conclusions:Computed tomography-guided 192Ir HDR ISBT offers safe, effective local control and rapid pain relief for pleural/chest wall tumours, demonstrating superior dosimetric conformity and lower toxicity to OARs. This minimally invasive approach is a viable option for patients unsuitable for surgery or external beam radiotherapy.