Predicting drug-drug interaction (DDI) events is critical for ensuring patient safety, optimizing therapeutic efficacy, and advancing drug discovery. Deep learning-based models have recently attracted considerable attention in this domain and achieved promising results. However, most existing approaches insufficiently account for both the chemical structural information of drugs and the multiplicity of interaction types, thereby limiting predictive accuracy. In this work, we present ChemDDI, a deep learning framework for DDI event prediction enabled by multi-view-enhanced chemical structural information. Specifically, ChemDDI employs multi-view graph- and image-based encoders to extract chemical structural information from three-dimensional conformations. Building on these chemically informed representations, ChemDDI incorporates multi-relational interaction information through Transformer-based graph neural networks and relational graph embeddings, while contrastive learning further aligns interaction features to enable robust DDI event prediction. Extensive experiments demonstrate that ChemDDI consistently outperforms state-of-the-art baselines, achieving substantial gains on rare interaction events. ChemDDI is available at https://github.com/gjin-DLOU/ChemDDI . .

Purpose To examine whether time-dependent diffusion MRI (Td-dMRI) and macromolecular proton fraction (MPF) mapping-derived quantitative metrics can effectively distinguish between cervical cancer with and without lymph node metastasis (LNM) before treatment. Materials and Methods In this prospective study of adults with clinically suspected cervical cancer who underwent Td-dMRI, MPF mapping, and pulsed gradient spin-echo diffusion-weighted imaging (DWIPGSE) examinations between October 2023 and June 2025, authors calculated Td-dMRI-derived parameters (cellularity, diameter, intracellular volume fraction [Vin], and extracellular diffusivity [Dex]), MPF, and DWIPGSE-derived parameter (pulsed gradient spin-echo apparent diffusion coefficient [ADCPGSE]). Through Ridge regression analysis, the authors identified independent predictors of LNM and developed a composite diagnostic tool using logistic regression analysis. To evaluate tool performance, the area under the receiver operating characteristic curve was determined. Results Among 98 female individuals with cervical cancer (mean age, 56.69 years ± 11.63 [SD]), participants who were LNM positive exhibited higher cellularity, Vin, and MPF but lower diameter, Dex, and ADCPGSE than their counterparts who were LNM negative (P < .001 to P = .007). Cellularity, maximum tumor diameter, and MPF were independent predictors of LNM status, with their combination yielding the best diagnostic performance (area under the receiver operating characteristic curve, 0.95; 95% CI: 0.89, 0.98). The performance of this combination surpassed that of individual imaging modality, including DWIPGSE (ADCPGSE), and MPF, as well as any individual parameter, including cellularity, Vin, diameter, and Dex. Conclusion Td-dMRI and MPF mapping were effective for predicting LNM in cervical cancer, with the combination of cellularity, maximum tumor diameter, and MPF showing the best diagnostic performance. Keywords: Time-Dependent Diffusion MRI, Macromolecular Proton Fraction, Cervical Cancer, Lymph Node Metastases © RSNA, 2026.

Purpose:Hafnium oxide nanoparticles have been established as effective radiosensitizers, however, tumor cells often develop resistance to single-modality radiotherapy, and the tumor microenvironment (TME) poses additional limitations to treatment efficacy. To address these challenges, we fabricated a doxorubicin and manganese oxide co-loaded hafnium oxide (MD-Hf) nanoplatform for synergistic radio-chemotherapy and evaluated its antitumor performance in cellular and animal models. Results:In MD-Hf nanoplatform, the HfO2 nanocrystal functions as the radiosensitizer and carrier, the Dox works for chemotherapy, while the manganese oxide coating layer are capable of modulating the TME by depleting glutathione (GSH) and converting H2O2 in to ·OH radicals. Moreover, the MnOx coating also allows the nanoplatform possessing TME-responsive Dox release. Upon exposure to X-rays, the MD-Hf exhibited evident toxicity to Panc 02 tumor cells, only 63.9±10.7% cell remain alive after irradiated with 2Gy X-ray, which is much lower than 86.7±6.33% of the group administrated with pure HfO2 NPs. In vivo studies further demonstrated superior therapeutic outcomes with the MD-Hf nanoplatform, as evidenced by markedly reduced tumor size and weight compared to treatment with HfO2 nanoparticles alone. RNA-seq analysis reveals the Dox can potentiate organelle damage, and the MnOx can even activate immune response, which further corroborates the multifunctionality of the integrated nanoplatform. Conclusion:The newly developed doxorubicin and manganese oxide co-loaded HfO2 nanoplatform significantly enhance radio-chemotherapeutic efficacy against pancreatic tumor cells, offering a promising strategy that may guide the future clinical development of HfO2-based radiotherapy.

RATIONALE AND OBJECTIVES:To investigate the value of time-dependent diffusion MRI (Td-dMRI) and macromolecular proton fraction (MPF) imaging in assessing the pathological grade of cervical cancer (CC). MATERIALS AND METHODS:A total of 92 CC patients, comprising 35 with high-grade (grade III) cancer and 57 with low-grade (grade I/II) cancer, who underwent Td-dMRI and MPF, were prospectively enrolled. Td-dMRI derived parameters including cellularity (cell density), diameter (tumor cell size), Dex (extracellular diffusivity), Vin (intracellular volume fraction), and three apparent diffusion coefficients (ADCPGSE, ADC17 Hz, ADC33Hz) and MPF derived parameter MPF (tissue macromolecular) were calculated and compared. Diagnostic performance was assessed via area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and decision curve analysis (DCA); internal validation was performed using 1000 bootstrap resamples to mitigate model optimism. Multiple pairwise AUC comparisons were adjusted using the DeLong test. RESULTS:Cellularity, Vin, and MPF were higher and diameter, Dex, ADCPGSE, ADC17 Hz, and ADC33 Hz were lower in high-grade group than in low-grade group (all P < 0.05). Cellularity, Vin, and MPF were independent predictors and their combination achieved optimal diagnostic efficacy (AUC = 0.960; 95% CI: 0.897-0.990; sensitivity = 97.14%; specificity = 82.46%), which was significantly higher than any individual parameter (AUC range: 0.685-0.923; all P < 0.05 after Bonferroni correction). Internal validation confirmed stable performance (AUC = 0.955; 95% CI: 0.939-0.960), and DCA demonstrated higher net benefit for patients. Vin strongly correlated with pathological nuclear fraction (r = 0.752, P < 0.001). CONCLUSION:Td-dMRI and MPF were effective methods of predicting pathological grade in CC, and the combination of cellularity, Vin, and MPF has the potential to serve as a new imaging marker, facilitating preoperative grading and personalized treatment.

BackgroundWe aimed to develop an interpretable radiomics–clinical model to predict overall survival (OS) in unresectable pancreatic cancer (PC).MethodsIn this retrospective cohort, 202 patients with unresectable PC were enrolled. A total of 1,130 radiomics features were extracted from a region of interest encompassing the largest primary lesion using 3D-Slicer. Least absolute shrinkage and selection operator (LASSO)-selected features associated with OS were used to construct a radiomics risk score (RS). Independent clinical predictors were identified through stepwise Cox regression. A nomogram integrating RS with independent clinical predictors was built.ResultsMedian OS (mOS) for the entire cohort was 20.3 months. From 1,130 baseline CT radiomics features, LASSO retained 12 prognostic descriptors, which were linearly combined to compute a radiomics RS. Stepwise Cox regression identified age, sex, and CA19-9 as independent clinical predictors. A nomogram integrating RS with these variables was constructed in the training set. In the validation set, the area under the receiver operating characteristic curve (AUC) reached 0.804, 0.812, and 0.794 for 1-, 2-, and 3-year OS, respectively.ConclusionAn interpretable radiomics–clinical nomogram provided accurate survival prediction in unresectable pancreatic cancer.

OBJECTIVES:To evaluate the performance of nomograms combining clinical factors, apparent diffusion coefficient (ADC), and radiomics features from functional MRI parametric maps in predicting deep myometrial invasion (DMI), high histopathological grade, and lymphovascular space invasion (LVSI) in early endometrial cancer (EC). METHODS:This multicenter study recruited 362 patients with EC undergoing preoperative MRI. Radiomics features were extracted from four functional MRI parametric maps and selected via a three-step process. High-risk factors were identified by logistic regression. Nonradiomics models (clinical high-risk factors plus ADC), radiomics models (radiomics features), and nomogram models combining both were developed and validated for predicting DMI, high histopathological grade, and LVSI. Performance was assessed by receiver operating characteristic analysis. RESULTS:Logistic regression identified age and ADC as predictors of DMI, tumor size and ADC of high-grade lesions, and CA125 and ADC of LVSI; 10, 13, and 12 radiomics features constituted the respective Radscores. In the external testing set, the areas under the curve (AUCs) of the nonradiomics, radiomics, and nomogram models were 0.691 (95% confidence interval [CI]: 0.578-0.804), 0.756 (95% CI: 0.665-0.848), and 0.785 (95% CI: 0.709-0.849) for predicting DMI; 0.718 (95% CI: 0.613-0.823), 0.876 (95% CI: 0.796-0.956), and 0.906 (95% CI: 0.846-0.948) for high histopathological grade; and 0.651 (95% CI: 0.545-0.756), 0.764 (95% CI: 0.653-0.875), and 0.808 (95% CI: 0.716-0.900) for LVSI. CONCLUSION:The nomogram combining radiomics features from functional MRI parametric maps with clinical factors and ADC performs promisingly in evaluating DMI, high histopathological grade, and LVSI in early EC, demonstrating potential for personalized management.

BackgroundTo evaluate the utility of restricted spectrum imaging (RSI) for predicting subtypes of non-small cell lung cancer (NSCLC).MethodsA total of 97 patients with NSCLC (30 with squamous cell carcinoma (SCC) and 67 with adenocarcinoma (AC)) were included. The parameters f1, f2, f3, apparent diffusion coefficient (ADC), and maximum standardized uptake value (SUVmax) were measured and compared between the two subtypes. Logistic regression analysis was used to identify independent predictors, and a combined diagnostic model was developed. The performance of the model was assessed using receiver operating characteristic (ROC) curve analysis, calibration curves, and decision curve analysis (DCA).ResultsCompared with the AC group, the SCC group exhibited significantly higher SUVmax, f2, and f3 values, and lower ADC and f1 values (all P < 0.05). Smoking status, f1, SUVmax, and ADC were independent predictors of NSCLC subtypes. The combined model demonstrated superior diagnostic accuracy (AUC = 0.909; sensitivity = 73.33%; specificity = 89.55%) compared with individual predictors (AUC = 0.693, 0.819, 0.767, and 0.742 for smoking status, f1, SUVmax, and ADC, respectively; all P < 0.01). Bootstrap resampling (1000 samples) validated the robustness of the model (AUC = 0.895). Calibration curves and DCA confirmed the model’s stability and clinical utility.ConclusionRSI can effectively differentiate NSCLC subtypes.

BACKGROUND:Accurately predicting synergistic drug combinations is critical for complex disease therapy. However, the vast search space of potential drug combinations poses significant challenges for identification through biological experiments alone. Nowadays, deep learning is widely applied in this field. However, most methods overlook the important role of protein-protein interaction networks formed by gene expression products and the pharmacophore information of drugs in predicting drug synergy. RESULTS:We propose MultiSyn, a multi-source information integration method for the accurate prediction of synergistic drug combinations. Specifically, we design a semi-supervised learning framework using an attributed graph neural network to integrate protein-protein interaction networks of gene expression products with multi-omics data, constructing initial cell line representations that incorporate multi-source information. Furthermore, we refine the initial cell line representation by adaptively integrating it with normalized gene expression profiles, enabling the extraction of cell line features that encapsulate global information. In addition, we decompose drugs into fragments containing pharmacophore information based on chemical reaction rules and construct a heterogeneous graph comprising atomic and fragment nodes. To enhance the capture of molecular structural information, we introduce a heterogeneous graph transformer to learn multi-view representations of heterogeneous molecular graphs. Extensive experiments show that MultiSyn outperforms several classical and state-of-the-art baselines in synergistic drug combination prediction tasks. CONCLUSIONS:This study provides a powerful tool for inferring promising synergistic drug combinations. By leveraging attention mechanisms and pharmacophore information, MultiSyn identifies key substructures that are critical for synergy. Further visualization and case studies validate its effectiveness in capturing biologically meaningful features and identifying potential drug combinations.

BackgroundPatients with locally advanced non-small cell lung cancer (NSCLC) who undergo concurrent chemoradiotherapy (CCRT) followed by consolidation immunotherapy show heterogeneous survival outcomes. Accurate prognostic prediction remains a major challenge in clinical practice. This study aimed to develop machine learning models to enhance personalized outcome prediction and guide precision immuno-radiotherapy.MethodsA total of 219 patients with locally advanced NSCLC were retrospectively enrolled. All patients received standard CCRT followed by consolidation immunotherapy. Prognostic variables were first selected using least absolute shrinkage and selection operator (LASSO) regression. A multivariate Cox proportional hazards model and a random survival forest (RSF) model were then constructed in the training cohort and validated in the independent cohort.ResultsLASSO regression identified four prognostic variables: Age, T stage, Stage, and Pathology. Multivariate Cox analysis confirmed Stage and Pathology as independent predictors of OS. The Cox model achieved a C-index of 0.62 and Area Under the Receiver Operating Characteristic Curve (AUC-ROC) of 0.748 and 0.736 for 1-and 2-year OS in the validation cohort. The RSF model demonstrated higher predictive accuracy, with a C-index of 0.67 and AUC-ROC of 0.79 and 0.78 for 1-and 2-year OS, respectively. Variable importance analysis indicated that Stage and Pathology were the most influential factors. Based on RSF-derived risk scores, patients were stratified into high-and low-risk groups, and the high-risk group showed significantly poorer survival.ConclusionThe RSF model demonstrated improved performance compared to the conventional Cox model in predicting survival and stratifying risk among patients with locally advanced NSCLC undergoing CCRT and consolidation immunotherapy.

To investigate the value of multiparametric 18F-FDG PET/MRI based on tri-compartmental restrictive spectrum imaging (RSI), amide proton transfer-weighted imaging (APTWI), and diffusion-weighted imaging (DWI) in the assessment of lymph node metastasis (LNM) of non-small cell lung cancer (NSCLC). A total of 152 patients (LNM-positive, 86 cases; LNM-negative, 66 cases) with NSCLC underwent chest multiparametric 18F-FDG PET/MRI were enrolled. 18F-FDG PET- derived parameter (SUVmax), RSI-derived parameters (f1, f2, and f3), APTWI-derived parameter (MTRasym(3.5 ppm)), DWI-derived parameter (ADC), and were calculated and compared. Logistic regression analysis was used to identify independent predictors, and combined diagnostics. Area under the receiver operating characteristic curve (AUC), calibration curves and decision curve analysis (DCA) were employed to assess the performance of the combined diagnostics. MTRasym(3.5 ppm), SUVmax, f2, and f3 were higher and ADC and f1 were lower in LNM-positive group than in LNM-negative group (all P < 0.05). Maximum lesion diameter, f1, MTRasym(3.5 ppm), SUVmax, and ADC were independent predictors of LNM status in NSCLC patients, and the combination of them had an optimal diagnostic efficacy (AUC = 0.978; sensitivity = 95.35

Accurate genotyping and prognosis of glioma patients present significant clinical challenges, often dependent on subjective judgement and insufficient scientific evidence. This study aims to develop a robust, noninvasive preoperative multi-modal MRI-based transformer learning model to predict IDH genotyping and glioma prognosis. This multi-centre study included 563 glioma patients to develop an interpretable classification model utilising various preoperative imaging sequences, including T1-weighted, T2-weighted, fluid-attenuated inversion recovery, contrast-enhanced T1-weighted, and diffusion-weighted imaging. The model employs a multi-task learning framework to extract and fuse radiomic, deep learning, and clinical features for IDH genotyping and glioma prognosis. Additionally, a multi-modal transformer strategy is integrated to analyse structural and functional MRI, thereby enhancing model performance. Experimental results indicate that the model demonstrates superior performance, surpassing previous research and other state-of-the-art methods. The model achieves an AUC of 91.40% in the IDH genotyping task and 93.37% in the glioma prognosis task. Group analysis reveals that the model exhibits higher sensitivity to IDH-mutant cases and more accurately identifies low-risk groups compared to medium- or high-risk groups. This study aims to achieve accurate IDH genotyping and glioma prognosis through effective classification method, offering valuable diagnostic insights for clinical practice and expediting treatment decisions.

PurposeThis study aims to construct an individualized glucose metabolism network using total-body 18F-FDG PET imaging, which provides a comprehensive view of glucose metabolism across various organs, to explore the role of inter-organ interactions in Diabetes Mellitus (DM).MethodsIn this study, we constructed covariance metabolic networks using static total-body PET images, normalized by lean body mass, from 36 patients with DM (DM group) and 36 age- and sex-matched healthy controls (HC group). Differences in network properties between the DM and HC groups were evaluated at both group and individual levels. In addition, correlation analysis was performed to explore the relationship between network properties and baseline clinical data in the DM subjects.ResultsWe observed that the same edges in the first three edges with the largest values were brain-subcutaneous adipose tissue (SAT) and brain-visceral adipose tissue (VAT) at both group and individual levels. There was a positive correlation between the brain-VAT and BMI and there was a negative correlation between the brain-SAT and age. The most perturbed organ was the brain at both group and individual levels, and there was a positive correlation between the strength of abnormality of brain and age.ConclusionThis study successfully used static total-body PET imaging to construct individualized glucose metabolism networks for patients with DM, identifying the brain-VAT and brain-SAT as the most significantly altered edge and the brain as the most affected organ. These findings provide novel insights into the role of the brain-white adipose tissue axis in glucose metabolism in DM.

Background:Non-gonadotroph macroadenomas may exhibit higher cavernous sinus invasiveness compared to gonadotroph macroadenomas. Magnetic resonance imaging (MRI) is the preferred modality for detecting pituitary adenomas, yet, conventional MRI cannot distinguish gonadotroph from non-gonadotroph macroadenomas. This study aimed to evaluate the efficacy of histogram analysis based on contrast-enhanced T1-weighted imaging (CE-T1WI) for differentiating gonadotroph from non-gonadotroph macroadenomas. Methods:A retrospective analysis was conducted on 58 gonadotroph and 60 non-gonadotroph macroadenomas, pathologically confirmed at Henan Provincial People's Hospital between January 2022 and September 2024. Using 3D Slicer software, regions of interest (ROIs) were delineated on the coronal section with the largest area on the CE-T1WI images for grayscale histogram analysis. Then, the ROIs were copied to T1-weighted imaging (T1WI) maps to yield the subtraction between CE-T1WI and T1WI. Eight histogram parameters were obtained from the CE-T1WI and the subtraction between CE-T1WI and T1WI, including: the 10th percentile (Perc.10%), 90th percentile (Perc.90%), kurtosis, mean, median, maximum, minimum, and skewness values. A combined parameter model incorporating statistically significant parameters was developed. Continuous variables were compared using the independent samples t-test or Mann-Whitney U test. Tumor invasiveness was assessed via Knosp grading. The diagnostic performance of significant parameters was assessed using receiver operating characteristic (ROC) curves with the area under the curve (AUC) calculations. Pearson analysis determined the correlations between histogram parameters and Ki-67/P53 expression levels. Results:Among the histogram parameters derived from the CE-T1WI and the subtraction between CE-T1WI and T1WI, the Perc.10%, Perc.90%, mean, median, maximum, and minimum values demonstrated statistically significant differences (P<0.001 for all) in distinguishing gonadotroph from non-gonadotroph macroadenomas. The histogram analysis derived from the subtraction between CE-T1WI and T1WI demonstrated better discrimination performance compared to that derived from CE-T1WI. For instance, Perc.10% derived from subtraction (AUC =0.979) had significantly greater AUC than that derived from CE-T1WI (AUC =0.838, P<0.05). The combined parameters achieved an AUC of 0.992, significantly outperforming individual parameters (P<0.05). Non-gonadotroph macroadenomas exhibited greater invasiveness than gonadotroph macroadenomas (Knosp grades 3 and 4: 58.3% versus 37.9%, P=0.027), and showed higher Ki-67/P53 expression levels (P=0.007, 0.042, respectively). Additionally, there were positive correlations between Perc.90%, maximum, minimum and Ki-67, as well as between Perc.10%, Perc.90%, minimum and P53 (P<0.05 for all). Conclusions:Histogram analysis of both CE-T1WI and the subtraction between CE-T1WI and T1WI could differentiate gonadotroph macroadenomas from non-gonadotroph macroadenomas. The CE-T1WI histogram parameters were correlated with the Ki-67 and P53 expression levels in the macroadenomas.

The intricate interplay between organs can give rise to a multitude of physiological conditions. Disruptions such as inflammation or tissue damage can precipitate the development of chronic diseases such as tumors or diabetes mellitus (DM). While both lung cancer and DM are the consequences of disruptions in homeostasis, the relationship between them is intricate. This study sought to investigate the potential influence of DM on lung cancer by employing total-body dynamic PET imaging. The present study proposes a framework for metabolic network analysis using total-body dynamic PET imaging of 20 lung cancer patients with DM (DM group) and 20 lung cancer patients without DM (Non-DM group), with the residuals of a third-order polynomial fit serving as an indicator of Pearson correlation. The framework successfully captured the deviation of the DM group from the Non-DM group at both the edge and organ levels. At the edge level, there was a significant difference in the lesion- left ventricle (LV) between the DM and Non-DM groups (P < 0.05). Furthermore, we discovered a positive correlation between the absolute value of Z-score (ZCC) of lesion - LV and the duration of DM (R = 0.680, P < 0.001). At the organ level, there was a significant difference in the kidney, brain, and abdominal fat between the DM and Non-DM groups (P < 0.05). This study demonstrated the feasibility of constructing metabolic networks to uncover complex alterations in lung cancer patients with DM. The findings contribute to understanding the systemic effects of DM on lung cancer metabolism and highlight the importance of personalized metabolic network analysis to comprehend the implications of concurrent diseases.

OBJECTIVES:The purpose of this study was to assess the performance of multiparametric MRI-based radiomic models in predicting the molecular subtypes of endometrial cancer (EC) patients. METHODS:A total of 310 patients with pathologically confirmed EC who underwent preoperative MRI were enrolled this retrospective study and randomly divided into training (n = 217) and testing (n = 93) cohorts. We extracted 22,640 radiomic features from intratumoral and 3-mm peritumoral regions of interest (ROIs) on MR images. Feature selection was performed using the Mann-Whitney U test, Max-Relevance and Min-Redundancy (mRMR) and the least absolute shrinkage and selection operator (LASSO). Twelve radiomic signatures (RSs) were constructed using logistic regression to predict four molecular subtypes (POLEmut, MMRd, NSMP, and p53abn). The performance of these RSs was assessed using receiving operating characteristic (ROC) curve analysis, and the area under the curve (AUC), sensitivity, specificity, and accuracy were calculated. RESULTS:In the testing cohort, the RSs based on intratumoral features for predicting the POLEmut, MMRd, NSMP and p53abn subtypes yielded AUCs of 0.764, 0.812, 0.893 and 0.731, respectively, whereas those based on peritumoral features yielded AUCs of 0.847, 0.836, 0.871 and 0.804, respectively. The RSs constructed by combining intratumoral and peritumoral features for predicting the POLEmut, MMRd, NSMP and p53abn subtypes had the AUCs of 0.844, 0.880, 0.943 and 0.801, respectively. CONCLUSION:The combination of intratumoral and peritumoral radiomic features from multiparametric MRI enables effective and noninvasive prediction of EC molecular subtypes.

Objectives To differentiate benign and malignant solitary pulmonary lesions (SPLs) by amide proton transfer-weighted imaging (APTWI), mono-exponential model DWI (MEM-DWI), stretched exponential model DWI (SEM-DWI), and 18 F-FDG PET-derived parameters. Methods A total of 120 SPLs patients underwent chest 18 F-FDG PET/MRI were enrolled, including 84 in the training set (28 benign and 56 malignant) and 36 in the test set (13 benign and 23 malignant). MTRasym(3.5 ppm), ADC, DDC, α, SUV max , MTV, and TLG were compared. The area under receiver-operator characteristic curve (AUC) was used to assess diagnostic efficacy. The Logistic regression analysis was used to identify independent predictors and establish prediction model. Results SUV max , MTV, TLG, α, and MTRasym(3.5 ppm) values were significantly lower and ADC, DDC values were significantly higher in benign SPLs than malignant SPLs (all P < 0.01). SUV max , ADC, and MTRasym(3.5 ppm) were independent predictors. Within the training set, the prediction model based on these independent predictors demonstrated optimal diagnostic efficacy (AUC, 0.976; sensitivity, 94.64%; specificity, 92.86%), surpassing any single parameter with statistical significance. Similarly, within the test set, the prediction model exhibited optimal diagnostic efficacy. The calibration curves and DCA revealed that the prediction model not only had good consistency but was also able to provide a significant benefit to the related patients, both in the training and test sets. Conclusion The SUV max , ADC, and MTRasym(3.5 ppm) were independent predictors for differentiation of benign and malignant SPLs, and the prediction model based on them had an optimal diagnostic efficacy.

Objective This study aims to evaluate the effectiveness of deep learning features derived from multi-sequence magnetic resonance imaging (MRI) in determining the O-6-methylguanine-DNA methyltransferase (MGMT) promoter methylation status among glioblastoma patients. Methods Clinical, pathological, and MRI data of 356 glioblastoma patients (251 methylated, 105 unmethylated) were retrospectively examined from the public dataset The Cancer Imaging Archive. Each patient underwent preoperative multi-sequence brain MRI scans, which included T1-weighted imaging (T1WI) and contrast-enhanced T1-weighted imaging (CE-T1WI). Regions of interest (ROIs) were delineated to identify the necrotic tumor core (NCR), enhancing tumor (ET), and peritumoral edema (PED). The ET and NCR regions were categorized as intratumoral ROIs, whereas the PED region was categorized as peritumoral ROIs. Predictive models were developed using the Transformer algorithm based on intratumoral, peritumoral, and combined MRI features. The area under the receiver operating characteristic curve (AUC) was employed to assess predictive performance. Results The ROI-based models of intratumoral and peritumoral regions, utilizing deep learning algorithms on multi-sequence MRI, were capable of predicting MGMT promoter methylation status in glioblastoma patients. The combined model of intratumoral and peritumoral regions exhibited superior diagnostic performance relative to individual models, achieving an AUC of 0.923 (95% confidence interval [CI]: 0.890 - 0.948) in stratified cross-validation, with sensitivity and specificity of 86.45% and 87.62%, respectively. Conclusion The deep learning model based on MRI data can effectively distinguish between glioblastoma patients with and without MGMT promoter methylation.

Purpose: Radiotherapy (RT) is currently recognized as an important treatment for glioblastoma (GBM), however, it is associated with several challenges. One of these challenges is the radioresistance caused by hypoxia, whereas the other is the low conversion efficiency of the strongly oxidized hydroxyl radical (center dot OH), which is produced by the decomposition of water due to high-energy X-ray radiation. These factors significantly limit the clinical effectiveness of radiotherapy. Results: To address these limitations, we developed a highly stable and efficient nanoplatform (MnO2/Pt@BSA). Compared to MnO2@BSA, this platform demonstrates high stability, a high yield of oxygen (O2), enhanced production of center dot OH, and reduced clearance of center dot OH. The system exhibited increased O2 production in vitro and significantly improved oxygen production efficiency within 100 s at the Pt loading of 38.7%. Furthermore, compared with MnO2, the expression rate of hypoxia-inducible factor (HIF-1 alpha) in glioma cells treated with MnO2/Pt decreased by half. Additionally, the system promotes center dot OH generation and consumes glutathione (GSH), thereby inhibiting the clearance of center dot OH and enhancing its therapeutic effect. Moreover, the degradation of the nanoplatform produces Mn2+, which serves as a magnetic resonance imaging (MRI) contrast agent with a T1-weighted enhancement effect at the tumor site. The nanoplatform exhibited excellent biocompatibility and performed multiple functions related to radiotherapy, with simpler components. In U87 tumor bearing mice model, we utilized MnO2/Pt nanocatalysis to enhance the therapeutic effect of radiotherapy on GBM. Conclusion: This approach represents a novel and effective strategy for enhancing radiotherapy in gliomas, thereby advancing the field of catalytic radiotherapy and glioma treatment.

Background Node Reporting and Data System (Node-RADS) was proposed and can be applied to lymph nodes (LNs) across all anatomical sites. This study aimed to investigate the diagnostic performance of Node-RADS in cervical cancer patients. Methods A total of 81 cervical cancer patients treated with radical hysterectomy and LN dissection were retrospectively enrolled. Node-RADS evaluations were performed by two radiologists on preoperative MRI scans for all patients, both at the LN level and patient level. Chi-square and Fisher’s exact tests were employed to evaluate the distribution differences in size and configuration between patients with and without LN metastasis (LNM) in various regions. The receiver operating characteristic (ROC) and the area under the curve (AUC) were used to explore the diagnostic performance of the Node-RADS score for LNM. Results The rates of LNM in the para-aortic, common iliac, internal iliac, external iliac, and inguinal regions were 7.4%, 9.3%, 19.8%, 21.0%, and 2.5%, respectively. At the patient level, as the NODE-RADS score increased, the rate of LNM also increased, with rates of 26.1%, 29.2%, 42.9%, 80.0%, and 90.9% for Node-RADS scores 1, 2, 3, 4, and 5, respectively. At the patient level, the AUCs for Node-RADS scores > 1, >2, > 3, and > 4 were 0.632, 0.752, 0.763, and 0.726, respectively. Both at the patient level and LN level, a Node-RADS score > 3 could be considered the optimal cut-off value with the best AUC and accuracy. Conclusions Node-RADS is effective in predicting LNM for scores 4 to 5. However, the proportions of LNM were more than 25% at the patient level for scores 1 and 2, which does not align with the expected very low and low probability of LNM for these scores.

Purpose: To investigate the potential of radiomics signatures (RSs) from intratumoral and peritumoral regions on multiparametric magnetic resonance imaging (MRI) to noninvasively evaluate HER2 status in breast cancer. Method: In this retrospective study, 992 patients with pathologically confirmed breast cancers who underwent preoperative MRI were enrolled. The breast cancer lesions were segmented manually, and the intratumor region of interest (ROIIntra) was dilated by 2, 4, 6 and 8 mm (ROIPeri2mm, ROIPeri4mm, ROIPeri6mm, and ROIPeri8mm, respectively). Quantitative radiomics features were extracted from dynamic contrast-enhanced T1-weighted imaging (DCE-T1), fat‐saturated T2‐weighted imaging (T2) and diffusion-weighted imaging (DWI). A three-step procedure was performed for feature selection, and RSs were constructed using a support vector machine (SVM) to predict HER2 status. Result: The best single-area RSs for predicting HER2 status were DCE_Peri4mm-RS, T2_Peri4mm-RS, and DWI_Peri4mm-RS, yielding areas under the curve (AUCs) of 0.716 (95% confidence interval (CI), 0.648–0.778), 0.706 (95% CI, 0.637–0.768), and 0.719 (95% CI, 0.651–0.780), respectively, in the test set. The optimal RSs combining intratumoral and peritumoral regions for evaluating HER2 status were DCE-T1_Intra + DCE_Peri4mm-RS, T2_Intra + T2_Peri6mm-RS and DWI_Intra + DWI_Peri4mm-RS, with AUCs of 0.752 (95% CI, 0.686–0.810), 0.754 (95% CI, 0.688–0.812) and 0.725 (95% CI, 0.657–0.786), respectively, in the test set. Combining three sequences in the ROIIntra, ROIPeri2mm, ROIPeri4mm, ROIPeri6mm and ROIPeri8mm areas, the optimal RS was DCE-T1_Peri4mm + T2_Peri4mm + DWI_Peri4mm-RS, achieving an AUC of 0.795 (95% CI, 0.733–0.849) in the test set. Conclusion: This study systematically explored the influence of the intratumoral region, different peritumoral sizes and their combination in radiomics analysis for predicting HER2 status in breast cancer based on multiparametric MRI and found the optimal RS.