Abstract Predicting response to induction chemotherapy (IC) and overall survival (OS) is critical for optimizing treatment in patients with locally advanced nasopharyngeal carcinoma (LANPC). This study aimed to develop and validate a multi-task deep learning model integrating pretreatment MRI and whole slide images (WSIs) to predict IC response and OS in LANPC. Pretreatment MRI and WSIs from 404 patients with LANPC were retrospectively collected to construct a multi-task model (MoEMIL) for the simultaneous prediction of early IC response and OS. MoEMIL employed multi-instance learning to process WSIs, PyRadiomics and a convolutional neural network (ResNet50) to extract MRI features, and fused multimodal features through a multi-gate mixture-of-experts architecture. Clustering-constrained attention multiple instance learning and gradient-weighted class activation mapping were applied for visualization and interpretation. MoEMIL effectively stratified patients into good and poor IC response groups, achieving areas under the curve of 0.917, 0.869, and 0.801 in the train, validation, and test sets, respectively, and outperformed the deep learning radiomics model, the pathomics model and TNM staging. The model also stratified patients into high- and low-risk OS groups ( P < 0.05 ). MoEMIL shows promise as a decision-support tool for early IC response prediction and prognostication in LANPC. Author Summary We have developed a deep learning model that integrates two types of medical images, including magnetic resonance imaging (MRI) and digital pathological slices, to simultaneously predict response to induction chemotherapy and prognosis in patients with locally advanced nasopharyngeal carcinoma. Current treatment decisions primarily rely on traditional tumor staging (TNM), which often fails to comprehensively reflect the complexity of the disease. Our model, named MoEMIL, was trained and tested on data from 404 patients across two hospitals and consistently outperformed both single-model approaches and TNM staging methods. By identifying patients who exhibit poor response to induction chemotherapy or higher prognostic risk, our tool can assist clinicians in achieving personalized treatment, enabling intensified management for high-risk patients and avoiding unnecessary side effects for low-risk patients. Additionally, we visualize the model’s reasoning process through heat map generation, which highlights the image regions exerting the greatest influence on prediction outcomes. This work represents a step toward more precise treatment for nasopharyngeal carcinoma; however, larger-scale prospective studies are required before the model can be integrated into routine clinical practice.

Although senolytics such as dasatinib and quercetin (D+Q) show promise in modulating aging, their tissue-specific efficacy and optimal intervention timing remain poorly understood. Given D+Q's potential off-target effects, incomplete senescent cell clearance and associated hematologic side effects, we performed an unbiased multitissue single-cell analysis in aged mice across different aging phenotypes and tissue contexts. Here through integrative transcriptomics, single-cell technologies, histopathology and molecular profiling, we investigated the influence of D+Q treatment on aging-related phenotypes at the tissue and cellular levels. Specifically, D+Q remodeled immunity by enhancing immune cell function and maintaining population stability, alleviated tissue inflammation and improved metabolic profiles. Furthermore, intervention initiated during early aging and prolonged treatment showed a greater tendency to mitigate readouts of aging compared to shorter, late-stage treatment. Our findings reveal that D+Q systematically attenuates several aging hallmarks in a tissue- and cell-type-specific manner, and support the possibility that early-initiated, long-term intervention may amplify efficacy.

BACKGROUND:To explore and compare the potential value of radiomics models based on contrast-enhanced computed tomography (CT) for noninvasive preoperative prediction of lymphovascular invasion (LVI) in laryngeal squamous cell carcinoma (LSCC). MATERIALS AND METHODS:This multicenter diagnostic study retrospectively enrolled patients with LSCC from three tertiary hospitals who underwent surgical treatment. Standardized preprocessing was performed on the CT images, followed by region-of-interest segmentation and extraction of traditional radiomics features and deep learning (DL) features. Features were selected using least absolute shrinkage and selection operator (LASSO) regression. Traditional radiomics models and deep learning radiomics (DLR) models were established using logistic regression, random forest, and multilayer perceptron algorithms, respectively. A transformer-based hybrid model was developed by integrating radiomics and DL features. The predictive performance of the three types of models was evaluated and compared using the area under the curve (AUC), decision curve analysis (DCA), sample probability distribution histograms, confusion matrices, calibration curves, net reclassification index (NRI), and integrated discrimination improvement (IDI). RESULTS:A total of 1024 patients were allocated to the training set (center1, n = 291), internal validation set ( n = 126), and external test sets (Center 2, n = 437; Center 3, n = 170). Three radiomics models and three DLR models were constructed, and the optimal performance was observed in the DLR_ Random Forest model (AUC: 0.812-0.867). The transformer hybrid model demonstrated superior predictive performance, with AUC values of 0.881, 0.843, 0.833, and 0.836 in the training, internal validation, and external test sets, respectively. DCA indicated a higher net benefit for the Transformer model, along with an improved NRI and IDI. CONCLUSION:Radiomics models based on CT images exhibit potential for noninvasive prediction of LVI in LSCC, with the transformer hybrid model achieving the highest diagnostic performance. This approach may provide clinicians with a preoperative decision support tool to optimize treatment strategies for patients with LSCC.

Manual interpretation of CT images for bone metastasis (BM) detection in primary cancer remains challenging. We present an automated Bone Lesion Detection System (BLDS) developed using CT scans from 2518 patients (9177 BMs; 12,824 non-BM lesions) across five hospitals. The system, developed on 1271 patients and tested on 1247 multicenter cases, demonstrates 89.1% lesion-wise sensitivity (1.40 false-positives/case [FPPC]) in detecting bone lesions on non-contrast CT scans, with 92.3% and 91.1% accuracy in classifying BM/non-BM lesions for internal and external test sets, respectively. Outperforming radiologists in lesion detection (40.5% sensitivity; 0.65 FPPC), BLDS shows lower BM detection sensitivity than junior radiologists, though comparable to trainees. BLDS improves radiologists' lesion-wise sensitivity by 22.2% in BM detection and reduces reading time by 26.4%, while maintaining 90.2% patient-wise sensitivity and 98.2% negative predictive value in real-world validation (n = 54,610). The system demonstrates significant potential to enhance CT-based BM interpretation, particularly benefiting trainees.

Introduction Precise diagnosis and treatment of diseases necessitate quantitative visualization and modulation of subcellular structures. The endoplasmic reticulum (ER), as one of the most essential organelles, presents a complex target due to its intricate morphology and diverse cellular roles. Regulating ER stress offers a promising strategy for treating diseases such as tumors. However, achieving accurate targeting and therapeutic intervention at the subcellular level remains a significant challenge. Thus, there is an urgent need for theranostic agents that can precisely target and modulate ER stress. Objectives This study proposes a novel AI-driven dual-targeting strategy combining “passive + active” mechanisms to efficiently design molecules that resolve the balance between passive ER enrichment and precise modulation. We aim to design multifunctional theranostic molecules that precisely target Grp78, a key biomarker of ER stress, at the atomic level, enabling concurrent imaging and modulation of ER stress. Methods A machine learning (ML)-based molecular fingerprints transfer method was developed for passive targeting based on identified subcellular targeting substructures. Meanwhile, a deep learning (DL)-based 3D molecular generation model, PM-1, was designed for active targeting through specific receptor interactions. By transferring key fingerprints and fluorescent motifs into PM-1-generated molecules, desired theranostic agents were produced. Their key properties were validated via dynamic simulations and quantitative calculations, followed by wet experiments. Results Guided by these strategies, we identified unreported ER-targeting rules by discovering key passive-targeting fingerprints derived from ML models, and generated diverse new structures with high affinity binding to Grp78. We successfully synthesized ABT-CN2, a multidimensional fluorescent agent that demonstrates cost-effective chemical structure (molecular weight <400), robust targeting capability (Pearson’s correlation coefficient = 0.93), and potential antitumor activity (IC50 = 53.21 μM). Conclusion This work presents a new paradigm for the intelligent design of fluorescent molecular probes with precise organelle-targeting capabilities for integrated diagnosis and therapy.

PURPOSE:To develop an integrative radiopathomic model based on deep learning to predict overall survival (OS) in locally advanced nasopharyngeal carcinoma (LANPC) patients. MATERIALS AND METHODS:A cohort of 343 LANPC patients with pretreatment MRI and whole slide image (WSI) were randomly divided into training (n = 202), validation (n = 91), and external test (n = 50) sets. For WSIs, a self-attention mechanism was employed to assess the significance of different patches for the prognostic task, aggregating them into a WSI-level representation. For MRI, a multilayer perceptron was used to encode the extracted radiomic features, resulting in an MRI-level representation. These were combined in a multimodal fusion model to produce prognostic predictions. Model performances were evaluated using the concordance index (C-index), and Kaplan-Meier curves were employed for risk stratification. To enhance model interpretability, attention-based and Integrated Gradients techniques were applied to explain how WSIs and MRI features contribute to prognosis predictions. RESULTS:The radiopathomics model achieved high predictive accuracy in predicting the OS, with a C-index of 0.755 (95 % CI: 0.673-0.838) and 0.744 (95 % CI: 0.623-0.808) in the training and validation sets, respectively, outperforming single-modality models (radiomic signature: 0.636, 95 % CI: 0.584-0.688; deep pathomic signature: 0.736, 95 % CI: 0.684-0.810). In the external test, similar findings were observed for the predictive performance of the radiopathomics, radiomic signature, and deep pathomic signature, with their C-indices being 0.735, 0.626, and 0.660 respectively. The radiopathomics model effectively stratified patients into high- and low-risk groups (P < 0.001). Additionally, attention heatmaps revealed that high-attention regions corresponded with tumor areas in both risk groups. CONCLUSION:The radiopathomics model holds promise for predicting clinical outcomes in LANPC patients, offering a potential tool for improving clinical decision-making.

ObjectiveTo develop and validate a radiomics model based on vertebral calcium-suppressed (CaSupp) images derived from dual-layer computed tomography (DLCT) for predicting chemotherapy-induced myelosuppression in patients with locally advanced nasopharyngeal carcinoma (LANPC).MethodsThis retrospective study included 150 LANPC patients treated with induction chemotherapy (IC). Radiomics features were extracted from lumbar vertebral CaSupp obtained from baseline DLCT scans. Models were developed to predict myelosuppression after the first chemotherapy cycle (IC - 1) and entire chemotherapy cycles (IC-n). The clinics, radiomics, and combined models were conducted via multivariate logistic regression. Models performance was evaluated by the area under the receiver operating characteristic curve (AUC). Clinical utility was analyzed with decision curve analysis.ResultsFor predict myelosuppression after IC - 1, the clinics, radiomics, and combined models had AUC values of 0.716, 0.825 and 0.859 in the train cohort, respectively; and AUC of 0.687, 0.752 and 0.790 in the test cohort, respectively. And for IC-n, the clinics, radiomics, and combined models exhibited AUC values of 0.771, 0.824, and 0.889 in the train cohort, respectively; and AUC of 0.652, 0.740 and 0.806 in the test cohort, respectively. For predicting myelosuppression after both IC - 1 and IC-n,the combined models demonstrated significantly higher AUC values than the clinics models for both IC - 1 and IC-n (all P<0.05).ConclusionsRadiomics model based on vertebral CaSupp images from DLCT could predict chemotherapy-induced myelosuppression in LANPC patients. This study highlights the potential of DLCT technology to provide quantitative bone marrow assessments and aid in personalized treatment planning. External validation and comparison with other imaging modalities are warranted in the future.

Background:Anemia negatively affects an individual's overall prognosis and quality of life, and thus represents a significant health burden. Dual-layer computed tomography (DLCT) detector imaging enables substance differentiation. This study aimed to compare the performance of DLCT parameters for different blood vessels in detecting anemia. Methods:DLCT parameter values [i.e., the computed tomography (CT) value, effective atomic number, and electron density] were retrospectively derived from the aortic arch, pulmonary artery, and portal vein of 240 patients. Differences in DLCT parameters between the anemia and normal groups were analyzed. Pearson correlation analysis and logistic regression models were employed to examine the relationships between the DLCT parameters and hemoglobin concentration. The diagnostic performance of DLCT parameters for anemia among different blood vessels was evaluated by receiver operating characteristic (ROC) analysis. Results:The anemia group (n=101) had significantly lower hemoglobin concentration than the normal group (n=139) (107.96±13.95 vs. 138.40±12.64 g/L, P<0.001), as well as significantly lower CT and electron density values for the three vessels (all P<0.05). The CT value and effective atomic number of the portal vein were significantly lower than those of the aortic arch and pulmonary artery (all P<0.05). The correlation of the CT value of the portal vein to hemoglobin concentration was significantly lower than that of the aortic arch (r=0.435 vs. 0.583, P=0.029) and slightly lower than that of the pulmonary artery (r=0.435 vs. 0.527, P=0.192). Regarding the correlation between electron density and hemoglobin concentration, there were no significant differences among the three blood vessels (all P>0.05). When using the CT value to detect anemia, the aortic arch had an area under the curve (AUC) value of 0.79, which was significantly higher than that of the portal vein (AUC =0.68, P=0.008) and slightly higher than that of the pulmonary artery (AUC =0.73, P=0.126). In relation to electron density, the aortic arch had an AUC value of 0.81, which was slightly higher than that of both the portal vein (AUC =0.77, P=0.239) and the pulmonary artery (AUC =0.75, P=0.095). Among the six CT predictors, the CT value of the portal vein had the lowest AUC value (AUC =0.68), and the value was significantly lower than that of the aortic arch (P=0.008), that of the electron density of the aortic arch (P=0.002), and that of electron density of the portal vein (P=0.007). The multivariable logistic regression showed that the CT value of the aortic arch, electron density of the pulmonary artery, and electron density of the portal vein were independent predictors of anemia. The logistic regression model that integrated the above three CT indicators showed the best performance (AUC =0.85) in predicting anemia, outperforming any single CT predictor of an individual vessel (all P<0.05). Conclusions:DLCT may assist in the detection of anemia. The DLCT parameters of the aortic arch demonstrated higher performance than those of the pulmonary artery and portal vein. Additionally, integrating different DLCT parameters (i.e., the CT value and electron density) of multiple vessels may improve diagnostic performance.

Background: Patients with microvascular invasion (MVI)-positive hepatocellular carcinoma (HCC) have shown promising results with adjuvant hepatic arterial infusion chemotherapy (HAIC) with FOLFOX after curative resection. The authors aim to develop an imaging-derived biomarker to depict MVI-positive HCC patients more precisely and promote individualized treatment strategies of adjuvant HAIC. Materials and methods: Patients with MVI-positive HCC were identified from five academic centers and utilized for model development (n=470). Validation cohorts were pooled from a previously reported prospective clinical study conducted [control cohort (n=145), adjuvant HAIC cohort (n=143)] (NCT03192618). The primary endpoint was recurrence-free survival (RFS). Imaging features were thoroughly reviewed, and multivariable logistic regression analysis was employed for model development. Transcriptomic sequencing was conducted to identify the associated biological processes. Results: Arterial phase peritumoral enhancement, boundary of the tumor enhancement, tumor necrosis stratification, and boundary of the necrotic area were selected and incorporated into the nomogram for RFS. The imaging-based model successfully stratified patients into two distinct prognostic subgroups in both the training, control, and adjuvant HAIC cohorts (median RFS, 6.00 vs. 66.00 months, 4.86 vs. 24.30 months, 11.46 vs. 39.40 months, all P<0.01). Furthermore, no significant statistical difference was observed between patients at high risk of adjuvant HAIC and those in the control group (P=0.61). The area under the receiver operating characteristic curve at 2 years was found to be 0.83, 0.84, and 0.73 for the training, control, and adjuvant HAIC cohorts, respectively. Transcriptomic sequencing analyses revealed associations between the radiological features and immune-regulating signal transduction pathways. Conclusion: The utilization of this imaging-based model could help to better characterize MVI-positive HCC patients and facilitate the precise subtyping of patients who genuinely benefit from adjuvant HAIC treatment.

Here we report an afterglow material, APT/TTQNPs. This material can be activated by superoxide anions (O2(center dot)-) to emit NIR afterglow. APT/TTQNPs possess high selectivity, high sensitivity, and high imaging signal-to-noise ratio to O2(center dot)-. We successfully use APT/TTQNPs for imaging LPS-induced arthritis in mice.

Obesity is associated with skeletal deterioration and increased fracture risk, but the underlying mechanism is unclear. Herein, it is shown that obese gut microbiota promotes skeletal deterioration by inducing bone marrow macrophages (BMMs) senescence and grancalcin (GCA) secretion. Obese mice and those receiving obese fecal microbiota transplants exhibit increased senescent macrophages and elevated GCA expression in the bone marrow. In a study of 40 participants, it is found that obese patients are associated with higher serum GCA levels. It is further revealed that obese gut-microbiota derived lipopolysaccharides (LPS) stimulate GCA expression in senescent BMMs via activating Toll-like receptor 4 pathway. Mice with depletion of the Gca gene are resistant to the negative effects of obesity and LPS on bone. Moreover, neutralizing antibody against GCA mitigates skeletal deterioration in obese mice and LPS-induced chronic inflammation mouse model. The data suggest that the interaction between gut microbiota and the immune system contributes to obesity-associated skeletal deterioration, and targeting senescent macrophages and GCA shows potential of protecting skeletal health in obese population.

Background: Virtual non-calcium (VNCa) imaging based on dual-energy computed tomography (CT) plays an increasingly important role in diagnosing spinal diseases. However, the utility of VNCa technology in the measurement of vertebral bone mineral density (BMD) is limited, especially the VNCa CT value at multiple calcium suppression levels and the slope of VNCa curve. This retrospective cross-sectional study aimed to explore the correlation between vertebral BMD and new VNCa parameters from dual-layer spectral detector CT. Methods: The dual-layer spectral detector CT and quantitative CT (QCT) data of 4 hydroxyapatite (HAP) inserts and 667 vertebrae of 234 patients (132 male and 102 female) who visited a university teaching hospital between April and May 2023 were retrospectively analyzed. The BMD values of 3 vertebrae (T12, L1, and L2) and inserts were measured using QCT, defined as QCT-BMD. The VNCa CT values and the slope lambda of the VNCa attenuation curve of vertebrae and inserts were recorded. The correlations between VNCa parameters (VNCa CT value, slope lambda) and QCT-BMD were analyzed. Results: For the vertebrae, the correlation coefficient ranged from -0.904 to 0.712 (all P<0.05). As the calcium suppression index (CaSI) increased, the correlation degree exhibited a decrease first and then increased, with the best correlation (r=-0.904, P<0.001) observed at the index of 25%. In contrast, the correlation coefficient for the inserts remained relatively stable (r=-0.899 to -1, all P<0.05). For the vertebrae, the values of 3 slopes lambda (lambda 1, lambda 2, and lambda 3) derived from the VNCa attenuation curve were 6.50 +/- 1.99, 3.75 +/- 1.15, and 2.04 +/- 0.62, respectively. Regarding the inserts, the lambda 1, lambda 2, and lambda 3 values were 11.56 [interquartile range (IQR): 2.40-22.62], 6.68 (IQR: 1.39-13.49), and 3.63 (IQR: 0.75-7.8), respectively. For the vertebrae, all 3 correlation coefficients between 3 slopes lambda and QCT-BMD were 0.956 (all P<0.05). For the inserts, the 3 correlation coefficients were 0.996, 0.998, and 1 (all P<0.05), respectively. Conclusions: A promising correlation was detected between VNCa CT parameters and QCT-BMD in vertebrae, warranting further investigation to explore the possibility of VNCa imaging to assess BMD.

Multifunctional molecules with both optical signal and pharmacological activity play an important role in drug development, disease diagnosis, and basic theoretical research. Aminopeptidase N (APN), as a representative tumor biomarker with anti-tumor potential, still lacks a high-precision theranostic probe specifically targeting it. In this study, a novel quaternity design strategy for APN theranostic probe was developed. This proposed strategy utilizes advanced machine learning and molecular dynamics simulations, and cleverly employs the strategy of conformation-induced fluorescence recovery to achieve multi-objective optimization and integration of functional fragments. Through this strategy, a unique “Off–On” theranostic probe, ABTP-DPTB, was ingeniously constructed to light up APN through fluorescence restoration, relying on conformation-induced effects and solvent restriction. Differ from the common diagnostic probes, the intelligent design with non-substrated linkage makes ABTP-DPTB for long-term in-situ imaging. The fabricated probe was used for detecting and inhibiting APN in various environments, with a better in vitro inhibitory than golden-standard drug bestatin.

Background Previous studies have demonstrated conflicting findings regarding the initial MRI patterns of radiotherapy-induced temporal lobe injury (RTLI) and the evolution of different RTLI patterns. The aim of this study was to evaluate the initial MRI pattern and evolution of RTLI in patients with nasopharyngeal carcinoma (NPC) by means of a large cohort study. Methods Data of patients with RTLI were retrospectively collected from two hospitals between January 2011 and December 2021. The injured lobes were categorized into three patterns based on initial MRI patterns: isolated white matter lesions (WMLs), isolated contrast-enhanced lesions (CELs), and combined WMLs and CELs. The latency period, MRI appearances, and temporal changes in WMLs and CELs were evaluated. Results A total of 913 RTLI patients with 1092 injured lobes were included in this study. The numbers of isolated WMLs, isolated CELs, and combined WMLs and CELs identified at the first MRI detection were 7(0.6%), 172(15.8%), and 913(83.6%), respectively. The evolution of bilateral RTLI was different in the same patient, and that of unilateral RTLI combined with WMLs and CELs also may occur asynchronously. The time intervals from the initial MRI detection of isolated WMLs, isolated CELs, combined WMLs and CELs to the last negative MRI scan were 8.6, 8.9 and 11.0 months, respectively. A significant difference was observed in the time intervals between the three patterns (H = 14.287, P = 0.001). And the time interval was identified as an independent factor influencing the initial MRI pattern of RTLI after Poisson regression ( P = 0.002). Conclusion Both WMLs and CELs could be the initial and only MRI abnormalities in patients with RTLI. This study is of great significance in accurately diagnosing RTLI early and providing timely treatment options. Additionally, it provides clinical evidence for guidelines on NPC, emphasizing the importance of regular follow-up of NPC patients.

To construct an integrative radiopathomics model for predicting progression-free survival (PFS) in nonmetastatic nasopharyngeal carcinoma (NPC) patients. 357 NPC patients who underwent pretreatment MRI and pathological whole-slide imaging (WSI) were included in this study and randomly divided into two groups: a training set (n = 250) and validation set (n = 107). Radiomic features extracted from MRI were selected using the minimum redundancy maximum relevance and least absolute shrinkage and selection operator methods. The pathomics signature based on WSI was constructed using a deep learning architecture, the Swin Transformer. The radiopathomics model was constructed by incorporating three feature sets: the radiomics signature, pathomics signature, and independent clinical factors. The prognostic efficacy of the model was assessed using the concordance index (C-index). Kaplan-Meier curves for the stratified risk groups were tested by the log-rank test. The radiopathomics model exhibited superior predictive performance with C-indexes of 0.791 (95

ObjectiveTo construct radiomics models based on MRI at different time points for the early prediction of cystic brain radionecrosis (CBRN) for nasopharyngeal carcinoma (NPC).MethodsA total of 202 injured temporal lobes from 155 NPC patients with radiotherapy-induced temporal lobe injury (RTLI) after intensity modulated radiotherapy (IMRT) were included in the study. All the injured lobes were randomly divided into the training (n = 143) and validation (n = 59) sets. Radiomics models were constructed by using features extracted from T2WI at two different time points: at the end of IMRT (post-IMRT) and the first-detected RTLI (first-RTLI). A delta-radiomics feature was defined as the percentage change in a radiomics feature from post-IMRT to first-RTLI. The radiomics nomogram was constructed by combining clinical risk factors and radiomics signatures using multivariate logistic regression analysis. Predictive performance was evaluated using area under the curve (AUC) from receiver operating characteristic analysis and decision curve analysis (DCA).ResultsThe post-IMRT, first-RTLI, and delta-radiomics models yielded AUC values of 0.84 (95% CI: 0.76–0.92), 0.86 (95% CI: 0.78–0.94), and 0.77 (95% CI: 0.67–0.87), respectively. The nomogram exhibited the highest AUC of 0.91 (95% CI: 0.85–0.97) and sensitivity of 0.82 compared to any single radiomics model. From the DCA, the nomogram model provided more clinical benefit than the radiomics models or clinical model.ConclusionThe radiomics nomogram model combining clinical factors and radiomics signatures based on MRI at different time points after radiotherapy showed excellent prediction potential for CBRN in patients with NPC.

Background The metastatic vascular patterns of hepatocellular carcinoma (HCC) are mainly microvascular invasion (MVI) and vessels encapsulating tumor clusters (VETC). However, most existing VETC‐related radiological studies still focus on the prediction of VETC status. Purpose This study aimed to build and compare VETC‐MVI related models (clinical, radiomics, and deep learning) associated with recurrence‐free survival of HCC patients. Study Type Retrospective. Population 398 HCC patients (349 male, 49 female; median age 51.7 years, and age range: 22–80 years) who underwent resection from five hospitals in China. The patients were randomly divided into training cohort ( n = 358) and test cohort ( n = 40). Field Strength/Sequence 3‐T, pre‐contrast T1‐weighted imaging spoiled gradient recalled echo (T1WI SPGR), T2‐weighted imaging fast spin echo (T2WI FSE), and contrast enhanced arterial phase (AP), delay phase (DP). Assessment Two radiologists performed the segmentation of HCC on T1WI, T2WI, AP, and DP images, from which radiomic features were extracted. The RFS related clinical characteristics (VETC, MVI, Barcelona stage, tumor maximum diameter, and alpha fetoprotein) and radiomic features were used to build the clinical model, clinical‐radiomic (CR) nomogram, deep learning model. The follow‐up process was done 1 month after resection, and every 3 months subsequently. The RFS was defined as the date of resection to the date of recurrence confirmed by radiology or the last follow‐up. Patients were followed up until December 31, 2022. Statistical Tests Univariate COX regression, least absolute shrinkage and selection operator (LASSO), Kaplan–Meier curves, log‐rank test, C‐index, and area under the curve (AUC). P < 0.05 was considered statistically significant. Results The C‐index of deep learning model achieved 0.830 in test cohort compared with CR nomogram (0.731), radiomic signature (0.707), and clinical model (0.702). The average RFS of the overall patients was 26.77 months (range 1–80 months). Data Conclusion MR deep learning model based on VETC and MVI provides a potential tool for survival assessment. Evidence Level 3 Technical Efficacy Stage 3

Obesity-induced chronic inflammation exacerbates multiple types of tissue/organ deterioration and stem cell dysfunction; however, the effects on skeletal tissue and the underlying mechanisms are still unclear. Here, we show that obesity triggers changes in the microRNA profile of macrophage-secreted extracellular vesicles, leading to a switch in skeletal stem/progenitor cell (SSPC) differentiation between osteoblasts and adipocytes and bone deterioration. Bone marrow macrophage (BMM)-secreted extracellular vesicles (BMM-EVs) from obese mice induced bone deterioration (decreased bone volume, bone microstructural deterioration, and increased adipocyte numbers) when administered to lean mice. Conversely, BMM-EVs from lean mice rejuvenated bone deterioration in obese recipients. We further screened the differentially expressed microRNAs in obese BMM-EVs and found that among the candidates, miR-140 (with the function of promoting adipogenesis) and miR-378a (with the function of enhancing osteogenesis) coordinately determine SSPC fate of osteogenic and adipogenic differentiation by targeting the Pparα-Abca1 axis. BMM miR-140 conditional knockout mice showed resistance to obesity-induced bone deterioration, while miR-140 overexpression in SSPCs led to low bone mass and marrow adiposity in lean mice. BMM miR-378a conditional depletion in mice led to obesity-like bone deterioration. More importantly, we used an SSPC-specific targeting aptamer to precisely deliver miR-378a-3p-overloaded BMM-EVs to SSPCs via an aptamer-engineered extracellular vesicle delivery system, and this approach rescued bone deterioration in obese mice. Thus, our study reveals the critical role of BMMs in mediating obesity-induced bone deterioration by transporting selective extracellular-vesicle microRNAs into SSPCs and controlling SSPC fate.

The accumulation and systemic propagation of senescent cells contributes to physiological aging and age-related pathology. However, which cell types are most susceptible to the aged milieu and could be responsible for the propagation of senescence has remained unclear. Here we found that physiologically aged bone marrow monocytes/macrophages (BMMs) propagate senescence to multiple tissues, through extracellular vesicles (EVs), and drive age-associated dysfunction in mice. We identified peroxisome proliferator-activated receptor α (PPARα) as a target of microRNAs within aged BMM-EVs that regulates downstream effects on senescence and age-related dysfunction. Demonstrating therapeutic potential, we report that treatment with the PPARα agonist fenofibrate effectively restores tissue homeostasis in aged mice. Suggesting conservation to humans, in a cohort study of 7,986 participants, we found that fenofibrate use is associated with a reduced risk of age-related chronic disease and higher life expectancy. Together, our findings establish that BMMs can propagate senescence to distant tissues and cause age-related dysfunction, and they provide supportive evidence for fenofibrate to extend healthy lifespan.