Find the papers that actually matter

This study aimed to identify neutrophil extracellular trap-related genes (NET-RGs), explore their prognostic significance, and predict drug interactions in peripheral T-cell lymphoma (PTCL). Differentially expressed NET-RGs (DE-NRGs) between PTCL and normal tissues were screened. Functional enrichment analysis was conducted. Bioinformatics and machine learning were used to identify hub genes and assess their diagnostic value. Univariate and multivariate analyses were used to evaluate prognostic roles. Correlation and immune infiltration analyses were performed to explore relationships with the tumor microenvironment (TME). Clinical data were collected from PTCL patients who received potential agents (lenalidomide) as first-line treatment. A total of 31 DE-NRGs were identified (18 upregulated and 13 downregulated), enriched in inflammatory response, extracellular matrix organization, and infection involvement. Four hub genes (AKT2, MAPK14, IRF1, and TNF) were identified as effective PTCL diagnostic markers. Higher AKT2/MAPK14 expression correlated with poorer overall survival (OS), while elevated TNF expression associated with better OS; AKT2 and TNF independently predicted OS. These genes were implicated in modulating TME remodeling. Potential therapeutic agents (e.g. capivasertib, lenalidomide) were predicted, and lenalidomide may represent a feasible initial treatment option for PTCL, with an objective response rate (ORR) of 40.0% and a maximum survival duration exceeding 50 months. NET-RGs play crucial roles in diagnosis, prognosis, and TME regulation, and lenalidomide, a putative TNF-targeting agent, may represent a feasible initial treatment option in PTCL.

FMS-like tyrosine kinase 3 (FLT3/CD135) regulates haematopoiesis and is frequently mutated as FLT3-internal tandem duplication (FLT3-ITD) in acute myeloid leukaemia (AML), associated with poor prognosis. Although FLT3 inhibitors show clinical benefits, resistance remains a challenge. This study hypothesises that antibody-drug conjugate (ADC) efficacy depends on distinct FLT3 trafficking mechanisms in FLT3-wt and FLT3-ITD cells. Confocal imaging showed that in THP-1 (FLT3-wt) cells, FLT3 mAb trafficked to lysosomes, while in MV4-11 (FLT3-ITD) cells, it accumulated in the Golgi. To evaluate the impact of this trafficking difference, we synthesised an anti-FLT3 mAb-MMAE, linked via a Val-Cit-PAB linker at the Fc N-glycan, which exhibited lower cytotoxicity in MV4-11 than THP-1 cells, indicating that the impaired lysosomal trafficking of FLT3-ITD limits drug release and reduces ADC potency. These findings highlight that effective lysosomal targeting is essential for ADC activity and suggest that optimising linker design or restoring lysosome trafficking may enhance FLT3-targeted ADC in AML.

KRAS-G12D mutations are common drivers of pancreatic and colorectal cancers, yet effective targeted therapies remain limited. This study describes the design, synthesis, and biological evaluation of two novel KRAS-G12D inhibitors, GD-2 and GD-4. Both compounds exhibited strong antiproliferative activity in AGS and ASPC1 cancer cell lines, with IC₅。 values ranging from 0.2 to 1.8 µM. The protein binding assay also demonstrated high affinity for KRAS-G12D, with dissociation constants (Kd) of 146 nM for GD-2 and 3.18 nM for GD-4. Mechanistic investigations revealed that both compounds significantly reduced downstream, as evidenced by a clear decrease in phospho-ERK expression. Additionally, molecular dynamics simulations confirmed stable binding interactions within the KRAS-G12D pocket. Collectively, these findings identify GD-2 and GD-4 as promising therapeutic candidates for KRAS-G12D-driven cancers.

TGF-β promotes progression and metastasis in later stages of tumour development, and inhibitors targeting TGF-β or its receptor have faced clinical limitations due to toxicity and poor selectivity. This study aimed to identify novel TGFβR1 inhibitors by screening the ChemDiv database using a structure-based virtual screening approach. Among the top-ranked compounds, 3282-0487 showed the highest potency. Its analogues were further evaluated, leading to four potent TGFβR1 inhibitors with sub-micromolar IC50 values. Molecular docking confirmed favourable binding interactions, and structure-activity relationship analysis highlighted key structural features contributing to inhibitory activity. Among these, compound 3282-0486 demonstrated the lowest IC50 values against colorectal cancer cells, inducing apoptosis and dose-dependent anti-migration effects. Its efficacy was further supported by changes in downstream TGFβR1 signalling, including p-Smad2, EMT markers, and PARP1 cleavage. Additionally, compound 3282-0486 exhibited selectivity for TGFβR1. Overall, these findings support compound 3282-0486 as a promising TGFβR1 inhibitor with therapeutic potential.

Canine mammary tumors (CMTs) are the most common neoplasms in intact female dogs, yet early detection remains challenging due to the lack of clinically validated, noninvasive biomarkers. This study aimed to develop a noninvasive diagnostic model for CMT detection by integrating serum autoantibody biomarkers with machine learning. Serum samples from 154 dogs with mammary tumors (31 benign, 123 malignant) and 39 healthy controls were analyzed using a custom multiplex immunoassay detecting autoantibodies against AGR2, HAPLN1, IGFBP5, and TYMS, normalized to anti-BSA levels. Median fluorescence intensity (MFI), standardized autoantibody ratios, and their combination, together with clinical variables, were used to train random forest classifiers. The model based on standardized autoantibody ratios achieved the best performance, with an AUC of 0.79 (sensitivity 75.3%, specificity 74.4%) for overall CMT detection; 0.78 (92.7%, 61.5%) for malignant CMTs; and 0.77 (82.2%, 71.8%) for early-stagemalignancies. Assuming a CMT prevalence of 0.5 in the hospital-referred population, the positive and negative predictive values ranged from 0.74-0.75 and 0.75-0.91, respectively. This proof-of-concept study demonstrates that a machine learning-assisted multiplex autoantibody assay offers a feasible noninvasive approach for CMT detection. Further validation in larger, independent cohorts is warranted to support clinical translation in veterinary oncology.

Seasonal influenza A viruses evolve rapidly, yet the emergence and molecular basis of resistance to the polymerase acidic (PA) inhibitor baloxavir marboxil (BXM), which is widely used in China, remain elusive. To address this, 3938 PA gene sequences were collected from influenza patients across mainland China between 2018 and 2025, from the national surveillance network and GISAID. By screening post-market mutations in the N-terminal domain of PA (PAN) that appeared in at least two samples and at a frequency below 50%, twenty-five single-point mutations were identified and additionally six linked mutations potentially associated with drug pressure. The impact of these mutations on BXM sensitivity was subsequently evaluated. Our analysis revealed the emergence of known mutations associated with reduced BXM sensitivity, including L28P, K34R, E198 K, although their prevalence remained low (2/3850, 0.05%). Notably, we identified a novel substitution, D27G, which conferred an approximately 12.4-fold reduction in BXM susceptibility compared with the wild-type virus and exhibited higher replication fitness than the canonical resistance mutation I38T, as demonstrated in human airway organoids. Molecular dynamics simulations further indicated that PA-D27G attenuates the interaction between PA and baloxavir acid, the active form of BXM. Epidemiological analysis showed that D27G mutation remained rare, being detected in four isolates (4/1247, 0.32%) in mainland China, and at a sporadic prevalence (<0.1%, 9/53132) across global isolates. In conclusion, these results demonstrate the early emergence of BXM-associated resistance in China and identify PA-D27G as a resistance-associated mutation with preserved viral fitness, underscoring the importance of continued genomic and epidemiologic surveillance.

Microscopy-based cancer cell classification traditionally relies on cell-based morphological features, while subcellular organelle organization remains underutilized. Existing machine learning methods often require manual preprocessing and handcrafted feature extraction, limiting scalability and introducing user bias. This study proposes an automated, interpretable, and organelle-focused deep learning framework for classifying breast cancer cell lines from high-resolution fluorescence microscopy images. We developed an end-to-end framework that incorporates patch-based sampling, sparsity filtering, and a channel-wise intermediate fusion strategy to independently extract and integrate organelle-specific features. Model interpretability was assessed using Grad-CAM visualizations and single-organelle classifier analyses. The framework was evaluated on fluorescence microscopy images from six breast cancer cell lines using 5-fold cross-validation. The proposed framework achieved a classification accuracy of 97.1 ± 1.1 %, performing comparably to or exceeding conventional handcrafted feature-based approaches while eliminating the need for manual segmentation and 3D rendering steps. Interpretability and classifier analyses revealed inter-organelle dependencies and mitochondria as the most informative contributors to classification decisions. Organelle morphology and spatial organization provide strong discriminative signals for cancer cell classification. The proposed framework offers a scalable, automated, and interpretable deep learning solution that advances microscopy-based phenotyping and supports broader applications in computational pathology and cellular informatics.

Radiation pneumonitis (RP) is a serious complication in lung cancer patients with pre-existing interstitial lung disease (ILD) undergoing radiotherapy. Accurate risk stratification is crucial for individualized management. But predictive models integrating multimodal data are lacking. This study aimed to develop a novel machine learning-based nomogram integrating clinical, dosimetric, and inflammatory predictors for RP risk assessment in this high-risk population. This retrospective study of 424 ILD patients collected clinical, dosimetric, and inflammatory data. Machine learning algorithms created composite dosimetric (D score) and inflammatory (Inflamm score) scores. A multivariable logistic regression nomogram was built incorporating these scores with clinical risk factors. Model performance was assessed using area under the curve (AUC), calibration curve, and decision curve analysis (DCA). RP occurred in 200 (47%) patients. Independent risk factors included higher performance status, Charlson comorbidity index (CCI), usual interstitial pneumonia (UIP) pattern, immunotherapy, concurrent chemoradiotherapy, more radiation sessions, and lower lung volume. The D score and Inflamm score were both independent predictors. The integrated nomogram (AUC = 0.929) showed excellent discrimination, significantly outperforming the clinical model (AUC = 0.86), D score (AUC = 0.758) (both p < 0.001), and Inflamm score (AUC = 0.910, p = 0.168). Calibration curve and DCA confirmed its strong calibration ability and clinical utility to identify high-risk patients early. The integrated nomogram combining clinical, dosimetric, and inflammatory predictors enables accurate, individualized RP risk assessment in lung cancer patients with ILD. It can guide adjustments to individualized radiotherapy plans or preventive interventions, supporting better patient selection, treatment decisions, and proactive follow-up.

The growing burden of multimorbidity challenges the healthcare system due to increased healthcare utilization and uncoordinated care. Identifying patients with multimorbidity who have high healthcare utilization is essential to improve management and reduce pressure on the healthcare system. This study aims to identify and characterize clusters of patients with multimorbidity based on both their chronic conditions and healthcare utilization patterns. A weighted K-means method was applied to a population of 1,184,334 individuals with two or more out of 33 chronic conditions, defined using diagnostic algorithms based on ICD-10 and ATC codes. Sociodemographic variables were applied to describe the identified clusters. Four clusters were identified based on chronic conditions and healthcare utilization. Cluster 1 had the highest healthcare utilization and a high burden of both somatic and mental conditions, combined with low social status. Cluster 2, consisting primarily of younger women with mental conditions, showed high use of psychological services, and few somatic conditions. The largest cluster, cluster 3, had low healthcare utilization and consisted of individuals with common, manageable conditions, and relatively high social status. Cluster 4 was defined by older individuals with complex somatic conditions requiring frequent contact with general practitioners and specialists. The identified clusters showed varying chronic condition patterns and levels of healthcare utilization. The findings underscore the importance of tailored strategies, particularly for multimorbidity patients with mental conditions taking social status into account, in order to improve care and manage resource use more effectively.

To explore the predictive value of machine learning-based multimodal MRI radiomics combined with clinical features in the efficacy of high-intensity focused ultrasound (HIFU) ablation of uterine fibroids. This study included 390 patients with uterine fibroids who underwent HIFU ablation. Patients were stratified into high and low ablation groups based on an 80% non-perfused volume ratio (NPVR) and randomly divided into training (70%) and test (30%) sets. Radiomics features were extracted from T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), and diffusion-weighted imaging (DWI). The most predictive features were selected via Recursive Feature Elimination (RFE) and the Least Absolute Shrinkage and Selection Operator (LASSO), and combined with clinical characteristics. Logistic Regression (LR), Random Forest (RF) and eXtreme Gradient Boosting (XGBoost), were constructed to predict ablation efficacy, with performance assessed using the area under the receiver operating characteristic curve (AUC). The results indicated that age, uterine fibroid location, and T2WI signal intensity were independent predictive factors (p < 0.05). The multimodal-clinical fusion XGBoost model exhibited the optimal performance. In the test set, this model achieved an AUC of 0.894, with an accuracy of 82.1%, sensitivity of 88.9%, and specificity of 74.1%.The calibration curve and decision curve analysis (DCA) confirmed that the predicted probabilities of the model were highly consistent with the actual risks, and stable clinical net benefits were achieved. The XGBoost model based on multimodal MRI and clinical features may serve as a reference for predicting HIFU ablation efficacy and optimizing treatment strategies.