Find the papers that actually matter

To develop a deep learning model to differentiate benign from malignant adrenal lesions. A total of 380 patients with 385 pathologically confirmed adrenal lesions (101 malignant, 284 benign) were retrospectively included. Adrenal lesions were manually segmented on CT images and analyzed in a deep learning pipeline aimed at differentiating benign from malignant lesions. Four predictive models were developed that incorporated combinations of radiological data (tumor size, and spontaneous attenuation) and non-radiological data (i.e., medical history and laboratory results). Data of 267 patients were used as a training set and those of 113 patients for the test set. The diagnostic capabilities of the four models were estimated using sensitivity, specificity, accuracy, and areas under the receiver operating characteristic curves (AUC) using histopathological findings as the gold standard. The reproducibility of manual segmentation was estimated using the Dice similarity coefficient after blinded resegmentation of 40 adrenal lesions by an independent radiologist. Segmentation reproducibility achieved a mean Dice similarity coefficient of 0.92 ± 0.03 (range: 0.72-0.97). The most accurate model, which combined clinical, biological, and radiological data, achieved 84.2% accuracy (95% confidence interval: 79.9, 88.6) and an AUC of 0.93 (95% confidence interval: 89.9, 97.9) in the test set for diagnosis of malignant adrenal lesion. A deep learning model integrating preoperative clinical, biological, and radiological features demonstrates high capabilities in differentiating benign from malignant adrenal lesions on initial CT examination.

Triple-negative breast cancer (TNBC) is an aggressive malignancy often characterized by chemoresistance, partly due to the overexpression of drug efflux transporters like ABCG2. To address this challenge, this study developed and evaluated a cRGD-modified, pH-sensitive liposomal system for the targeted co-delivery of docetaxel (DTX) and siRNA against ABCG2 (si-ABCG2). The synthesized nanoparticles (DTX/siRNA/cRGD-PLPs) exhibited optimal physicochemical properties, including a mean particle size of approximately 241.7 nm, efficient co-loading of DTX and siRNA, and pH-responsive cargo release, while protecting the siRNA from degradation in serum. Also, DTX/siRNA/cRGD-PLPs maintained homogeneous size distributions over the storage period and induced minimal hemoglobin release, with hemolysis rates remaining below safety threshold. These liposomes demonstrated enhanced, time-dependent uptake into TNBC cell lines HCC1937 and MDA-MB-231. In vitro, the DTX/siRNA/cRGD-PLPs formulation was significantly more effective at inhibiting cell viability, proliferation, migration, and invasion, and at inducing apoptosis, compared to the free drug combinations and other controls. Moreover, the dual-payload co-delivery liposomes (DTX/siRNA/cRGD-PLPs) exerted superior anti-tumor effects relative to single-agent formulations. In an MDA-MB-231 xenograft mouse model, the liposomal treatment was well-tolerated and resulted in marked tumor growth inhibition, which was associated with reduced cell proliferation (Ki67) and increased apoptosis (Caspase-3) within the tumor tissue. This targeted co-delivery system shows significant potential for improving TNBC treatment by synergistically enhancing therapeutic efficacy and overcoming chemoresistance.

Analyzing gene expression data is essential for predicting and detecting diseases, including cancer. The data is very repetitive and noisy, which makes it hard to find important information about illnesses. In the past decade, several traditional machine learning and feature selection models have been developed for cancer type classification from gene expression data. Rather than introducing new deep learning primitives, this work presents a principled integration framework that combines sparsity-aware representation learning, structure-inducing spatial embedding, and hierarchical multi-scale attention. This paper presents a method for cancer gene classification based on Hierarchical Attention Assisted Feature Pyramid Network (HA-FPN). The work involved two publicly available datasets. The proposed methodology starts with dimensionality reduction through a Variational Sparse Autoencoder (VSAE), followed by an updated DeepInsight algorithm for image conversion of the input. Next, the classification technique is constructed using the proposed HA-FPN model. Moreover, the improved gradient descent optimization (IGDO) is utilized to change the hyperparameter of the classification model. In addition, the results demonstrate that the model combined with an IGDO outperforms the currently existing methods in terms of accuracy, precision, recall, and F1-score. The method that is presented efficiently brings out different aspects of the data through t-SNE computations. Moreover, the proposed approach is very robust, and hence, it can reach high performance levels on two different datasets.

Aristolochic acids (AAs) are established human carcinogens strongly associated with upper tract urothelial carcinoma (UTUC). However, the multi-target oncogenic network beyond their genotoxic mechanism remains incompletely elucidated. This study employed an integrated computational approach combining network toxicology, machine learning, molecular docking, and molecular dynamics (MD) simulations to systematically explore the potential molecular mechanisms of AA-induced UTUC. We identified 97 shared potential targets of AAs and UTUC. Enrichment analyses revealed their significant involvement in lipid metabolism, xenobiotic detoxification, and cancer-related pathways such as PI3K-Akt signaling. Topological analysis of the protein-protein interaction network and a nested cross-validation machine learning model highlighted five core genes: CASP3, EGFR, PARP1, PTGS2, and HSP90AA1. Molecular docking predicted high binding affinities of AA with these core targets, particularly for PTGS2 (-9.3 kcal/mol) and EGFR (-8.2 kcal/mol). Subsequent 100-ns MD simulations and Molecular Mechanics/Poisson-Boltzmann Surface Area (MM/PBSA) calculations confirmed the structural stability and spontaneous binding (ΔG_bind = -55.68 kcal/mol) of the AA-EGFR complex. Our multi-omics analysis suggests that AAs may promote UTUC not only via canonical DNA adduct formation but also potentially through direct interactions with key signaling proteins, implicating a synergistic mechanism involving both genotoxic and non-genotoxic pathways. These findings provide a theoretical foundation for novel preventive and therapeutic strategies against AA-associated UTUC.

Neoadjuvant therapy (NAT) has emerged as a standard treatment strategy for locally advanced breast cancer (BC), yet robust biomarkers for response prediction remain elusive. Here, we established a real-world NAT cohort of 1161 Chinese BC patients, including 1145 cases with matched clinicopathological data and targeted sequencing, to systematically evaluate genomic features associated with NAT outcomes. We identified both cross-subtype and subtype-specific genomic associations with treatment response. PI3K-pathway alterations emerged as a consistent feature of resistance across subtypes, whereas mutations such as ERBB2 in HER2+ disease and MAP3K1 in triple-negative breast cancer were associated with subtype-specific response patterns. Regimen-level analyses further showed that some genomic associations were treatment-context dependent across chemotherapy-, endocrine-, anti-HER2-, and immunotherapy-containing regimens. Among patients with non-pathological complete response (non-pCR), genomic profiling further refined risk stratification for distant recurrence by revealing subtype-specific prognostic alterations, including TOP3B and SETD2. Furthermore, a machine-learning model integrating genomic and clinicopathological features showed favorable performance for NAT response prediction. Overall, our study provides a comprehensive genomic framework for response prediction and recurrence risk assessment, supporting more precise stratification and biomarker-guided treatment optimization in Asian breast cancer patients.

Triple-negative breast cancer (TNBC) is an aggressive subtype lacking effective targeted therapies. Although immune checkpoint inhibitors such as pembrolizumab have improved clinical outcomes in a subset of patients, limited response rates and durability highlight the need for alternative strategies. Tumor-infiltrating lymphocyte (TIL)-based adoptive cell therapy represents a promising approach; however, the absence of reliable biomarkers for TIL functionality remains a major obstacle. In this study, we isolated TILs and autologous tumor cells from nine TNBC patients and classified TILs as reactive or non-reactive based on cytotoxic activity. Transcriptomic profiling combined with weighted gene co-expression network analysis and machine learning approaches identified a 14-gene signature that robustly distinguished reactive from non-reactive TILs. Transcription factor analysis revealed that Friend leukemia integration 1 (FLI1) regulates a large proportion of genes upregulated in non-reactive TILs. Notably, pharmacological inhibition of FLI1 using TK216 restored the cytotoxic function of non-reactive TILs and reduced expression of immunosuppressive genes. These findings identify a functional gene signature associated with TIL reactivity and suggest that targeting FLI1 may represent a strategy to enhance TIL-based immunotherapy in TNBC.

Benzo(a)pyrene (BaP) and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), as typical environmental carcinogens, are widely present in tobacco smoke and air pollution. Their combined exposure is an important cause of high incidence of laryngeal cancer. However, the molecular mechanism of their synergistic carcinogenesis remains unclear. To fill this gap, this study employed an integrated strategy combining network toxicology, multi-dimensional transcriptomics (bulk and single-cell), machine learning, molecular simulation, and cell function verification to systematically explore the mechanism by which BaP and NNK combined exposure induce laryngeal cancer. Through multi-dimensional data mining and machine learning algorithms, the core molecule FADS1 was identified. Functional enrichment analysis revealed that FADS1-related genes mainly participate in lipid metabolism reprogramming and tumor malignant phenotype regulation pathways. Molecular docking and 100 ns kinetic simulation confirmed that both BaP and NNK can stably bind to FADS1, with a stronger binding affinity for BaP (ΔG = -9.0 kilocalories/mole), and the binding mode is mainly based on van der Waals forces and hydrophobic interactions. Cell experiments demonstrated that combined exposure of BaP and NNK can significantly upregulate the expression of FADS1 in laryngeal cancer cells, enhance cell proliferation, migration, and invasion abilities, while silencing FADS1 can effectively reverse these malignant phenotypes. In summary, this study clarified the key role of FADS1 in mediating the synergistic promotion of laryngeal cancer by BaP/NNK, providing experimental support for understanding the carcinogenic mechanism of environmental compound pollutants, and also offering potential targets for risk assessment, early diagnosis, and precise prevention and control of air pollution and tobacco exposure-related laryngeal cancer.

Multi-parametric quantitative magnetic resonance imaging (mqMRI) holds significant clinical potential through multi-parametric tissue characterization, yet its adoption is hindered by prolonged scan time and sensitivity to non-ideal signal conditions, especially in high-resolution whole-brain protocols. To address these challenges, we propose a novel signal encoding method integrating phase-modulated three dimensional steady-state free precession with multiple overlapping-echo detachment (3D SSFP-MOLED). This method simultaneously encodes six physiological parameters (M0, T1, T2, T2*, B1+, ΔB0) into k-space by controlling overlapping echo detachment in signal acquisition. A physics-constrained synthetic data pipeline was developed to simulate MR signal evolutions with realistic field variations (ΔB0, B1+ inhomogeneities), enabling robust training of network for real-time parameter mapping. Whole-brain parametric maps (1×1×2 mm³ resolution) can be delivered within 3 minutes with only 2x parallel acquisition acceleration. Validation was performed on phantom, healthy volunteers, and clinical cases with tumors/hemorrhage. Experimental results show that our method can achieve rapid multi-parametric quantitation with high accuracy and reproducibility. By synergizing adaptive signal encoding, physics-informed synthetic training, and reproducible deep learning reconstruction, this work establishes a new paradigm for efficient and reliable mqMRI in clinical signal processing applications.

Epidermal growth factor receptor (EGFR) is a subclass of tyrosine kinase receptor and it plays an important role in cell growth, proliferation and differentiation in non-small cell lung cancer (NSCLC). Targeting EGFR was proved as an effective approach for the treatment of NSCLC. Herein, we presented the discovery and development of a highly potent and effective EGFR inhibitor, SH-1092 as a potent third generation EGFR inhibitor through cyclization strategy. A methyl group was inserted to the scaffold to finetune the potency and drug-like property. SH-1092 exhibited significant EGFR inhibition with an IC50 of 0.96 nM and it inhibited H1975 (EGFR T790M/L858R) cell growth at an IC50 of 8 nM. In addition, it displayed high metabolic stability in vitro and in vivo. Overall, it represents a promising clinical candidate for NSCLC.

The epidermal growth factor receptor (EGFR) is a critical oncogenic driver in colorectal cancer, establishing the need for novel small-molecule inhibitors. We designed and synthesized two new triazole-grafted quinazolinones, T6 and T7, utilizing click chemistry to efficiently construct their hybrid architecture. Biological evaluations demonstrated that both compounds possess multi-panel antiproliferative activity across nine NCI cancer subtypes, with mean growth inhibition (GI%) exceeding 100% against CNS cancer (128.54% and 87.98%), melanoma (122.75% and 136.30%), and colon cancer (114.34% and 88.22%) for T6 and T7, respectively. In the five-dose screening, T6 and T7 showed notable activity against HT29 colon cancer cells, with GI₅₀ values of 0.53 μM and 0.39 μM, respectively. Both compounds exhibited markedly lower cytotoxicity against normal WI-38 fibroblasts (IC50: 156.59 μM for T6 and 191.80 μM for T7) compared to the reference kinase inhibitor dasatinib (IC50: 37.87 μM), supporting a favorable in-vitro safety profile. Both quinazolinones effectively inhibited EGFR kinase activity with IC50 values of 0.198 μM and 0.131 μM, respectively, compared to 0.046 μM for the reference inhibitor erlotinib. Mechanistic studies revealed that T6 and T7 induce G₀/G₁ cell cycle arrest and significantly trigger apoptosis in HT29 cells, with total apoptosis rates of 67.40% for T6 and 89.62% for T7, versus 30.16% in untreated controls. Both compounds also demonstrated notable anti-migratory effects, with T6 limiting wound closure to 16.85% and T7 to 33.71%, compared to 57.18% in untreated HT29 cells over 48 h. Molecular docking suggested favorable initial EGFR binding of the synthesized quinazolinones, while molecular dynamics analysis supported improved dynamic stability of T7 relative to T6. The results support T6 and T7 as viable and safe candidates for further development in colorectal cancer treatment.