Liver Tumor Prediction using Attention-Guided Convolutional Neural Networks and Genomic Feature Analysis

MethodsX, Journal Year: 2025, Volume and Issue: unknown, P. 103276 - 103276

Published: March 1, 2025

The task of predicting liver tumors is critical as part medical image analysis and genomics area since diagnosis prognosis are important in making correct decisions. Silent characteristics interactions between genomic imaging features also the main sources challenges toward reliable predictions. To overcome these hurdles, this study presents two integrated approaches namely, - Attention-Guided Convolutional Neural Networks (AG-CNNs), Genomic Feature Analysis Module (GFAM). Spatial channel attention mechanisms AG-CNN enable accurate tumor segmentation from CT images while providing detailed morphological profiling. Evaluation with three control databases TCIA, LiTS, CRLM shows that our model produces more output than relevant literature an accuracy 94.5%, a Dice Similarity Coefficient 91.9%, F1-Score 96.2% for Dataset 3. More considerably, proposed methods outperform all other different datasets terms recall, precision, Specificity by up to 10 percent including CELM, CAGS, DM-ML, so on.•Utilization (AG-CNN) enhances region focus accuracy.•Integration (GFAM) identifies molecular markers subtype-specific classification.

Language: Английский

---

Emerging roles of cuproptosis in liver diseases

Digestive and Liver Disease, Journal Year: 2025, Volume and Issue: unknown

Published: April 1, 2025

Intracellular copper levels should be maintained within a controlled range to obtain homeostasis. Cuproptosis, newly discovered form of cell death, occurs when excessive ions bind the lipoylated enzymes in tricarboxylic acid cycle, which leads protein aggregation, proteotoxic stress, and ultimately death. Herein, we summarize current knowledge regarding metabolism, discovery molecular mechanism cuproptosis. In addition, discuss implications cuproptosis pathogenesis various liver diseases, including hepatocellular carcinoma (HCC), Wilson disease (WD), metabolic-associated fatty (MAFLD), fibrosis, hepatic ischemia-reperfusion injury (HIRI) drug-induced (DILI). Understanding can not only provide deeper insights into diseases but also open up new avenues for development targeted therapies.

Language: Английский

---

The molecular mechanism and therapeutic landscape of copper and cuproptosis in cancer

Signal Transduction and Targeted Therapy, Journal Year: 2025, Volume and Issue: 10(1)

Published: May 8, 2025

Abstract Copper, an essential micronutrient, plays significant roles in numerous biological functions. Recent studies have identified imbalances copper homeostasis across various cancers, along with the emergence of cuproptosis, a novel copper-dependent form cell death that is crucial for tumor suppression and therapeutic resistance. As result, manipulating levels has garnered increasing interest as innovative approach to cancer therapy. In this review, we first delineate at both cellular systemic levels, clarifying copper’s protumorigenic antitumorigenic functions cancer. We then outline key milestones molecular mechanisms including mitochondria-dependent independent pathways. Next, explore cuproptosis biology, well interactions mediated by between cells immune system. also summarize emerging opportunities targeting discuss clinical associations cuproptosis-related genes. Finally, examine potential biomarkers put forward existing challenges future prospects leveraging Overall, review enhances our understanding landscape cancer, highlighting copper- or cuproptosis-based therapies treatment.

Language: Английский

---

A prognostic signature of cuproptosis and TCA-related genes for hepatocellular carcinoma

Frontiers in Oncology, Journal Year: 2022, Volume and Issue: 12

Published: Oct. 17, 2022

Hepatocellular carcinoma (HCC) is the most common malignant tumor of liver. Cuproptosis a newly defined form cell death. Copper ion induces death by binding to tricarboxylic acid cycle (TCA). The effect cuproptosis-related and TCA-related genes on clinical prognosis HCC still unclear. In this study, we explores genetic changes that affect TCA process their potential therapeutic value in patients.The cuproptosis were obtained from articles molecular signatures database. eight related constructed using last absolute shrinkage selection operator (LASSO), Receiver Operating Characteristic (ROC) curves used evaluate signature. addition, analyzed downstream functional enrichment immune infiltration explore cuproptosis-inducing drugs immunotherapeutic responses. All these analyses validated multiple datasets International Cancer Genome Consortium (ICGC).TCA copper malnutrition-related (CDKN2A, IDH1, OGDHL, IDH3G, IDH3B, GLS, DLAT, LIPT1) finally included. According risk score, they divided into high-risk low-risk groups. Survival analysis showed overall survival (OS) group was significantly lower than group. We established prognostic feature predict OS patients with HCC. Based stage, nomogram. Functional revealed pathways organelle division cycle. Different scores had different abundances cells (including macrophages regulatory T-cells) antigen-presenting co-stimulation). Moreover, drug sensitivity eleschomol PD-L1 better status TP53 somatic mutation also closely score.In new prediction signature process, which can effectively patients.

Language: Английский

---

Identification and validation of cuproptosis-related LncRNA signatures as a novel prognostic model for head and neck squamous cell cancer

Cancer Cell International, Journal Year: 2022, Volume and Issue: 22(1)

Published: Nov. 11, 2022

Head and neck squamous cell cancer (HNSCC) is a common malignant cancer. We aimed to explore prognostic cuproptosis-related lncRNAs (CRLs) risk models for HNSCC.The transcriptome profiles clinical data were obtained from the TCGA database, 19-cuproptosis-related genes (CRGs) acquired previous studies. Then, model based on seven CRLs was established. analysed its value evaluate prognosis, drug sensitivity, tumour immune functions of patients with HNSCC. Finally, we used quantitative reverse transcription polymerase chain reaction (qRT‒PCR) validate CRLs.We established 7-CRL signature. Kaplan‒Meier (K-M) curve analysis demonstrated significantly preferable prognosis in low-risk group. Multivariate Cox regression revealed that score could serve as an independent factor. Nomogram, ROC curve, principal component indicated signature presented significant predictive capability. Moreover, most high-risk group showed lower levels IC50 certain chemotherapy drugs, such cisplatin, cytarabine, docetaxel, doxorubicin, etoposide, gemcitabine, methotrexate, paclitaxel, dasatinib. expression AP001372.2, MIR9-3HG, AL160314.2, POLH-AS1, AL109936.2 upregulated, while AC090587.1 WDFY3-AS2 downregulated HNSCC lines compared normal by qRT‒PCR.The be novel biomarker predicting

Language: Английский

---

Development and validation of a cuproptosis-associated prognostic model for diffuse large B-cell lymphoma

Frontiers in Oncology, Journal Year: 2023, Volume and Issue: 12

Published: Jan. 12, 2023

Diffuse large B-cell lymphoma (DLBCL) is a highly heterogeneous disease. Therefore, more reliable biomarkers are required to better predict the prognosis of DLBCL. Cuproptosis novel identified form programmed cell death (PCD) that different from oxidative stress-related (e.g., apoptosis, ferroptosis, and necroptosis) by Tsvetkov colleagues in recent study released Science. copper-dependent PCD closely tied mitochondrial metabolism. However, prognostic value cuproptosis-related genes (CRGs) DLBCL remains be further elucidated. In present study, we systematically evaluated molecular changes CRGs found them associated with prognosis. Subsequently, based on expression profiles CRGs, characterized heterogeneity identifying two distinct subtypes using consensus clustering. Two isoforms exhibited survival, biological functions, chemotherapeutic drug sensitivity, immune microenvironment. After differentially expressed (DEGs) between CRG clusters, built model Least absolute shrinkage selection operator (LASSO) Cox regression analysis validated its analysis, Kaplan-Meier curves, receiver operating characteristic (ROC) curves. addition, risk score can clinical characteristics, levels infiltration, Furthermore, nomogram incorporating features was generated optimize stratification quantify assessment. Compared International Prognostic Index (IPI), has demonstrated accuracy survival prediction. gene through external experiments. conclusion, signature serve as potential predictor patients may provide new insights into cancer therapeutic targets.

Language: Английский

---

Cuproptosis- and m6A-Related lncRNAs for Prognosis of Hepatocellular Carcinoma

Biology, Journal Year: 2023, Volume and Issue: 12(8), P. 1101 - 1101

Published: Aug. 8, 2023

Cuproptosis and N6-methyladenosine (m6A) have potential as prognostic predictors in cancer patients, but their roles hepatocellular carcinoma (HCC) are unclear. This study aimed to screen a total of 375 HCC samples were retrieved from the TCGA database, lncRNAs related cuproptosis m6A obtained through correlation analysis. To construct risk assessment model, univariate Cox regression analysis LASSO employed. Analyze regulatory effect relevant models on tumor mutation load (TMB) immune microenvironment. A five (AC007405.3, AL031985.3, TMCC1-AS1, MIR210HG, TMEM220-AS1) with independent overall survival-related by survival regression. TP53 CTNNB1 three genes found most mutations high-risk group patients. The low TMB had worst survival, whereas low-risk high best survival. KEGG pathway revealed that was enriched cell cycle, oocyte meiosis, senescence, glycolysis/glucose production pathways. We constructed reliable cuproptosis- m6A-related lncRNA model for prognosis HCC. may provide new insights into managing further research is needed validate it.

Language: Английский

---

A Pump-Free SERS Microfluidic Chip Based on Identification-Competition Strategy for Ultrasensitive and Efficient Simultaneous Detection of Liver Cancer-Related microRNAs

Biomedical Optics Express, Journal Year: 2024, Volume and Issue: 15(11), P. 6469 - 6469

Published: Oct. 16, 2024

In this study, we present a pump-free SERS microfluidic chip capable of detecting liver cancer-related miR-21 and miR-155 concurrently with ultra-sensitivity high efficiency. We employed Fe 3 O 4 @cDNA-AuNPs@Raman reporter@H composite structure recognition competition strategy. When the target miRNAs (miR-21 miR-155) are in test liquid, they specifically compete nucleic acid complementary strand(H) reporter@H, causing AuNPs to competitively detach from surface , resulting decrease signal. Consequently, enables detection more rapidly accurately complex environments. This method offers an approach for simultaneous efficient holds promising applications early diagnosis cancer.

Language: Английский

---

Analysis of the correlation between non-alcoholic fatty liver disease and the risk of colorectal neoplasms

Frontiers in Pharmacology, Journal Year: 2022, Volume and Issue: 13

Published: Nov. 9, 2022

This study aims at assessing the potential association between non-alcoholic fatty liver disease (NAFLD) and colorectal neoplasms (CRN). PubMed, Cochrane Library, Embase were searched for cohort studies. 14 studies with a total population of 38,761,773 included meta-analysis after selection. The results showed that NAFLD is related to an increased risk CRN (OR = 1.23; 95% CI: 1.14-1.32; I2 70.7%, p < 0.001). In subgroup analysis, found be independent factor adenoma (CRA) 1.29; CI 1.15-1.45; 66.4%) cancer (CRC) 1.13; 1.12-1.15; 69.4%). There no close correlation smoking status patients CRN. Interestingly, bioinformatics analysis revealed there overlap dysregulated gene sets among NAFLD, CRC, two recently identified regulated cell death types, ferroptosis cuproptosis, respectively. Our meta- shows increases Ferroptosis cuproptosis may critical links CRN, These findings here support necessary considered as emerging

Language: Английский

---

Pan‐cancer image segmentation based on feature pyramids and Mask R‐CNN framework

Medical Physics, Journal Year: 2024, Volume and Issue: 51(8), P. 5427 - 5440

Published: March 4, 2024

Abstract Background Cancer, a disease with high mortality rate, poses great threat to patients' physical and mental health can lead huge medical costs emotional damage. With the continuous development of artificial intelligence technologies, deep learning‐based cancer image segmentation techniques are becoming increasingly important in detection accurate diagnosis. However, tasks, there differences efficiency between large small objects limited effects on individual sizes. The previous frameworks still have room for improvement multi‐scale collaboration when segmenting objects. Purpose This paper proposes method train learning framework using feature pyramid processing dataset improve average precision (AP) index, realizes cooperation target segmentation. Objective Pan‐Cancer Histology Dataset Nuclei Instance Segmentation Classification (PanNuke) was selected include approximately 7500 pathology images cells from 19 different types tissues, including five classifications cancer, non‐cancer, inflammation, death, connective tissue. Methods First, uses whole‐slide pan‐cancer histology nuclei instance classification dataset, combined mask region convolutional neural network (Mask R‐CNN) improved loss function segment detect each cellular tissue cancerous sections. Second, address problem non‐synergistic object at scales segmentation, scheme pyramids process adopted as part extraction module. Results Extensive experimental results this show that yields 0.269 AP boost about 4% compared original Mask R‐CNN framework. Conclusions It is effective feasible use data set effect

Language: Английский

---