Cancer Informatics最新文献

Objectives: The objective of the Prediction Augmented Screening Initiative (PASI) pilot application was to design and implement a clinical tool to optimize the lung cancer screening (LCS) workflow for providers. The Boston Informatics Group (BIG) at the Department of Veterans Affairs (VA) developed the Enabling Technologies for Rapid Learning Health Systems Platform (ENTHRALL) to support delivery of knowledge in a Learning Health System (LHS) framework. The BIG leveraged ENTHRALL to implement the PASI pilot application on a very short timeline. The application uses VA data to estimate patients' benefit from LCS based on National Cancer Institute (NCI) models, allowing proactive outreach to patients with high predicted benefit from LCS.

Methods: The application was designed utilizing ENTHRALL infrastructure, including optimized nightly data pulls to gather patient information, Natural Language Processing to extract smoking history, and a user interface (UI). Cross-functional collaboration allowed the use of the NCI's peer-reviewed prediction algorithm to provide daily patient benefit scores.

Results: The UI displays patients in descending order of benefit, delivering a prioritized list to providers. Clinicians can fill in information and track patient status to assist with their outreach activities. For the pilot, only patients meeting USPSTF LCS criteria (the current field standard) were displayed. Five VA stations were included.

Conclusions: Utilizing the VA BIG's ENTHRALL framework for an LHS, the group demonstrated their ability to design and deliver a new application within 3 months of inception, which was successfully utilized at 5 VA hospitals. The VA's capability to rapidly build clinically relevant applications will help it become an LHS tailored to current problems impacting the Veteran. Due to the success of the pilot, the clinical research team got approval to expand their study. The BIG is working on a non-pilot build.

Background: Carbonic anhydrase 9 (CA9) plays a crucial role in pH regulation and adaptation under hypoxic conditions in the tumor microenvironment. Despite its known involvement in the progression of specific cancers, a comprehensive pan-cancer examination of the prognostic value and biological implications of CA9 has not been performed. This study systematically explored the diverse roles of CA9 across multiple cancer types.

Methods: Bioinformatics methods were applied via extensive datasets from TCGA, GTEx, CPTAC, CancerSEA, and the public literature. We systematically analyzed the associations between CA9 expression profiles and various clinical parameters, prognosis, immune infiltration, immune-related genes, TMB, MSI, and tumor stemness scores. Additionally, a single-cell functional analysis was conducted.

Results: CA9 was significantly upregulated in 29 out of 33 cancer types, indicating high discriminatory ability between tumor and normal tissues. Elevated CA9 expression correlated with poor OS and PFIs in multiple cancers, such as GBMLGG, CESC, LUAD, KIPAN, GBM, THYM, LIHC, THCA, PAAD, and KICH. In 39 cancers, CA9 expression was predominantly negatively correlated with the infiltration of 22 immune cell infiltrations. It was also associated with TMB in 12 tumors and with MSI in 9. Single-cell analysis revealed positive links between CA9 and essential processes such as hypoxia, metastasis, angiogenesis, and stemness.

Conclusion: This study provides compelling evidence that CA9 is a potential pan-cancer prognostic marker and diagnostic tool. The associations of CA9 with immune components and determinants of immunotherapy response indicate the importance of CA9 in advancing cancer research and personalized treatment strategies.

Objectives: Multimodal deep learning models have the potential to significantly improve survival predictions and treatment planning for cancer patients. These models integrate diverse data modalities using early, intermediate, or late fusion techniques. However, many existing multimodal models either underperform or show only marginal improvements over unimodal models. To establish the true efficacy of multimodal survival prediction models, it is critical to demonstrate consistent and substantial advantages over unimodal counterparts.

Methods: In this paper, we introduce the Robust Multimodal Survival Model (RMSurv), a novel discrete late fusion model that leverages synthetic data generation to compute time-dependent weights for various modalities. RMSurv utilizes up to 6 distinct data modalities from The Cancer Genome Atlas Program (TCGA) non-small cell lung cancer and the TCGA pan-cancer datasets to predict overall survival over a period of 10 years. The key innovations of RMSurv are the calculation of time-dependent late fusion weights using a synthetically generated dataset and a new statistical feature normalization technique to enhance the interpretability and accuracy of discrete survival predictions. We evaluate the performance of the proposed method and several alternatives with cross validation using the concordance index, and vary the number of modalities included. We also create a late fusion simulation to highlight the complex relationships of multimodal fusion.

Results: In our experiments, RMSurv outperforms the best unimodal model's Concordance index (C-Index) by 0.0273 on the 6-modal TCGA Lung Adenocarcinoma (LUAD) dataset. Existing late and early fusion methods improved the C-index by only 0.0143 and 0.0072, respectively. RMSurv also performs best on the combined TCGA non-small-cell lung cancer dataset and the TCGA pan-cancer dataset.

Conclusions: These advancements underscore RMSurv's potential as a powerful approach for survival prediction, establishing robust multimodal benefits, and setting a new benchmark for survival prediction models in pan-cancer settings.

The complexity of cancer has long challenged the medical community, driving the need for improved early detection and treatment. Artificial intelligence (AI) has profoundly impacted oncology research in recent decades, resulting in innovative diagnostic and therapeutic approaches. This review synthesizes the critical applications of AI in oncology, focusing on 4 key areas: medical imaging, digital pathology, robotic surgery, and drug discovery. We highlight the role of AI in cancer diagnosis and treatment by reviewing key studies and machine learning methods, and we address the field's current technical and ethical challenges. AI models have significantly enhanced the accuracy of medical imaging by efficiently detecting lesions and disease sites, leading to earlier and more precise diagnoses. In digital pathology, AI tools aid in risk prediction and facilitate the examination of extensive tissue sample sets for patterns and markers, simplifying the pathologists' tasks. AI-powered robotic surgery provides different levels of automation, leading to precise and minimally invasive procedures that not only improve surgical outcomes but also lower readmission rates, hospital stays, and infection risks. Moreover, AI expedites the process of discovering cancer therapies by identifying potential lead compounds, predicting drug reactions, and repurposing current medications. In the past decade, several AI-developed drugs have successfully entered clinical trials. These significant advancements underscore the expanding role of AI in shaping the future of cancer diagnosis and treatment. Although standardization, transparency, and equitable implementation must be addressed, AI brings hope for more personalized and effective therapies.

Objective: The aim of this study was to investigate the role and mechanism of CD79A⁺ B cells in mediating the microenvironment of breast cancer and the relationship with the prognosis of breast cancer.

Methods: Single-cell RNA sequencing and bulk RNA sequencing analysis were combined to annotate breast cancer cell subtypes, perform cell communication and trajectory analysis. CD79A-related signature was constructed by LASSO and multivariate Cox analysis. CD79A⁺ B cell subsets in the tumor microenvironment were explored by immunoanalysis and multiple immunofluorescence analysis.

Results: There were communication relationships between CD79A⁺ B cells and multiple cell types. A prognostic risk signature containing 6 genes was constructed by combining the TCGA dataset. The immune profile analysis showed that the low-risk group showed a higher immune response. In addition, multiple immunofluorescence analysis showed an attraction between CD79A⁺ B cells and tumor cells, and patients with high CD79A⁺ B cells expression had significantly higher survival rates.

Conclusion: This study comprehensively explored the heterogeneity of CD79A⁺ B cells through transcriptome analysis and chromatin analysis, which contributes to an in-depth understanding of the function of CD79A⁺ B cells in biological processes as well as the molecular mechanism of breast carcinogenesis, providing a theoretical basis for treatment and prevention.

Objective: Osteosarcoma (OS) is a rare and complex form of cancer that mostly affects children and adolescents. Pain is a common symptom for patients in OS which causes significant unhappiness and persistent aches. To date, there is minimal knowledge on the mechanisms underlying OS induced pain and few treatment options for patients. Previous genetic studies have demonstrated that the panel of four genes, artemin (ARTN), persephin (PSPN), glial cell line-derived neurotropic factor (GDNF), and neurturin (NRTN) are associated with the regulation of pain processing in OS and analgesic responses.

Methods: In the present study, by utilising a scRNA-seq OS dataset, we aimed to measure the gene expression levels of four pain related genes, and compare them between the different cell types in human OS tissues and cell lines.

Results: Within a complex and diverse range of cell types in OS tissues, including osteoblastic OS cells, carcinoma associated fibroblasts (CAFs), B cells, myeloid cells 1, myeloid cells 2, NK/T cells, plasmocytes, ARTN and NRTN genes had the highest expression in Osteoblastic OS cells, GDNF gene had a peak expression in carcinoma associated fibroblasts, and PSPN gene in endothelial cells. In addition, all four genes showed deferential pattern of expression in 16 OS cell lines.

Conclusion: Future studies should investigate the potential to target deferentially expressed pain-related genes in specific cell types of OS for therapeutic benefit to improve the quality of life for patients living with pain caused by OS.

Background: Gastric cancer is aggressive with poor prognosis due to high invasion and metastasis rates, a hallmark of cancer. The Snail family (SNAI1 and SNAI2) drives EMT, enabling epithelial cells to gain migratory and invasive traits.

Methods: We used "limma" package to identify genes with differential expression between high and low levels of SNAI1/SNAI2 in TCGA stomach adenocarcinoma dataset, intersecting these with cancer invasion and metastasis genes obtained from 5 databases. Using Cox regression analysis, we developed a risk score model and created a nomogram incorporating clinical data. The model's prognostic accuracy was validated with survival and ROC analyses in both TCGA and GEO datasets. Additionally, we performed WGCNA and constructed a ceRNA network to investigate gene interactions, and used CIBERSORT analysis to evaluate immune cell composition in the tumor microenvironment.

Results: We developed 5 and 9 risk signatures and nomograms incorporating clinical data. Survival analysis showed high-risk patients had worse overall survival than low-risk patients. WGCNA identified a lightyellow module associated with SNAI1 and SNAI2 expressions, emphasizing extracellular matrix organization. CeRNA network analyses found 6 common hub genes linked to SNAI1 and SNAI2. Immune profiling showed that SNAI1 expression was related to 8 types of immune cells, while SNAI2 was connected to 6, indicating their roles in influencing the tumor microenvironment.

Conclusion: This study highlights the significant prognostic impact of SNAI1 and SNAI2 in stomach adenocarcinoma, linking their high expression to poorer survival and aggressive tumor behavior, while also identifying potential therapeutic targets through comprehensive computational analysis.

Objectives: Early skin cancer detection in primary care settings is crucial for prognosis, yet clinicians often lack relevant training. Machine learning (ML) methods may offer a potential solution for this dilemma. This study aimed to develop a neural network for the binary classification of skin lesions into malignant and benign categories using smartphone images and clinical data via a multimodal and transfer learning-based approach.

Methods: We used the PAD-UFES-20 dataset, which included 2298 sets of lesion images. Three neural network models were developed: (1) a clinical data-based network, (2) an image-based network using a pre-trained DenseNet-121 and (3) a multimodal network combining clinical and image data. Models were tuned using Bayesian Optimisation HyperBand across 5-fold cross-validation. Model performance was evaluated using AUC-ROC, average precision, Brier score, calibration curve metrics, Matthews correlation coefficient (MCC), sensitivity and specificity. Model explainability was explored using permutation importance and Grad-CAM.

Results: During cross-validation, the multimodal network achieved an AUC-ROC of 0.91 (95% confidence interval [CI] 0.88-0.93) and a Brier score of 0.15 (95% CI 0.11-0.19). During internal validation, it retained an AUC-ROC of 0.91 and a Brier score of 0.12. The multimodal network outperformed the unimodal models on threshold-independent metrics and at MCC-optimised threshold, but it had similar classification performance as the image-only model at high-sensitivity thresholds. Analysis of permutation importance showed that key clinical features influential for the clinical data-based network included bleeding, lesion elevation, patient age and recent lesion growth. Grad-CAM visualisations showed that the image-based network focused on lesioned regions during classification rather than background artefacts.

Conclusions: A transfer learning-based, multimodal neural network can accurately identify malignant skin lesions from smartphone images and clinical data. External validation with larger, more diverse datasets is needed to assess the model's generalisability and support clinical adoption.

Objective: Hepatocellular carcinoma (HCC) is the most common primary liver cancer mainly caused by hepatitis viral infection. Early stage diagnosis is still challenging due to its asymptomatic behavior so there is an urgent need for effective biomarkers. This study aimed to identify effective diagnostic biomarker or therapeutic target for HCC.

Method: Label-free quantitative mass spectrometry was performed to analyze protein expression in HCC and control tissues. Protein-protein interaction (PPI) analysis was done using the STRING database and hub proteins were identified by Cytohubba. The survival analysis and expressions profiling of hub proteins were performed by using GEPIA. Functional and pathway enrichment analysis were carried out using Gene Ontology (GO) and Kyoto Encyclopedia of Gene and Genome (KEGG).

Results: A total of 1539 proteins were identified, of which 116 were differentially expressed proteins (DEPs). PPI network analysis revealed 10 hub proteins; EGFR, GAPDH, HSP90AA1, MMP9, PTPRC, CD44, ANXA5, PECAM1, MMP2, and CDK1. Among these, GAPDH, MMP9, ANXA5, HSP90AA1, and CDK1 were significantly associated with low survival rate (p ⩽ .05). Moreover, MMP9 and CDK1 were showed significantly increased expression in tumor tissues as compared to control (p ⩽ .05). The GO analysis based on biological process, cellular components and molecular function indicated that DEPs were enriched in stress response, vesicle and extracellular space, protein binding and enzyme activity. The KEGG pathway analysis showed that the thyroid hormone synthesis pathway is the most enriched.

Conclusion: The hub proteins GAPDH, HSP90AA1, MMP9, ANXA5, and CDK1 demonstrated significant prognostic potential, could be used as promising theragnostic biomarkers for HCC.

Hepatitis B virus (HBV) causes liver cancer, which is the third most common cause of cancer-related death worldwide. Chronic inflammation via HBV in the host hepatocytes causes hepatocyte remodeling (hepatocyte transformation and immortalization) and hepatocellular carcinoma (HCC). Recognizing cancer stages accurately to optimize early screening and diagnosis is a primary concern in the outlook of HBV-induced hepatocyte remodeling and liver cancer. Genomic signatures play important roles in addressing this issue. Recently, machine learning (ML) models and bioinformatics analysis have become very important in discovering novel genomic signatures for the early diagnosis, treatment, and prognosis of HBV-induced hepatic cell remodeling and HCC. We discuss the recent literature on the ML approach and bioinformatics analysis revealed novel genomic signatures for diagnosing and forecasting HBV-associated hepatocyte remodeling and HCC. Various genomic signatures, including various microRNAs and their associated genes, long noncoding RNAs (lncRNAs), and small nucleolar RNAs (snoRNAs), have been discovered to be involved in the upregulation and downregulation of HBV-HCC. Moreover, these genetic biomarkers also affect different biological processes, such as proliferation, migration, circulation, assault, dissemination, antiapoptosis, mitogenesis, transformation, and angiogenesis in HBV-infected hepatocytes.