Machine Learning-Based Pathomics Model to Predict the Prognosis in Clear Cell Renal Cell Carcinoma DOI Creative Commons
Xiangyun Li, Xiaoqun Yang, Xianwei Yang

et al.

Technology in Cancer Research & Treatment, Journal Year: 2024, Volume and Issue: 23

Published: Jan. 1, 2024

Clear cell renal carcinoma (ccRCC) is a highly lethal urinary malignancy with poor overall survival (OS) rates. Integrating computer vision and machine learning in pathomics analysis offers potential for enhancing classification, prognosis, treatment strategies ccRCC. This study aims to create model predict OS ccRCC patients. In this study, data from patients the TCGA database were used as training set, clinical serving validation set. Pathological features extracted H&E-stained slides using PyRadiomics, was constructed non-negative matrix factorization (NMF) algorithm. The model's predictive performance assessed through Kaplan-Meier (KM) curves Cox regression analysis. Additionally, differential gene expression, ontology (GO) enrichment analysis, immune infiltration, mutational conducted investigate underlying biological mechanisms. A total of 368 patients, comprising two subtypes (Cluster 1 Cluster 2) successfully NMF KM revealed that 2 associated worse OS. 76 genes identified between subtypes, primarily involving extracellular organization structure. Immune-related genes, including CTLA4, CD80, TIGIT, expressed 2, while VHL PBRM1 along mutations PI3K-Akt, HIF-1, MAPK signaling pathways, exhibited mutation rates exceeding 40% both subtypes. learning-based effectively predicts differentiates critical roles immune-related CTLA4 pathways offer new insights further research on molecular mechanisms, diagnosis,

Language: Английский

Uncovering the Potential of Pathomics: Prognostic Prediction and Mechanistic Investigation of Pancreatic Cancer DOI
Rixin Su, Xiaohong Zhao, Fabiao Zhang

et al.

Published: Jan. 1, 2025

Language: Английский

Citations

0
MALDI imaging combined with two-photon microscopy reveals local differences in the heterogeneity of colorectal cancer DOI Creative Commons

Arora Bharti,

Kulkarni Ajinkya,

Markus M. Andrea

et al.

npj Imaging, Journal Year: 2024, Volume and Issue: 2(1)

Published: Sept. 23, 2024

Abstract Colorectal cancer (CRC) remains a leading cause of cancer-related mortality worldwide, accentuated by its heterogeneity and complex tumour microenvironment (TME). The role TME on pathophysiology is pivotal, especially the influence components extracellular matrix (ECM), such as collagen. We introduce novel multimodal imaging strategy to unravel spatial CRC integrating features from two-photon laser scanning microscopy (2PLSM) histology with proteomics signatures matrix-assisted desorption ionization-mass spectrometry (MALDI MSI). Our study first correlate structural coherence collagen fibres nuclei distribution profile tissue peptide signatures, offering insights into proteomic landscape within regions high (HND), well chaotic organised use this approach distinguish patient tissues originating left-sided colorectal (LSCC) right-sided (RSCC). This discriminative signature highlights architecture in progression. Complementary m/z values several proteins associated ECM, plectin, vinculin, vimentin, myosin, have shown differentially intensity distributions between LSCC RSCC. findings demonstrate potential combining information identify molecular different retrieve new pathophysiology.

Language: Английский

Citations

1
Development of a neoadjuvant chemotherapy efficacy prediction model for nasopharyngeal carcinoma integrating magnetic resonance radiomics and pathomics: a multi-center retrospective study DOI Creative Commons
Yiren Wang, Huaiwen Zhang, Huan Wang

et al.

BMC Cancer, Journal Year: 2024, Volume and Issue: 24(1)

Published: Dec. 5, 2024

This study aimed to develop and validate a predictive model for assessing the efficacy of neoadjuvant chemotherapy (NACT) in nasopharyngeal carcinoma (NPC) by integrating radiomics pathomics features using particle swarm optimization-supported support vector machine (PSO-SVM). A retrospective multi-center was conducted, which included 389 NPC patients who received NACT from three institutions. Radiomics were extracted magnetic resonance imaging scans, while derived histopathological images. total 2,667 254 initially extracted. Feature selection involved intra-class correlation coefficient evaluation, Mann-Whitney U test, Spearman analysis, least absolute shrinkage operator regression. The PSO-SVM constructed validated 10-fold cross-validation on training set further evaluated an external validation set. Model performance assessed area under curve (AUC) receiver operating characteristic curve, calibration curves, decision analysis. Eight significant (five pathomics) identified. radiopathomics achieved superior compared models based solely or features. AUCs 0.917 (95% CI: 0.887–0.948) internal 0.814 0.742–0.887) validation. Calibration curves demonstrated good agreement between predicted probabilities actual outcomes. Decision analysis showed that provided higher clinical net benefit over wider range risk thresholds other models. effectively integrates features, offering enhanced accuracy utility NPC. approach robust underscore its potential personalized treatment planning, supporting improved decision-making patients.

Language: Английский

Citations

0
Machine Learning-Based Pathomics Model to Predict the Prognosis in Clear Cell Renal Cell Carcinoma DOI Creative Commons
Xiangyun Li, Xiaoqun Yang, Xianwei Yang

et al.

Technology in Cancer Research & Treatment, Journal Year: 2024, Volume and Issue: 23

Published: Jan. 1, 2024

Clear cell renal carcinoma (ccRCC) is a highly lethal urinary malignancy with poor overall survival (OS) rates. Integrating computer vision and machine learning in pathomics analysis offers potential for enhancing classification, prognosis, treatment strategies ccRCC. This study aims to create model predict OS ccRCC patients. In this study, data from patients the TCGA database were used as training set, clinical serving validation set. Pathological features extracted H&E-stained slides using PyRadiomics, was constructed non-negative matrix factorization (NMF) algorithm. The model's predictive performance assessed through Kaplan-Meier (KM) curves Cox regression analysis. Additionally, differential gene expression, ontology (GO) enrichment analysis, immune infiltration, mutational conducted investigate underlying biological mechanisms. A total of 368 patients, comprising two subtypes (Cluster 1 Cluster 2) successfully NMF KM revealed that 2 associated worse OS. 76 genes identified between subtypes, primarily involving extracellular organization structure. Immune-related genes, including CTLA4, CD80, TIGIT, expressed 2, while VHL PBRM1 along mutations PI3K-Akt, HIF-1, MAPK signaling pathways, exhibited mutation rates exceeding 40% both subtypes. learning-based effectively predicts differentiates critical roles immune-related CTLA4 pathways offer new insights further research on molecular mechanisms, diagnosis,

Language: Английский

Citations

0