Development and Application of a Novel Machine Learning Model Predicting Pancreatic Cancer-Specific Mortality DOI Open Access

Yongji Sun,

Sien Hu, Xiawei Li

et al.

Cureus, Journal Year: 2024, Volume and Issue: unknown

Published: March 29, 2024

Precise prognostication is vital for guiding treatment decisions in people diagnosed with pancreatic cancer. Existing models depend on predetermined variables, constraining their effectiveness. Our objective was to explore a novel machine learning approach enhance prognostic model predicting cancer-specific mortality and, subsequently, assess its performance against Cox regression models. Datasets were retrospectively collected and analyzed 9,752 patients cancer surgery performed. The primary outcomes the of carcinoma at one year, three years, five years. Model discrimination assessed using concordance index (C-index), calibration Brier scores. Survival Quilts compared clinical use, decision curve analysis done. demonstrated robust one-year (C-index 0.729), three-year 0.693), five-year 0.672) mortality. In comparison models, exhibited higher C-index up 32 months but displayed inferior after 33 months. A subgroup conducted, revealing that within subset individuals without metastasis, showcased significant advantage over cohort metastatic cancer, outperformed before 24 weaker 25 This study has developed validated learning-based predict outperforms model.

Language: Английский

Single-cell transcriptome analysis revealed heterogeneity in glycolysis and identified IGF2 as a therapeutic target for ovarian cancer subtypes DOI Creative Commons

Jinting Ji,

Fangfang Bi,

Xiaocui Zhang

et al.

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

Published: July 31, 2024

As the most malignant tumor of female reproductive system, ovarian cancer (OC) has garnered increasing attention. The Warburg effect, driven by glycolysis, accounts for cell proliferation under aerobic conditions. However, metabolic heterogeneity linked to glycolysis in OC remains elusive.

Language: Английский

Citations

2
Regulatory T cells-related gene in primary sclerosing cholangitis: evidence from Mendelian randomization and transcriptome data DOI Creative Commons
Jianlan Hu,

Youxing Wu,

D. Zhang

et al.

Genes and Immunity, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 5, 2024

The present study utilized large-scale genome-wide association studies (GWAS) summary data (731 immune cell subtypes and three primary sclerosing cholangitis (PSC) GWAS datasets), meta-analysis, two PSC transcriptome to elucidate the pivotal role of Tregs proportion imbalance in occurrence PSC. Then, we employed weighted gene co-expression network analysis (WGCNA), differential analysis, 107 combinations 12 machine-learning algorithms construct validate an artificial intelligence-derived diagnostic model (Tregs classifier) according average area under curve (AUC) (0.959) cohorts. Quantitative real-time polymerase chain reaction (qRT-PCR) verified that compared control, Akap10, Basp1, Dennd3, Plxnc1, Tmco3 were significantly up-regulated mice yet expression level Klf13, Scap was lower. Furthermore, infiltration functional enrichment revealed significant associations hub Tregs-related with M2 macrophage, neutrophils, megakaryocyte-erythroid progenitor (MEP), natural killer T (NKT), scores autophagic death, complement coagulation cascades, metabolic disturbance, Fc gamma R-mediated phagocytosis, mitochondrial dysfunction, potentially mediating onset. XGBoost algorithm SHapley Additive exPlanations (SHAP) identified AKAP10 KLF13 as optimal genes, which may be important target for

Language: Английский

Citations

2
Endoplasmic reticulum stress in breast cancer: a predictive model for prognosis and therapy selection DOI Creative Commons
Bin Yang, Shu Wang, Yanfang Yang

et al.

Frontiers in Immunology, Journal Year: 2024, Volume and Issue: 15

Published: Feb. 19, 2024

Background Breast cancer (BC) is a leading cause of mortality among women, underscoring the urgent need for improved therapeutic predictio. Developing precise prognostic model crucial. The role Endoplasmic Reticulum Stress (ERS) in suggests its potential as critical factor BC development and progression, highlighting importance models tailored treatment strategies. Methods Through comprehensive analysis ERS-related gene expression BC, utilizing both single-cell bulk sequencing data from varied subtypes, we identified eight key genes. LASSO regression machine learning techniques were employed to construct model, validated across multiple datasets compared with existing predictive accuracy. Results developed ERS-model categorizes patients into distinct risk groups significant differences clinical prognosis, confirmed by robust ROC, DCA, KM analyses. forecasts survival rates high precision, revealing immune infiltration patterns responsiveness between groups. Notably, discovered six druggable targets Methotrexate Gemcitabine effective agents high-risk treatment, based on their sensitivity profiles addressing lack active BC. Conclusion Our study advances research establishing link ERS prognosis at molecular cellular levels. By stratifying risk-defined groups, unveil disparities cell drug response, guiding personalized treatment. identification opens new avenues targeted interventions, promising enhance outcomes paving way therapy.

Language: Английский

Citations

1
Comprehensive machine learning-based preoperative blood features predict the prognosis for ovarian cancer DOI Creative Commons

Meixuan Wu,

Sijia Gu,

Jiani Yang

et al.

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

Published: Feb. 26, 2024

Abstract Purpose Significant advancements in improving ovarian cancer (OC) outcomes have been limited over the past decade. To predict prognosis and improve of OC, we plan to develop validate a robust signature based on blood features. Methods We screened age 33 features from 331 OC patients. Using ten machine learning algorithms, 88 combinations were generated, which one was selected construct risk score (BRS) according highest C-index test dataset. Results Stepcox (both) Enet (alpha = 0.7) performed best dataset with 0.711. Meanwhile, low RBS group possessed observably prolonged survival this model. Compared traditional prognostic-related such as age, stage, grade, CA125, our combined model had AUC values at 3, 5, 7 years. According results model, BRS can provide accurate predictions prognosis. also capable identifying various prognostic stratifications different stages grades. Importantly, developing nomogram may performance by combining stage. Conclusion This study provides valuable machine-learning that be used for predicting individualized

Language: Английский

Citations

1
Development and Application of a Novel Machine Learning Model Predicting Pancreatic Cancer-Specific Mortality DOI Open Access

Yongji Sun,

Sien Hu, Xiawei Li

et al.

Cureus, Journal Year: 2024, Volume and Issue: unknown

Published: March 29, 2024

Precise prognostication is vital for guiding treatment decisions in people diagnosed with pancreatic cancer. Existing models depend on predetermined variables, constraining their effectiveness. Our objective was to explore a novel machine learning approach enhance prognostic model predicting cancer-specific mortality and, subsequently, assess its performance against Cox regression models. Datasets were retrospectively collected and analyzed 9,752 patients cancer surgery performed. The primary outcomes the of carcinoma at one year, three years, five years. Model discrimination assessed using concordance index (C-index), calibration Brier scores. Survival Quilts compared clinical use, decision curve analysis done. demonstrated robust one-year (C-index 0.729), three-year 0.693), five-year 0.672) mortality. In comparison models, exhibited higher C-index up 32 months but displayed inferior after 33 months. A subgroup conducted, revealing that within subset individuals without metastasis, showcased significant advantage over cohort metastatic cancer, outperformed before 24 weaker 25 This study has developed validated learning-based predict outperforms model.

Language: Английский

Citations

1