Untargeted Lipidomic Biomarkers for Liver Cancer Diagnosis: A Tree-Based Machine Learning Model Enhanced by Explainable Artificial Intelligence DOI Creative Commons
Cemil Çolak, Fatma Hilal Yağın, Abdulmohsen Algarni

и другие.

Medicina, Год журнала: 2025, Номер 61(3), С. 405 - 405

Опубликована: Фев. 26, 2025

Background and Objectives: Liver cancer ranks among the leading causes of cancer-related mortality, necessitating development novel diagnostic methods. Deregulated lipid metabolism, a hallmark hepatocarcinogenesis, offers compelling prospects for biomarker identification. This study aims to employ explainable artificial intelligence (XAI) identify lipidomic biomarkers liver develop robust predictive model early diagnosis. Materials Methods: included 219 patients diagnosed with healthy controls. Serum samples underwent untargeted analysis LC-QTOF-MS. Lipidomic data univariate multivariate analyses, including fold change (FC), t-tests, PLS-DA, Elastic Network feature selection, significant candidate lipids. Machine learning models (AdaBoost, Random Forest, Gradient Boosting) were developed evaluated utilizing these differentiate cancer. The AUC metric was employed optimal model, whereas SHAP utilized achieve interpretability model’s decisions. Results: Notable alterations in profiles observed: decreased sphingomyelins (SM d39:2, SM d41:2) increased fatty acids (FA 14:1, FA 22:2) phosphatidylcholines (PC 34:1, PC 32:1). AdaBoost exhibited superior classification performance, achieving an 0.875. identified 40:4 as most efficacious predictions. d41:2 d36:3 lipids specifically associated risk low-onset elevated levels lipid. Conclusions: demonstrates that lipidomics, conjunction machine learning, may effectively detection results suggest metabolism are crucial progression provide valuable insights incorporating lipidomics into precision oncology.

Язык: Английский

Untargeted Lipidomic Biomarkers for Liver Cancer Diagnosis: A Tree-Based Machine Learning Model Enhanced by Explainable Artificial Intelligence DOI Creative Commons
Cemil Çolak, Fatma Hilal Yağın, Abdulmohsen Algarni

и другие.

Medicina, Год журнала: 2025, Номер 61(3), С. 405 - 405

Опубликована: Фев. 26, 2025

Background and Objectives: Liver cancer ranks among the leading causes of cancer-related mortality, necessitating development novel diagnostic methods. Deregulated lipid metabolism, a hallmark hepatocarcinogenesis, offers compelling prospects for biomarker identification. This study aims to employ explainable artificial intelligence (XAI) identify lipidomic biomarkers liver develop robust predictive model early diagnosis. Materials Methods: included 219 patients diagnosed with healthy controls. Serum samples underwent untargeted analysis LC-QTOF-MS. Lipidomic data univariate multivariate analyses, including fold change (FC), t-tests, PLS-DA, Elastic Network feature selection, significant candidate lipids. Machine learning models (AdaBoost, Random Forest, Gradient Boosting) were developed evaluated utilizing these differentiate cancer. The AUC metric was employed optimal model, whereas SHAP utilized achieve interpretability model’s decisions. Results: Notable alterations in profiles observed: decreased sphingomyelins (SM d39:2, SM d41:2) increased fatty acids (FA 14:1, FA 22:2) phosphatidylcholines (PC 34:1, PC 32:1). AdaBoost exhibited superior classification performance, achieving an 0.875. identified 40:4 as most efficacious predictions. d41:2 d36:3 lipids specifically associated risk low-onset elevated levels lipid. Conclusions: demonstrates that lipidomics, conjunction machine learning, may effectively detection results suggest metabolism are crucial progression provide valuable insights incorporating lipidomics into precision oncology.

Язык: Английский

Процитировано

0