Scholarlink

<p>To explore the application value of a multi - omics fusion model based on exosome metabolomics and the expression levels of programmed death − 1/programmed death - ligand 1 (PD-1/PD-L1) in the early prediction of the efficacy of immunotherapy for advanced liver cancer, screen key biomarkers affecting the response to immunotherapy, construct and validate a highly efficient prediction model, and provide a basis for clinical individualized treatment decisions. A total of 824 patients with advanced liver cancer who received PD-1/PD-L1 inhibitor treatment were selected. They were randomly divided into a training set and a test set at a ratio of 7:3. General clinical data, exosomal metabolomics data, and PD-1/PD-L1 expression levels of the patients were collected. Key predictive indicators were screened through univariate analysis and multivariate Logistic regression analysis. Multi - omics fusion prediction models were constructed. The performance of the models was evaluated using accuracy, precision, recall, F1 score, and the area under the receiver operating characteristic curve (AUC). Decision curves and calibration curves were plotted to validate the clinical practicality of the models. Multivariate analysis showed that Child - Pugh classification, extra-hepatic metastasis, neutrophil - to - lymphocyte ratio, sphingosine − 1 - phosphate, glutamine, and the expression level of programmed death - ligand 1 were independent predictive factors (all <i>P</i> &lt; 0.05). Among the constructed models, the random forest fusion model had the best performance. Its AUCs in the training set and test set were 0.757 and 0.751 respectively, and its precision, accuracy, recall, F1 score and other indicators were also superior to those of other models. The multi-omics fusion model based on exosome metabolomics and PD-1/PD-L1 expression has shown good early prediction efficacy for the efficacy of immunotherapy in advanced liver cancer. The random forest model performs the best, which provides a single-center preliminary reference for clinicians to identify patients who may benefit from immunotherapy at an early stage, and its clinical value in optimizing treatment plans and improving patients’ prognosis needs to be further confirmed by multi-center external validation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-omics integration model of exosome metabolomics and PD-1/PD-L1 expression for early prediction of immunotherapy efficacy in advanced liver cancer

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月30日

查看全文
  • Caixia Li,
  • Dandan Du,
  • Yanbao Liu,
  • Xuan Wang,
  • Yuchen Feng,
  • Panpan Cui

摘要

To explore the application value of a multi - omics fusion model based on exosome metabolomics and the expression levels of programmed death − 1/programmed death - ligand 1 (PD-1/PD-L1) in the early prediction of the efficacy of immunotherapy for advanced liver cancer, screen key biomarkers affecting the response to immunotherapy, construct and validate a highly efficient prediction model, and provide a basis for clinical individualized treatment decisions. A total of 824 patients with advanced liver cancer who received PD-1/PD-L1 inhibitor treatment were selected. They were randomly divided into a training set and a test set at a ratio of 7:3. General clinical data, exosomal metabolomics data, and PD-1/PD-L1 expression levels of the patients were collected. Key predictive indicators were screened through univariate analysis and multivariate Logistic regression analysis. Multi - omics fusion prediction models were constructed. The performance of the models was evaluated using accuracy, precision, recall, F1 score, and the area under the receiver operating characteristic curve (AUC). Decision curves and calibration curves were plotted to validate the clinical practicality of the models. Multivariate analysis showed that Child - Pugh classification, extra-hepatic metastasis, neutrophil - to - lymphocyte ratio, sphingosine − 1 - phosphate, glutamine, and the expression level of programmed death - ligand 1 were independent predictive factors (all P < 0.05). Among the constructed models, the random forest fusion model had the best performance. Its AUCs in the training set and test set were 0.757 and 0.751 respectively, and its precision, accuracy, recall, F1 score and other indicators were also superior to those of other models. The multi-omics fusion model based on exosome metabolomics and PD-1/PD-L1 expression has shown good early prediction efficacy for the efficacy of immunotherapy in advanced liver cancer. The random forest model performs the best, which provides a single-center preliminary reference for clinicians to identify patients who may benefit from immunotherapy at an early stage, and its clinical value in optimizing treatment plans and improving patients’ prognosis needs to be further confirmed by multi-center external validation.