Radiogenomic analysis for predicting lymph node metastasis and molecular annotation of radiomic features in pancreatic cancer
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.Machine learning and radiomics can predict lymph node metastasis in pancreatic cancer. This approach identifies key features linked to proliferation, aiding in preoperative assessment.
Area Of Science
- Oncology
- Radiology
- Bioinformatics
Background
- Pancreatic cancer often involves lymph node metastasis, impacting prognosis.
- Accurate preoperative prediction of metastasis is crucial for treatment planning.
Purpose Of The Study
- To develop a preoperative prediction model for lymph node metastasis in pancreatic cancer.
- To identify molecular information associated with key radiomic features.
Main Methods
- Utilized two cohorts (151 and 54 patients) of pancreatic cancer.
- Extracted radiomic features and applied 10 machine learning algorithms (77 combinations) for metastasis prediction.
- Performed Weighted Gene Coexpression Network Analysis (WGCNA) and molecular pathway enrichment analysis.
Main Results
- The best radiomics model (StepGBM and Enet combination) achieved AUCs of 0.84 (training) and 0.85 (validation).
- Identified 15 core radiomic features predictive of lymph node metastasis.
- Proliferation-related processes were identified as key molecular alterations linked to these features.
Conclusions
- Machine learning-based radiomics effectively predicts lymph node metastasis status in pancreatic cancer.
- Radiomic features are associated with proliferation-related molecular alterations, offering insights into tumor biology.
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.