Related Experiment Video
Updated: Oct 29, 2025
10:26
Author Spotlight: A 3D Digital Model for the Diagnosis and Treatment of Pulmonary Nodules
Published on: May 19, 2023
2.1K
Prediction model for malignant pulmonary nodules based on cfMeDIP-seq and machine learning.
Jian Qi1,2, Bo Hong1,3, Rui Tao4
1Anhui Province Key Laboratory of Medical Physics and Technology, Institute of Health and Medical Technology, Hefei Institutes of Physical Science, Chinese Academy of Sciences, Hefei, China.
Cancer Science
|July 12, 2021
Summary
A novel cell-free methylated DNA immunoprecipitation and high-throughput sequencing (cfMeDIP-seq) method detects lung cancer. This technique identified differentially methylated regions (DMR) in cell-free DNA (cfDNA) for a non-invasive diagnostic model.
Area of Science:
- Genomics
- Molecular Biology
- Oncology
Background:
- Lung cancer diagnosis often relies on invasive procedures.
- Early detection of lung tumors is crucial for improving patient outcomes.
- Cell-free DNA (cfDNA) in blood offers a potential source for non-invasive biomarkers.
Purpose of the Study:
- To develop and validate a non-invasive method for detecting malignant lung tumors.
- To identify specific methylation patterns in cfDNA associated with lung cancer.
- To build a predictive model for distinguishing lung tumors from normal controls.
Main Methods:
- Employed cell-free methylated DNA immunoprecipitation and high-throughput sequencing (cfMeDIP-seq) to analyze whole-genome methylation of cfDNA.
- Identified differentially methylated regions (DMR) between lung tumor samples and healthy controls.
- Constructed a random forest prediction model utilizing the top 300 DMR.
Main Results:
- The cfMeDIP-seq technique successfully detected whole-genome methylation patterns in cfDNA.
- Significant differentially methylated regions (DMR) were identified between lung tumors and normal controls.
- The random forest model achieved high sensitivity (91.0%) and specificity (93.3%) with an AUROC of 0.963 in distinguishing malignant lung tumors.
Conclusions:
- A non-invasive prediction model for lung cancer has been successfully developed using cfDNA methylation.
- The cfMeDIP-seq approach combined with machine learning shows promise for early lung cancer detection.
- This method offers a sensitive and specific tool for identifying malignant pulmonary nodules.
