JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Development and validation of growth prediction models for multiple pulmonary ground-glass nodules based on CT features, radiomics, and deep learning

  • 0Department of Diagnostic Radiology, National Cancer Center/National Clinical Research for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Translational Lung Cancer Research +

|

July 17, 2025

|

July 17, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

This study developed models to predict the growth of multiple pulmonary ground-glass nodules (GGNs). The Clinical-Radiomics model showed the best accuracy in predicting GGN growth, outperforming other models.

Area Of Science

  • Pulmonary Medicine
  • Radiology
  • Oncology

Background

  • Previous models for pulmonary ground-glass nodules (GGNs) growth prediction primarily focused on solitary nodules.
  • Understanding the natural history of multiple GGNs is crucial for precise monitoring and intervention.
  • This study investigates multiple GGNs, comparing CT features, radiomics, and deep learning (DL) for growth prediction.

Purpose Of The Study

  • To investigate the natural history of multiple pulmonary GGNs.
  • To develop and validate growth prediction models using CT features, radiomics, and DL.
  • To compare the predictive performance of different models for multiple GGN growth.

Main Methods

  • Retrospective review of patients with at least two persistent GGNs and >= 3 years of follow-up.
  • GGN growth defined by diameter, volume increase, or solid component changes.
  • Construction and validation of Clinical, Radiomics, DL, Clinical-Radiomics, and Clinical-DL models using AUC for performance assessment.

Main Results

  • 732 GGNs from 231 patients were analyzed; 37.2% showed growth.
  • Lobulation, vacuole, initial volume, and mass were identified as risk factors for GGN growth.
  • The Clinical-Radiomics model achieved the highest AUC (0.908), outperforming other models in predicting GGN growth.

Conclusions

  • Multiple pulmonary GGNs exhibit indolent growth patterns.
  • The Clinical-Radiomics model demonstrated superior accuracy in predicting the growth of multiple GGNs.
  • This model facilitates more precise identification of GGNs requiring close monitoring or early intervention.

Keywords: