Related Experiment Video
Updated: Apr 15, 2026
06:03
Integrating Augmented Reality Tools in Breast Cancer Related Lymphedema Prognostication and Diagnosis
Published on: February 6, 2020
7.3K
Fully Automated AI-Based Lymph Node Measurements in Chest CT: Accuracy and Reproducibility Compared with Multi-Reader
Andra-Iza Iuga1, Heike Carolus2, Liliana Lourenco Caldeira1
1Institute of Diagnostic and Interventional Radiology, Faculty of Medicine and University Hospital Cologne, University of Cologne, 50937 Cologne, Germany.
Diagnostics (Basel, Switzerland)
|April 14, 2026
Summary
Fully automated artificial intelligence (AI) accurately measures lymph nodes (LN) in chest CT scans, showing strong agreement with expert radiologists. This AI workflow enhances reproducibility in oncologic imaging, reducing inter-reader variability for staging and therapy monitoring.
Area of Science:
- Radiology
- Artificial Intelligence
- Medical Imaging
Background:
- Accurate lymph node (LN) measurement is crucial for cancer staging and treatment monitoring.
- Manual LN measurements in CT scans suffer from significant inter-reader variability.
- A need exists for reproducible and automated methods for LN measurement.
Purpose of the Study:
- To evaluate the accuracy and reproducibility of a fully automated AI workflow for LN measurement in contrast-enhanced chest CT.
- To compare AI-based measurements against multi-reader manual measurements as the reference standard.
- To assess the stability of AI measurements across different hardware systems.
Main Methods:
- Sixty thoracic LNs from seven patients were measured by 13 radiologists across two rounds.
- A fully automated 3D Convolutional Neural Network (CNN) segmented LNs to derive short and long axis diameters.
- Agreement was assessed using Friedman testing, intraclass correlation coefficients (ICCs), and concordance correlation coefficients (CCCs).
Main Results:
- Manual measurements exhibited significant inter-reader variability (p < 0.01).
- AI-based measurements showed no significant difference from the ground truth (median of all manual measurements) (p > 0.01).
- Good agreement was found between AI and manual measurements, with CCC values of 0.86 for short axis and 0.79 for long axis.
Conclusions:
- Fully automated AI-based LN measurement in chest CT demonstrates strong agreement with multi-reader consensus.
- The AI workflow offers high numerical stability and reproducibility.
- Automated LN measurement has the potential to standardize oncologic imaging assessment.
