Related Experiment Video
Updated: May 10, 2026
05:18
Radiation Planning Assistant - A Web-based Tool to Support High-quality Radiotherapy in Clinics with Limited Resources
Published on: October 6, 2023
Validation of automatic contour propagation for 4D treatment planning using multiple metrics
M Peroni1, M F Spadea, M Riboldi
1Department of Electronics, Information and Bioengineering, Politecnico di Milano, via Golgi 39, 20133 Milano, Italy. marta.peroni@psi.ch.
Technology in Cancer Research & Treatment
|June 11, 2013
Summary
Clinical validation of deformable image registration and contour propagation in 4D radiotherapy planning revealed metric discrepancies. Combining Volume Difference (VD) and Surface Distances (SD) offers accurate failure detection for contour propagation algorithms.
Area of Science:
- Medical Physics
- Radiotherapy
- Image Analysis
Background:
- Deformable image registration and contour propagation are crucial for 4D lung radiotherapy planning.
- Accurate contouring is essential for precise dose delivery and treatment efficacy.
- Evaluating the performance of these methods requires robust clinical validation metrics.
Purpose of the Study:
- To provide insights into the clinical validation of deformable image registration and contour propagation methods in 4D lung radiotherapy.
- To compare the performance of multiple quantitative metrics for assessing contour propagation accuracy.
- To determine the most reliable metrics for detecting failures in contour propagation algorithms.
Main Methods:
- Analysis of Volume Difference (VD), Dice Similarity Coefficient (DSC), Positive Predictive Value (PPV), and Surface Distances (SD).
- Utilized three patient datasets with ground-truth volumes generated using the Simultaneous Truth And Performance Level Estimation (STAPLE) algorithm from five expert outlines.
- Evaluated metric sensitivity to image artifacts and anatomical structures like the esophagus and spinal cord.
Main Results:
- Significant discrepancies were found in quality assessments provided by different metrics across all cases.
- Metrics showed varying sensitivity, particularly with image artifacts and tubular structures.
- Volume Difference (VD) did not account for positional differences, and Dice Similarity Coefficient (DSC) had limitations due to symmetry and volume dependency.
- Positive Predictive Value (PPV) was robust but could not detect volume inclusions.
- Surface Distances (SD) captured shape but could be sensitive to local contour differences.
Conclusions:
- No single metric adequately assesses contour propagation accuracy in 4D radiotherapy planning.
- The combination of Volume Difference (VD) and Surface Distances (SD) provides the necessary accuracy for failure detection in contour propagation algorithms.
- This combined approach enhances the reliability of validating contour propagation in 4D lung radiotherapy.
