Related Experiment Video
Updated: Jul 1, 2025
05:18
Author Spotlight: Improving Radiation Therapy Access with Radiation Planning Assistant
Published on: October 6, 2023
1.3K
Experience of Implementing Deep Learning-Based Automatic Contouring in Breast Radiation Therapy Planning: Insights
Byung Min Lee1, Jin Sung Kim2, Yongjin Chang3
1Department of Radiation Oncology, Yonsei Cancer Center, Yonsei University College of Medicine, Seoul, Republic of Korea; Department of Radiation Oncology, Uijeongbu St. Mary's Hospital, Catholic University of Korea, Seoul, Republic of Korea.
Summary
An auto-contouring system for radiation therapy improved contour accuracy for organs at risk and target volumes. However, careful quality management is essential to mitigate automation bias risks in clinical practice.
Area of Science:
- Radiation Oncology
- Medical Imaging
- Artificial Intelligence in Healthcare
Background:
- Auto-contouring systems are increasingly used in radiation therapy to streamline treatment planning.
- Evaluating the clinical utility and impact of these systems is crucial for safe and effective implementation.
Purpose of the Study:
- To assess the impact and clinical utility of an auto-contouring system in radiation therapy treatments.
- To compare auto-contours with manually finalized contours for breast radiation therapy patients.
Main Methods:
- Retrospective analysis of 2428 adjuvant breast radiation therapy patients before and after auto-contouring system implementation.
- Comparison of auto-contours and final contours using Dice Similarity Coefficient (DSC) and 95% Hausdorff Distance (HD95).
- Evaluation of organs at risk and target volumes for segmentation accuracy.
Main Results:
- Post-implementation, organs at risk (heart, esophagus, spinal cord, contralateral breast) showed significantly improved DSC and decreased HD95.
- Target volumes (CTVn_L2, L3, L4, internal mammary node) also demonstrated improved DSC and decreased HD95.
- Lungs exhibited inaccurate segmentation; final contours were generally larger than auto-contours.
Conclusions:
- Auto-contouring systems enhance contour accuracy for many structures in radiation therapy.
- Increased reliance on automated settings highlights the potential for automation bias.
- Integration of risk assessments and quality management is vital for optimal system use.
