Related Experiment Video
Updated: Jul 6, 2025
08:17
Dynamic Lung Tumor Tracking for Stereotactic Ablative Body Radiation Therapy
Published on: June 7, 2015
15.7K
Efficient quality assurance for isocentric stability in stereotactic body radiation therapy using machine learning
Sana Salahuddin1,2, Saeed Ahmad Buzdar3, Khalid Iqbal4
1Institute of Physics, The Islamia University of Bahawalpur, Bahawalpur, Pakistan. sana.salahuddin@iub.edu.pk.
Radiological Physics and Technology
|December 31, 2023
Summary
Machine learning models accurately predict isocentric stability for stereotactic body radiation therapy (SBRT) quality assurance (QA). Decision Tree and Random Forest models show high accuracy, enhancing safety and efficiency in patient treatments.
Area of Science:
- Medical Physics
- Machine Learning in Radiation Oncology
Background:
- Manual quality assurance (QA) for stereotactic body radiation therapy (SBRT) is time-consuming and prone to errors.
- Accurate isocentric stability is critical for effective and safe SBRT delivery.
Purpose of the Study:
- To develop and evaluate machine learning (ML) models for predicting isocentric stability during SBRT QA.
- To support medical physicists in improving the accuracy and efficiency of QA procedures.
Main Methods:
- Collected 247 Winston-Lutz (WL) tests from TrueBeam linac SBRT QA using the RUBY phantom.
- Extracted statistical features using IsoCheck EPID software.
- Trained and evaluated five ML models: Logistic Regression, Decision Tree (DT), Random Forest (RF), Naive Bayes, and Support Vector Machines.
Main Results:
- Decision Tree (DT) and Random Forest (RF) models achieved the highest test accuracy (93.5%–99.4%) and Area Under Curve (AUC) (90%–100%).
- The DT model demonstrated superior precision, recall, and F1 scores in predicting isocenter stability deviations.
- ML models effectively predicted QA outcomes across collimator, gantry, and table tests.
Conclusions:
- ML-based QA assessment can significantly improve early error prediction in SBRT.
- Implementing ML models enhances the safety and effectiveness of patient radiation treatments.
- DT and RF models offer a robust solution for automated isocentric stability QA.
