Related Experiment Video
Updated: Mar 22, 2026
05:05
Four-Dimensional CT Analysis Using Sequential 3D-3D Registration
Published on: November 23, 2019
8.6K
Automatic assessment of average diaphragm motion trajectory from 4DCT images through machine learning
Guang Li1, Jie Wei2, Hailiang Huang1
1Department of Medical Physics, Memorial Sloan Kettering Cancer Center, New York, USA.
Biomedical Physics & Engineering Express
|April 26, 2016
Summary
This study introduces a machine learning method to automatically estimate diaphragm motion from 4D CT scans. The approach accurately predicts average diaphragm motion trajectory (ADMT), aiding respiratory motion assessment in patients.
Area of Science:
- Medical Imaging
- Computational Biology
- Radiotherapy Physics
Background:
- Accurate assessment of respiratory motion is crucial for radiotherapy planning.
- Four-dimensional computed tomography (4DCT) provides dynamic respiratory information.
- Diaphragm motion significantly influences upper abdominal tumor motion.
Purpose of the Study:
- To develop an automated method for estimating average diaphragm motion trajectory (ADMT) using 4DCT.
- To facilitate clinical assessment of respiratory motion and its variations.
- To enable retrospective analysis of diaphragm motion during treatment.
Main Methods:
- Developed a machine learning algorithm utilizing differential volume-per-slice (dVPS) curves derived from 4DCT.
- Applied discrete cosine transform (DCT) for frequency domain analysis of dVPS curves.
- Employed multiple linear regression (MLR) to predict ADMT based on frequency coefficients, validated with leave-one-out cross-validation.
Main Results:
- Seven lowest DCT frequencies effectively approximated patient dVPS curves (R = 91%-96% in MLR fitting).
- Mean prediction error for ADMT was 0.3 ± 1.9 mm (left) and 0.0 ± 1.4 mm (right).
- Combined 4DCT data (simulation and treatment) yielded the lowest prediction error (0.2 ± 1.6 mm).
Conclusions:
- The developed frequency-analysis-based machine learning technique automatically predicts ADMT with acceptable accuracy.
- This volumetric approach is robust and unaffected by lung tumors.
- Provides a valuable tool for evaluating diaphragm motion in clinical practice.
