A novel approach for estimating lung tumor motion based on dynamic features in 4D-CT

Area of Science:

• Medical Imaging
• Radiotherapy Physics
• Computational Biology

Background:

• Limited 4D-CT data (5 phases) hinders accurate lung tumor motion assessment for radiotherapy planning.
• Missing respiratory phases (10-90%) in 4D-CT scans pose challenges for precise treatment.
• Accurate tumor boundary and motion information is crucial for defining the planned target volume (PTV).

Purpose of the Study:

• To develop an automated interpolation method for deriving lung tumor contours from limited 5-phase 4D-CT data.
• To improve the accuracy of radiotherapy planning by reconstructing missing respiratory phases.
• To evaluate the efficacy of the Uniform Angular Interval-Dynamic Mode Decomposition (UAI-DMD) method.

Main Methods:

• Applied Dynamic Mode Decomposition (DMD), a data-driven technique, to analyze high-dimensional 4D-CT data.
• Developed Uniform Angular Interval (UAI) sampling to generate suitable snapshot vectors for DMD analysis of 3D tumor motion.
• Validated the UAI-DMD method on 4D-CT data from ten lung cancer patients.

Main Results:

• The UAI-DMD method accurately approximates deformable lung tumor boundary surfaces and nonlinear motion trajectories.
• The estimated tumor centroid accuracy was within 2 mm, outperforming traditional B-Spline interpolation (3 mm).
• The methodology shows potential for reconstructing 20-phase respiratory motion from 10-phase 4D-CT data.

Conclusions:

• The UAI-DMD method offers a robust solution for interpolating missing respiratory phases in 4D-CT data.
• This technique enhances the accuracy of lung tumor motion modeling for radiotherapy.
• Improved PTV estimation through accurate tumor motion reconstruction can lead to more effective lung cancer treatment.