Optimizing principal component models for representing interfraction variation in lung cancer radiotherapy.
Ahmed M Badawi1, Elisabeth Weiss, William C Sleeman
1Department of Radiation Oncology, Virginia Commonwealth University, 401 College Street, P.O. Box 980054, Richmond, Virginia 23298, USA.
Medical Physics
|October 23, 2010
Summary
Principal component analysis (PCA) effectively models lung cancer radiotherapy variations. Retrospective models achieved millimeter accuracy, with prospective models stabilizing after a few measurements.
Area of Science:
- Medical Physics
- Radiotherapy
- Image Analysis
Background:
- Interfractional anatomical variation poses challenges in lung cancer radiotherapy.
- Accurate modeling of these variations is crucial for optimizing treatment delivery.
- Active breath-hold techniques aim to minimize respiratory motion during treatment.
Purpose of the Study:
- To optimize modeling of interfractional anatomical variation during active breath-hold radiotherapy for lung cancer.
- To evaluate the accuracy of principal component analysis (PCA) models in reconstructing GTV and lung anatomy.
- To compare retrospective and prospective PCA modeling approaches.
Main Methods:
- 12 lung cancer patients undergoing active breath-hold radiotherapy were analyzed.
- Weekly CT scans with intrafraction repeats were acquired at end-inspiration.
- Deformable image registration propagated GTV and lung contours; PCA modeled variability.
Main Results:
- Retrospective PCA models (W2W and allscans) reconstructed anatomy with average errors of 0.7-1.1 mm.
- 3-5 dominant modes were sufficient to represent at least 95% of anatomical variability.
- Prospective models achieved stable accuracy after 4-5 measurements, with errors around 0.8-1.2 mm.
Conclusions:
- PCA models accurately reconstruct GTV and lung anatomy within millimeters during active breath-hold radiotherapy.
- Retrospective models using 3-4 modes provided reliable anatomical reconstruction.
- Prospective models demonstrated similar accuracy after a limited number of measurements, supporting adaptive radiotherapy.
