TU‐C‐213CD‐11: A Universal Margin Reconstruction Method for Estimating Anatomy Induced Proton Range Uncertainties

• Published in: Medical Physics Vol. 39; no. 6; p. 3904
• Main Authors: Yu, Z, Klopp, A, Kudchadker, R, Mourtada, F, Eifel, P, Park, P, Dong, L
• Format: Conference Proceeding Journal Article
• Language: English
• Published: American Association of Physicists in Medicine 01-06-2012
• Subjects: Anatomy; Cancer; Computed tomography; Medical imaging; Medical treatment planning; Photons; Proton therapy; Protons
• ISSN: 0094-2405; 2473-4209
• DOI: 10.1118/1.4735938

Purpose: Unlike photon therapy, proton beams are affected by range uncertainties caused by variations of anatomy in the beam path. Appropriate target margins must be used during treatment planning to account for range uncertainties. Currently, there are no clinical data to quantify the magnitude of interfractional anatomical uncertainties for different beam directions. The purpose of this study was to develop a ray‐tracing based method to determine margins to account for anatomical variations. The method was applied for nodal boost treatments in the pelvic region. Methods: CT scans obtained weekly during chemoradiation for intact cervical cancer were utilized for this study. Common and external iliac nodal contours were transferred from planning to weekly images using rigid alignment. A ray‐tracing method was used to create 2D water equivalent thickness matrices for both distal and proximal surfaces of targets in the beam direction. Positive and negative differences between weekly and planning matrices were used to determine required planning margins in the distal and proximal directions, respectively. We required 90% of the treatment fractions to be covered for population margin definition. Results: For common iliac nodes, the angles with the least and most anatomical variations for the distal margin were at 90‐degrees (LtLat) and 45‐degrees (LAO), which required 0.5 and 1.6cm margins, respectively. For proximal margins, the least and most anatomical variations were at 180‐degrees (PA) and 315‐degrees (RAO) with 1.1 and 2.9cm margins, respectively. For external iliac nodes, PA and lateral beams required 0.8 and 0.4cm distal margins and 1.0 and 1.1cm proximal margins, respectively. Conclusions: We have developed a fast margin determination method using historical clinical data. We also demonstrated that inter‐fractional anatomical variations are asymmetric and dependent on beam angle. Beams passing through the bowel region where ranges are impacted by gas content and weight loss have much greater anatomical variations.