Program Information
A Knowledge-Based Radiation Therapy (KBRT) Algorithm for the Treatment-Planning of Simultaneous Integrated Boost (SIB)target Volumes
x chen1 2*, O Nwankwo3 4* , D Sihono3 , F Wenz3 , G Glatting4 5 , (1) Master program in Medical Physics, Universitatsmedizin Mannheim, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany (2)Radiation Oncology Department, National Cancer Center/Cancer Hospital,Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, P.R. China (3)Department of Radiation Oncology, Universitatsmedizin Mannheim, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany (4)Medical Radiation Physics/Radiation Protection, Universitatsmedizin Mannheim, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany (5)Medical Radiation Physics, Department of NuclearMedicine, Ulm University, Ulm, Germany * = Equal contribution
Presentations
SU-E-FS4-8 (Sunday, July 30, 2017) 1:00 PM - 1:55 PM Room: Four Seasons 4
Purpose: To present a knowledge-based radiation therapy (KBRT) treatment-planning algorithm for the treatment planning of simultaneous integrated boost (SIB) target volumes.
Methods: A reference database containing 110 clinical SIB plans for head and neck cancer(HN) was subdivided into a learning set (95 plans) and a validation set (15 plans). The SIB plans contained a planning target volume(PTV1) and a boost volume(PTV2). The relationship between the positions and the doses received by organs at risk (OARs) in HN SIB plans were derived from the analysis of the learning set using multi-level modeling. The position-dose relationships were defined by two functions (f1&f2). The functions related the distances (d1&d2) of each OAR voxel to target volumes (PTV1&PTV2)and the dose received by the voxel. These organ-specific models of position-dose relationships were developed for four OARs.
Results: The doses of OARs in SIB plans depend on both distances d1 and d2. Distance d1 determines voxel doses up to a threshold distance, which is henceforth referred to as the cutoff distance. Voxel dose beyond the cutoff distance is determined by the d2 value of the voxel. The cutoff distances for the left parotid, right parotid and brain stem are 7.5mm, 7.2mm and 7.2mm respectively. Only d1 is required to model the dose to the spinal cord because the contribution of d2 is negligible. Based on a metric introduced in a previous work (organ agreement scores), the doses predicted by the model for the brain stem, left parotid, right parotid and spinal cord of the 15 test subjects are on the average within -0.5%±3.6%, -0.4%±5.0%, 2.0%±3.3% and 0.3%±2.4% of the reference values, respectively.
Conclusion: A KBRT algorithm for the treatment planning of HN SIB target volumes is introduced. The presented approach can be used also to model SIB target volumes of other treatment regions.
Contact Email: