Combined Radiotherapy Using a Method for Biologically Informed IMRT Planning

Abstract

Listen

Translate article

Radiotherapy (RT) regimens routinely employ multiple courses that vary by delivery modality and/or fractionation schedule as part of initial or follow up treatment. The combination of RTs ideally should consider the biological effect of the total composite dose distribution. Biological effect can be quantified by biological effective dose (BED) which is a function of the RT modality, dose fractionation, and tissue radiosensitivity. This work describes a method of optimizing a follow up intensity modulated RT (IMRT) plan that considers BED from prior RT. Required inputs for this method are the desired composite BED objectives (BEDref) for target volume(s) and organs at risk (OARs). First, the physical dose in each CT voxel from prior RT (Dprior) is converted to BED (BEDprior). The difference between BEDref and BEDprior gives the optimal BED distribution for the follow up IMRT plan (BEDIMRT). Next, the optimal physical dose (DIMRT) corresponding to BEDIMRT is calculated for each voxel as a function of the number of IMRT fractions (nIMRT). Commercial IMRT treatment planning systems require optimization objectives to be specified for contoured structures, not individual voxels, so a workaround is proposed: 1) for each structure, BEDref is converted to physical dose (Dopti) as a function of nIMRT, and 2) for each voxel, the difference of Dopti and DIMRT is calculated. The result is a "base dose distribution" (Dbase) that is used with standard inverse optimization tools to optimize an IMRT plan so that the sum of Dbase and the IMRT dose approaches Dopti for each voxel. This method was implemented using a custom software workflow to calculate Dbase as a function of Dprior and input BEDref values. Validation was performed by comparing the workflow output with hand calculated Dbase values at randomly selected points within clinically significant regions. Satisfactory agreement was considered to be within 5%. Feasibility of the method was demonstrated using an RT trial protocol for glioblastoma multiforme that combines permanent low dose rate brachytherapy (BT) using collagen tiles embedded with 131Cs seeds and follow-up IMRT (20 fractions). The Dbase distribution was obtained from the software as a function of the BT dose and BEDref for one patient model. Comparison of Dbase values from the software and hand calculated values resulted in agreement within 5% for 22/25 points and within 10% for 3/25 points. Discrepancies >5% (max 9.4%) were attributed to data precision resolution effects inherent in the DICOM standard. The resulting IMRT plan was deemed clinically acceptable and the composite dose distribution achieved dosimetric objectives for all target and OARs. A method for optimizing a follow up IMRT plan that considers dose from prior RT to produce a desired composite BED distribution was derived and proof of concept was successfully demonstrated. Further validation of this method with additional patients and multimodality RT protocols will be performed.