(Tuesday, 7/16/2019) 10:00 AM - 10:30 AM
Room: Exhibit Hall | Forum 5
Purpose: To develop an objective and evidence-based method by analyzing geometric and dosimetric changes in setup cone-beam CTs to predict HN cancer treatment replanning.
Methods: We retrospectively analyzed the change of CBCT images of 10 head and neck patients with clinical replanning. In our study, three major metrics, including the target’s image similarity and the target’s dosimetric measurements and the OAR’s dosimetric measurements, are utilized for decision making in replanning. However, due to their purposes in treatment, tumor targets and organs at risk (OAR) are processed differently. CTV and PTV contours are copied rigidly from the planning CT to individual CBCTs through clinical setup. OAR contours are copied from the planning CT to CBCTs through deformable registrations between the planning CT and CBCTs. Then the treatment plan rigidly copied from the planning CT to CBCTs through clinical setup is used for computing dose distribution in CBCTs. For CTV, the geometric similarity between 1st CBCT and the following CBCTs are analyzed. For PTV and OAR, we analyze CBCT DVH variations (D95, D90, D10, D5) between 1st CBCT and the following CBCTs. In addition, for OARs, the certain dose constraint variations in CBCTs are analyzed.
Results: Our study shows than Pearson’s image correlation coefficient less than 0.75 and the dosimetric variation greater than 5% will be a good indicator for head and neck cancer treatment replanning. By applying our method to CBCTs of 6 randomly chosen HN patients, it shows the above criteria are very effective in deciding when to replan.
Conclusion: Trend analysis of CBCT’s geometric and dosimetric changes provides an objective solution for HN cancer replanning to supersede either the breakpoint-based replanning strategy or the spotcheck-based replanning strategy.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
