Experience Based Quality Control in IMRT Treatment Planning of High Risk Post-Prostatectomy Prostate Cancer with RapidPlan

Loading...
Thumbnail Image
Penn collection
Master of Science in Medical Physics Thesis Projects
Degree type
Discipline
Subject
RapidPlan
knowledge based planning
KBP
IMRT
prostate cancer
radiation therapy
RTOG 0621
0621
Medicine and Health Sciences
Oncology
Other Physics
Physics
Funder
Grant number
Copyright date
Distributor
Related resources
Contributor
Abstract

Purpose: To develop a knowledge based planning (KBP) model with RapidPlan (Varian Medical Systems, Palo Alto, USA) for the treatment of high risk post-prostatectomy prostate cancer. The model was trained on a knowledge database of high quality treatment plans from the national clinical trial RTOG 0621, then tested as a QA tool. Methods: An initial dosimetric analysis was carried out to identify high quality plans from clinical trial RTOG 0621. Treatment plans for patients enrolled in the trial were scored according to the system used by the Imaging and Radiation Oncology Core (IROC) of the National Clinical Trials Network (NCTN) of the NCI to assess adherence to the trial protocol. Of the 80 plans enrolled in the trial 39 were chosen for the training sample. Another subset of 8 plans, orthogonal to the training sample, was chosen for the validation sample to ensure that the model accurately predicts dose volume histograms (DVHs) for all critical structures. The validation plans were then re-optimized with the model in order to test its effectiveness as a tool for planning QA. DVHs of the re-optimized plans were compared with those of the original clinical plans. Normal tissue complication probabilities and tumor control probabilities were calculated with the Lyman-Kutcher-Burman (LKB) model before and after re-optimization to determine the effect on patient outcome. Results: The RapidPlan prostate model was shown to accurately predict estimated DVH bands for all plans in the validation sample that matched the geometry of the training sample. Three treatment plans in the validation sample were geometric outliers with respect to the training sample leading to inaccuracies in the model predictions for the cone down phase of these treatment plans. All of the re-optimized plans showed increased dose sparring to the bladder and rectum respectively without lose of target coverage. The average reduction in NTCP was 0.34 ± 0.21 % for the bladder and 0.11 ± 0.25 % for the rectum with corresponding p-values of 0.116 and 0.668. The average TCP for the prostate bed decreased slightly from 97.05 % to 96.54 % with a p-value of 0.149. Due to limited statistics the changes reported in these numbers are not statistically significant as indicated by the p-values. Although the average values are inconclusive the model was effectively used to identify sub-optimal treatment plans which were improved through re-optimization with the model. For treatment plan 0621c0027 the NTCP decreased from 0.35 % to 0.06 % for the bladder and from 0.10 % to 0.06 % for the rectum while the TCP increased from 96.78 % to 96.87 %. Conclusions: The RapidPlan prostate model developed in this study is an effective tool for monitoring the quality of IMRT treatment plans for high-risk post prostatectomy prostate cancer.

Advisor
Date Range for Data Collection (Start Date)
Date Range for Data Collection (End Date)
Digital Object Identifier
Series name and number
Publication date
2017-01-01
Volume number
Issue number
Publisher
Publisher DOI
Journal Issue
Comments
Recommended citation
Collection