Purpose: To present an exhaustive dosimetric comparison between three geometric optimization methods and our inverse-planning simulated annealing (IPSA) algorithm, with two different prescriptions for high-dose-rate (HDR) boost of the prostate. The objective of this analysis was to quantify the dosimetric advantages of the IPSA algorithm compared with more standard geometric optimizations.
Methods and materials: Between September 1999 and June 2001, 34 patients were treated to a dose of 40-44 Gy by external pelvic fields, followed by an HDR boost of 18 Gy in 3 fractions. The first 4 patients were treated with HDR using geometric optimization, and anatomy-based inverse-planning dose optimization was used for the remaining 30 patients. We retrospectively used the data from these 30 patients to create HDR dose distributions according to five different dose optimization protocols, including our IPSA algorithm. The various geometric optimization procedures differed in the way the dwell positions were activated and plan normalization was performed. Dose-volume histograms from all these plans were analyzed and multiple implant quality indexes extracted.
Results: The IPSA algorithm provided better clinical tumor volume prescription dose coverage than did the geometric optimizations. The average prostate volume receiving 100% of the prescribed dose (V100) was 96.3% and 94.5% for IPSA with two different prescriptions compared with 92.1%, 92.6%, and 88.8% for the three geometric optimization schemes. The average urethra V150 value was 0.0% and 0.7% for IPSA with two different prescriptions, and the three geometric optimization protocols generated average values of 22.9%, 33.9%, and 38.8%. The bladder and rectal dose-volume histograms were similar, although the latest version of the IPSA algorithm slightly decreases the dose to these organs at risk because of organ-specific dose constraints included in the objective function.
Conclusion: We found that planning an HDR prostate boost could be performed in a fast, secure, and effective manner with the IPSA algorithm. We demonstrated that our inverse-planning algorithm produces superior HDR plans than more conventional geometric optimizations for adenocarcinoma of the prostate. The organs at risk protection included in the objective function is a major feature of the algorithm and should allow us to escalate the HDR dose to the prostate without increasing undesirable side effects.