Impact of prostate catheter displacement in inverse planning--simulated annealing and geometric optimization

The aim of this study was to compare inverse planning-simulated annealing (IPSA) with geometric optimization (GO) in high-dose-rate (HDR) prostate brachytherapy, to assess variations in dosimetric indices associated to catheter displacement.

We retrospectively studied the dosimetric effect of catheter displacement in 20 patients treated with HDR prostate brachytherapy with salvage intention. The catheter loadings, with optimized dwell positions, from the first fraction were transferred to the catheter positions in the second fraction.

Median catheter displacement was 8.7 ± 3.3 mm (range, 2.7 ± 1.1 mm-14.7 ± 1.7 mm). D90% median variations for IPSA and GO were -10% with a maximum of -59%, and -29% with a maximum of -63%, respectively. V100% median variation was -11% with a maximum of -37% for IPSA, and -20% with a maximum of -37% for GO. V150/V100 implant median variations were 15% and 9% for IPSA and GO with maximum values of 65% and 47%, respectively (p < 0.05). No significant differences were observed for V200/V100 indices, nor were any significant differences found for organs at risk. Correlation between mean catheter displacement and dosimetric indices was found only in the planning target volume D90% and V100%; linear regression slopes were -2.0% per mm and -2.6% per mm for IPSA vs. -2.4% per mm and -3.5% per mm for GO.

IPSA does not present any additional risk compared with GO in HDR prostate brachytherapy when catheter uncertainties are taken into account. Moreover, IPSA optimization preserves planning target volume coverage better than GO, suggesting that it may be superior when catheter displacement is considered.