Bi‐objective optimization of catheter positions for high‐dose‐rate prostate brachytherapy

Abstract

Bi-objective simultaneous optimization of catheter positions and dwell times for highdose-rate (HDR) prostate brachytherapy, based directly on dose-volume indices, has shown promising results. However, optimization with the state-of-the-art evolutionary algorithm MO-RV-GOMEA so far required several hours of runtime, and resulting catheter positions were not always clinically feasible. The aim of this study is to extend the optimization model and apply GPU parallelization to achieve clinically acceptable computation times. The resulting optimization procedure is compared with a previously introduced method based solely on geometric criteria, the adapted Centroidal Voronoi Tessellations (CVT) algorithm. Methods: Bi-objective simultaneous optimization was performed with a GPU-parallelized version of MO-RV-GOMEA. This optimization of catheter positions and dwell times was retrospectively applied to the data of 26 patients previously treated with HDR prostate brachytherapy for 8-16 catheters (steps of 2). Optimization of catheter positions using CVT was performed in seconds, after which optimization of only the dwell times using MO-RV-GOMEA was performed in 1 min. Results: Simultaneous optimization of catheter positions and dwell times using MO-RV-GOMEA was performed in 5 min. For 16 down to 8 catheters (steps of 2), MO-RV-GOMEA found plans satisfying the planning-aims for 20, 20, 18, 14, and 11 out of the 26 patients, respectively. CVT achieved this for 19, 17, 13, 9, and 2 patients, respectively. The P-value for the difference between MO-RV-GOMEA and CVT was 0.023 for 16 catheters, 0.005 for 14 catheters, and <0.001 for 12, 10, and 8 catheters. Conclusions: With bi-objective simultaneous optimization on a GPU, high-quality catheter positions can now be obtained within 5 min, which is clinically acceptable, but slower than CVT. For 16 catheters, the difference between MO-RV-GOMEA and CVT is clinically irrelevant. For 14 catheters and less, MO-RV-GOMEA outperforms CVT in finding plans satisfying all planning-aims.

Catheters and dose optimization using a modified CVT algorithm and multi‐criteria optimization in prostate HDR brachytherapy

“…In the article by van der Meer et al., 8 Multi‐Objective Real‐Valued Gene‐pool Optimal Mixing Evolutionary Algorithm (MO‐RV‐GOMEA) is used to optimize the catheter placement and the plans together. A comparison with our original CVT (oCVT) is also performed.…”

“…This aspect has been partially answered with the new CVT optimization, which increases the acceptance rate even for a low number of catheters. With GOMEA, 8 the optimization process for a fixed number of catheters and multiple plans takes about 5 min. In the case of CVT, a multiple numbers of catheters are optimized, and plans are immediately optimized, which take less than 40 s in total.…”

“…Hence, it could be of great interest to decrease the number of catheters used without compromising the plan's quality.It has been shown that using patient-specific insertion is feasible and allows for better consideration of the anatomy, such as reducing the radiation to healthy organs. [6][7][8] Moreover, there is a need for optimal catheter placement when a smaller number of catheters are used, such as in prostate focal therapy or hemi-gland therapy. [9][10][11] Other research has been done to develop new algorithms to determine catheter number or placement, ranging from compress sensing-based algorithm 12 to mixedinteger programming.…”

Analytical HDR prostate brachytherapy planning with automatic catheter and isotope selection

“…Hybrid inverse planning optimization (HIPO), for example, is a popular option that uses stochastic simulated annealing to optimize catheter locations followed by a deterministic dose volume histogram‐based optimization of dwell times 21 . Other heuristics and linear programming‐based models for catheter selection have also been presented and shown to improve dosimetry 22–25 . In 2021, Wang et al.…”

“…21 Other heuristics and linear programming-based models for catheter selection have also been presented and shown to improve dosimetry. [22][23][24][25] In 2021, Wang et al proposed a fast-iterative shrinkage thresholding algorithm (FISTA) with a group sparsity term and adaptive heuristic group sparsity weighting to perform simultaneous dwell time optimization and catheter selection. 26 They concluded that this was an effective method for catheter selection and improved OAR sparing in the final plans over manual catheter placement.…”