Needle Placement using Patient-Specific, Skew-Line Needle Guides for Prostate HDR Brachytherapy: An Open-Loop Phantom Study

Purpose: Recent advances in real time 3D ultrasound, EM tracking, needle position optimization, and rapid prototyping support a radically new prostate high dose rate (HDR) brachytherapy workflow that uses custom needle guides to insert needles. The purpose of this study is to evaluate one part of this workflow: the needle placement error associated with using a rapidly prototyped (3D printed) needle guide to insert skew-line needle arrangements and the effect this error has on 5 treatment quality. Method: We created four gelatin phantoms to simulate prostate cancer cases. We used Needle Planning by Integer Program (NPIP) to generate a skew-line needle arrangement for each phantom. A dose distribution was computed for each needle arrangement using Inverse Planning by Integer Program (IPIP). We 3D printed a needle guide to accommodate each needle arrangement 10 and used it to insert the needles into the phantoms. We did not use any feedback or sensing information to alter the trajectory of the needles into a more desirable arrangement. We evaluated the implanted needle arrangements by placement error, critical structure puncturing, and the final dose distribution. Results: Three phantoms were successfully implanted without puncturing any non-target struc15 ture. One case punctured the urethra, and we considered this case a failure. The RMS distance between the planned and implanted needles was 3.6 mm, 3.0 mm, 4.5 mm, and 2.3 mm, respectively. Most of this error was along the needle insertion direction, i.e., z-direction. The RMS distance between the planned and implanted needles was 1.4 mm, 1.7 mm, 1.5 mm, and 0.9 mm, respectively, when only the xand y-direction error was included. Despite these placement errors, 20 a satisfactory dose distribution for our clinic was achieved for every case. The maximum difference in target coverage between the planned and implanted needle arrangement was 2%. Conclusion: Without the use of real-time feedback, needle insertion using custom needle guides leads to substantial placement errors. These placement errors do not seem to affect the ability to meet clinical dose objectives. Future work will explore these needle guides in the context of a more 25 complete brachytherapy delivery system.