6C-3 Singular Spectrum Analysis for Detecting Brachytherapy Seeds with Angle Variation

Abstract

Prostate brachytherapy uses small, titanium-shelled radioactive seeds to locally treat prostate cancer. Unfortunately, during the implantation procedure, needles inserted under the guidance of transrectal-ultrasound (TRUS) cause gland movement, which often results in seed misplacement and suboptimal dosimetry. In a previous study, an algorithm based on singular spectrum analysis (SSA) was described to determine actual seed locations. The SSA algorithm demonstrated promise for reliably imaging seeds when their orientation was optimal, i.e., with the seed axis normal to the ultrasound beam. In this present study, the performance of the SSA algorithm was tested when the orientation of the seed deviated up to 22 degrees from normality. Experimental data from a single seed inserted in a gel pad and another inserted in a piece of beef were collected with a single-element, 5-MHz transducer. Simulations were designed and evaluated with the SSA algorithm. Finally, success in seed detection was expressed quantitatively through developed scoring metrics. These metrics also were computed for conventional B-mode images used for comparison. The results showed that the SSA algorithm was able to detect the seed accurately up to a deviation of ap8 degrees.