Detection accuracy of I-125 seeds and needle tips, in MRI prostate tissue mimicking material: Phantom design, fabrication, and testing

Abstract

Accurately identifying needle tip and seed positions for low dose rate prostate brachytherapy on MRI images is challenging. Uncertainties in locating needle tip positions can lead to misplacement of seeds compared to planned coordinates. Furthermore uncertainty in establishing true seed positions on the images, leads to uncertainty in the dose distributions.

In this study, a novel phantom has been designed for the analysis of I-125 seed and needle tip detection and tip image distortion. The phantom utilises a gel that mimics prostate tissue in MRI, to evaluate the uncertainty in establishing seed and needle tip positions. Reults are reported for the IsoSeed (Bebig) source, in clinically relevant seed arrangements, and for a novel nitinol needle.

The choice of MRI sequence impacts the accuracy of detecting the needle tips and seeds. This is most prevelant when the seeds are in clusters, at the boundary of the prostate and at 90˚ to the long axis of the scanner. Detected needle tip position, when the MRI metal artefact correction algorithm was used, was measured consistently inferior to the actual position (mean tip at -2.3 ± 1.5 mm (k = 2), p = 0.03).

We have demonstrated the design of a phantom that can be used to quantitatively assess seed and needle tip positions simultaneously, to establish the accuracy of detection, or presence of artefacts on MRI.