Performance of an Enhanced Afterloader with Electromagnetic Tracking Capabilities for Channel Reconstruction and Error Detection.

The reconstruction and error detection performance of a Flexitron research unit equipped with a special check-cable in which is integrated an EM sensor (NDI Aurora V3) that enables tracking and reconstruction capability was assessed. Reconstructions of a 24 cm long catheter were performed using two methods: continuous fixed speed check cable backward stepping and stepping through each dwell position every 1mm. The ability of the system to differentiate between two closely located (parallel) catheters was investigated by connecting catheters to the afterloader and moving it from its axis with an increment of 1mm. A robotic arm was used to move the catheter between each reconstruction. Reconstructions were obtained with a locally weighted scatter-plot smoothing algorithm. To quantify the reconstruction accuracy, distances between two catheters were computed along the reconstruction track with a 5mm step. Reconstructions of curve catheter paths were assessed through parallel and perpendicular phantom configuration to the EM field generator. Indexer length and lateral error were simulated and ROC analysis was made. Using a 50cm/s check cable speed does not allow accurate reconstruction. A slower check cable speed results in better reconstruction performance and smaller standard deviations. The optimum operating distance from the field generator (50-300mm) resulted in a lower absolute mean deviation from the expected value (0.2pm0.1mm) versus being positioned on the edge of the electromagnetic sensitive detection volume (0.6pm0.3mm). Reconstructions of curved and circular paths with a check cable speed under 5cm/s gave a 0.8pm0.5mm error, or better. All indexer and lateral shifts of 1mm were detected with a check cable speed of 2.5cm/s or under. A speed under 5cm/s is recommended for straight and curved catheter reconstructions. The check cable can also be used to detect common shift errors.