Investigating the use of image thresholding in brachytherapy catheter reconstruction

This study evaluated the accuracy of image thresholding in the reconstruction of catheters in brachytherapy treatment planning systems. Six test cases including four planar catheter configurations, an interstitial prostate and an intracavitary treatment plan were made use of in this study. The four planar CT scanned catheter arrangements included; catheters placed approximately 1, 0.5 cm apart, catheters closely arranged in a plan (<0.5 cm apart) and a loop arrangement. The intracavitary plan consisted of catheters arranged inside a mould configuration. All reconstruction methods were based on tracking wire markers placed inside the plastic catheters. Each of these catheter arrangements was reconstructed using an existing window adjustment technique (manual reconstruction) in the treatment planning system followed by a CT-based automated thresholding technique available in the same planning system. A corresponding reconstructed catheter was created using a segmented catheter structure using image thresholding from another planning system within the same department. Co-ordinates from all the reconstructed catheters were compared against each other to assess the geometric shift between manual and threshold based reconstruction on each transaxial image using in-house software and the maximum variations were recorded for assessment. It was observed in general that automated thresholding technique could assist in catheter reconstruction for catheters which are greater than 0.5 cm apart. The segmented thresholding method reported smaller variations when compared to the manual reconstruction using window adjustment technique. Automated reconstruction saves time in the brachytherapy planning, however it was noted that it is not feasible for closely spaced catheters. Segmented catheter reconstruction although time consuming, did provide a better alternative in most cases.