Abstract: |
In prostate brachytherapy, the physical advantages of high dose rate (HDR) treatment with remote afterloading versus low dose rate (LDR) permanent implant include more conformal dose distribution. In practice, catheters placed in the patient are often not rectilinear or parallel to each other as intended. If the treatment planning system does not account for this feature the resulting treatment plan - while adequate on paper - may result in erroneous dose delivery. There are basically three approaches to catheter digitization: (1) Parallel reconstruction: one digitizes only one point for each catheter at the central slice, and assumes that the catheters are parallel to each other and horizontal; (2) Straight reconstruction: both ends of the catheters are digitized, thus assuming the catheters are straight while at an angle to each other; (3) Slice-by-slice reconstruction: every catheter is digitized on all slices to obtain its exact geometry. The in-house treatment planning system currently in use at Memorial Sloan-Kettering Cancer Center makes use of slice-by-slice reconstruction. With isodose distributions and dose volume histograms, we studied the problems encountered when approximate reconstruction methods are used. Dose distribution resulting from parallel reconstruction is significantly different from "real" distribution. In contrast, dosimetric difference between straight and slice-by-slice reconstructions appears minor. This is because quite often catheters are not parallel to each other, but each individual catheter is fairly straight. We suggest that parallel reconstruction is too inaccurate for dosimetric planning, while slice-by-slice reconstruction may be too time-consuming. Straight (two point) reconstruction represents a balance between accuracy and efficiency. |