Comparison of MRI- and CT-based post-implant dosimetric analysis of transperineal interstitial permanent prostate brachytherapy.

Abstract

The purpose of this work was to investigate how a recently developed MRI-based post-implant dosimetric analysis technique for ultrasound guided transperineal interstitial permanent prostate brachytherapy (TIPPB) compared with the currently accepted CT-based technique. The study was based upon 3-mm MRI and CT scans of 15 patients who had received either 125I or 103Pd implantation. All images were acquired on post-operative day 1 and within 1 hr of each other. Prostate volumes were determined by the same physician. Sources were digitized and calculations performed using an in-house treatment planning system with a nearest neighbor seed sorting routine and AAPM TG43 formalism. Prostate volume, geometric source distribution spread (rcom), dose volume histogram (DVH), and tumor control probability (TCP) calculations were performed from both image sets. Differences in source localization were evaluated by comparing source spread and prescription isodose volumes. Differences in dosimetric analysis were evaluated through prostate-specific DVH and TCP comparisons. Prostate volume as determined from MRI was larger than that of CT by an average of +9.1% (R = 0.70). Calculated rcom was smaller by an average of -0.9 mm (R = 0.81). Isodose volumes at 80, 90, 100, and 150% of the prescription dose differed by an average of +2.5, +2.9, -2.9, and +4.8%, respectively (R = 0.97, 0.98, 0.98, and 0.91). Percentage volume of the prostate encompassed by 80, 100, and 150% of the prescription dose differed by an average of -0.9, -0.9, and -0.1%, respectively (R = 0.34, 0.35, and 0.35). TCP differed by an average of -0.8% (R = 0.37). The results of this study further support our initial findings that MRI may be used to reliably localize the implanted sources for TIPPB. This study also demonstrated that MRI-based post-implant dosimetric analysis is possible. However, it is evident that differences in prostate localization from MRI to CT can result in significantly different assessments of prostate volume coverage. There is clearly a need to further quantify the differences between these two imaging modalities in this application and address whether greater accuracy in describing the dose-volume relationship based on improvements in visualization of the prostate gland from MRI will translate into improved correlation with treatment outcome.