Magnetic resonance imaging in the radiation treatment planning of localized prostate cancer using intra-prostatic fiducial markers for computed tomography co-registration

Abstract

To assess the feasibility, and potential implications, of using intra-prostatic fiducial markers, rather than bony landmarks, for the co-registration of computed tomography (CT) and magnetic resonance (MR) images in the radiation treatment planning of localized prostate cancer. All men treated with conformal therapy for localized prostate cancer underwent routine pre-treatment insertion of prostatic fiducial markers to assist with gross target volume (GTV) delineation and to identify prostate positioning during therapy. Six of these men were selected for investigation. Phantom MRI measurements were obtained to quantify image distortion, to determine the most suitable gold alloy marker composition, and to identify the spin-echo sequences that optimized both marker identification and the contrast between the prostate and the surrounding tissues. The GTV for each patient was contoured independently by three radiation oncologists on axial planning CT slices, and on axial MRI slices fused to the CT slices by matching the implanted fiducial markers. From each set of contours the scan common volume (SCV), and the scan encompassing volume (SEV), were obtained. The ratio SEV/SCV for a given scan is a measure of inter-observer variation in contouring. For each of the 18 patient-observer combinations the observer common volume (OCV) and the observer encompassing volume (OEV) was obtained. The ratio OEV/OCV for a given patient-observer combination is a measure of the inter-modality variation in contouring. The distance from the treatment planning isocenter to the prostate contours was measured and the discrepancy between the CT- and the MR-defined contour recorded. The discrepancies between the CT- and MR-defined contours of the posterior prostate were recorded in the sagittal plane at 1-cm intervals above and below the isocenter. Phantom measurements demonstrated trivial image distortion within the required field of view, and an 18K Au/Cu alloy to be the marker composition most suitable for CT-MRI image fusion purposes. Inter-observer variation in prostate contouring was significantly less for MR compared to CT. The mean SEV/SCV ratio was 1.58 (confidence interval (CI): 1.47-1.69) for CT scans and 1.37 (CI: 1.33-1.41) for MR scans (paired t-test; P=0.036). The overall magnitude of contoured GTV was similar for MR and CT; however, there were spatial discrepancies in contouring between the two modalities. The greatest systematic discrepancy was at the posterior apical prostate border, which was defined 3.6 mm (SD 3.5 mm) more posterior on MR- than CT-defined contouring. Prostate contouring on MR is associated with less inter-observer variation than on CT. In addition, we have demonstrated the feasibility of using intra-prostatic fiducial markers, rather than bony landmarks, for the co-registration of CT and MR images in the radiation treatment planning of localized prostate cancer. This technique, together with on-line correction of treatment set-up according to the fiducial marker position on electronic portal imaging, may enable a reduction in the planning target volume (PTV) margin needed to account for inter-observer error in target delineation, and for prostate motion.