PO94

Purpose Prostatic edema following transperineal interstitial permanent prostate brachytherapy implantation is commonly evaluated based on either prostate or implant volume. The current study compares the edema time course between the MR-delineated prostate contour and the CT-localized stranded seeds, enabling pairwise comparison in the presence of individual patient variation. In addition, unique identification of seeds enables the characterization of stranded implant dynamics. Materials and Methods Twenty patients were implanted with stranded Iodine-125 seeds (0.5 U strength) to the prostate at a prescribed dose of 145 Gy, following standard procedure. Pelvic scans were performed using computer tomography (CT) and magnetic resonance imaging (MRI) (T2-weighted fast spin-echo and balanced steady-state free precession (bSSFP)) on the day of implantation (D0), D3, D10, and D30 (30 days post-implant). A Prostate Coordinate System, based on the MR-delineated prostate contour, served as a common coordinate system across all time points. MR(bSSFP)-CT rigid registration was performed based on the mutual information metric. A strand reconstruction software uniquely matched individual seeds to the strand configuration in the preplan. The relative edema, normalized to D30, was calculated for MR-based contours and CT-based seed positions. Correlated movement of seeds within a strand were quantified: strand movement was calculated from the shift in the strand center-of-mass; strand length was determined as the total length of the line segments connecting sequential seeds in a strand. Simulation of the stranded seed model was performed. Initial D0 seed positions were moved based on the observed strand characteristic movement and compared against actual D30 seed positions. Results Prostatic edema resulted in swelling of the prostate, which peaks at D0 and mostly resolves by D30. The contour- and seed- based relative edema were similar and correlated (p < 0.01) in the lateral and ant-pos directions. The edema magnitudes differed noticeably in the sup-inf direction with no statistically significant correlation (p = 0.11). The average strand movement was 0.09, 0.12, and 0.26 cm in the Lateral, Ant-Pos, and Sup-Inf directions respective, resulting in a more compact seed distribution. The movement was largest between D0 and D3 and smallest between D10 and D30. Conversely, the strand length was relatively constant during the initial time points, followed by a length contraction of 5% between D10 and D30. Thus, the stranding material initially limits independent seed movement (i.e. strands moved as a whole) and subsequently loses integrity over time, allowing for strand contraction. Simulation of the stranded seed model reproduced the observed relative edema, particularly in the Sup-Inf strand direction. The average residual distance between simulated and actual D30 seed positions was 0.27 cm. For comparison, the actual seed movement was 0.38 cm and the residual from a loose seed model was 0.29 cm. Conclusions The study characterized edema resolution based on stranded seeds in permanent prostate brachytherapy. Comparison of contour-based and seed-based relative edema for the same patient cohort revealed statistically significant differences in the strand direction. Dynamic strand-specific behaviours pointed to the potential impact of stranding material integrity. Simulation of stranded seed behaviour reproduced the observed relative edema, presenting a plausible explanation of the dynamics of stranded seeds during the course of edema resolution. Prostatic edema following transperineal interstitial permanent prostate brachytherapy implantation is commonly evaluated based on either prostate or implant volume. The current study compares the edema time course between the MR-delineated prostate contour and the CT-localized stranded seeds, enabling pairwise comparison in the presence of individual patient variation. In addition, unique identification of seeds enables the characterization of stranded implant dynamics. Twenty patients were implanted with stranded Iodine-125 seeds (0.5 U strength) to the prostate at a prescribed dose of 145 Gy, following standard procedure. Pelvic scans were performed using computer tomography (CT) and magnetic resonance imaging (MRI) (T2-weighted fast spin-echo and balanced steady-state free precession (bSSFP)) on the day of implantation (D0), D3, D10, and D30 (30 days post-implant). A Prostate Coordinate System, based on the MR-delineated prostate contour, served as a common coordinate system across all time points. MR(bSSFP)-CT rigid registration was performed based on the mutual information metric. A strand reconstruction software uniquely matched individual seeds to the strand configuration in the preplan. The relative edema, normalized to D30, was calculated for MR-based contours and CT-based seed positions. Correlated movement of seeds within a strand were quantified: strand movement was calculated from the shift in the strand center-of-mass; strand length was determined as the total length of the line segments connecting sequential seeds in a strand. Simulation of the stranded seed model was performed. Initial D0 seed positions were moved based on the observed strand characteristic movement and compared against actual D30 seed positions. Prostatic edema resulted in swelling of the prostate, which peaks at D0 and mostly resolves by D30. The contour- and seed- based relative edema were similar and correlated (p < 0.01) in the lateral and ant-pos directions. The edema magnitudes differed noticeably in the sup-inf direction with no statistically significant correlation (p = 0.11). The average strand movement was 0.09, 0.12, and 0.26 cm in the Lateral, Ant-Pos, and Sup-Inf directions respective, resulting in a more compact seed distribution. The movement was largest between D0 and D3 and smallest between D10 and D30. Conversely, the strand length was relatively constant during the initial time points, followed by a length contraction of 5% between D10 and D30. Thus, the stranding material initially limits independent seed movement (i.e. strands moved as a whole) and subsequently loses integrity over time, allowing for strand contraction. Simulation of the stranded seed model reproduced the observed relative edema, particularly in the Sup-Inf strand direction. The average residual distance between simulated and actual D30 seed positions was 0.27 cm. For comparison, the actual seed movement was 0.38 cm and the residual from a loose seed model was 0.29 cm. The study characterized edema resolution based on stranded seeds in permanent prostate brachytherapy. Comparison of contour-based and seed-based relative edema for the same patient cohort revealed statistically significant differences in the strand direction. Dynamic strand-specific behaviours pointed to the potential impact of stranding material integrity. Simulation of stranded seed behaviour reproduced the observed relative edema, presenting a plausible explanation of the dynamics of stranded seeds during the course of edema resolution.