Quantitative Assessment of CT Artifact Severity for Cs-131 Brachytherapy Seed Designs with and without a Radiographic Marker
Abstract
Purpose
Current LDR brachytherapy seeds contain internal radiopaque markers to enable visibility with radio-graphic imaging. However, this high-Z material introduces artifacts on CT that degrade image quality and may affect seed localization and dosimetric accuracy. This work quantitatively evaluates artifact reduction between Cs-131 seed designs with and without gold markers.
Methods
CT images were analyzed from three experimental datasets containing Cs-131 seeds with and without a gold marker. Seeds were embedded in bolus material for the first two datasets, while in the third dataset the seeds were placed in bovine tissue to approximate a tissue-equivalent environment. Regions of interest (ROIs) were defined around individual seeds and corresponding background regions. The ρ-index characterizes the magnitude of CT artifact severity and was derived herein from extreme-value statistics of local Hounsfield Unit perturbations near the seed. The ρ values from all ROIs were pooled across datasets and grouped by seed design. Group means and standard deviations were calculated to glean statistical significance.
Results
Gold seed ROIs exhibited substantially higher ρ values than no-gold ROIs. Across all datasets, gold seeds produced ρ=15.4±4.1, while no marker produced ρ=6.8±3.6. These were statistically different (p<0.0001, one-sided Mann–Whitney U test). The reduction in ρ values for the no-marker seeds was observed consistently across all three datasets, including both bolus and tissue-equivalent environments. This indicates that the decrease in artifact severity was not dependent on the surrounding medium and reflects an intrinsic improvement in seed design. The limited overlap between the gold and no-gold marker distributions further supports the robustness of this effect.
Conclusion
These results demonstrated that eliminating the internal gold marker in Cs-131 seeds substantially reduced CT imaging artifacts and improved visualization of the seed and neighboring tissues. The ρ-index provided a robust quantitative metric for comparing seed designs and assessing CT artifact severity.