Related Experiment Videos
Prostate brachytherapy seed identification on post-implant TRUS images
Ben H Han1, Kent Wallner, Gregory Merrick
1Department of Radiation Oncology, University of Washington, Seattle, Washington 98108, USA.
Medical Physics
|May 30, 2003
Summary
Transrectal ultrasound (TRUS) shows limited reliability for identifying brachytherapy seeds post-implant. Physician identification accuracy varied significantly, with a median of only 43% of seeds correctly identified, questioning its use for dosimetry.
Area of Science:
- Medical Imaging
- Radiation Oncology
- Medical Physics
Background:
- Transrectal ultrasound (TRUS) is considered for intraoperative dosimetry in brachytherapy due to its availability.
- Current methods often rely on CT scans, which lack intraoperative convenience.
- The feasibility of using standard 2D TRUS for post-implant seed identification requires rigorous evaluation.
Purpose of the Study:
- To assess the feasibility and reliability of identifying brachytherapy seeds using standard two-dimensional transrectal ultrasound (TRUS) images immediately after implantation.
- To evaluate the inter-observer agreement among physicians in identifying seeds on TRUS images.
- To determine the accuracy of TRUS-based seed identification compared to the actual number of implanted seeds.
Main Methods:
- Ten patients undergoing I-125 or Pd-103 brachytherapy were studied.
- Transverse 6 MHz TRUS images were acquired immediately post-implant.
- Four physicians independently identified potential seeds on the images using overlaid grids, without knowledge of the actual seed count or type.
- Manual comparison of overlays was performed to assess identification accuracy and observer agreement.
Main Results:
- The median number of actual implanted seeds was 60 (range: 44-108).
- Physicians identified a median of 43 seeds per patient (range: 26-82), averaging 74% accuracy.
- Individual physician accuracy varied widely (median 44% to 91%).
- Inter-observer agreement was low, with only 20% of identified spots recognized by all four physicians.
- Five instances of over-identification (more seeds marked than implanted) occurred.
Conclusions:
- Standard two-dimensional TRUS shows significant limitations in reliably identifying brachytherapy seeds post-implant.
- The accuracy and consistency of seed identification among physicians were suboptimal.
- TRUS-based seed identification for intraoperative post-implant dosimetry should be approached with skepticism due to current reliability issues.