Intraoperative 3D reconstruction of prostate brachytherapy implants with automatic pose correction

Area of Science:

• Medical Physics
• Radiotherapy
• Image-guided surgery

Background:

• Prostate brachytherapy requires precise seed placement for effective radiation dose delivery.
• Intraoperative seed localization is crucial for real-time dose assessment and treatment optimization.
• Existing algorithms like REDMAPS offer rapid seed localization but can be sensitive to imaging pose errors.

Purpose of the Study:

• To enhance the accuracy and robustness of prostate brachytherapy seed reconstruction.
• To develop an automatic pose correction (APC) process integrated with the REDMAPS algorithm.
• To improve seed localization performance even with significant imaging pose variations.

Main Methods:

• Integration of an automatic pose correction (APC) process with the REDMAPS seed localization algorithm.
• Utilizing reconstructed seeds as fiducials to iteratively correct imaging pose errors.
• Minimizing overall projection error to refine seed matching and reconstruction.

Main Results:

• Achieved a high average seed matching rate of over 99.4%.
• Reduced reconstruction error to 0.5 mm or less.
• Demonstrated matching solution optimality exceeding 99.8% in simulations and clinical studies.
• Significantly improved seed reconstruction accuracy, especially with larger pose errors.

Conclusions:

• The combined REDMAPS and APC approach significantly enhances the accuracy of prostate brachytherapy seed localization.
• This method is robust to considerable imaging pose errors, increasing its clinical applicability.
• Improved seed localization accuracy supports more precise dose delivery and treatment planning in prostate brachytherapy.