How does CBCT reconstruction algorithm impact on deformably mapped targets and accumulated dose distributions?

Area of Science:

• Medical Physics
• Radiotherapy
• Image Reconstruction

Background:

• Cone-beam CT (CBCT) is crucial for image-guided radiotherapy.
• Standard Feldkamp-Davis-Kress (FDK) reconstruction may introduce artifacts affecting downstream analysis.
• Iterative reconstruction algorithms offer potential improvements in image quality.

Purpose of the Study:

• To quantitatively compare deformable image registration (DIR) and deformable dose accumulation (DDA) between standard FDK-CBCT and a novel iterative CBCT reconstruction algorithm.
• To assess the impact of different CBCT reconstruction methods on dosimetric accuracy and contouring in prostate cancer radiotherapy.

Main Methods:

• CBCT datasets were reconstructed using both FDK and iterative algorithms for 323 treatment fractions in 10 prostate cancer patients.
• Planning CT images were registered to CBCT datasets, and daily doses were computed and mapped to the planning CT.
• Comparisons included voxel dose differences, clinical constraint adherence, and contour accuracy using three DIR algorithms against ground truth.

Main Results:

• Average daily dose differences were within 1 cGy, but maximum point dose differences after deformable mapping reached 1.33 Gy.
• Maximum normalized DDA differences per fraction were up to 0.80 Gy.
• The iterative CBCT algorithm improved contour accuracy, with increased Dice similarity (0.10) and decreased distance metrics (2.5 mm for mass center, 3.3 mm for Hausdorff).

Conclusions:

• The iterative CBCT reconstruction algorithm yields different mapped volumes of interest, deformed doses, and cumulative doses compared to conventional FDK-CBCT.
• These differences suggest that iterative reconstruction may enhance the accuracy of image-guided radiotherapy for prostate cancer patients.