Room: Exhibit Hall | Forum 1
Purpose: A novel iterative CBCT reconstruction algorithm using patient-specific scatter corrections has been shown to improve image quality, in particular with regard to image noise reduction and enhancement of soft-tissue boundaries. Here we perform quantitative analysis of differences in deformable dose accumulation (DDA) computed on CBCT datasets reconstructed using the standard (Feldkamp-Davis-Kress: FDK-CBCT) and iterative (iterative-CBCT) algorithms.
Methods: Ten prostate cancer patients were selected for this retrospective analysis. An in-house system was used to perform automatic, B-spline-based deformable image registration (DIR) and DDA between planning CT (pCT) and CBCT images for 361 fractions. Dose calculation was performed on deformed CT (dCT) image set which was generated from DIR between each CBCT image set and the pCT. Dose constraints for targets and organs-at-risk were evaluated from DDA based on FDK-CBCT and iterative-CBCT image sets. Direct dose comparisons were performed for daily doses before and after mapping with the pCT dataset, and DDA based on FDK-CBCT and iterative-CBCT.
Results: For 7 patients, dose differences over 2 Gy were noted between DDA distributions based on FDK-CBCT and iterative-CBCT datasets; for instance minimum PTV doses varied as much as 8.5 Gy. Maximum differences in overall cumulative doses 15.87Â±6.54 Gy, with the greatest variation observed in the soft-tissues within the pelvic region where DIR results were found to be different. Maximum daily dose variation between FDK-CBCT and iterative-CBCT image sets increased from 0.25Â±0.09 Gy to 1.40Â±0.40 Gy after doses mapping to pCTâ€™s.
Conclusion: Although daily dose distributions are similar, displacement vector fields (DVF) between FDK-CBCT and iterative-CBCT datasets were observed to be different, resulting in discrepancies in accumulated dose distributions, especially in high dose gradient regions. The improved image quality from iterative-CBCT is a promising approach to potentially yield more accurate adaptive accumulated dose distributions relative to FDK-CBCT.
Funding Support, Disclosures, and Conflict of Interest: Work supported in part by a grant from Varian Medical Systems, Palo Alto, CA.
Not Applicable / None Entered.