Room: Exhibit Hall | Forum 4
Purpose: To assess the intra-fraction variation of dose delivered to organs-at-risk (OARs) using a kV-CBCT scanner and estimate the role of this technology in adaptive high-dose-rate brachytherapy (HDRBT) of cervical cancer.
Methods: Planning CT-images, pre-treatment kV-CBCT-images, and data of 19 cervical cancer patients and 57 HDRBT sessions were retrospectively reviewed. The kV-CBCT-images were manually registered with CT-images using the applicators as landmarks. The OARs (bladder, rectum and sigmoid) were contoured in each image-set. Image-quality of both scanners was evaluated and compared using a Catphan-504 phantom. The subjective perception of image-quality was assessed by the radiation oncologist using a visual-grading scale. Correlations between image-quality and patientâ€™ BMI, and the impact of image-quality on intra-fractional OARs-volume variation were explored in univariate analysis by linear-regression test.The variations of OARs doses D2cc and D0.1cc were evaluated by the relative-percentage-difference between the doses calculated on CT- and kV-CBCT-images. Possible correlations between OARs-dose and volume variation were investigated by linear-regression test.
Results: Although kV-CBCT-images had lower performance than CT regarding the low contrast resolution, uniformity and CNR, the image-quality was sufficient to define the applicators but not all OARs. Negative weak, yet statistically-significant correlations were found between kV-CBCT image-quality scores for bladder and rectum and patientâ€™ BMI (p=0.005 and p=0.019 respectively).Average intra-fractional variations of OARs-volume were: 2%Â±34.9% for bladder, 1.6%Â±46.9% for rectum, and 8.2%Â±62% for sigmoid. No statistical significant correlation was found between image-quality and intra-fraction variations of OARs-volumes.We found -2.7%, -1.2%, and 6.3% difference in mean D2cc dose for bladder, rectum and sigmoid respectively, between planning-CT and pre-treatment-kV-CBCT. Large standard deviations for all OARs were noticed.
Conclusion: The mean OARs-dose difference in cervix HDRBT was found to be small, although large random variations were observed in individual patients. A kV-CBCT-scan acquired before treatment can detect unfavorable anatomical changes, warranting adaptive treatment planning.